52007SC0106

[pic] | COMMISSION OF THE EUROPEAN COMMUNITIES |

Brussels, 31.1.2007

SEC(2007) 106

COMMISSION STAFF WORKING DOCUMENT Accompanying the

COMMUNICATION FROM THE COMMISSION Sector Inquiry under Article 17 of Regulation (EC) No 1/2003 on retail banking (Final Report) {COM(2007) 33 final}

DG Competition Report on

Retail Banking Sector Inquiry

TABLE OF CONTENTS

Introduction 6

The importance of the retail banking sector in Europe 6

The Commission’s sector inquiry into retail banking 7

The methodology of the sector inquiry 7

Format of the report 8

Part A: Competition in the market for current accounts and related services 10

A.1. Market characteristics of retail banking in Europe 11

A.1.1. Supply side characteristics of retail banking markets 11

A.1.2. Demand side characteristics of retail banking markets 14

A.1.3. Differing distribution models for retail banking products 14

A.1.4. Regulation of retail banking 15

A.2. Market structures and financial performance in retail banking 18

A.2.1. Market concentration and integration in retail banking 18

A.2.2. Financial performance of the retail banking sector 22

A.3. Credit registers and banking competition 29

A.3.1. Economics of credit data markets and credit registers 29

A.3.2. Markets for credit information in the Member States 30

A.3.3. Potential competition issues 38

A.3.4. Conclusions 45

A.4. Cooperation 48

A.4.1. Scope of the inquiry's analysis of cooperation 48

A.4.2. Co-operative banks 49

A.4.3. Savings banks 52

A.4.4. Competition issues raised by cooperation among banks 55

A.4.5. Conclusions 59

A.5. Setting of policies and prices 61

A.5.1. Banks' practice of product tying 61

A.5.2. Setting of prices on current accounts 71

A.5.4. Conclusions 80

A.6. Customer choice and mobility 82

A.6.1. Relationship between customer mobility and competition 82

A.6.2. Switching costs in retail banking 84

A.6.3. The sector inquiry’s evidence on customer mobility in retail banking 89

A.6.4. Empirical relationship between customer mobility and market performance 94

A.6.5. Conclusions 98

Part B: Competition in the market for payment cards and payment systems 101

B.1 Introduction 102

B.1.1. Cashless payments in the EU 102

B.1.2. Creating a single market for payment services 104

B.1.3. Organisation of POS Card Payment Systems 105

B.1.4. Fees paid in a POS card payment system 107

B.1.5. The inquiry’s data and methodology on payment cards 107

B.2. Market concentration and integration 110

B.2.1. Concentration in acquiring and issuing markets 110

B.2.2. The relationship between market concentration and merchant fees 114

B.2.3. Integration of card Payment Systems 114

B.2.4. Competition issues 118

B.2.5. Conclusions 120

B.3. Cardholder fees 121

B.3.1. Credit cards 122

B.3.2. Debit cards 124

B.3.3. Correlation between cardholder and interchange fees 125

B.3.4. Conclusions 126

B.4. Merchant fees 128

B.4.1. MSCs for Visa and MasterCard credit cards 129

B.4.2. MSCs for debit cards 130

B.4.3. The effect of blending and surcharges on the setting of MSCs 132

B.4.4. Conclusions 134

B.5. Interchange Fees 136

B.5.1. Types of interchange fee applied in international networks 136

B.5.2. Credit card interchange fees in international networks 137

B.5.3. Debit card interchange fees 139

B.5.4. Industry and academic views on the role of interchange fees 142

B.5.5. Competition issues arising from interchange fees 143

B.5.6. Conclusions 144

B.6. Profitability of the payment cards industry 146

B.6.1. Methodology for the inquiry’s profitability analysis 146

B.6.2. The profitability of credit card business 147

B.6.3. Profitability of debit card business 152

B.6.4. Findings on profitability in the payment cards industry 155

B.6.5. Conclusions 157

B.7. Free funding periods for payment card transactions 159

B.7.1. Main parameters affecting funding periods for card transactions 159

B.7.2. Analysis of free funding period and net float per card brand 160

B.7.3. Analysis of free funding period and net float in the Member States 161

B.7.4. Conclusions 164

B.8. Membership and governance rules 165

B.8.1. Membership conditions in payment card networks 165

B.8.2. Network joining fees 167

B.8.3. Governance in Card Payment Systems 168

B.8.4. Conclusions 171

B.9. Cross-border Competition in Acquiring 172

B.9.1. Merchants acquired cross-border 172

B.9.2. Conclusions 174

B.10. Payment infrastructures 176

B.10.1. The typical operation of payment infrastructures 176

B.10.2. Methodology and scope of the inquiry into payment infrastructures 178

B.10.3. Main features of surveyed payment systems 180

B.10.4. Ownership and management of payment infrastructures 181

B.10.5. Access issues 181

B.10.6. Fees charged to users 182

B.10.7. Competition analysis 185

B.10.8. The European context: towards a Single Euro Payments Area 188

B.10.9. Conclusions 189

B.11. Multilaterally agreed interchange fees for ATMs and for non-card payments 191

B.11.1. Interchange fees for ATM withdrawals 191

B.11.2. Multilateral interchange fees for other non-card payments 192

B.11.3. Conclusions 195

Part C: Conclusions and possible next steps 197

C.1. Conclusions and possible next steps 198

C.1.1. Findings on retail banking market structure and performance 198

C.1.2. Findings and recommendations on current accounts and related services 199

C.1.3. Findings and recommendations on payment cards and payment systems 203

Glossary 209

Introduction

Competitive financial services markets that serve European consumers and businesses efficiently contribute to economic growth and, therefore, to the achievement of the Lisbon goals. Against this background the Commission in June 2005 decided to open sector inquiries into two important areas of the financial services sector: retail banking and business insurance.

The instrument of sector inquiries has its legal basis in Article 17 (1) of Regulation 1/2003, according to which the Commission may conduct an inquiry into a particular sector of the economy or into particular types of agreements across various sectors, where the trend of trade between Member States, the rigidity of prices or other circumstances suggest that competition may be restricted or distorted within the common market. Sector inquiries represent an important element in the Commission’s modernised approach to competition policy which abolished the notification system and opted for a more pro-active antitrust practice.

In December 2005, following extensive consultation with stakeholders, the Commission set out its future strategy on financial services in the White Paper Financial services policy 2005-2010 [1]. The retail banking sector inquiry makes an important contribution to this strategy , which identified as priorities the extension of better regulation principles into all policy making and the strengthening of competition among providers, particularly in retail business.

To underpin the development of a single market for financial services and harness the full potential benefits of the Euro, the European banking industry is creating a Single Euro Payment Area (SEPA). The Commission as well as the European Central Bank strongly support[2] and are working closely with industry on the development of SEPA.

The importance of the retail banking sector in Europe

Despite growth and diversification in the financial services sector, retail banking – banking services to consumers and small firms – remains the most important sub-sector of banking, representing over 50% of total banking activity in Western Europe[3]. The Commission estimates that in 2004 retail banking activity in the European Union generated gross income of €250-275 billion, equivalent to approximately 2% of total EU GDP. As a whole the banking sector in the European Union directly provides over three million jobs.

In essence retail banks[4] provide three basic services to consumers and small businesses: saving; borrowing; and payment services. Retail banking is carried out by a wide range of providers. These range from small banks that supply only retail services to medium-sized or very large banks that may operate in a range of banking activities (e.g. private or corporate banking).

The European payment cards industry is large and provides the means for a significant part of retail sales in Europe. Total sales volumes with point-of-sale card transactions in the EU in 2005 were more than €1350 billion. It is estimated that businesses in the EU paid more than €2 billion in fees in 2005.

The SEPA project aims to create an integrated market for payment services which is subject to effective competition and where there will be no distinction between cross-border and national payments within the euro area. This should make cross-border payments as easy and affordable as domestic payments. Successful implementation of the SEPA project could save the EU economy between €50 and €100 billion per year.[5]

The Commission’s sector inquiry into retail banking

In its sector inquiry the Commission has examined two complementary aspects of retail banking: firstly, the markets for payment cards and payment systems; and secondly, the markets for current accounts and related services. The detailed analysis and findings from both parts of the inquiry are presented together in this technical annex. The key findings of the report and recommendations are contained in the Final Report of the sector inquiry, which the Commission presents as a Communication to the Council and the European Parliament.

In line with their original objectives, the sector inquiries have identified competition concerns that may require investigation and remedy under the European competition rules. The Commission intends to provide a sound basis for a coherent approach to antitrust practice carried out by the National Competition Authorities (NCAs) and the Commission. Should there be evidence, after further investigation, that particular practices or arrangements violate EC or national competition law in an individual case, these practices or arrangements can be addressed by individual antitrust action.

In addition, the findings of the inquiry into retail banking, and in particular payment card systems, will provide valuable evidence to inform the future development of the SEPA project. In particular the enquiries aim to show how differing forms of organisation, structure and governance of payment systems in the EU can produce differing competitive outcomes. The evidence gathered for the enquiry suggests that the characteristics of some payment systems lead to significantly higher prices for firms and consumers in some Member States. As work continues to develop the appropriate principles and structures to support SEPA and its Payment Cards Framework (PCF), significant consideration should be given to the findings of the Commission’s retail banking enquiry.

The methodology of the sector inquiry

The core of the evidence base for the retail banking sector inquiry is provided by detailed pan-European market surveys of: (i) issuing and acquiring banks in the payment cards market; and (ii) banks providing retail banking services to consumers and small and medium enterprises (SMEs). Both of these market surveys were based on a large, statistically robust sample of around 250 banks, relating to their activities from 2000 onwards. The methodologies for the market surveys firstly on payment cards and payment systems and secondly on current accounts and related services, are described in detail in Chapter 3 of Interim Report I and Chapter 1 of Interim Report II of the sector inquiry.

Specifically in relation to payment systems, the Commission also conducted extensive market surveys of clearing and settlement (retail payment) systems and payment card networks throughout the EU25. The inquiries therefore provide a full and up-to-date overview of the structure of national and pan-European payment systems, and the potential competition issues arising from their operations.

In addition to the considerable evidence gathered on retail banking providers and payment systems, the Commission has examined the roles of several other important players in retail banking. Specifically in the course of the sector inquiry the Commission has gathered information from bank associations, banking regulators, national central banks and credit registers. The Commission looked into the practice of national competition authorities (NCAs) with the help of an inquiry organised by the NCAs. The Commission has also taken account of the views of consumers and small business through surveys of the relevant national associations and official sources such as the Eurobarometer surveys.

Finally, public discussion and consultation have played an important role in shaping the findings and recommendations of the retail banking sector inquiry. The first interim report of the inquiry, on payment cards and payments systems, was published in April 2006 and subject to a ten week public consultation. The second interim report, on current accounts and related services, was published in July 2006 and subject to a twelve week public consultation. A public hearing was held in July 2006 to present the preliminary findings of the inquiry and discuss them with market participants. These public consultation exercises produced extensive and valuable feedback for the Commission reflected in the Final Report of the sector inquiry. Non-confidential comments as well as a summary of the public feedback are published on DG COMP's website.[6]

Format of the report

The technical annex to the Final Report of the retail banking sector inquiry is structured as follows:

- Part A sets out the main findings and analysis concerning the market for current accounts and related services;

- Part B sets out the main findings and analysis concerning the market for payment cards and payment systems; and

- Part C summarises the inquiry's conclusions and possible next steps.

Part A: Competition in the market for current accounts and related services

A.1. Market characteristics of retail banking in Europe

This section of the report discusses the markets for current accounts and related services in the European Union; the bulk of core retail banking business. The central retail product of the current account is used as the starting point for the competition analysis here, and the scope is extended to consider other core retail banking products supplied to consumers and SMEs. These related products include credit products such as mortgages and loans to consumers and SMEs; and savings products such as deposit accounts.

The inquiry has found that the retail banking sector in the European Union remains largely fragmented along national lines and integration is far from complete. How far the economics of retail banking in the EU will lead to full integration remains an open question. There is evidence of integration and converging performance at the regional level, for example in the Benelux and Nordic countries. Overall, however, retail banking markets in the EU Member States display several common characteristics. This chapter surveys the principal characteristics of these banking markets, discussing:

- supply side characteristics of retail banking markets;

- demand side characteristics of retail banking markets;

- differing distribution models for retail banking products; and

- regulation of retail banking.

The following chapter summarises the sector inquiry’s findings on market structures and financial performance in retail banking.

A.1.1. Supply side characteristics of retail banking markets

Market structures differ considerably among Member States. This applies to the degree of market concentration as well as to the identity of leading players. Though concentration can be described as modest in most Member States, some countries such as Belgium, the Netherlands, Finland and Sweden have significantly higher concentration ratios. Retail banking in the Benelux and Nordic countries is also characterised by significantly more cross border activity and, consequently, a higher degree of market integration. Other countries such as Germany or Spain are dominated by savings or co-operative banks with a strong regional focus. Subsidiaries of foreign banks have a major market presence predominantly in the new Member States.

Fragmented market infrastructures

Some basic aspects of the infrastructure for European retail banking remain fragmented, which in turn entrenches the current fragmentation of product markets. Relevant aspects of market infrastructure are clearing and settlement (payment systems) and credit registers; and legal infrastructure such as tax policies, regulation and consumer protection regimes.

In relation to market infrastructure, firstly t he organisation and management of payment infrastructures varies significantly from country to country. Whereas payment systems are run by central banks on a non-profit basis in some Member States, others are operated by joint ventures of banks in various forms. Consequently, access conditions and fee structures differ considerably, and widespread entry barriers remain (discussed below in part B). However, the advent of SEPA should enable a fully integrated EU market for retail payment services. Secondly, there are major differences in the market structure and operation of credit registers across the EU (discussed below in chapter A.3). This fragmentation has consequences for the volume and type of customer data that is available to credit providers and for the ability of credit providers to access registers, especially in other Member States.

In relation to legal infrastructure, tax policies on company earnings, VAT treatment and capital gains vary between Member States. These varying tax regimes clearly influence the investment decisions of banks (for example on whether and how to enter new markets) and the consumption, saving and borrowing decisions of retail banking customers. Banking regulation is discussed below in more detail. It is worth noting here that while prudential rules have been largely harmonised at European level, significant differences remain in areas such as the ownership structures and the geographic scope of certain banks. Lastly, consumer protection rules for retail banking still vary considerably across the Member States, which raises the cost of entering new markets and maintains market fragmentation.

Traditionally high level of cooperation

Widespread cooperation between market participants is a common characteristic of retail banking markets. Banks co-operate in a variety of areas, including the interconnection and operation of payment systems; ownership or membership of credit registers; and the joint development/promotion of new products and services. As the inquiry shows, such co-operative activities of banks can provide a means for limiting competition. This can materialise in restricted access to networks and systems, discriminatory fee structures or in higher fees for consumers.

Multi-market contacts

Retail banks normally offer a wide range of products, so that the industry is not only characterised by networks and cooperation, but also by multi-market contacts of suppliers. Though some suppliers are specialised and only offer one product or a very limited range of products (e.g. mortgages or online current account services), the major players normally offer the full range of retail products. In theory multi-market contacts may induce collusion because retaliation against cheating firms can take place on all shared markets[7]. However, whether or not multi-market contacts may facilitate co-ordinated behaviour in retail banking markets can only be established on a case-by-case basis.

Varying degrees of price transparency

The literature on retail banking cites examples where prices for particular products are transparent and relatively easy to compare.[8] This typically applies to products such as deposits or mortgages, where the interest rate is a good proxy for the ‘price’ (though some significant charges or fees may apply to mortgages). However, there are several products where prices are clearly not very transparent; for example, current accounts. The effective price charged by banks for providing current accounts may be reflected in the interest rates and fees applied to the account, and also in the level of charges for payment services. Thus it may be hard for consumers and SMEs to assess the effective ‘price’ of a particular current account, and harder still to compare products across several suppliers with differentiated products.

Significant barriers to entry

Some characteristics of retail banking markets such as economies of scale, the importance of a local branch network, limited innovation and obstacles to customer mobility can function as entry barriers. Several comments submitted by banks in the context of the public consultation stated that entry mainly occurs by means of acquiring an existing customer base with a branch network and possibly an established brand.

Prudential rules and supervision can be used to hinder entry, for instance, takeovers, mergers or entry by foreign banks.[9] In addition, regulation in some EU countries restricts entry, mergers or takeovers concerning certain types of credit institutions (e.g. savings banks). Other entry barriers result from market structures and the conduct of participants, particularly with respect to cooperation agreements and the functioning of networks such as payments systems or credit bureaus. In the context of networks natural, regulatory and behavioural barriers can be distinguished. Whereas natural barriers are the result of the ‘inherent’ economies of scale of networks such as payment systems, access to networks may be also rendered difficult by artificial barriers such as regulatory provisions or incumbents’ behaviour. Fee structures that disadvantage smaller banks or newcomers may be the result of natural or artificial barriers.

A.1.2. Demand side characteristics of retail banking markets

The demand side characteristics, of retail banking can result in less efficient market performance. Such inefficiencies arise from two principal factors: firstly, information asymmetry, where banking consumers lack full information, reduces the intensity of price competition; and secondly, high switching costs, where the level of informational and transactional costs of changing some banking products (notably current accounts), discourage consumers from leaving their current provider. The combination of these two factors tends to reduce the mobility of customers in the market for products such as current accounts. Levels of mobility may also be artificially reduced by some banks’ practices, such as the tying of retail banking products and the imposition of high exit fees (discussed in chapter A.5).

The inquiry’s data suggests that in the majority of Member States annual switching rates for current accounts are low and stable at 5 to 10 per cent per year.[10] This suggests that the overall effects of information asymmetry and switching costs (whether intrinsic or artificial) are fairly similar across the EU, notwithstanding some differences between Member States. This observed low level of customer mobility has important consequences for the operation of retail banking markets in the EU and the intensity of competition because it seems to render market entry difficult For example, it is a common industry view that, for full service retail banking, ‘greenfield’ entry into other Member State markets tends to be more risky and less successful than entry through merger and acquisition. Among other factors, low customer mobility appears to be one explanation. Since retail banking customers are relatively immobile it is difficult for a greenfield operation to win large numbers of customers through price competition and thus acquire significant scale in a commercially viable time-frame.

A.1.3. Differing distribution models for retail banking products

Distribution models for retail banking products vary to some degree across Member States. The branch network remains the primary channel, although is increasingly complemented or substituted by other channels including internet and telephone. Distribution models also vary according to the product or products being sold. Banks also increasingly enter specific product markets, such as credit cards and deposit accounts, particularly when going cross-border. This approach can eliminate the need for a branch network and enable all operations to be carried out via internet or phone, significantly reducing costs.

However, for ‘full service’[11] retail banking a branch network remains vital since customers typically want and need face-to-face contact with bank staff to discuss their finances. As banks confirmed in their responses to the sector inquiry, a widespread branch network remains a critical factor in expansion and accruing significant market share in full service retail banking.[12] Large domestic players have an inherent advantage over smaller players through their branch operations. This does not necessarily mean there are competition problems in the market. However it does mean that over the medium term, existing branch networks will help to set an upper-bound on the market share of smaller players.

A.1.4. Regulation of retail banking

There has been growing emphasis on strengthening competition in the retail banking sector. National competition authorities (NCAs) are becoming increasingly active in the retail banking sector, launching competition investigations. These investigations have in certain countries helped to develop a broad policy agenda for public authorities and the banking industry to take forward. In addition, special exclusions for the banking sector from full antitrust law have been removed in most Member States. A second important change at Member State level is the shift in institutional competence for national merger approval in banking. Scrutiny has generally moved from regulators and central banks to competition authorities.[13]

Alongside the increasing scope of antitrust enforcement over the banking sector in Europe, regulators at Member State and EU level increasingly encourage self-regulation to deliver efficient market outcomes. In its White Paper, Financial services policy 2005-2010 , the European Commission emphasised the use of efficient alternatives to legislation, including greater use of competition enforcement and more reliance on self-regulation, where appropriate and proportionate.

The literature increasingly shows that competition can not be viewed as in any way inconsistent with banking sector stability.[14] Against a background of increasing stability and more effective supervision[15] in the banking sector, the emphasis of policy in the European Union and other advanced economies is quite rightly on increasing competition. The Commission encourages Member State authorities to ensure that competition policy applies fully to the banking sector and to ensure that the policy environment is favourable to tough competition.

There is an important tension in the regulatory framework for retail banking. On the one hand, governments will wish to see strong competition driving an efficient, innovative banking sector to deliver value for consumers and businesses. On the other hand, regulation of the banking sector – especially pro-competitive regulation – requires expert sectoral knowledge. This knowledge is likely to be deepest with the banking regulator. Therefore an institutional framework is required which creates incentives for the competition authority and banking regulators to cooperate and for banking regulators to promote competition.

The sector inquiry's Interim Report II noted that while the scope of direct state intervention in the retail banking sector has narrowed, governments continue to intervene in the banking sector through other means. Member States have an obligation to abstain from measures that have the potential to distort competition in the common market and deprive Articles 81 and 82 EC of their effects by, for example requiring or reinforcing anticompetitive behaviour or by delegating regulatory powers (Article 3(1)(g) EC, Article 10(2) EC applied in conjunction with and Articles 81 EC or 82 EC). Similarly, in the case of public or privileged companies, the Commission has to ensure that Member States do not enact or maintain in force state measures that reinforce or lead these companies to engage in anticompetitive behaviour (Article 86(1) applied in conjunction with Articles 81 EC or 82 EC).

The Commission scrutinises advantages provided to certain financial institutions by means of State aid control in order to ensure a level playing field for all market participants and to enhance undistorted competition (Articles 87 to 89 EC). In particular, the Commission ensures that public and private institutions operate under similar conditions by removing unlimited state guarantees or fiscal advantages favouring particular banks; and by applying the so-called market economy investor test (MEIP). One of the most significant distortions results from tax preferences conferred on a limited number of banks for distributing savings products. From a competition viewpoint the Commission is particularly concerned about discriminatory fiscal privileges which favour specific banks.[16]

Finally, compared to other sectors in the EU the incidence of cross-border mergers and acquisitions (M&A) in banking remains fairly low. Empirical evidence suggests that on balance foreign bank entry tends to enhance consumer welfare. A range of policy measures could help provide a more supportive environment for cross-border M&A, including more streamlined and effective banking supervision, and the removal of obstacles to corporate expansion and reorganisation on a pan-European basis.

Structure of Part A of the technical annex

The analysis of markets for current accounts and related services is structured as follows:

- chapter 2 examines market structures and the financial performance of the European retail banking sector.

- chapter 3 discusses the competition issues in EU retail banking arising from credit registers;

- chapter 4 discusses cooperation among banking associations and special groups of banks;

- chapter 5 examines the setting of banks’ prices and policies, particularly for current accounts; and

- chapter 6 examines customer mobility and choice in retail banking.

A.2. Market structures and financial performance in retail banking

As discussed in the opening chapter, the retail banking sector is large and important for the long-term performance of the European economy. Based on the evidence of the market survey, the Commission estimates that retail banking generates gross income of between 250-275 € billion[17] in 2004; equivalent to around 2% of EU GDP. Of this total, the vast majority of activity is captured under the heading of ‘current accounts and related services’, the subject of this part of the technical annex.

This chapter is structured as follows:

- Section 1 examines market concentration and integration in retail banking; and

- Section 2 surveys the financial performance of the retail banking sector.

A.2.1. Market concentration and integration in retail banking

The Commission has estimated concentration ratios of European retail banking at national level.[18] The estimates were based on data given by roughly 250 retail banks in the EU25. This sample approach – despite high coverage rates with respect to individual responses and countries – implied that the calculated individual intra-sample shares of certain volume indicators such as retail income had to be extrapolated with the help of external data in order not to overestimate individual 'market shares' (for details see Interim Report II, chapter 2 and 4). The possibility of over- or underestimations, however, could not be fully excluded due to data limitations.

The Commission decided to publish its estimates despite such limitations because they help add to the limited stock of data on retail banking concentration in the EU. Most concentration measures are based on total bank assets (not on retail activity indicators) and individual credit institutions (not on group consolidated data). In view of respective remarks in the public consultation, it has to be repeated, however, that the inquiry's concentration data must be interpreted with care. They are estimates that serve illustration purposes. They do not replace a proper antitrust or merger case analysis including the required definition of relevant markets, a complete coverage of these markets and subsequent calculation of market shares.

The Commission used several indicators – total retail income, income on current accounts and number of current accounts – to estimate market concentration and to cross-check results (for details see chapter 4 of Interim Report II). In general, taking all three measurements into account the average CR3 ratio across all EU25 countries (weighted by Member State population) is around 50%. The ratios of the New Member States are a bit higher with CR3 ratios of about 55-60%. The most concentrated countries typically include – independent of the measure – Belgium, the Netherlands, Finland, Lithuania and Sweden. The least concentrated ones also tend to be the same independent of the measure and include Italy, Spain, Latvia and, in particular, Germany. CR5 ratios are, of course, higher but do not change the concentration patterns and country orders with the exception of countries such as the UK and France which tend to have average CR3 and high CR5s ratios.

Using gross retail banking income, the following combined market shares for the leading three (CR3) and five (CR 5) retail banks in the various Member States were estimated[19]. Figure 1 presents the concentration estimates.

Figure 1:

[pic]

* Descending order based on CR3.

These estimates support the hypothesis that European retail banking markets in general are moderately concentrated at national level. They also confirm the perception that Belgium and the Netherlands on the one hand and the Nordic countries on the other, have more concentrated retail markets than the European average, in particular when taking into account the CR3 measures.

Some countries, most prominently Germany, but also others such as Spain and Italy, have comparatively low CR3 and CR5s. One explanation for this phenomenon is that large parts of the national retail activities in these countries are covered by a high number of savings and co-operative banks that are legally independent entities (for details see chapter A.4.). In Germany, for instance, almost 500 savings banks cover up to 50% of some of the national retail activities. Consequently, even the leading banks in Germany are significantly below a 10% share of the national retail market.

As highlighted in Interim Report II, European retail markets remain extremely fragmented. With the exception of the Benelux and the Nordic countries, there are very few players that have a leading market share in two or more Member States. In general, the number of non-domestic banks among the leading banks in the Member States is limited. Figure 2 shows the balance between domestic and foreign banks in the top five banks in each Member States, measured by gross total retail income. The results show firstly that foreign banks tend to have much stronger market positions in the New Member States than in the EU15. For example, the inquiry's data suggests that in France, Italy, Spain, Sweden and the UK the top five banks by market share are all domestic. By contrast, in the Czech Republic and Slovakia the top five banks are under foreign control, as are four of the top five in Hungary and Poland.

Figure 2: Domestic and foreign[20] banks in the top five[21] banks per country, 2004

[pic]

In general, market structures and, in particular, concentration ratios do not present a uniform picture. Consequently, a differentiated analysis is required before any wider conclusions can be drawn from these concentration estimates.

A.2.2. Financial performance of the retail banking sector

This section examines the financial performance of the banking sector, specifically:

- gross income from retail banking;

- profitability and cost-income ratios in retail banking;

- determinants of profitability in retail banking; and

- long-term trends in banking sector profitability.

Interim Report II set out detailed calculations for the Member States on several aspects of their banking sector performance: overall volume of activity; gross income by product line; profitability and cost-income ratios. Readers should note that the detailed methodology for these calculations and related caveats are contained in chapter 5 Interim Report II; they are not repeated here. The data relate to retail banking services to consumers and SMEs[22], with the former group generating around 80% of total gross income. This section summarises the main findings of this examination and, where appropriate, draws conclusions on the differing models of retail banking and market structures in the Member States.

Gross income from retail banking

Table 3 below summarises the proportion of gross income from consumers that banks generate by product line. Based on a weighted mean of all Member States mortgages are clearly the major source of income, generating just over 30% of gross retail income from consumers. However in the New Member States this share is significantly lower. Current accounts generated more than 25% of gross income in the EU25, and more than one-third of income for banks in the New Member States. Across the EU25, savings accounts and consumer loans both generated between 17% and 18% of banks’ gross income, respectively, while credit cards generated over 7% of banks’ gross income from consumers.

Table 3: Gross income share by consumer product line, weighted average, 2004

NMS Average | 34.95% | 23.45% | 20.85% | 15.62% | 5.13% | 100% |

EU-25 Average | 27.87% | 17.16% | 17.66% | 30.06% | 7.25% | 100% |

Note: Country-level estimates are gross income weighted averages across banks surveyed in the country.

The estimates for EU-15, New Member States and EU-25 are country-level averages weighted by population.

Source: Commission’s “Retail Banking Survey”, 2005-2006.

The inquiry has found wide national variations in banks’ income for specific product lines. Table 4 below summarises the gross income per customer data – aggregated for EU25, EU15 and the New Member States – for the consumer retail banking products in the Commission’s market survey. Comparisons across a range of retail products show that banks’ income per customer is typically twice as high in the EU15 as it is in the new Member States.

Table 4: Gross income per consumer (€) by product line, weighted average, 2004

Current accounts | Deposits and savings | Consumer loans | Mortgages | Credit cards |

EU-15 Average | 133 | 69 | 421 | 1,126 | 64 |

NMS Average | 48 | 41 | 88 | 442 | 66 |

EU-25 Average | 119 | 64 | 367 | 1,015 | 65 |

Note: Country-level estimates are weighted averages by market share across banks surveyed in the country.

The estimates for EU-15, New Member States and EU-25 are country-level averages weighted by population.

Source: Commission’s “Retail Banking Survey”, 2005-2006.

Profitability and cost-income ratios in retail banking

This section summarises the inquiry’s overall findings on profitability and costs in retail banking across the EU25. Using market survey data, Figure 5 presents the country level weighted averages of the ratio of pre-tax profit to gross income (for all retail banking activity) from 2002 to 2004.[23]

Figure 5: Profitability ratio, 2002-2004, weighted average

(Profit before tax as a share of total retail income)

[pic]

Source: Commission’s “Retail Banking Survey”, 2005-2006

Most Member States show average pre-tax profitability close to the weighted EU average (20 to 30%). We can identify two groups of countries according to their sustained pre-tax profit record during the period. A first group of five Member States (Ireland, Spain and the three Nordic countries) reported sustained pre-tax profitability ratios of about 40% and are always above the EU average. A second group of Member States (containing Germany, Austria, and Belgium) reported low profitability throughout. The remaining Member States report profit ratios around the EU average during 2002-2004. It is notable that during the observation period, the volatility of profits is higher where profitability is lower.

Based on operating costs as a share of total retail income, the inquiry found a wide dispersion across Member States in banks’ cost bases. On average banks’ operating costs in 2004 accounted for 63% of total retail income. Banks in Spain and Ireland had the lowest cost ratios (45-50% on average), while banks in Germany, Austria and the Netherlands had the highest ratios (75-80% on average).[24] This pattern is unsurprising since a low cost-income ratio is largely the corollary of a high pre-tax profitability ratio, and vice versa.

Determinants of profitability in retail banking

The sector inquiry data presented above relating to profitability of retail banking and banks’ cost income ratios require further qualification. In their responses to the public consultation many stakeholders highlighted a range of influences on banks’ profitability, over the short- and long-term. These influences on retail banking profitability are discussed below.

In terms of macroeconomic influences, firstly the economic cycle clearly influences profitability, particularly over the short observation period of 2002-2004. Member States which experienced fairly rapid economic growth during the period should experience rising demand for banking services (particularly credit products) and a fairly low share of non-performing loans (which tend to materialise when the economy slows). By contrast Member States where growth was slow would generally find lower demand, constrained margins and a higher share of bad debts.[25] For example, Ireland recorded the fastest economic growth in the EU15 between 1998 and 2005 and also displayed the highest bank profitability in the Commission's market survey of 2002-2004. Economic growth in Italy and Germany was among the slowest in the EU15 during this period, and bank profitability was also low.

Secondly, the overall savings ratio and development of credit markets are additional macroeconomic determinants of banking profitability. Simply put, lending is more profitable for banks than taking deposits. Thus markets where demand for credit is strong – whether through mortgages, loans or credit cards – would be expected to provide greater opportunity for banks to generate profits.[26]

Three industry-specific factors also influence the observed level of profitability and costs in retail banking. Firstly, the level of banking sector competition will influence profitability. Factors including entry barriers, high market concentration, high switching costs and the scale advantages deriving from branch networks may all serve to weaken competition, and large banks in some Member States may be able (separately or collectively) to exercise market power in the setting of prices and margins. Thus notwithstanding macroeconomic influences on banking profitability, high and sustained profits may signify the exercise of significant market power. For example, the conjunction of high market concentration and high profitability observed in some Member States might raise concerns about the extent of banks' market power; particularly where there is additional evidence of obstacles to market entry and fair competition.[27] Conversely, sustained low or negative profitability in banking may also indicate weak competition, suggesting distortions in the process of market exit and entry. Such a pattern may be observed in Member States where there are regulatory obstacles to cross-border entry or historically high levels of public ownership or State aid have distorted market performance.[28]

Secondly, distribution models differ across Member States. Models based heavily on a branch network rather than remote channels such as internet and telephone will typically have a higher cost base. Moreover, a greater reliance IT tools rather than staff in areas such as credit scoring and internal risk management will tend to reduce costs further. Thirdly, the corporate and capital structures of retail banks still vary widely across the EU. Large shares of retail banking activity in Member States such as Germany, France, Italy and Spain are undertaken by cooperative banks and savings banks.[29] Many of these banks have explicit social objectives, including widening access to finance for consumers and SMEs. These objectives, coupled with a cooperative ownership structure, are likely to create a different set of profit incentives to shareholder-owned private banks.

Long-term trends in banking sector profitability

Since the coverage of banks’ information systems limited the time horizon of the market survey to a handful of years, the Commission examined the long-term of banking sector profitability using data collected by the OECD. As described in Interim Report II, the methodology for the OECD’s figures[30] covering 1981-2003 is consistent with the method used above to analyse the 2002-2004 period. However, there are three important differences in the scope of the data. Firstly, since six EU Member States[31] are not member countries of the OECD, there is no long-term data on their banking sector. Secondly, the OECD’s own data describe all banking activity by country, of which retail banking comprises over 50 per cent of the total (measured by gross income). Thirdly, the OECD’s data describe the global performance of banks according to their nationality, whereas the Commission’s data describes their performance in one specific Member State. Despite these differences in scope, a comparison of the Commission and OECD profitability estimates for 2002 and 2003 shows a marked convergence. This is unsurprising since the majority of banks have significant operations only in one EU Member State and retail banking activity forms the bulk of the OECD’s dataset.

On this basis, the Commission concludes that the long-term trend of profitability is upwards in the EU banking sector as whole. Based on operating profits as a share of gross income from all banking activity , banks in almost every Member State have become more profitable since the 1980s. From the clear overall trend of rising pre-tax profitability, it can also be inferred that retail banking profitability has risen over the long-term.[32] By the end of the period, average profitability in most Member States was 20 to 40 per cent of gross banking income. Moreover, the profitability ratios for each Member State were consistent with those observed in the Commission’s market survey, which focused specifically on retail banking.

A.3. Credit registers and banking competition

Credit registers operate in almost all Member States and collect various kinds of financial information on individuals. The data stored on credit registers is provided by banks and other firms providing credit. Members of the registers are able to access this data for commercial purposes such as bank lending, subject to data protection rules. Banks and credit providers require access to good quality credit data in order to overcome information asymmetry when they set prices for new or potential borrowers. Thus credit registers are an important element of retail banking market infrastructure. To ensure strong competition among credit providers in retail banking markets it is vital that credit registers enable open and non-discriminatory access to credit data.

This chapter surveys the operation of credit registers in the EU and highlights potential competition issues. The information presented here is based on the literature on credit data markets and on the Commission’s extensive market survey of credit registers throughout the EU, conducted in the course of the sector inquiry.

This chapter is structured as follows:

- section 1 outlines the economics of credit data markets and credit registers;

- section 2 surveys the markets for credit data in the Member States;

- section 3 examines potential competition issues arising from credit registers; and

- section 4 concludes.

A.3.1. Economics of credit data markets and credit registers

The credit information industry displays strong network effects. The initial high fixed cost structure and near-zero marginal costs create strong forces for concentration in the industry.[33] Moreover, as the market coverage of a credit register increases, so does the register’s value to its clients and potential clients.

Common problems in markets with network effects are standardisation problems, path dependency, switching costs and lock-in to one network or technology. Many of these features are observable in credit reporting competition and they contribute to market concentration. Monopoly and oligopoly are common market forms in credit data markets; particularly where oligopoly is accompanied by very small niche players.[34] While typically one or a small number of large players satisfy the demand from banks and retailers in a particular Member State, small niche players specialize on specific areas such as tenant reporting.

Credit data sharing tends to have positive economic effects. First, credit data reduces the information asymmetry between a bank and its potential customer, which is likely to result in lower default rates (as does the sharing of positive and negative information compared to negative information only).[35] Second, credit information sharing acts as a borrower discipline device: borrowers know that if they default, this fact becomes public knowledge and their reputation with other lenders is affected. This could make it more expensive or even impossible for the customer to obtain credit. Third, credit reporting helps reduce problems of adverse selection[36], generally ensuring greater credit availability on better conditions. Finally, data sharing is also linked to consumer mobility: it has been argued that banking markets where databases are more active show more consumer mobility.[37] In summary, banks and other providers of credit require access to good quality credit information in order to price accurately for borrowers, and a greater availability of credit data tends to improve banking market performance.[38]

The European Court of Justice (ECJ) recently issued a judgement on the compatibility of credit registers with banking competition, following a request for a preliminary ruling from the Spanish competition court (the Tribunal Supremo). On the particular case, Asnef-Equifax v Ausbanc,[39] the ECJ judgement stated that Article 81 (1) EC must be interpreted as meaning that a credit reporting system " does not, in principle, have as its effect the restriction of competition within the meaning of that provision. This holds provided that the relevant market or markets are not highly concentrated, that the system does not permit lenders to be identified and that the conditions of access and use by financial institutions are not discriminatory, in law or in fact ."

A.3.2. Markets for credit information in the Member States

This section reviews the structure and regulatory framework for credit information markets in the EU. The section discusses:

- public and private credit registers in the Member States

- cross-border information sharing through credit registers; and

- the legal framework for credit information sharing in the EU.

Public and private credit registers in the Member States

There are three types of credit reporting systems evident in the Member States: (i) dual systems (combining public and private credit registers); (ii) private systems; and (iii) public systems. The breakdown by Member State shows:

- Dual systems in ten countries: Austria, Czech Republic, Germany, Italy, Latvia, Lithuania, Portugal, Slovakia, Slovenia, Spain;

- Private systems in twelve countries: Denmark, Estonia, Finland, Greece, Cyprus,[40] Hungary, Ireland, Malta, Netherlands, Poland, Sweden, UK; and

- Public systems in two countries: Belgium and France.

Table 6 presents an overview of the system in each Member State and the type of data held by each register.

Table 6: Credit reporting systems in Europe

Public credit register (PCR) | Private credit register (PRCR) |

Name of PCR | Data on Individuals? | Reporting Threshold* (Euro) | Information shared** | Data on Individuals? | Information shared** |

Austria | Grosskreditevidenz | Yes | 350.000 | + | Yes | +/- |

Belgium | Centrale des Crédits aux Particuliers | Yes | 200 | +/- | No | n/a |

Cyprus | None | n/a | n/a | n/a | Yes | - |

Czech Republic | Central Register of Credits | No | n/a | n/a | Yes | +/- |

Denmark | None | n/a | n/a | n/a | Yes | - |

Estonia | None | n/a | n/a | n/a | Yes | - |

Finland | None | n/a | n/a | n/a | Yes | - |

France | FICP | Yes | Unknown | - | No | n/a |

Germany | Evidenzzentrale für Millionenkredite | Yes | 1.500.000 | +/- | Yes | +/- |

Greece | None | n/a | n/a | n/a | Yes | - |

Hungary | None | n/a | n/a | n/a | Yes | +/- |

Ireland | None | n/a | n/a | n/a | Yes | +/- |

Italy | Centrale dei Rischi | Yes | 75.000 | +/- | Yes | +/- |

Latvia | Register of Debtors | Yes | 150 | - | Yes | +/- |

Lithuania | Loan Risk Database | Yes | 14.500 | +/- | Yes | - |

Luxembourg | None | n/a | n/a | n/a | No | n/a |

Malta | None | n/a | n/a | n/a | Yes | - |

Netherlands | None | n/a | n/a | n/a | Yes | +/- |

Poland | None | n/a | n/a | n/a | Yes | +/- |

Portugal | Servicio de Centralização de Riscos de Credito | Yes | 50 | +/- | Yes | +/- |

Slovakia | Register of Bank Loans and Guarantees | No | n/a | n/a | Yes | Unknown |

Slovenia | Credit Register | Yes | 0 | +/- | Yes | Unknown |

Spain | Central de Información de Riesgos | Yes | 6.000 | +/- | Yes | +/- |

Sweden | None | n/a | n/a | n/a | Yes | +/- |

United Kingdom | None | n/a | n/a | n/a | Yes | +/- |

Note: PCR is a public credit register and PRCR is a private credit register. FICP is the Fichier National des Incidents de Remboursement des Crédits aux Particuliers . The names of private credit registers in each Member State are not provided. The signs '+' and '-' denotes positive and negative information.

* The reporting threshold is the value above which credits must be declared to the register or debt amount outstanding threshold above which a borrower must be reported.

** Consistent with the definitions used in Interim Report II, 'negative' information describes defaults (e.g. late payments, arrears and bankruptcies). 'Positive' information describes total amounts and types of loans, accounts currently open and active, balances and credit limits.

There are several reasons for this diversity of credit reporting systems across the EU. Firstly, it may reflect the emphasis by policymakers on different objectives in different Member States, such as monitoring financial stability (e.g. the German public credit register Evidenzzentrale) or reducing over-indebtedness (e.g. the Belgian PCR). Secondly, attitudes differ concerning the use of personal data for commercial purposes. Countries have adopted different modes of information exchange. Thirdly, the structure of the credit reporting system will depend on – and also in turn, influence – the development of credit markets. It is notable that credit reporting systems tend to be less advanced in the New Member States, where retail banking markets are still maturing. Fourthly, as discussed below, the legal and regulatory frameworks for data protection, banking secrecy and credit data sharing still differ considerably across the Member States.

Incentive structures embodied in credit registers

Credit registers can be publicly or privately run. The Commission’s market survey covered nine public credit registers, all of which are run on a non-profit basis. Of the 20 private credit registers surveyed, 16 are run on a for-profit basis. Table 7 outlines the corporate status of the credit registers in the Commission’s market survey.

Table 7: For-profit versus non-profit registers

Responses (total) | 29 |

- for-profit information sharing | 16 |

- not-for-profit information sharing | 13 |

- of which: central banks | 9 |

As is argued below, the corporate structure of a credit register can create incentives towards full or partial data sharing, and thus affect the intensity of competition in downstream retail banking markets.

Public credit registers are owned by national central banks and run on a non-profit basis. Reporting to these institutions is mandated by law for specific items of data. In view of the governance of these registers, there is little risk that they are used directly as an instrument to distort banking market competition. However, this does not always imply that foreign banks will have full access to these institutions on a cross-border basis (which is discussed further below).

Private credit registers are primarily run on a for-profit basis, although their ownership structures vary. Approximately half of the PRCRs surveyed are run by specialist credit information providers and generally owned by a holding company.[41] For-profit institutions have a reputation to lose if they make partial decisions in favour of big market players. In these companies, banks are unlikely to have systematic influence on the policy of the credit bureau. Meanwhile, banks and other financial institutions might have more leverage in non-profit associations, foundations or other entities to influence access policy than they have in independent for-profit institutions. In non-profit environments, the opportunity and reputation costs of not accepting a (foreign) member are lower than in for-profit environments.[42]

Cross-border information sharing through credit registers

Credit information markets remain fragmented along national lines. Only a few credit bureaus conduct cross-border reporting, albeit for low volumes of data. The main reason for this low level of cross-border data sharing is the lack of demand (and to some extent, supply) for cross-border lending to retail customers. However there also are regulatory barriers in some Member States, which further limit the development of cross-border data sharing. The lack of cross-border credit reporting may create problems for consumers who are mobile and seek to borrow in more than one Member State. Currently they are generally required to build up separate credit histories in order to obtain credit, whether in the form of a mortgage, personal loan or credit card. Moreover, consumers who move permanently to another Member State might have problems accessing credit and telecom services, because their credit cannot be transferred to their new residence state.

One market-led initiative to extend cross-border data sharing is being sponsored by the Association of Consumer Credit Information Suppliers (ACCIS), an international association for credit information providers. To help its members to provide pan-European credit information services to their clients, ACCIS has developed a model contract for cross-border data exchange. The model contract embodies the key principle of reciprocal data sharing and permits individual credit registers to decide whether to engage in bilateral cross-border data exchange. Public authorities in the EU are also working to expand the scope of cross-border data sharing. In 2003, seven Member States signed a memorandum of understanding (MoU) on cross-border data exchange among credit providers, through the network of public credit registers.[43] The MoU covers data relating to lending where the value of the credit exceeds €25 000 (a fairly high threshold for the purposes of retail banking providers). The MoU came into force in May 2005 in the seven signatory countries: Austria, Belgium, France, Germany, Italy, Portugal and Spain. However, the most recent information suggests that Austria and France have still to fully implement the terms of the MoU.

Legal framework for credit information sharing in the EU

Credit information sharing is based upon national laws and regulations governing data protection, banking and credit reporting; and the European data protection directive. The legal analysis of these laws shows that there is still divergence between the individual countries. Among the strictest regimes of credit reporting regulations are France, Germany and Ireland, on the lower end are Belgium and the Czech Republic.[44]

Table 8: Laws applying to information exchange and data protection

Country | Law |

Austria | Bundesgesetz über den Schutz personenbezogener Daten |

Belgium | Law of 8 December 1992 on Privacy Protection in Relation to the Processing of Personal Data |

Cyprus | Processing of Personal Data (Protection of the Person) Law |

Czech Rep. | Act on the Protection of Personal Data |

Denmark | Act on Processing of Personal Data |

Estonia | Personal Data Protection Act |

Finland | Personal Data Act |

France | Loi N° 78-17 du 6 Janvier 1978 relative à l'informatique, aux fichiers et aux libertés |

Germany | Federal Data Protection Act |

Greece | Law on Protection of individuals with regard to the processing of Personal Data |

Hungary | Protection of Personal Data and Disclosure of Data of Public Interest |

Ireland | Data Protection Act |

Italy | Protection of Individuals and other subjects with regard to the processing of personal Data |

Latvia | Law on Personal Data Protection |

Lithuania | Law on Legal Protection of Personal Data of 1996 |

Luxembourg | Protection des Personnes a l’egard du traitement des donnees a caractere personnel |

Malta | Data Protection Act |

Netherlands | Personal Data Protection Act |

Poland | Law on the Protection of Personal Data |

Portugal | Act on the Protection of Personal Data |

Slovakia | Act on Personal Data Protection |

Slovenia | Law on Personal Data Protection (1990), Personal Data Protection Act (1999) |

Spain | Organic Law 15/1999 of 13 December on the Protection of Personal Data |

Sweden | Personal Data Act |

UK | Data Protection Act |

Source: Jentzsch (2006)

European data protection laws typically cover the same core provisions: they attribute specific rights to the individual (access, correction, etc.), the supervisory authority (oversight, audits) and the credit reporting industry (notification, accuracy, registration). Moreover, many implementing regulations contain retention time periods for specific data items (such as bankruptcies) and guarantee extra protection for sensitive information such as race, political or philosophical belief, and religion as well as health data. Many countries do have clauses in their data protection acts that state that complete information must be shared (Table 9).

Table 9: Requirement of Completeness

Country | Complete info requested in law? | Clause |

Austria | right DS | Bundesgesetz über den Schutz personenbezogener Daten, § 27 |

Belgium | yes | Law of 8 December 1992 on Privacy Protection in Relation to the Processing of Personal Data, Art. 4 |

Czech Republic | n/a |

Denmark | yes, indirect | Act on Processing of Personal Data, Part 4 (24) |

Finland | yes | Personal Data Act, Sect. 9 (2) |

France | right DS | Loi N° 78-17 du 6 Janvier 1978 relative à l'informatique, au1 fichiers et au1 libertés, Sect. 36 |

Germany | indirect | Federal Data Protection Act |

Greece | indirect | Law on Protection of individuals with regard to the processing of Personal Data |

Ireland | yes | Data Protection Act, 3 Sect 2 |

Italy | right DS | Protection of Individuals and other Subjects with regard to the Processing of Personal Data, Art. 13 (3) |

Latvia | yes | Law on Personal Data Protection, Sect. 10(4) |

Lithuania | yes | Law on Legal Protection of Personal Data, Art. 3 |

Netherlands | right DS | Personal Data Protection Act, Art 36 |

Poland | right DS | Law on the Protection of Personal Data, Art. 32 |

Portugal | yes | Act on the Protection of Personal Data, Sect. 1 (d) |

Slovakia | indirect | Act on Personal Data Protection of 2002, Art. 12 |

Slovenia | right DS | Personal Data Protection Act, Art. 19 (1.) |

Spain | yes | Organic Law 15/1999 on the Protection of Personal Data, Title II, Art. 4 |

Sweden | yes | Personal Data Act, Sect. 9 |

UK | indirect | Data Protection Act |

Notes: No ‘complete information’ clause was found in the German, Greek and UK laws. “Right DS” refers to the right of the data subject to correct information if it is inaccurate or incomplete. “Indirect” means that there are statements that indirectly refer to complete information such as “inaccurate” or “misleading” – incomplete information can be misleading in terms of the credit risk it reflects.

International information sharing is regulated by the European Data Protection Directive (Directive 95/46/EC). Since all Member States have implemented this directive, data protection rules should not act as an obstacle to cross-border information flows.

A.3.3. Potential competition issues

Competition barriers in credit registers can arise from several sources, all of which can restrict access to credit data and thereby weaken competition in credit markets. This section considers three sets of competition issues in relation to credit registers:

- unfair or discriminatory access conditions;

- partial data sharing; and

- regulatory barriers.

It is shown below that the operation of some credit registers or the regulatory framework for credit data markets may weaken banking competition in several ways, by:

- disadvantaging or excluding foreign banks;

- disadvantaging non-shareholders;

- disadvantaging smaller players; and

- disadvantaging or excluding some types of non-bank credit provider.

The most basic test which should be applied to assess whether a credit register’s operation is consistent with competition rules is whether the register provides non-discriminatory access to credit data to all relevant credit providers. This provision is established in Article 8 of the Proposal for a Directive on Consumer Credit.[45]

Unfair or discriminatory access conditions

However these access conditions may also have a basis in law, where credit reporting is mandatory for banks (and other credit providers) and the data can be accessed through a public credit register. These access conditions comprise (i) membership criteria for the credit register; and (ii) the fee structure for membership and use of the credit register.

Membership criteria for the credit register

Public and private credit registers stipulate a range of criteria which their members or clients must meet in order to gain access to the register. These criteria vary across public and private credit registers and across Member States. The majority of credit registers responding confirmed that in order to access the database, the interested entity must meet some or all of the following criteria:

- undertaking credit granting activity;

- holding a banking license;

- having a physical presence in the Member State;

- compliance with reciprocity agreements; and

- compliance with data protection laws.

Most credit registers required members to be credit providers[46], to hold a bank license or be defined by law as ‘reporting institution’. In a minority of replies it was mentioned that the reporting institution needs a physical presence in the country. The vast majority of registers replied that members must adhere to reciprocity principles as well as to data protection principles. There was no direct evidence that foreign banks established in the home country of the credit register were subject to any per se discrimination from membership criteria.[47] From the perspective of retail banking market performance, two types of membership criteria highlighted above might be seen as restricting competition. Firstly, the requirement for a banking license in order to access the credit register could have the effect of excluding non-bank credit providers such as credit card companies and finance companies. Such requirements might also exclude companies that provide services involving the credit risk of deferred payments, such as retailers, mail order companies or telecom companies. Access of these institutions to central databases must be judged in the light of data protection.[48]

Secondly, requiring a physical presence in a particular Member State in order to access the credit register is likely to weaken banking market competition. Providers in another Member State wishing to supply credit cross-border will be unable to obtain reliable data on their prospective client, increasing lending risk and increasing prices. Such a restriction would appear to apply whether the account supplied cross-border was located in the home country of the provider or of the borrower (e.g. if supplied remotely by an internet bank). Explicit national presence required to access the register was stated by the Austrian Central Bank Register, the credit registers of the Bank of Spain, Bank of Portugal and the Bank of Latvia.

Fee structures

The vast majority of credit registers in Europe, whether public or private, charge fees to their members. However, fee structures and levels vary greatly across the Member States. The sector inquiry has already examined specific instances where fee setting in payment systems can foreclose entry or restrict competition in retail banking and payment card markets.[49] Similar risks arise in relation to fee setting for access to credit registers. However it should be noted that because of the lower scale of economic activity conducted through credit registers, operational costs and fees tend to be lower in absolute terms than in payment systems.

The inquiry has examined the fee structures and fee levels of public and private credit registers in the Member States, based on responses to the market survey. Fee structures for credit registers can comprise: one-off joining fees; ongoing membership fees; and per-transaction fees for consulting the credit register. The inquiry has found that three aspects of the fee structure for credit registers can weaken competition in retail banking markets. These aspects are:

- high joining fees;

- discriminatory volume-based transaction fees; and

- high fixed transaction fees for access to the register.

The level of joining fees for credit registers varies first according to whether the register is public or private. All of the public credit registers responding to the Commission’s market survey reported a zero joining fee. Of the twenty private credit registers sampled, twelve reported zero joining fees and one private register reported a nominal joining fee of €90. The remaining seven private registers reported fees in excess of €1000. Among this latter group there is a surprising degree of divergence. Three of the four highest joining fees are observed among credit registers owned and managed by banks, which charge joining fees of €75 000, €25 000 and €13 500.[50] The relatively high level of joining fees charged by these private credit registers raises two potential competition concerns. Firstly, such fees may be used as a tool to extract rent from new entrant banks requiring access to credit data; or secondly, high joining fees may be set with the intention of discouraging membership of the register and reducing the contestability of clients of the incumbent banks.

Transaction fees for access to credit registers vary according to the type of data extracted, with positive and full data generally having higher prices than negative data. Fees for access to most credit registers in Europe are generally in the range of zero to €1 per consultation, whether positive or negative data is requested. However, some private credit registers charge per transaction fees significantly above these levels. In such cases, fees for consulting a credit register may weaken banking competition.

Three private credit registers apply relatively high transaction fees, and two of these registers operate in the same Member State. The average cost of obtaining positive and negative data on clients in this Member State is among the highest reported in the market survey, at around €2 per credit report inquiry. Moreover, the cost of consulting the register varies significantly according to the volume of usage. One private credit register in this Member State (which is jointly owned by a holding company and a group of banks) reported that the average per transaction fee varied from 0,46 € for its largest clients to €10.95 for its smallest clients, according to a volume-based price schedule. The latter level of transaction fee might discourage smaller players from making full use of the credit register.

A credit register in one Member State charges a fee of €2.24 to €2.44 per consultation, and also requires the highest joining fee reported in the market survey. Thus it is unlikely that the high joining fee is not used to cross-subsidise low access fees for members of this register. Moreover, this credit register is owned and operated by domestic banks, which are able to set per transaction fees for access to client data; data which principally relates to the client base of the incumbent banks. The banks operating this credit register might have an incentive to set high fees of consulting the register, in order to raise the costs of competitors and discourage them from making full use of the information held in the register.

Partial data sharing

As discussed in Section 1 of this chapter, greater availability of credit data will tend to reduce information asymmetry between lenders and clients and reduce the informational advantage of large banks over new entrants. The evidence suggests that overall, consistent with data protection and competition rules, greater credit data availability strengthens banking market competition. However the structure of some credit data markets and the operation of some credit registers may significantly limit the availability of credit data, thereby reducing the ability of banks to compete for rivals’ customers.

Two forms of partial data sharing (or composition effects) may potentially inhibit banking market competition:

- low market coverage rates of credit registers; and

- incomplete reporting by credit register members.

Low market coverage rates of credit registers

The market coverage of a credit register describes the share of all (retail) borrowers in a particular Member State on which the register has data. The share of market coverage will depend on four factors: (i) the share of credit institutions in the market that provide credit reports; (ii) whether reporting institutions provide positive, negative or full data;[51] (iii) the reporting standards used, since these still vary from country to country[52]; and (iv) the reporting threshold applied, i.e. the de minimis value of credits reported to the register.

Using data gathered in the market survey, the Commission has estimated the market coverage of credit registers across the EU. The numerator of this ratio – the number of individual borrowers on whom data is held – is based on data provided by credit registers. The denominator of the ratio – the total number of retail banking clients in a Member State – is estimated using the total population of a country above fifteen years of age as a proxy for economically active population.[53] The estimate coverage rates of various credit registers are shown below in Figure 10 on an anonymised basis.

Figure 10: Coverage rates of various credit registers in the EU

[pic]

The data show a striking divergence in the level of coverage of credit registers across the EU. Twelve of the credit registers described below have coverage rates of five per cent or less. Of these twelve registers, six[54] are public registers to which credit reporting is compulsory for credit institutions in a given Member State. Therefore the very low coverage of these PCRs results from their limitation to negative-only data and/or use of high reporting thresholds (where information can also be positive and negative). Of the six private registers with very low coverage rates, many contain only negative data.

The remaining thirteen credit registers shown above have estimated market coverage ratios between 35 and 80 per cent, with nine registers estimated above 50 per cent coverage. Of these registers, three[55] are PCRs which hold positive data. Of the remaining ten, there are at least two registers from a single Member State, pointing to competition among credit data providers and overlapping coverage of borrowers. The clear majority of the registers showing high coverage are in the EU15. Poland is the only New Member State for which credit data market coverage is estimated above 35 per cent. As discussed above, the extent of credit data market coverage is likely to influence the strength of competition in retail banking. Where the coverage of credit registers is low, larger banks are likely to have an advantage since their extensive client book will enable them to build more accurate risk models than smaller players and new entrants. This advantage is likely to be compounded where credit registers in a particular Member State hold only negative data, since available credit data is limited to a small, adversely selected pool of borrowers. Such a position in a Member State may be entirely consistent with data protection rules and the framework of contracts governing the operation of credit registers. Nonetheless it is likely to weaken retail banking competition.

Incomplete reporting by credit register members

Many non-bank institutions such as credit card providers share their information voluntarily. In many countries, positive and negative information is exchanged over credit bureaus. However, in some countries credit bureaus collect only negative information. Some of the systems are dual systems where the public credit register could potentially be a substitute for the private credit bureau, but only for the reporting institutions. This is the case if the register collects information on consumers starting from a relatively low threshold. Low thresholds exist in Belgium and France, Latvia, Portugal and Spain – again this is only possible for institutions that are obliged to report.[56]

The customer data which banks hold – particularly detailed positive data – are a significant asset and potentially a source of competitive advantage. This advantage may arise from better risk scoring models or more effective marketing and cross-selling. In an environment where information sharing is voluntary, banks may have incentives to disclose or withhold specific types of information. For instance, if banks disclose negative information and with it incomplete positive information[57], this can blur the true picture of a consumer’s credit risk, placing rival lenders at a competitive disadvantage. A recent study by the Federal Reserve Bank in the United States highlighted the extent of incomplete information reporting in the banking industry. The Federal Reserve study found that in 1999 around 70 percent of consumers sampled had missing credit limits on one or more of their revolving accounts. Action was taken by private and public sector to improve the accuracy and consistency of reporting. Nevertheless, a sample taken in June 2003 showed credit limits still missing from 14 percent of revolving accounts; and omissions affecting the records of 46 percent sampled consumers.[58]

Incomplete reporting by lenders should not arise in public credit registers where full disclosure is a legal obligation. In private credit registers, however, there may be conflicting incentives. Specialist private credit registers will naturally seek full disclosure from lenders, since more and richer information increases the value of the register to its clients. Lenders, however, face a strategic decision on the benefit of joining the register compared to the ‘cost’ of enabling rival banks to access valuable client information. The larger a lender’s market share, the less attractive this trade-off is likely to appear. Thus some private credit registers may accommodate larger banks by waiving the requirement for full disclosure of data. One specialist private credit register responding to the Commission’s market survey indicated that it was common practice among lenders not to report the balances owed by customers or their overdraft limit. Another specialist private credit register confirmed that over twenty of its clients chose to share only negative information; and that three lenders refused to share data relating to specific products (such as mortgage and personal loans).[59]

Where a private credit register is owned and managed by the main domestic banks, a different set of incentives arise. The credit register has no strong commercial incentive[60] to ensure full information disclosure by lenders. Meanwhile, the main banks may have incentives to withhold information and to free-ride on the full disclosure of their rivals. The Commission’s market survey has identified at least one credit register owned by incumbent banks where this conflict of interest results in incomplete information sharing. Monitoring by the register appears to be weak, with the management unable to estimate what proportion of the data held was positive, negative, or both. This situation also raises some concerns about the strength of data protection safeguards in this register.

Regulatory barriers

Regulatory barriers also raise competition issues. Entry into credit information markets can be explicitly foreclosed by specific regulations, or implicitly foreclosed by a particular interpretation of relevant laws. France is an example of the former case. France has established a public and centralised system of information sharing. All reporting institutions (banks), must contribute data on incidents to the register and only these institutions are able to access the register. Currently, there are no other credit registers active in the country. The interpretation of laws in the country by authorities such as the Commission Nationale de l'Informatique et des Libertés (CNIL) prevents sharing of positive information. This provides an advantage to incumbent banks, which are able to build accurate and efficient credit scoring models based on their existing client book. By contrast, new entrants are only able to access an adversely selected pool of borrowers for whom only negative information is available. This situation does not appear to constitute an infringement of competition law. However one effect is to maintain the information advantage of incumbent banks over smaller banks and new entrants.

In Spain the interpretation of existing laws has acted as a de facto entry barrier for some credit data providers. As discussed above, a recent ECJ ruling on the ASNEF-Equifax case appears to remove any legal obstacles to that player entering the Spanish credit data market. However the initial notification from ASNEF-Equifax of its intention to enter the market was made in 1999 and the company has been effectively foreclosed from the market since then.

A.3.4. Conclusions

The analysis has highlighted three sets of issues in relation to credit registers which can weaken competition in retail banking markets: unfair access conditions; partial data sharing; and regulatory barriers. These competition issues should be addressed through competition law and, where appropriate, through other measures.

Application of competition law

Credit data markets are generally fragmented along national lines and cross-border data sharing is limited. Where a credit register holds a dominant position in a Member State credit data market and sets access conditions that may be regarded as unfair or discriminatory (e.g. accepting only members that are banks established in that Member State), such conduct might constitute an abuse of dominance[61] under Article 82 of the EC Treaty. Alternatively, where a credit register is run by a consortium or joint undertaking of banks in a Member State, the application of unfair access conditions might constitute an infringement of Article 81 of the EC Treaty. Clearly, however, such legal assessments could only be made following investigation of a specific case and based on the full facts.

Full enforcement of data protection rules

The sector inquiry has identified concerns in relation to compliance with national data sharing rules. For example, there appear to be instances where some parties fail to disclose full information on their clients to a credit register and where the credit register does not exercise close scrutiny of the information provided by its members. The evidence suggests that such problems may arise in only a small number of Member States. In such cases, compliance with data protection rules should be investigated by the appropriate national regulators. Full compliance by credit registers with data protection rules is vital to protect the integrity of individuals’ data and to ensure their fair access to credit. Moreover, it will strengthen banking competition.

Assessment of non-discriminatory access to credit registers

The principle of non-discriminatory reciprocal access to credit registers is enshrined in the rules of most, if not all credit registers in Europe. The Proposal for a Directive on Consumer Credit seeks to embed this principle in European law, including for cross-border access to credit registers. However, the sector inquiry has highlighted that this principle is not yet fully operational. Firstly, there are still regulatory obstacles in some Member States to cross-border data sharing. Secondly, aspects of the fee structure of some credit registers may be regarded as discriminatory and disadvantageous to small players and new entrants. Thirdly, even at domestic level the principle of reciprocal data sharing is not always fully enforced by credit registers, and the risk of abuse appears to be highest where registers are owned by incumbent banks. Therefore a future examination of credit registers may be warranted to assess their compatibility with an open and competitive European credit market.

Review of the regulatory framework for credit data sharing in some Member States

Finally, practice in credit data sharing still varies considerably across the EU, particularly in relation to the treatment of positive and negative data and the threshold level of credits to be reported. The frameworks for data protection and credit data sharing are sensitive matters and require careful scrutiny by Member State governments. Authorities should note that credit data sharing regimes with high reporting thresholds or based on the exchange of only negative data are likely to favour large incumbents at the expense of smaller players and particularly new entrants. Therefore national authorities that are seeking to enhance competition in the banking sector and improve the efficiency of credit markets may wish to consider reforms to their regulatory framework for credit data sharing. Such reforms may prove particularly helpful in the New Member States, where the coverage of credit registers is generally low and credit markets are still maturing.

A.4. Cooperation among banks: Savings and co-operative banks

The banking industry in general and retail banking in particular is characterised by a high level of cooperation between market players. Cooperation frequently occurs with respect to the ownership and management of payments systems or credit registers as well as the development of codes of conducts and other forms of self regulation. Usually, cooperation takes place at the level of banking associations, but it can also go wider; for instance, when nationwide payment systems or credit registers require cooperation between specific associations.

A large part of the rather extensive co-operative activity in the banking sector can be explained by the industry’s standardisation and compatibility requirements. In order to handle non-cash payments, for instance, banks have to agree on issues such as technology formats or mutual cost compensation. From the first clearing houses in the 19th or even 18th century to the current preparations of the SEPA banks have, therefore, relied on a variety of networks for cooperation.

This chapter is structured as follows:

- Section 1 discusses the scope of the inquiry's analysis of cooperation;

- Section 2 discusses cooperation among cooperative banks;

- Section 3 discusses cooperation among savings banks;

- Section 4 examines the competition issues raised by cooperation among banks; and

- Section 5 concludes.

A.4.1. Scope of the inquiry's analysis of cooperation

Cooperation between competitors always bears the risk that co-ordination goes beyond what is strictly necessary to achieve economic benefits and affects the competitive behaviour of the co-operating parties, for instance, with respect to price setting or market presence. Effective market competition is in particular impeded if independent banks with a significant combined market position engage in anticompetitive behaviour that limits competition among the co-operating parties or hinders third parties (such as potential competitors) from market entry. Insofar as such issues arise in the context of payment systems and credit registers, they are dealt with in chapters B.10. and A.3. respectively.

Cooperation among banks may go beyond the joint operation of platforms and concern wider areas such as common guarantee schemes, joint marketing and business strategies or regional market sharing. This is particularly the case for savings and cooperative banks which, due to their origins, tend to have closer ties than other banks. These specific types of banks or ‘sub-sectors’ cover a significant proportion of the retail banking activities in Europe and play an important role in several Member States such as Germany, France, Italy, Spain or Austria.

Due to the variety of ownership patterns, company structures, cooperation areas and regulatory provisions within the Member States, fact finding, even with regulatory bodies or national competition authorities, is difficult and an overall competition assessment is impossible. However, on the basis of an additional fact finding exercise following publication of the interim report, the Commission has now a clearer picture including potential competition issues that may arise in these contexts.

This inquiry can only give an overview and not replace an in-depth competition analysis on a case-by-case basis. The Commission, however, will carry out such investigations following this inquiry in cases that indicate the existence of competition problems with a community interest. The same applies if regulatory aspects or other forms of state intervention are involved. Depending on the issues in question, national authorities may be well placed to act. In any event, the Commission and the competent authorities of the Member States will apply the Community competition rules in close cooperation to ensure an effective enforcement of Articles 81 and 82 of the Treaty.

A.4.2. Co-operative banks

Co-operative banks play a major role in France where the three co-operative banking groups (Crédit Agricole including Crédit Lyonnais, Banques Populaires and Crédit Mutuel) together account for roughly around 50% of the national retail sector, with, however, significant variations across the relevant product markets[62]. Considering that following restructuring the French savings banks were also transformed into a co-operative form and that they have partially merged with the Groupe Banque Populaire[63], the co-operative sector, including the Groupe Caisse d’Epargne (GCE), has an even stronger position on the French retail banking market.

Apart from France, co-operative banks have significant market positions in countries such as the Netherlands, Finland, Italy and Austria. In the Netherlands Rabobank has a leading share of almost 40% in some national retail markets such as private savings and lending to small and medium sized enterprises.[64] In Finland, OP Bank group has shares of over 30% in some retail markets and/or segments such as loans and deposits.[65] In Italy the Banche popolari and Credito cooperativo combined account for roughly 25% of the whole Italian banking sector.[66] Meanwhile in Austria, two associations, Raiffeisenbanken and Volksbanken, hold a roughly estimated combined share of around 20-25% of the whole banking sector and in relation to saving deposits.[67] In Germany, the roughly 1300 co-operative banks (mainly Volks– and Raiffeisenbanken) also play an important role, in particular regarding their local presence. Their combined share of the German retail banking market is difficult to assess but based on figures for the German banking market in general is estimated at about 10-15%.[68] Co-operative banks also exist in other Member States where their scope of activities and market presence, however, appears to be more limited.

Co-operative banks go back to the 19th century when they were founded across Europe, mostly in rural areas, by common people and small businesses. Co-operative banks originated from the idea of take borrowing and lending into the hands of the members and, thereby, providing financial services for the members at bearable costs. This original and principal purpose of co-operative banks remains to a large extent relevant. One usual feature of co-operative banks therefore is the customers’ ownership (membership): either all or a large part of a co-operative bank’s customers (natural and/or legal persons) constitute its owners and members.[69]

Another common feature is the “one man one vote principle”, independent of the number of shares an individual member holds. Consequently, it is impossible for individuals or other institutions to gain decisive influence over a co-operative bank. Some co-operative banks permit more than one vote, however, there is always a strict limitation on voting rights to preserve the basic co-operative principle and, thereby, prevent the acquisition of controlling rights. A takeover of co-operative banks or banking groups by other banks is, therefore, excluded. Some co-operative banking groups, on the other hand, are quite active in acquiring controlling stakes in other banks. For instance, the central institute of the Austrian co-operative Raiffeisenbanken, the Raiffeisen Zentralbank Österreich AG (RZB) – more precisely its fully consolidated subsidiary the Raiffeisen International Bank-Holding AG – has over the last 10-15 years acquired a significant number of banks in Central and Eastern Europe.[70] The French Crédit Agricole, to give another example, is not only the leading retail bank in France but also one of the leading European banks with subsidiaries in several countries and a targeted expansion strategy outside France.[71]

Whereas the described ownership and voting principles are similar and practically define co-operative banks, company structures differ significantly. Some co-operative banks have transformed completely into a consolidated group or developed ‘group-like’ structures that allow for consolidation according to national supervisory rules. Others remain legally and economically independent credit institutions, at least with respect to the local banks carrying out the retail business. In practice, however, hybrid structures are predominant. Notwithstanding other differences these structures follow similar patterns. For example, the local or regional retail banks typically own (or are members of) one or more central institutions. These central institutions, firstly, carry out infrastructure and assistance tasks such as providing payment systems and IT platforms, running the protection scheme (guarantee fund) and supporting individual member banks regarding business strategies, accounting, marketing and similar areas. Secondly, the central institutions often also function as a true group holding for subsidiaries that offer financial services other than retail banking (investment and corporate banking, leasing, insurances etc.) or retail banking in non-domestic countries (e.g. the new Member States). In some cases, however, the central bank also is the controlling body for the retail business of the local or regional banks. This is, for instance, given with full cross-guarantee schemes, as in the case of the Dutch Rabobank, where instructive powers by the central institution are required for the participants of the scheme to be accepted as a consolidated group from the viewpoint of prudential rules.[72]

Within the described principal structures, there is a sliding scale from the most centralised co-operative banks, such as Rabobank, to the most decentralised groups such as the co-operative banks in Germany. As said above, most company structures are, however, hybrids with varying degrees of centralisation. The French co-operative banks are, for instance, perceived as groups, also by the competition authorities, which for the purpose of merger evaluations regard the groups (not the individual local retail banks) as the decisive competitive entities for the supply of retail banking products[73]. This is despite an apparently varying degree of independence of the local banks within the different groups[74]. Other co-operative banks such as the Finish OP Bank group and the Austrian Raiffeisenbanken have structures somewhere in between with a strong consolidated sub-group. Even less centralised groups such as the German Volks- and Raiffeisenbanken have, though comparatively small, consolidated sub-groups for activities other than retail banking.

It is often difficult to judge whether in any individual case the domestic retail banking activities are controlled by the central institution, co-ordinated within the whole group or network or carried out independently by the individual banks with the central institution only providing auxiliary services. These distinctions are relevant, for instance, regarding the potential application of antitrust rules, but can only be analysed on a case by case basis.

There is rather limited regulation, and in most countries general banking and supervisory rules apply to co-operative banks. Specific provisions normally exist only with respect to the definition of co-operative banks; auditing; the obligations of the member owners regarding own funds and similar issues; or the one-man-one-vote principle[75]. In Austria, cooperation between co-operative banks are, at least to a certain extent exempted from antitrust rules.[76] Some co-operative banks in Europe support social, regional, environmental or other non-profit projects; however, on a voluntary basis.

In France the so-called regional or territorial principle is applied by the co-operative banking groups[77] effectively limiting the retail activities of the regional co-operative banks to a territory defined in their statutes. Some of these statutory definitions are based on the respective provisions in the French Code Monétaire et Financier in its sections on co-operative banks[78]. According to information submitted to the Commission following the interim report, German co-operative banks do not have legislative or statutory provisions that restrict the geographic area of activities. To what an extent a regional principle de facto exists – as an effect of the regional focus of co-operative banks or as a result of individual agreements and concerted practices – is difficult to judge for all Member States on the basis of this inquiry.

One aspect of state measures regarding (also) co-operative banks is currently under investigation by the Commission. It concerns the special rights given by France to three banking groups, among them Crédit Mutuel and Caisses d’Épargne, to distribute the ‘livret A’ and the ‘livret bleu’. These are tax-free savings products intended to increase savings and finance social housing. The Commission has extended a state aid procedure with respect to the question of overcompensation for the service provided to the State (by Crédit Mutuel) and opened an infringement procedure regarding the special rights as such and their effects on the freedom of establishment and the freedom to provide services.[79]

A.4.3. Savings banks

Savings banks play a major role in Germany and Spain where they together account for roughly 30% to 50% of the different national retail activities, with variations depending on the relevant product markets.[80] Furthermore, savings banks have significant market positions in Luxembourg (up to almost 50% in national personal banking according to own assessment)[81]; Austria (roughly 20% of the whole banking sector)[82]; Sweden (around 20-30 % in the different relevant retail markets)[83]; France (roughly 10 to over 20% with variations regarding relevant products)[84] and Italy. They are also present in several other Member States where they, however, play a less significant role. Finally, in countries such as Hungary, the savings banks developed into a normal commercial bank group with a corporate structure retaining only the historic name.

The origins of savings banks go back to the 19th, sometimes 18th, century when they were founded for social purposes, mainly to promote savings and guaranteed deposits for poorer people. Due to their historic background savings banks are all – at least to a certain extent - entrusted with a kind of social or welfare tasks, be it directly or by means of foundations. Social obligations are often mentioned in the relevant savings banks laws which, for instance, may determine that all surpluses that are not allocated to reserves or to shareholders have to be channelled to community or other social projects.[85] The use of the name 'savings bank' in countries with respective legislation, therefore, is often conditional upon the fulfilment of such social or public obligations that are described as typical for the savings banks in question. Clear definitions of these tasks as well as implementing rules are, however, often lacking. It is therefore not always clear whether these tasks or obligations are enforceable or enforced. Nonetheless, savings banks – like co-operative banks – are banks that offer the usual range of retail banking services in competition with other banks and have, therefore, to be regarded as undertakings from the perspective of competition law.

Despite common roots and in contrast to co-operative banks, savings banks do not have a common ownership pattern. They are still de facto or de jure publicly owned or controlled[86] in Member States such as Germany[87], Luxembourg and to a substantial extent, Spain[88]. Meanwhile they are wholly privatised in other countries (mainly in the new Member States); restructured into a co-operative group form (France); or consist of hybrid structures with private and publicly owned/controlled savings banks alongside (e.g. Austria, Italy).

The company structures of savings banks groups also vary. However, with the exception of the Banque et Caisse d’Epargne de l’Etat Luxembourg, truly consolidated and centrally controlled groups hardly seem to exist[89]. Also other savings banks have centralised institutions that provide payment infrastructures and other platforms or ancillary services. These central institutions sometimes also function as a holding for subsidiaries that offer non-retail banking products. However, normally savings banks state their independence, at least with respect to the local retail business. On the other hand, they often co-operate in a variety of fields such as product development, marketing, business strategies and other projects.

The Austrian savings banks are worth mentioning because they constitute a special, rather complex case. The Austrian savings banks have – via their central institution that also holds shares in some but few of the individual savings banks – very close co-operative ties, including a common guarantee scheme, information exchange and common business and marketing policy. According to Austrian law (Bankwesengesetz), the savings banks are defined as a ‘credit institution group’ that, among other things, allows for the consolidation of capital and the application of national merger rules (concentration privilege) to certain contractual agreements under the national competition law. Furthermore, since 2002, cooperation between members of a credit institution group according to the Austrian Banking Act - as ErsteBank/savings banks - is exempted from Austrian antitrust rules[90]. However, the Austrian Competition Court recently decided that Article 81 EC Treaty is applicable to the Austrian savings banks, thereby not defining them as a group under European competition rules. [91]

This example not only shows the complexity of classifying company structures for competition purposes, it also indicates that for savings banks in some Member States specific regulation exists. This applies in particular to publicly owned, controlled or otherwise state influenced banks. In Germany, for instance, the name ‘Sparkasse’ according to national banking law (section 40, Kreditwesengesetz) is, with few exceptions, reserved to banks based on public law. In principle, therefore, the German banking law prevents the sale of savings banks with the name 'Sparkasse' to private acquirers. However, in the context of a particular case, an agreement recently reached between the Commission and Germany allows for privatisation of a German savings bank and the continuation of the name 'Sparkasse' by the potentially privatised bank.[92]

The application of the regional principle by German savings banks – formally abolished in other Member States such as Austria and Spain – is based on legislation, i.e. the savings banks laws of the Länder. According to the information of the Commission, not all of the respective laws contain an explicit ban on business extension into another territory. However, the regional principle always seems to imply at least a strong focus on a defined territory. In France, as described above in the context of co-operative banks, the regional principle or 'territorialité' is applied and governed by the statutes of the savings banks. Whether the savings banks in other Member States de facto apply the regional principle or respect 'territorialité, is not clear.

A.4.4. Competition issues raised by cooperation among banks

This section studies competition issues arising from cooperation among banks, specifically:

- entry barriers;

- the role of regulation and State intervention;

- potential competition restrictions by company behaviour; and

- possible benefits of cooperation and impediments of effective competition.

Entry barriers

In Interim Report II the Commission described the high degree of fragmentation of retail banking markets in areas such as market structures, identity of main players, concentration and pricing. Market entry is difficult for several reasons, such as the necessity of local branch networks, the immobility of customers or difficulties to access to platforms such as payment systems or credit registers.

According to market participants, entry is economically viable only at a certain minimum scale (translating, for instance, into a certain minimum share per relevant market) and – due to the importance of personal customer-bank relationships – on the basis of established local branch networks. These market characteristics tend to render green field investments difficult and to favour entry by means of acquiring retail banks that are established nationwide or in an attractive regional market (e.g. larger cities).

Consequently, it appears particularly difficult to enter a market where a large share is held by banks or banking groups that cannot be acquired because of ownership structures (e.g. co-operative banks and some savings banks) and/or regulatory restrictions (e.g. some savings banks). For instance, France – where co-operative banks including the savings banks account for roughly 60 to over 70% of the different retail product markets – is generally viewed as a difficult market for foreign banks to enter.

Co-operative banks are practically immune against takeovers because of the ‘one man one vote’ rule or similar vote limitations. In Italy where, for instance, also co-operative bank groups with considerable size that are, however, sheltered from takeovers have emerged, the governor of the Bank of Italy recently stated that a reflection was necessary on how to adapt the corporate governance rules typical of co-operative banks to reality.[93] Similarly, savings banks can be sheltered from takeovers because of public ownership. Ownership as such does not raise competition problems. It may, however, explain why in some market entry is particularly difficult. Moreover, if certain ownership and company structures are based on regulation and/or combined with state intervention or companies’ conduct distorting or restricting competition, the competition assessment can change.

The role of regulation and State intervention

As described above, publicly owned or controlled savings banks are often subject to specific and detailed national regulation. Some regulatory elements have repeatedly triggered complaints by third parties alleging the infringement of Treaty rules concerning the free movement of capital or the right of establishment and freedom to provide services. One example was the recently closed infringement procedure on the already mentioned Article 40 of the German banking legislation (Kreditwesengesetz).

Another infringement procedure (already described above) concerns the the tax-free 'livret A’ and the ‘livret bleu’ products distributed by, for instance, the French Crédit Mutuel and Caisses d’Épargne. The Commission opened an infringement procedure regarding the special rights given by the French government to the distributing banks and their effects on the freedom of establishment and the freedom to provide services.[94] The case raises also state aid issues with respect to possible overcompensation for the service provided to the State (by Crédit Mutuel).

Over the last decade, publicly owned banks have been subject to various state aid investigations and decisions which concerned, for instance, privileged access to tier one capital or rescue and restructuring aid cases.[95] The Commission also took action to abolish the general state guarantees for public banks – including savings banks – in Germany (2005) and Austria (2007)[96]. Such state guarantees reduced the beneficiaries’ refinancing costs and enabled them to expand their business at the expense of competitors. With their abolition the Commission has opened the way for more effective, undistorted competition in the European banking sector.

As complaints regularly show, however, the Commission must continue to monitor state intervention potentially favouring certain banks. Recently, for instance, the Commission has opened state aid proceedings against Austria because of an alleged circumvention of the guarantee abolishment in case of a savings bank.[97]

Potential competition restrictions by company behaviour

Savings and co-operative banks co-operate closely with respect to the domestic retail banking business. This raises the key question of whether competition is appreciably restricted between independent market players, i.e. between companies that do not control each other; are not controlled by a third company; and do not form a group (one economic entity) in the sense of merger control. In other words, cooperation, depending on its subject and scope, may reduce or even exclude competition between companies that are supposed to compete with each other. In some countries the regional principle, for example, excludes actual and/or potential competition between independent companies by means of reserved territories. Even if there may not be in all cases a strict ban on entering other territories - permitting marginal competition, for instance, with respect to internet banking - the regional principle is likely to have the effect of excluding or extremely limiting competition between the companies concerned.

Another example of cooperation that may restrict competition is joint pricing; for instance, in the context of common marketing and advertising campaigns, or joint interest calculation for loans by means of internal guidelines, excluding competition for the products concerned. Other measures such as the exchange of sensitive business data can have similar effects. These examples show that, while they remain independent economic agents in some respects (including from the perspective of competition law), some savings and co-operative banks tend to act as one economic entity regarding essential parameters such as product and service development, regional scope of activity or price setting.

Competition authorities in several Member States have shown a certain tolerance regarding cooperation within savings or co-operative banks. This is, first, explainable by economic benefits some forms of cooperation bring about (see below). Secondly, there seems to be such a large variety of country-specific and detailed arrangements that the picture may not be always completely transparent, even to the national authorities. Thirdly, specific legislation, state measures and political ties with local, regional or central governments seem to render antitrust investigations and action difficult[98]. Finally, certain forms of bank cooperation may be explicitly exempted from national antitrust rules. This is, for instance, the case in Austria, where the national competition law was amended in 2002 to allow for certain forms of cooperation within the so-called decentralised sector comprising savings and co-operative banks[99].

However, there are signs that competition authorities may take a somewhat stricter approach in future. One example is the recent decision of the Austrian Competition Court ruling that certain elements of the cooperation between Austrian savings banks infringed Article 81(1) and did not fulfil the criteria of Article 81(3)[100]. Another one is the withdrawal of a joint advertisement campaign by the German savings banks, following the Bundeskartellamt's concerns about joint pricing of the products (“Leuchtturmprodukte”) in question.[101] Similarly, the Tribunal de Defensa de la Competencia in Spain has opened an inquiry into alleged anticompetitive practices, in particular market sharing, of regional banks including savings banks.[102]

Possible benefits of cooperation and impediments of effective competition

Certain forms and areas of cooperation may be indispensable for bringing about additional efficiencies and consumer benefits. Cooperation, for instance, is necessary to agree on common standards and infrastructures for the operation of networks such as payment systems. Moreover, cooperation which improves the risk or cost management of small banks and thereby enables the latter to compete with large banks, can also benefit consumers; for instance, in rural regions where only one or two small banks operate.

On the other hand, potential benefits resulting from certain areas and forms of cooperation cannot justify all potential competition restrictions. In particular, severe competition restrictions such as market sharing or price fixing are unlikely to be outweighed by economic benefits. Moreover, it is highly questionable whether these types of competition restrictions bring about significant benefits.

Even if individual cooperation agreements bring about economic benefits, the effects on market competition have to be thoroughly analysed on a case-by-case basis. Among other factors, this depends to a large extent on the combined market strength of the banks restricting competition. Certainly, quick and general answers are not possible.

Finally, competition restrictions may be the result of legislation or other state measures. Under these circumstances it has to be assessed whether state measures lead to or reinforce anticompetitive behaviour. Member States have an obligation to abstain from measures that have the potential to distort competition in the common market and deprive Articles 81 and 82 EC of their effects by, for example requiring or reinforcing anticompetitive behaviour or by delegating regulatory powers (Article 3(1)(g) EC, Article 10(2) EC applied in conjunction with and Articles 81 EC or 82 EC). Similarly, in the case of public or privileged companies, the Commission has to ensure that Member States do not enact or maintain in force state measures that reinforce or lead these companies to engage in anticompetitive behaviour (Article 86(1) applied in conjunction with Articles 81 EC or 82 EC). In that context, the potential justification of competition restrictions by public interest objectives according to Article 86(2) Treaty may play a role. The analysis of whether or not a derogation from the prohibitions contained in Article 86(1) is necessary for the achievement of a purpose of general economic interest is, however, complex and must be carried out on a case-by-case basis.

A.4.5. Conclusions

Certain forms and areas of cooperation are indispensable for bringing about efficiencies and consumer benefits. Cooperation, for instance, is necessary to agree on common standards and infrastructures for the operation of networks such as payment systems. On the other hand, benefits resulting from certain areas and forms of cooperation cannot justify all potential competition restrictions. In particular, severe competition restrictions such as market sharing or price fixing are unlikely to be outweighed by economic benefits. Even if individual cooperation agreements bring about economic benefits, the effects on market competition have to be thoroughly analysed on a case-by-case basis.

The Commission intends to further evaluate certain competition issues arising in the context of close banking cooperations. This evaluation has to be carried out by means of a thorough analysis on a case by case basis. Should it turn out that one or more of these cooperations raise antitrust issues, the Commission would take up those cases with a Community dimension. The same applies if regulatory aspects or other forms of state measures are involved. Depending on the issues in question, national authorities may be well placed to deal with certain cases.

Issues for follow-up analysis

The Commission intends to investigate:

- company structures and areas of cooperation among those savings banks and co-operative banks that play a substantial role in one or more retail banking market(s);

- behaviour that results in substantial competition restrictions among the participants and on the market;

- economic benefits arising from these types of cooperation;

- State measures potentially requiring, leading to or reinforcing anticompetitive behaviour; and

- regulation and state intervention potentially infringing other Treaty provisions (e.g. free movement of capital, freedom of establishment) and/or potentially distorting competition by State aid favouring certain companies.

A.5. Setting of policies and prices

Using data gathered in the market survey, this chapter examines the setting of banks’ prices and policies, particularly in relation to current accounts. The chapter is structured as follows:

- section 1 examines banks’ practice of product tying; and

- section 2 analyses price variability in relation to current accounts.

A.5.1. Banks' practice of product tying

This section discusses:

- the definition of product tying;

- possible anticompetitive effects of product tying;

- the Commission’s market survey data on product tying;

- general compatibility of product tying with competition law; and

- actions taken by Member States in relation to product tying.

The definition of product tying

Interim report II discussed banks’ practices of product tying and bundling.[103] They are distinct practices. Bundling occurs where two or more products are sold together in a package, although each product is also available separately. Tying occurs when two or more products are sold together in a package, and at least one of these products is not sold separately. That is to say, the customer is forced to buy extra products in order to secure the single product they wanted. This discussion here focuses specifically on the practice of product tying since, unlike bundling, it involves coercing customers to take on additional – and often unnecessary – products.

Product tying is a common strategy for retail banks throughout the EU. Because it is relatively expensive and difficult for banks to win new customers, they often decide to focus their growth strategy on increasing cross-selling to existing customers. Product tying offers a simple way of increasing cross-selling. Such product ties are found in a range of core retail banking products, e.g.:

- selling a current account to a consumer buying a mortgage or personal loan;

- selling payment protection insurance or life insurance to a mortgage customer; or

- selling a current account to an SME taking out a business loan.

Possible anticompetitive effects of product tying

From a competition view point, product tying in retail banking may weaken competition in several ways. Firstly, since it binds customers into buying more products from the same bank, product tying raises switching costs and therefore is likely to reduce customer mobility. Secondly, by binding customers into buying several products from the same bank, tying is likely to discourage the entry of new players, especially mono-line providers..[104] Thirdly, by introducing additional – perhaps unnecessary – products into the transaction, tying reduces price transparency and comparability among providers.[105] The Commission is concerned that the possible anticompetitive effects described above will be strongest in markets where one or more large banks tie products.

Product tying by one or more undertakings in a particular Member State may constitute an exclusionary abuse under Article 82 EC, where such undertakings have a dominant position in a product market that is subject to tying. The possible competition issues arising are discussed below in general terms. Clearly the assessment of a particular tying practice would depend on the specifics of the case; e.g., the products being tied; the extent of dominance; and the extent and effect of product tying.

The Commission’s market survey data on product tying

This section presents an analysis of tying practices in the mortgage and loan markets on the basis of data gathered in the Commission’s market survey. The section discusses:

- the methodology for collecting the data and possible caveats;

- patterns of tying at Member State level; and

- the impact of tying by large banks on the conduct of smaller banks.

The methodology for collecting the data and possible caveats

The survey covers an estimated 75-85% of retail banking activity in the EU: therefore the market shares described here are upper-bound limits of the share of particular firms in each Member State. In some cases, depending on the sample coverage, true product market shares may be considerably lower than the figures presented here. Clearly a full assessment of the relevant product market(s) would be needed to provide the basis for an infringement case.

In its market survey of retail banks the Commission asked banks whether they tied certain sets of retail banking products together. The aim of these questions was to identify which banks offered only tied products to their customers; i.e. those banks undertaking ‘forced’ tying. From the specification of these questions the Commission believes that banks responded to the question based on whether or not they practiced tying for the relevant product combinations. Thus the market survey responses should provide an accurate reflection of the extent to which retail banks in the Member States currently tie their products.

Table 11 below shows the results of the market survey on questions concerning tying of retail banking products. The percentages shown below are weighted for each Member State based on their intra-sample percentage share of all customers holding the lead product (e.g. mortgages or personal loans, rather than current accounts).[106] The data shown below relate to four sets of potentially tied products:

- a mortgage which is tied to a current account;

- a non-mortgage consumer loan which is tied to a current account;

- an SME loan which is tied to an SME current account; and

- a mortgage which is tied to a life insurance policy.

Table 11: Sampled banks reporting product tying, weighted by banks’ combined % share of customer numbers in the lead product market

Mortgage + current account | Consumer loan + current account | SME loan+ current account | Mortgage + life insurance |

Austria | 0% | 0% | 29% | 0% |

Belgium | 44% | 0% | 73% | 0% |

Cyprus | 35% | 30% | 100% | 6% |

Czech Republic | 32% | 41% | 100% | 0% |

Denmark | 93% | 62% | 53% | 0% |

Finland | 75% | 85% | 13% | 0% |

France | 86% | 71% | 91% | 0% |

Germany | 5% | 0% | 11% | 0% |

Greece | 69% | 89% | 81% | 39% |

Hungary | 100% | 80% | 100% | 0% |

Ireland | 0% | 0% | 0% | 0% |

Italy | 48% | 58% | 73% | 2% |

Lithuania | 100% | 100% | 100% | 0% |

Malta | 48% | 39% | 67% | 0% |

Netherlands | 0% | 55% | 58% | 38% |

Poland | 29% | 26% | 85% | 25% |

Portugal | 100% | 100% | 100% | 54% |

Slovakia | 100% | 100% | 100% | 2% |

Slovenia | 24% | 38% | 100% | 11% |

Spain | 67% | 65% | 91% | 0% |

Sweden | 1% | 17% | 13% | 0% |

UK | 1% | 1% | 0% | 0% |

EU25 | 39% | 35% | 63% | 6% |

Source: Commission’s “Retail Banking Survey”, 2005-2006

Note: Data for countries with less than 3 valid observations are not separately reported but are included in the EU aggregate.

Two caveats should be noted here on banks’ answers concerning tying. The first is that banks’ responses reflect a snapshot of their current policy, and thus the practice applied to prospective new customers. Banks were not asked to confirm that they have always tied the products in question. It may thus be possible that a bank which currently ties mortgages and current accounts has only recently adopted this policy and has a large stock of ‘untied’ mortgages. The converse case is also possible, whereby banks no longer tie but have a large stock of ‘tied’ customers. Secondly, the simple yes/no answer banks were asked to provide may not apply strictly to all their current transactions. There may be a small number of exceptions where – perhaps as the result of negotiation – a bank that generally ties its products will excuse individual customers from the requirement to purchase a current account. Moreover, there may be many cases where banks which do not always tie their products may still in practice offer some customers only a tied combination of mortgage and current account.[107]

Patterns of tying at Member State level

The data show some clear patterns at Member State level. The practice of current account tying appears to be widespread in the EU retail banking sector, whether purchased alongside a mortgage, consumer loan or SME loan. In six Member States all banks surveyed report tying current accounts to all three lead products. In five more Member States – Denmark, France, Greece, Hungary and Spain – a majority of banks tied all three products on a weighted share basis. There is a small number of Member States – all in the EU15 – where tying is rare or not practised.

The overall incidence of tying appears to be highest for SME loans, where a weighted average of 63% of EU banks reported tying a current account. Significant levels of current account tying were also seen in the mortgage market and personal loans markets, with EU-level weighted averages of 39% and 35% respectively. By contrast the incidence of tying life insurance policies to mortgages was much less common: on a weighted average basis, only 6% of banks in the EU reported this practice. Portugal was the only Member State where a weighted majority of banks tying was observed.

Clearly the incidence of tying is generally higher in the New Member States, where the retail banking sector is less developed, than in the EU15. There are several possible explanations for this pattern, such as banks’ response to higher credit risk. An alternative explanation for the higher incidence of tying in the New Member States may be the attitudes of financial regulators. It may be that in their desire to promote banking sector growth and development, regulators have thus far exert less scrutiny on the possible anticompetitive effects of tying.

The impact of tying by large banks on the conduct of smaller banks

As argued above, the practice of tying products is likely to raise switching costs and reduce the mobility of customers. Where such tying is widespread (according to the Commission’s weighted market share measure) this will be because the largest banks practice tying. And where the largest banks tie their products this changes the incentive structure for smaller banks. Specifically small banks seeking to expand will find it difficult to encourage switching from customers in either the lead product market or the tied product market because of the increased switching costs. With less opportunity to expand their customer base at the expense of larger banks, the obvious strategy for smaller banks is to focus on retaining existing customers and increasing the extent of cross-selling. In short, the Commission’s hypothesis is that where the largest bank ties its products, the incentive for smaller banks is to mirror this conduct by tying similar sets of products.

The Commission has checked this hypothesis by examining the conduct of the smaller banks in each Member State in response to the tying policy of the largest bank in the sample. Size is measured by the number of customers held by each bank in the relevant lead product market. Thus where the largest mortgage bank in each Member State ties its products, the Commission has also studied the conduct of its competitors in the sample. The criterion for convergent behaviour is whether more than half (measured by weighted mortgage market share) of smaller banks also tie their mortgages. The summary below shows that in eleven Member States, the largest mortgage bank in the Commission’s sample tied a current account to its mortgage. The data show that in nine of these eleven Member States the majority of smaller banks also followed this policy of tying. The two Member States where most banks did not mirror the tying policy of the largest bank were Belgium and Italy. Nonetheless the practice of banks in the clear majority of the relevant Member States supported the Commission’s hypothesis.

Mortgages: do most smaller banks follow the leader by tying? |

Yes | Denmark, Finland, France, Greece, Hungary, Latvia, Lithuania, Portugal, Slovakia |

No | Belgium, Italy |

The Commission carried out a similar analysis for personal loans, where again the largest bank was selected according to its volume of personal loan customers. In ten of the Member States the largest player tied a current account to its personal loan. The results showed that in seven Member States a majority of the competing banks mirrored the leaders’ policy of personal loan tying. Although slightly less clear-cut than the observed pattern on mortgages this also suggests that smaller banks face a strong incentive to match the tying policies of the largest bank.

Personal loans: do most smaller banks follow the leader by tying? |

Yes | Finland, France, Greece, Hungary, Lithuania, Portugal, Slovakia |

No | Denmark, Italy, Netherlands |

The evidence also suggests that some foreign banks adapt their tying and bundling strategy according to domestic competitive conditions. Where tying is widespread among domestic banks these foreign entrants tend to follow suit. This suggests that the commercial rewards to foreign entrants are greater from following this type of conduct by incumbents rather than competing on a new product model. Thus a high proportion of product tying in a particular Member State is likely to be self-reinforcing, weakening the impact of new entrants on competition.

General compatibility of product tying with competition law

As discussed above, banks’ practice of product tying may in some circumstances constitute an infringement of EC competition law. This section sets out some general issues for consideration, specifically:

- possible definition of relevant product markets;

- possible definition of geographic markets;

- relevant provisions of EC competition law; and

- possible justifications for and efficiencies arising from product tying.

The issues below relating to product tying are discussed in general terms only. Any specific assessment of a potential competition infringement could be made only following thorough investigation; including to define relevant product and geographic markets, market positions and the extent of competition, and beneficial and anticompetitive effects of conduct.

Possible definition of relevant product markets

For the purposes of this survey of banks’ tying practice, the following retail banking products are taken as proxies for relevant product markets:

- Personal current accounts: which in view of their typical range of complex payment functions have no close substitutes (though for saving and overdraft activities can be substituted by cash deposits and personal loans respectively).

- SME current accounts: which in view of their typical range of complex payment functions have no close substitutes (though for saving and overdraft activities can be substituted by cash deposits and SME loans and credit lines respectively).

- Mortgages: which because of their large scale and long term nature have few close substitutes, though personal loans may be an alternative for low value borrowing.

- Personal loans: which have a range of substitutes depending on the value and duration of borrowing. For short term lower value borrowing current accounts, credit cards and loans from credit unions may be substitutes. For longer term higher value borrowing there are few substitutes, though equity release and remortgaging products may be alternatives for homeowners.

- SME term loans: which have a range of substitutes including fixed and open credit lines, asset finance and leasing and various forms of capital market finance.

Clearly the definition and use of particular banking products varies across Member States, as does the availability of complementary and substitute products. Any specific investigation of product tying would require a full assessment of relevant product markets and market positions of providers.

Possible definition of geographic markets

For the purposes of this survey, individual Member States are taken as proxies for relevant geographic markets for the products in question. However it is possible that the geographic market for some products and services may be regional or even local. [108] Factors such as the general preference of banking customers for local suppliers, the significance of a dense branch network and the need for the bank to be physically close to its customers[109] were mentioned by the Commission as criteria for defining the geographic scope of relevant banking markets in previous merger decisions. Clearly the definition of relevant geographic markets may vary across Member States. Any specific investigation of product tying would require a full assessment of relevant geographic markets.

Relevant provisions of EC competition law

In relation to Article 82 EC, several criteria must all be fulfilled for product tying to be regarded as an exclusionary abuse of dominance by a single firm. Firstly, the supplier would have to be shown to be dominant in the lead product market. Secondly, the lead product and the tied products would have to be two distinct products[110]. Thirdly, evidence would be required that the tying practice was likely to distort or foreclose competition in the tied product market. Finally, the tying practice would have to be shown not be justified objectively or on the basis of efficiencies.

It is also possible that product tying might be found to be an abuse of collective dominance. Article 82 prohibits any abuse by one or more undertakings of a dominant position, and this is also applicable when two or more undertakings together hold a dominant position. In addition to meeting at least one of the criteria for collective dominance, clearly the criteria for an abuse of dominance (specified above) must also be met to constitute an infringement of Article 82 under collective dominance.

Possible justifications for and efficiencies arising from product tying

According to explanations provided by industry to the inquiry, product tying in retail banking may be justified or create efficiencies for the following reasons:

- reducing credit risk;

- creating economies of scope; and

- the technical difficulty of product unbundling.

Firstly, banks have argued that the practice of tying current accounts – particularly for larger credits such as mortgages and SME loans – is a means of reducing the bank’s credit risk. Such a justification would rely on two assumed sets of efficiency: (i) the efficiencies arising from coercing a customer to pay their main income into the tied current account for the duration of the credit; and (ii) of efficiencies resulting from limiting the customer’s access to other sources of credit.

Banks also argue that a tied current account enables them to monitor the customer’s finances effectively and so reduce credit risk. This justification rests on the assumption that other effective means of risk monitoring were not available to the bank (particularly a customer’s credit record comprising positive or negative credit data). This justification also rests on the assumption that ongoing credit risk assessment via the current account is of equal importance to accurate initial risk assessment, i.e. the decision to grant the credit and the terms agreed.

Secondly, banks might justify the tying of products such as mortgages and current accounts on the basis that they generate significant economies of scope, which could be passed on to consumers. In other words, a bank could offer the tied combination of products more cheaply than had they been bought separately.[111] It seems irrefutable that such a practice might lead to cost savings for a bank compared to selling products untied; and that – depending on the intensity of competition – these savings could be passed on to the consumer. However, it is not clear that this tying practice is generally economically more efficient than selling the products untied. Where a consumer genuinely wishes to buy a mortgage and a current account from the same bank, the tying practice may be efficient for both bank and consumer. However, most customers seeking a mortgage or loan will already have a functioning current account; many held with another bank. The fact that the same consumer can now obtain an additional current account – even fairly cheaply – does not appear to create significant marginal benefit for the customer.[112] In such cases the main effect of tying would not be to create economies of scope to pass on to consumers, but to capture any consumer surplus they may enjoy on the lead product.

Thirdly, the technical difficulty of product unbundling might also be used to justify tying products; for example, where the costs of unbundling both products would be uneconomic for a bank. The Commission asked several banks whether this was the case, particularly in relation to tied current accounts. One large bank in a new Member State reported technical difficulties in granting a mortgage without a linked current account.[113] However it is not obvious why servicing a mortgage, personal loan or SME loan from a current account with another bank should be technically or economically unviable.[114] Typically servicing such credits requires only one routine transaction per month; a transaction which would be normal traffic in a domestic payment system. Moreover, the implementation of the Single Euro Payments Area will make such routine payments still easier – not only within one Member State but across borders. Thus the advent of SEPA should ensure that any remaining ‘technical’ requirements[115] for a tied current account are soon removed.

Actions taken by Member States in relation to product tying

The Commission asked national competition authorities and banking regulators in the Member States about relevant practice in the area of product tying and any previous enforcement activity. The Commission is aware of relevant practice in three Member States: the United Kingdom; Hungary and Ireland.

In the UK, in 2002 the Competition Commission (CC) investigated the supply of banking services to SMEs and found evidence of ‘complex monopoly’ among the main UK clearing banks. In order to strengthen competition and improve SMEs’ ability to switch, the CC recommended several measures, including limiting the tying of products such as loans and current accounts.[116] As a result, in autumn 2002 the main UK clearing banks made undertakings which included committing not to directly or indirectly require SMEs to open or maintain a current account as a condition of the granting or maintaining a loan or deposit account. Exemptions to this commitment were only made where a business current account tied to an SME loan was: (i) a legal requirement by UK law; (ii) necessary because of technical constraints[117]; or (iii) necessary to enable the Bank to take a fixed charge over the book and other debts of that SME. The Office of Fair Trading is currently reviewing competition in the UK SME banking market, including the compliance with and impact of the 2002 undertakings made by the main UK retail banks.

In December 2005 the Hungarian Competition Authority (GVH) concluded its sector inquiry into the mortgage market. This investigation considered product tying among other issues. The GVH found that one bank had a 52% share of the mortgage market. However in view of considerable market entry, intense competition and the leading bank’s declining market share, the GVH found that the bank did not have a dominant position within the Hungarian mortgage market. Thus the tying of current accounts to the bank’s mortgages could not be deemed to be an exclusionary abuse under Article 82 of the EC Treaty.

Notwithstanding the finding of an absence of dominance, the GVH found evidence of widespread tying in the Hungarian mortgage market. The GVH found this practice economically unjustified and argued that banks’ motivation to tie was principally to raise profits. The GVH noted that tying a current account to a mortgage might increase banks’ monitoring ability but made clear that a customer’s ability to repay the credit was independent of whether or not they had a current account with the same bank. Overall therefore the GVH found tying in the Hungarian mortgage market to distort and restrict competition. The GVH therefore recommended to Hungary’s financial supervisor that the practice of mortgage tying should be prohibited or restricted.

In Ireland a new Statutory Consumer Protection Code (issued by the financial supervisor) in July 2006 makes a prohibition on tying of all financial services products (including retail banking products) to customers. The prohibition on making the sale of one product contingent on the purchasing of another product is described in Chapter 2 Rule 4 of the Code.[118] The general principles and existing provisions carried forward from the current codes and handbooks must be complied with immediately. The Financial Regulator expected that all necessary changes would be in place by the end of August 2006.

A.5.2. Setting of prices on current accounts

The price setting[119] behaviour of banks and the evolution of prices over time are potential indicators of the degree of competition and integration in the sector. Banks can develop their pricing strategy for retail products along different but interrelated parameters, such as interest rates, payment fees and other fees charged for various services. For example, banks may opt for a strategy of offering high deposit rates on saving accounts while simultaneously charging substantial fees for the daily management of and operations on these accounts. The converse strategy could also be pursued.[120] As a result, comparing prices across banks and Member States is difficult.

Setting prices for new clients

The economic literature[121] suggests that if suppliers can discriminate between new customers and repeat customers, they will charge lower prices to attract new customers. However, once customers are locked in to a banking relationship, the supplier can charge higher prices, since customers will tend to factor their switching costs into any decision to change supplier. (The interrelated issues of customer mobility and switching costs are explored in the next chapter.)

In its market survey, the Commission asked respondent banks to indicate whether they offered preferential conditions to new current account customers. The survey results suggest that banks generally offer special inducements to new current account customers. Only in a few Member States (e.g. Netherlands and Belgium) do the majority of surveyed banks not offer specific inducements to new current account customers. The typical incentives offered by banks to new customers (start-ups as well as switchers) are the following:

- free banking (e.g. free current account, free payment cards), discounts on transactions and account management fees for a specific period (generally ranging from 3 to 18 months);

- preferential interest rates and reduced, or even zero, overdraft rates for a limited period of time[122];

- discounts on other products and services (such as loans and mortgages, insurance products, fuel or telecomm services); or

- cash incentives, which are also offered to existing customers introducing new clients.

In some cases the value of financial inducements for new current account customers can be considerable. For example, some UK banks provide cash incentives worth around €75 to new customers. When set against banks' average annual gross income per customer (the EU average was €119 in 2005) it is clear that current accounts for new customers are in some cases cross-subsided by higher prices for established customers. This cross-subsidisation can also be seen as inter-temporal, since new customers will eventually stop receiving preferential terms and thus face higher prices.

Analysis of price dispersion among different producers for similar products is one of the elements can help evaluate the extent of competition in financial services markets. Although some characteristics of retail banking markets make it particularly difficult to compare similar products and to construct reliable indicators of price dispersion, simple comparisons of the pricing behaviour of banks can give some initial indications on the degree of competition in the market.

On the basis of the information collected in the Commission’s market survey the following section analyses price variability[123] within and across Member States for selected services, specifically:

- account management fees;

- closing fees;

- excess borrowing fees;

- fees for ATM withdrawals; and

- fees for credit transfers.

Concerning interest rates, Appendix II of Interim Report II presents a detailed comparison of harmonised retail rates. This analysis found that in the euro area there is greater variation in interest rates for deposits than for loans. Households seem to face more dispersed rates than non-financial corporations. Preliminary evidence also suggests that some euro area Member States tend to cluster in groups displaying relatively high or low intermediation margins and spreads. The analysis concluded that a range of factors may explain these patterns of price dispersion and emphasised the difficulty of isolating the effects of specific factors on interest rate differences across countries or products. Overall, the analysis suggested that a substantial part of the divergences would originate from national regulatory regimes and from other structural determinants.

It is first worth noting that the intensity of use of the accounts varies across the 25 Member States.[124] This point was also emphasised by banks responding to the public consultation. In their responses to the Commission’s market survey, several banks indicated that current account transactions are not charged separately, but are included in a package.

Account management fees

Account management fees are the fixed fees that banks charge for the maintenance of a current account, irrespective of financial balance or transaction volumes. In order to enable comparisons between banks’ revenues deriving from such packages, the inquiry examined banks’ average gross fee income from current accounts on an annual basis. The estimates are generated by dividing the total income reported by banks for current account management by the total number of current accounts[125].

Figure 12 shows for each Member State the highest and lowest annual revenues per customer for account management fees. The figure also show the 25th and 75th percentiles, with the bar showing the degree of heterogeneity of prices for 50% of the sample. The EU25 weighted average (approximately €14) is also reported for reference.

Figure 12:

[pic]

Source: Commission’s “Retail Banking Survey”, 2005-2006

Note: Truncation at €140.

The income data reported by the banks indicate that the level of account management fees varies significantly across Member Sates: the figures appear particularly high in some countries[126] (€40 and €90 in Germany and Italy, respectively), whereas in several Member States (Cyprus, Denmark, Latvia, Lithuania, Netherlands, Sweden) average fees are lower than €2.5.

In relation to price variability, the pricing strategies of banks surveyed vary both within and across the Member States. Four countries (Czech Republic, Germany, Italy and Luxembourg) show high variability of annual fees earned by the surveyed banks for current account management.

Closing fees

Closing fees are the direct financial cost that customers must pay when closing their bank account. Closing fees can be levied on various products including current accounts, timed deposit accounts or investment funds. The Commission's market survey asked banks to specify the typical level of closing charges for current accounts. The results of the survey are displayed below in Figure 13.

Figure 13:

[pic]

Source: Commission’s “Retail Banking Survey”, 2005-2006

Note: Truncation at €65.

As can be seen from the figure, practice varies considerably across the Member States. In more than half of the Member States customers are not charged for closing their current account. In those Member States where a closing charge is generally charged to customers for the termination of the current account relationship, the levels of fees show an overall high degree of variability. In countries such as Italy and Luxembourg, maximum charges levied by the sampled banks for closing a current account reach €60 and €100, respectively. However it should be noted that following a ruling from Italy's competition authority in September 2006, members of Associazone Bancaria Italiana were required to remove closing fees from current accounts.[127]

Excess borrowing fees

Excess borrowing fees for current accounts are charged by banks when a customer exceeds their agreed credit limit and when payments are made while a customer is beyond their credit limit. In the inquiry's market survey banks were asked to indicate total annual income from excess borrowing fees. These results were then compared to the total levels of (non-interest) fee income reported by banks for current accounts. Clearly, excess borrowing fees are only one source of fee income for current accounts.

In general across the EU banks' total income from excess borrowing fees represents a fairly low share of the total gross fee income on current accounts; in most Member States the level in 2004 was below 10 per cent of total current fee income.[128] However, in France, Spain, the UK and Cyprus, more than a quarter of banks' current account fee income derives from excess borrowing fees. Thus excess borrowing fees generally play little role in the revenue stream of banks in New Member States and are a significant revenue stream in only a handful of EU15 countries. This finding raises the question of how far banks in different Member States use these fees on the one hand simply to cover the costs resulting from customers exceeding their borrowing limits; and on the other hand as an additional source of income.

It is noticeable that the large private banks and the savings and co-operative banks in both France and Spain generate significant shares of their non-fee income from this source. In France, for example, the largest banks generated between 30 and 70 per cent of current account fee income in 2004 from excess borrowing fees. Meanwhile in the UK the proportion of fee income for the largest five banks was between 16 and 58 per cent. In the UK, the OFT is currently carrying out a fact-finding exercise on current account default charges levied by UK retail banks.[129]

Fees for ATM withdrawals

Banks across the EU apply various pricing formulae for payments services, separately or in combination with other services. These formulae include explicit pricing for a single product, in the form of transaction related fees; fees for a package of products; charges for currency conversion; and other elements. In addition, there are some less visible prices including value dating practices and cross-subsidisation with other products. Additionally, for each payment instrument, the concrete pricing formulae may depend on a number of variables (e.g. the channel used, the amount and urgency of the order, the time at which the order has been submitted and so on) that make comparisons complex.

In relation to both ATM withdrawals and credit transfers, fees levied to customers vary among banks and across Member States, not only in terms of level but also in terms of pricing structure. Some banks indeed charge a fixed amount per transaction, others levy a percentage of the transacted amount (generally establishing also a minimum amount), others apply a mixed structure, combining these two components. For ATM withdrawals, most banks also charge different fees according to the type of payment card used (i.e. debit or credit card).

The inquiry's comparison is based on a simulation of banks' fees for €100[130] ATM 'off-us'[131] withdrawal with a debit card. Average fees for ATM withdrawals are weighted based on the total number of current accounts per bank and Member State. Because of reporting problems the comparison covers only euro area Member States. [132] Figure 14 presents the results of the simulation. The figure also show the 25th and 75th percentiles, with the bar showing the degree of heterogeneity of prices for 50% of the sample. The EU25 weighted average fee (approximately €1.14) is also reported for reference.

Figure 14:

[pic]

Source: The Commission’s sector inquiry in retail banking, 2005-2006.

Note: Truncation at €8.

Banks in five countries (Belgium, Finland, Germany, the Netherlands and Spain) report high variability of fees for off-us ATM withdrawals. In Austria, Finland, Ireland, Luxembourg, Netherlands and Portugal, the weighted average fee is lower than half of the EU12 weighted average (€1.14); whereas Member States such as Germany and Spain report figures higher than €2.8 per transaction. The remaining Member States present a certain degree of homogeneity. It is interesting to note that in France and Italy, where the highest and lowest fees differ substantially, the 50 percent of sampled banks around the median charge fairly similar fees.

It is interesting to note that for off-us ATM withdrawals Member States generally report weighted average fees that are lower than the simple average values. This suggests that larger banks (in terms of number of current accounts) tend to charge lower ATM fees than do smaller banks.

Fees for credit transfers

Credit transfers are payments that are made between bank accounts at the instruction of the payer. The inquiry has conducted compared prices for credit transfers using a similar methodology to that discussed above for ATM withdrawals. The comparison is based on a simulation of fees charged for a €100 credit transfer, for both 'on us' and 'off us' transactions[133]. The results of the simulation exercise are reported in Figures 15 and 16 respectively.

Figure 15:

[pic]

Figure 16:

[pic]

Source: The Commission’s sector inquiry in retail banking, 2005-2006.

The results show that for a €100 domestic credit transfer, the weighted average fees in EU12 are respectively €1.30 and €2 for on-us and off-us transactions. The weighted average fees show high variation across Member States. For on-us credit transfers, fees range from zero to €9.77; and for off-us credit transfers, from zero to €10.85.

Concerning price variability, in seven Member States the majority of surveyed banks do not charge fees for on-us credit transfers (i.e. transfers between accounts within the same institution). In the remaining five Member States (France, Greece, Ireland, Portugal and Spain) customers are generally charged a fee for on-us transactions, which varies from bank to bank. The results show that of the euro area Member States, on-us credit transfers are most expensive in Greece. Fees in Ireland are also clearly above the EU12 average.

For off-us credit transfers, banks in five Member States (Austria, Belgium, Finland, Germany and the Netherlands) reported zero or very low weighted average fees. Weighted average fees were highest in Greece and Luxembourg, where the 50 per cent of banks around the sample median charged fees of €7.30 to €17.00 and €0.56 to €10, respectively. The weighted average fee for an off-us credit transfer in Greece (€10.85) was more than five times the EU12 average.[134] Banks in France, Italy and Ireland reported mean weighted fees around the EU12 average of €2. However price variation in all three Member States was large, with the highest reported fees all in excess of €10.

A.5.4. Conclusions

The characteristics of the retail banking industry make it difficult to compare similar products and construct reliable indicators that will enable an evaluation of competitive structure. Nevertheless, the pricing behaviour of banks provides some initial indications on the degree of competition in the market. Based on the inquiry’s market survey, a commonly observed that banks tend to compete less aggressively for switchers than for new-to-market customers, though the intensity varies across Member States.

In relation to current accounts the inquiry has examined a range of fees and found evidence of significant variation in prices within and across Member States. Such patterns are evident in relation to the pricing of several parameters including account management fees; closing charges; excess borrowing fees; fees for ATM withdrawals; and fees for credit transfers.

Notwithstanding significant differences between Member States for some types of current account fee, convergence can also be observed in the pricing behaviour of banks in several Member States. For example, banks in Italy and Luxembourg reported the highest levels of fees for both account management and closing an account. Similarly, excess borrowing fees typically generate a small share of banks' total fee income on current accounts; less than 10 per cent in more than half of the Member States. However such fees generate over a quarter of the current fee income for banks in France, Spain, the UK and Cyprus. For credit transfers banks in around half of the Member States surveyed do not charge fees for domestic transactions, while banks in Greece typically charge fees well in excess of the euro area average.

The inquiry’s market survey suggests that in most Member States the majority of banks tie a current account to mortgages, personal loans and SME loans. Moreover, where the largest bank in a Member State ties its products, the inquiry’s data suggests that the majority of its competitors, including foreign entrants, choose to follow suit. From a competition view point, product tying in retail banking may weaken competition in three ways. Firstly, tying raises switching costs and therefore is likely to reduce customer mobility. Secondly, by binding customers into buying several products from the same bank, tying is likely to discourage the entry of new players and growth of smaller players. Thirdly, by introducing additional – perhaps unnecessary – products into the transaction, tying reduces price transparency and comparability among providers.

The Commission is concerned that possible anticompetitive effects will be strongest in markets where one or more large banks tie products. Product tying by one or more undertakings in a particular Member State may constitute an exclusionary abuse of dominance under Article 82 EC, where such undertakings have a dominant position. Clearly the assessment of a particular tying practice would depend on the specifics of the case.

A.6. Customer choice and mobility

Retail banks typically compete on a range of product characteristics such as quality of services, price (the interest rates and fees for particular products), location and reputation. Customers consider all of these characteristics when choosing the most attractive offer. Thus more efficient providers, who will offer cheaper or better quality services, should see their market shares rise as customers tend to choose their products. For competition to be effective customers need to have clear information with which to choose the best offer on the market, and they need to be able to switch providers when a significantly better offer appears. In this way customer choice and mobility in retail banking markets exerts competitive pressure on existing and potential suppliers to continually improve their performance. Therefore customer mobility should be seen as an important contributor to competitive retail banking markets.

The inquiry does not consider a high level of customer mobility as an end in itself, nor as a simple measure of success. The inquiry has examined the issue of customer mobility solely with a view to identifying and providing evidence on unnecessary obstacles to customer mobility. Helping to remove such obstacles should, over time, help to strengthen competition in retail banking throughout Europe.

Interim Report II of the sector inquiry examined in detail the issue of customer mobility, particularly in relation to current accounts. The report studied how often customers switch their current account; the average length of banking relationships; and the average number of retail banking products bought by customers. In addition, the report conducted a preliminary statistical analysis of the relationship between these variables and the financial performance of retail banks, to examine whether the level of customer mobility, market concentration or cross-selling generally increases banks’ ability to exercise market power.

This chapter refines the preliminary analysis and findings set out in the interim report, based on further analysis and responses from stakeholders to the public consultation. The chapter is structured as follows:

- Section 1 discusses the relationship between customer mobility and competition;

- Section 2 examines switching costs in retail banking;

- Section 3 outlines the inquiry’s evidence on customer mobility in retail banking;

- Section 4 examines the empirical relationship between customer mobility and market performance; and

- Section 5 concludes.

A.6.1. Relationship between customer mobility and competition

There are a number of market characteristics that influence the level of customer choice and mobility on a particular market. These characteristics include the nature of the business and the products, market dynamics, the pace of product innovation, the competitive structure, customer preferences. Theoretically, in a homogeneous industry under perfect competition there is no case for switching, as all service providers provide the same service quality at the same price. However, under imperfect competition the actual level of customer mobility in a given industry is determined by a complex interaction of the above characteristics.

There are certain special characteristics of retail banking that influence customer preferences when choosing a provider. The complexity and long-term perspective of some products makes customers favour service providers with whom they can develop a relationship. In this relationship trust and reliability are important, and can weigh heavily alongside the financial implications of choosing one bank over another.[135] Proximity also matters. Even though new efficient delivery channels (phone, internet, ATM, etc.) rapidly gain share in a multi-channel system, direct contact through a bank branch remains the main form for maintaining customer confidence and establishing new account relationships. Therefore, although internet banking offers attractive rates and is expanding rapidly as a distribution channel, internet services generally remain a compliment rather than a substitute to branch banking[136] - at least in the foreseeable future.

Banks are fully aware that continuous provision of high quality services is a key to customer loyalty and commercial success. Respondents in the consultation – particularly from industry – cited surveys on customer satisfaction in individual Member States. In such surveys the proportion of customers expressing satisfaction with their bank ranged from 56% (Italy)[137] to 94% (Netherlands).[138] Two points should be made in relation to such surveys. Firstly, since they are conducted using different methodologies, their results are likely to be valid on a time-series basis but not comparable across countries (unless carried out across several countries). Secondly, consumer satisfaction is contingent on market context. Consumers in one Member State may be satisfied with prices and service levels that customers in another Member State would find unacceptable.

Nonetheless, it is likely that a large proportion of banking customers – probably the majority in most Member States – would describe themselves as satisfied with their current bank. These customers may well believe that their bank offers among the best quality products and services and has competitive pricing. For these customers the question of switching bank (and its related costs) does not arise. Thus the scope of the inquiry’s analysis of customer mobility is consumers and SMEs who are not fully satisfied with their current provider or are seeking to change bank for other reasons.

The ability of customers to switch provider easily and their propensity to switch can have a significant influence on the competitive process. In markets where customers can change provider freely and a high proportion do so regularly, there is pressure on existing and potential providers to continually improve their performance. Obstacles that reduce customers’ ability to switch provider correspondingly reduce the competitive pressure on firms to retain existing customers and limit the number of potential new customers. In the case of banking, where providers are aware that customers are unlikely to switch provider, banks may be able to extract rent over the long-term from their existing customers.

Not only market characteristics determine customer mobility, but customer mobility also influences market characteristics – effectively the competition structure. Customer mobility as a feedback mechanism places competitive pressure on banks to win and retain customers. Therefore, the issue of customer mobility has profound implications for the intensity and nature of competition in the retail banking industry.

A.6.2. Switching costs in retail banking

Switching costs are costs that existing customers incur when they change their suppliers. Before deciding to switch their business to a competitor the customer has to decide whether the benefits of such a move outweigh the costs, including the financial and other costs arising from changing providers. Thus the presence of switching costs will, other things being equal, reduce the propensity of customers to change bank.

This section discusses:

- sources of switching costs in retail banking;

- effects of switching costs in retail banking; and

- possible measures to reduce switching costs in retail banking.

Sources of switching costs in retail banking

The sector inquiry identified four sources of switching costs that are likely to reduce the ability of customers to switch bank:

- administrative burden;

- information asymmetry and low price transparency;

- bundling and tying; and

- closing charges.

Administrative burden

Administrative burden is a transactional switching cost that occurs when the change of service provider is implemented. Switching banking providers requires work and effort from customers. The scale of the administrative effort required will vary according to the banking product in question. Switching current accounts is a complex operation because of the range of everyday functions that are conducted through the account. Filling in the necessary forms for opening the new account, closing the old one, transferring balances, transferring direct debits, setting up payment instructions, informing customers about the new account number requires time and effort on behalf of the customer switching. Moreover, a customer’s ongoing reliance on a single account to receive wages and pay bills and make everyday transactions means that there may be greater risks to switching provider. In particular, switching current accounts will create an obligation to customers to transfer their payment arrangements from one bank to another. This complex operation may deter customers who would otherwise switch accounts.

Information asymmetry and low price transparency

In retail banking, the relationship between bank and customer has economic and psychological value. A banking relationship often results in a better understanding by the bank of the credit quality of its customers. This information on credit quality may be lost when a customer switches banks. Low credit risk customers are pooled together with higher credit risk customers when changing bank and consequently charged higher interest rates. Therefore, one opportunity cost of switching bank is the foregone capitalised value of their previously established relationship.[139]

In addition, the price information provided to retail banking customers on their current account and other products may be inadequate or complex – making it difficult to compare prices and choose between banks. Complexity forces customers to make a substantial investment of time in searching for the best supplier. There is a good reason to believe that a customer would search for an alternative supplier only if the expected gain would offset the (expected) search costs. As part of the sector inquiry, questionnaires were sent to a number of consumer associations in the EU to seek their views in relation to retail banking. Lack of price transparency was mentioned in several replies[140]. This information complexity and low transparency from the supply side can also reduce customer mobility.

Bundling and tying of banking products

Tying and bundling are common practices in retail banking. In some cases they may enable banks to offer a range of products that are better suited to customer needs, while generating savings in production, distribution and transaction costs that can be passed on to the customer in the form of lower prices.[141]

However in some cases product bundling and particularly tying may weaken banking competition. Since it offers a way to differentiate otherwise identical individual services, bundling may make it difficult for the customer to compare the prices of its current bank with those of the competitors. For this reason prices might be kept at a higher level than otherwise. In addition to potential problems of price transparency, the more services are sold to the customer, the more their switching costs increase, making it more difficult to change service provider. Tying is discussed specifically from a competition law perspective in Chapter A.5.

Closing charges

Banks may charge customers for terminating various services. In some cases these charges may be used to cover the administrative cost to the bank of filling in forms or providing certain documents; in other cases they may be used to compensate for interest rate risk exposure (e.g. where a term loan or deposit is closed early). There is also the possibility that banks might levy closing charges in order to deter customers from closing their account. Since closing charges are explicit financial charges, they are the easiest type of switching cost to quantify in retail banking. In most countries banks do not apply closing charges to current accounts. However, this practice is common in other Member States such as Austria, Slovakia and Slovenia. Levels of closing charges for current accounts are considered in detail in Chapter A.5.

Early termination or partial repayment of consumer loans triggers varying levels of charges in Member States. In case of mortgages the picture appears more complicated due to long-term nature and varying legal requirements of mortgage products. The product also has more variations in pricing. In the case of floating rate mortgages the early repayment fee is often zero, or less than with fixed rate mortgages, in accordance with remaining interest rate exposures and risks. In Belgium and France, where such fees cannot be charged, banks tend to tie additional products to the mortgage. This helps explain the much higher cross-selling ratios for mortgages in both countries (selling an extra 3.53 and 3.27 products respectively, compared to an EU average of 1.97 products).

Effects of switching costs in retail banking

The inquiry’s Interim Report II outlined three sets of effects that could result from high levels of switching costs in the retail banking industry. Firstly, switching costs might increase banks’ market power, enabling them to set higher prices to established customers who appear locked in to a banking relationship[142]. This effect is examined in further detail in Section 4 of this chapter and, based on the analysis of sector inquiry market data, appears to have empirical support.

Secondly, Interim Report II argued that switching costs in banking might discourage market entry, since it may become uneconomic for new entrants to provide a sufficiently competitive offer to induce customers to switch. Here a distinction needs to be made between firstly, new entrants into niche markets such as credit cards and personal loans; and secondly, new providers of full service retail banking. It appears that new providers are able to enter and rapidly acquire scale in some niche product markets throughout the EU, depending on the particular circumstances[143]. However, for full service banking, market practice suggests that switching costs tend to limit prospects for market entry. Indeed, foreign banks seeking to establish a large full service operation in another Member State typically do so through merger and acquisition. Greenfield entry into full service banking is generally perceived as risky and low customer mobility makes it difficult to acquire scale in a commercially viable time period.

Thirdly, it was assumed that high switching costs in retail banking would tend to discourage product innovation. The reasoning was that markets with lower customer mobility generally offer lower potential returns to successful innovation, so reducing the pace of innovation. It is difficult to assess how far this theoretical concern applies in practice to retail banking.

Possible measures to reduce switching costs in retail banking

Chapter 7 of Interim Report II surveyed a range of measures already taken by Member States in order to reduce switching costs and enable customer mobility in retail banking.

Possible measures to reduce the administrative burden of changing current accounts include:

- switching regulations: provisions requiring banks to observe certain procedures and deadlines when transferring a customer’s account details to a new bank; and

- switching codes: voluntary undertakings between banks which have similar features to switching regulations, though are delivered through industry self-regulation.

The design of switching regulations and codes can significantly influence their effectiveness. For example, one common criticism of the switching code recently introduced in Ireland is that the requirement to close the ‘old’ account shortly after switching discourages customers from using the service as many customers prefer a trial period. The Dutch switching code, by contrast, has a default thirteen month transition period during which both accounts remain open.

Possible measures to reduce problems of information asymmetry and low price transparency might include:

- providing transparent comparable information ex ante on prices of banking products[144];

- disclosing prices and charges applied ex post by banks for particular products; and

- increasing the financial awareness and product knowledge of customers.

Interim Report II also identified closing charges as a possible obstacle to customer mobility in the current account market. The level of such charges appeared most problematic in Italy. However, recent Italian legislation abolished the fees related to closing of open ended contracts, such as bank, securities and savings accounts in August 2006[145], and the Italian banking industry has recently introduced several related initiatives to ease customer mobility.[146]

As set out in its White Paper Financial Services Policy 2005-2010 , the Commission has set up an Expert Group on Customer Mobility in relation to Bank Accounts. This expert group is working to facilitate the opening and switching of bank accounts on a domestic and cross-border basis. Separately the Commission will prepare a White Paper on Mortgage Credit, for publication in the first half of 2007. Some of the issues under consideration in the mortgage market are closely related to customer mobility and price transparency; for example, issues on pre-contractual information and advice, the average percentage rate of charge (APRC) and early repayment fees. Both of these initiatives are led by DG Internal Market and Services, assisted by other Commission services including DG Competition.

A.6.3. The sector inquiry’s evidence on customer mobility in retail banking

This section summarises some of the key findings of the sector inquiry in relation to customer mobility and patterns of customer behaviour. The section examines:

- market churn and the length of banking relationships;

- the extent of cross-selling in retail banking; and

- the practice of multi-banking by customers.

Market churn and the length of banking relationships

Customer mobility in retail banking is likely to vary from product to product. There are indications that more shopping around is done regarding taking or switching mortgages and loans, due to higher potential price advantages.[147] The interim report looked specifically at mobility related to current accounts. First, current accounts are the most widely held and used retail banking product. Second, unlike some other products such as customer loans which expire after a set time-period, current accounts are open-ended. There was reason to believe that it is not feasible to obtain data on mortgage or loan switching from respondent banks. Third, and partly because of the open-ended nature of the current account relationship, current accounts play a gateway role: banks often use current account as the basis for cross-selling other products to their customers.

Customer mobility by definition relates to customer behaviour, yet the sector inquiry was collecting data from banks and not customers. Customer mobility therefore could only be illustrated by the churn and longevity indicators with certain important limitations. Churn tries to capture the share of customers who change providers in a given year. Longevity is a measure of the average age of existing banking relationships. The sector inquiry gave the following results for current account number growth, churn controlled for market growth, and longevity in EU Member States:

Table 17: Number of current accounts growth, Churn and Longevity. Weighted average. Year 2005. Consumers and SMEs

Growth | Churn after control | Longevity |

Belgium | 3.39% | -1.67% | 5.27% | 8.90% | 10.04 | 9.99 |

Cyprus | 6.28% | 16.81% | 10.33% | 13.00% | 6.65 | 4.63 |

Czech Republic | 1.76% | -0.07% | 8.61% | 10.70% | 7.91 | 7.87 |

Denmark | -0.20% | 3.71% | 10.02% | 15.43% | 12.06 | 9.75 |

Estonia |

Finland | 0.67% | -0.84% | 4.23% | 6.27% | 17.44 | 13.98 |

France | 1.82% | 2.31% | 6.84% | 12.26% | 11.06 | 8.39 |

Germany | 2.55% | 0.01% | 8.46% | 15.15% | 11.55 | 9.85 |

Greece | 8.34% | 12.75% | 2.36% | 3.55% | 4.34 | 5.23 |

Hungary | 3.43% | 12.33% | 10.41% | 17.59% | 6.26 | 4.29 |

Ireland | 5.04% | 4.08% | 5.44% | 6.95% | 8.13 | 10.14 |

Italy | -1.54% | 1.45% | 7.68% | 11.23% | 9.39 | 8.23 |

Latvia | 8.82% | 6.31% | 6.74% | 7.13% | 3.11 | 4.81 |

Lithuania | 6.99% | 9.02% | 7.73% | 3.34% | 6.23 | 4.46 |

Luxembourg | -2.98% | 1.50% | 6.46% | 11.29% | 7.20 | 6.45 |

Malta | 9.06% | 4.93% | 5,39% | 6,49% | 8.83 | 6.64 |

Netherlands | 0.24% | 1.72% | 4.17% | 8.88% | 14.33 | 10.45 |

Poland | 4.30% | 1.09% | 9.11% | 17.00% | 6.18 | 4.04 |

Portugal | 2.02% | 1.70% | 11,88% | 14.34% | 11.21 | 8.87 |

Slovakia | 6.26% | 2.40% | 10.81% | 15.80% | 4.49 | 5.54 |

Slovenia | 1.02% | 1.21% | 5.97% | 10.89% | 7.02 | 3.06 |

Spain | 3.61% | 1.58% | 12.12% | 10.34% | 6.91 | 6.02 |

Sweden | 1.66% | 1.55% | 5.62% | 8.80% | 11.82 | 12.33 |

United Kingdom | 3.64% | 0.81% | 5.07% | 13.72% | 10.66 | 7.66 |

NMS Average | 4.17% | 3.28% | 9.02% | 14.82% | 6.28 | 4.67 |

EU-25 Average | 2.42% | 1.72% | 7.78% | 12.63% | 9.74 | 7.93 |

Notes: Growth rate: number of current accounts end of year / number of current accounts beginning of year.

Churn after control: (new currents accounts + closed current accounts) / (2* number of current accounts beginning of year).- absolute value of growth rate/2

Longevity in years

The estimates for EU-15, New Member States and EU-25 are country-level averages weighted by population.

Source: Commission’s “Retail Banking Survey”, 2005-2006.

It should be emphasised that longevity measures the average age of existing active current accounts; not their expected lifetime. We have two reasons to believe that the expected lifetime of these accounts is significantly higher than their current 'longevity'. Firstly and evidently, most accounts currently open will remain to be open for some time in the future.[148] Secondly, an estimate of the expected lifetime of current accounts could be deduced from the churn measure used in the inquiry. One could take the average yearly consumer churn of 7.8% and control it with 1.2-2.4% for demography (see below). If in one year between 5.4% and 6.6% of bank customers switch service provider, than all customers switch on average in every 15-18 years. The inquiry’s estimates on the length of existing current account relationships could not control for the effects of market growth, therefore in countries where the proportion of new accounts are higher, as in the New Member States, the difference between the expected lifetime of current accounts and current longevity figures could be significantly higher than in other member states.[149]

Demography was a known factor increasing the churn indicator, and was also not controlled for. Basically, the churn rate overestimated the rate of customers moving accounts by the average of birth and death rates. This figure is estimated to fall between 1.2% and 2.4% in different countries for consumers; and between 5.6% and 13.6% for SMEs.[150] SME demography figures are higher in new member states, a pattern that is also observable in the churn statistics. Some part of consumer switching takes place due to changes in life situations, e.g. new job, move, marriage. There is no comparative data on EU25 regarding this element either.

The extent of cross-selling in retail banking

The degree of cross-selling in an industry is often perceived as an indicator of the competitive structure of a market. Retail banking is a multi-product activity. Customers often consume from a single bank a wide range of financial products including current accounts, mortgages, deposits, credit cards, insurance, and other financial products. Customers can very well be purchasing multiple products from a single provider due to the existence of economies of scope that allow the customer to reach a higher level of satisfaction from concentrating its purchases.

Cross-selling can be measured as the average number of products that customers purchasing a specific product are altogether purchasing from the same bank. This specific product line used as a reference is called the “hook” product. Table 18 reports country level weighted cross-selling averages with respect to several hook products, for customers and SMEs respectively.

Table 18: Cross-selling ratio. Weighted average 1 . Year 2005.

[pic]

1 Banks weight: consumer cross-selling

2 Ratio of the sum of all cross-sold products to the sum of all hook products

3 The estimates for EU-15, New Member States and EU-25 are country-level averages weighted by population.

Source: Commission’s “Retail Banking Survey”, 2005-2006

Mortgages are the hook product that link to a higher consumption of other products in every country for retail consumers. On average, in the EU25 a consumer taking a mortgage from a financial institution consumes 2.97 products altogether. The average number of products consumed by SMEs tends to be higher than for individual consumers. Credit lines are the most effective hook product among SMEs. Loans are also a very effective form of cross-selling to SMEs. On average, an SME consumes altogether 2.81 products from a bank which offers her a loan. Cross-selling also differs largely across countries and tends to be lower in the New Member States.

The practice of multi-banking by customers

There is evidence that customers opening new current account do not always immediately close their old one. ECB data suggest[151] that in several Member States there are more active current accounts than inhabitants. The same data also suggest that in the UK and Finland there are around two current accounts per capita. This points at the growing trend of multi-banking: having parallel accounts at different institutions.

The choice of customers to maintain more than one current account will be influenced by the cost structure set by banks in that market. In some Member States, maintaining a current account is free (i.e. there are no account management fees) and therefore many customers who change banks decide to keep their old current account, at least for a substantial period. In other Member States, however, account management fees are substantial and comprise a significant share of banks’ revenues from current accounts. Chapter A.5. reports, using 2004 data, that several banks in Italy charged annual account management fees of €90 per year. This type of price structure clearly raises the cost of multi-banking by customers and reduces the number of accounts held by customers. [152]

A common response from banks in the public consultation was that the Commission’s analysis of customer mobility failed to capture the effect of multi-banking, and that its positive effects on competition were significant in some markets. In fact, multi-banking appeared in the Commission’s data as a contribution to overall market growth. Interim Report II presented churning rates for 2005, before and after controlling for overall market growth. Weighted current account market growth rates in 2005 – including the effect of increasing multi-banking type of mobility – were 2.42% for consumers[153] and 1.72% for SMEs. Much of this overall growth will result from demographic change or increasing market penetration. It cannot be excluded that in some Member States in particular, the spread of multi-banking may significantly increase competition. However, across Europe it appears to play at most a modest role.

A.6.4. Empirical relationship between customer mobility and market performance

Using data gathered in the market survey the sector inquiry has tried to analyse the empirical relationship between customer mobility and market performance in retail banking. The purpose of this analysis is to understand how competitive outcomes vary according to different market characteristics, including market concentration, customer mobility and the level of cross-selling. This section discusses:

- the inquiry’s interim findings on customer mobility and market performance; and

- further findings based on multivariate analysis at bank level.

The inquiry’s interim findings on customer mobility and market performance

Interim Report II analysed the market survey data using univariate correlations. On this basis the report set out some preliminary findings, based on an analysis of the data at country level and at firm (i.e. bank) level. At country level, the interim report found:

- a weak negative correlation between churn and market concentration, and no relationship with profitability;

- a clear positive correlation between longevity and the level of cross-selling; and

- no strong correlation between the level of cross-selling and market concentration or profitability.[154]

At bank level, the interim report found:

- a significant positive relationship between longevity and bank profitability;

- a significant negative relationship between consumer churn and bank profitability;

- a significant negative relationship between customer mobility and market concentration; and

- a positive and significant relationship between cross-selling and market share; but

- no robust positive relationship between cross-selling and profits.

Overall, bank level analysis provided clearer and more robust results, which may be expected given the greater number of observations. The overall picture from the preliminary analysis was that measures of customer mobility (specifically higher market churn and lower longevity) tended to be negatively correlated with measures of market power (specifically profit ratios and market share). These correlations between customer mobility and market power indicators tended to have higher levels of significance when longevity was used as a mobility indicator.

However, as was made clear in Interim Report II, the evidence was suggestive, the analysis had not controlled for the effects on customer mobility of third variables absent in the analysis. Therefore, the findings were presented as tentative, to be developed further on the basis of multivariate correlation analysis.

Further findings based on multivariate analysis at bank level

Building on the preliminary findings of Interim Report II, the inquiry has conducted a multivariate analysis of the relationship between customer mobility and market performance. In the public consultation on Interim Report II, many industry stakeholders commented that the quantitative analysis of this relationship failed to take account of (i) the direct effect of switching costs on customer mobility; (ii) levels of customer satisfaction in explaining customer mobility; (iii) the impact of differing levels of banking sector stability on market performance. These comments are valid. However, it has not been possible to incorporate these variables (switching costs, customer satisfaction and banking stability) in the multivariate analysis. The reason, in all cases, is that objective and reliable measures are not available for use in the multivariate analysis.[155] This does not mean that the inquiry disregards that such factors have explanatory power: as the correlation analysis below shows, the factors analysed here tend to explain only 50 to 60 per cent of total variance. In the regressions below we have used country dummies that should help in controlling some of this cross-country variability that exists in these variables. Nonetheless, the Commission doubts that multivariate analysis can reliably estimate the direct effects of switching costs or customer satisfaction on observed levels of customer mobility; or the effect of financial stability on retail banking market performance.

An important issue is whether higher bank profitability is related to the observed consumption patterns of cross-selling and customer mobility; or, as the conduct-structure-performance paradigm suggests, industry profitability is positively correlated with market concentration and market share of the bank. We evaluate these relationships in Table 19 by running regressions of bank profitability on the bank’s market share, cross-selling, churning and longevity, controlling for country effects. The profitability and market share measures used here refer to all retail banking activities; i.e. the sum of consumer and SME retail activity. Therefore, we estimated a model including measures both for SMEs and consumers as regressors.[156]

Table 19: Correlation estimates for retail banking profitability

[pic]

The results show a positive and significant relationship between profitability and market share (regression 1). However, when we add to this basic specification our measures of churning, longevity and cross-selling this positive relationship between market share and profitability is no longer statistically significant. On the other hand, we do observe a strong negative relationship between higher mobility – via lower longevity for SMEs or higher churning for consumers – and lower bank profitability.

In particular the measure of consumer churn shows a strong negative correlation with retail banking profitability. The correlation coefficient in regression 5 suggests that a 1 percentage point increase in the level of 'churn' in bank's customer book will reduce its pre-tax profit margin by 1.6 percentage points.[157] However, this result may overstate the true effect of consumer churn on banks' profit margins. The reason for possible overestimation is that – compared to the results of the univariate analysis – in this multivariate correlation analysis, the coefficients of SME churn and consumer longevity change direction; the former becoming positive and the latter negative. It appears likely that this result is caused by collinearity between the consumer and SME variables for churn and longevity.[158] Thus part of the higher negative coefficient for consumer churn results from the countervailing positive coefficient[159] for SME churn; an effect which is evident in regressions 2 and 5. Overall, however, there appears to be a strong and significant negative relationship between consumer churn and profitability: univariate analysis in Interim Report II estimated the correlation coefficient at -1.05, whereas multivariate analysis excluding SME churn produces a correlation coefficient of -0.84.[160]

Overall, the results of the bank-level analysis confirm that higher churn and lower longevity are related to lower pre-tax profitability in retail banking. This effect is robust and statistically significant, suggesting that banks face greater pressure on margins where customers are more mobile.

Contrary to what one would expect, the relationship between profitability and cross-selling or market share is not significant. Market share is not a significant determinant of bank profitability after controlling for other characteristics of the bank such as its amount of cross-selling or customer mobility. This is not to say that particular banks in highly concentrated markets do not exercise market power to raise their margins.[161] However, the overall results at bank level across the EU indicate that the dynamics of competition in the industry – particularly the level of customer churn – play a greater role than market structure in determining the scope for banks to exercise market power. Therefore one conclusion from the inquiry is that simple, proportionate steps to enable customer mobility will enhance competition in retail banking.

A.6.5. Conclusions

It is likely that a large proportion of banking customers – probably the majority in most Member States – would describe themselves as satisfied with their current bank. For these customers the question of switching bank (and its related costs) does not arise. Thus the scope of the inquiry’s analysis of customer mobility is consumers and SMEs who are not fully satisfied with their current provider or are seeking to change bank for other reasons.

The inquiry’s analysis suggests that typically between 5.4% and 6.6% of current account customers in the EU will change provider per year. However, industry surveys suggest that the proportion of unsatisfied customers is typically much higher. For this group, the level of switching costs will be an important consideration. The sector inquiry has identified four sources of switching costs that are likely to reduce the ability of consumers to switch bank: administrative burden; information asymmetry and low price transparency; bundling and tying; and closing charges.

The evidence gathered by the sector inquiry suggests that high levels of switching costs in the retail banking industry may weaken competition in two ways. Firstly, switching costs may increase banks’ market power, enabling them to set higher prices for established customers who appear locked in to a banking relationship. Secondly, high switching costs and low customer mobility may limit prospects for market entry in full service retail banking, notably through greenfield operations.

Building on the preliminary findings of Interim Report II, the inquiry has conducted a multivariate analysis of the relationship between customer mobility and market performance. Owing to measurement problems, variables such as customer satisfaction, switching costs and financial stability could not be incorporated directly into the quantitative analysis. However, the inquiry has estimated the impact of customer mobility in the current account market on banks’ ability to exercise market power (using total retail banking profitability as a proxy). Multivariate analysis at bank level suggests that a one percentage point increase in the level of market churn corresponds to a similar reduction in banks' pre-tax profitability ratio. This effect is robust and statistically significant, suggesting that banks face greater pressure on profit margins where customers are more mobile. Therefore one conclusion from the inquiry is that simple, proportionate steps to reduce switching costs will enhance competition in retail banking.

As set out in its White Paper Financial Services Policy 2005-2010 , the Commission has set up an Expert Group on Customer Mobility in relation to Bank Accounts. This expert group is working to facilitate the opening and switching of bank accounts on a domestic and cross-border basis. Separately the Commission will prepare a White Paper on Mortgage Credit, for publication in the first half of 2007. Some of the issues under consideration in the mortgage market are closely related to customer mobility and price transparency; for example, issues on pre-contractual information and advice, the average percentage rate of charge (APRC) and early repayment fees. Both of these initiatives are led by DG Internal Market and Services, assisted by other Commission services including DG Competition.

Part B: Competition in the market for payment cards and payment systems

B.1 Introduction

This section of the report presents the inquiry’s main findings in relation to payment cards and payment systems. The operation of these infrastructures may raise a range of competition issues, many of which are discussed in this report. This chapter provides an overview of the operation of payment systems and the scope of the Commission's inquiry, prior to the examination of specific competition issues in subsequent chapters.

The chapter is structured as follows:

- section 1 outlines the main means of cashless payments used in Europe;

- section 2 discusses the creation of a single market for payment services;

- section 3 outlines the organisation of POS payment cards systems;

- section 4 describes the fees paid in a POS card payment system; and

- section 5 discusses the inquiry’s data and methodology on payment cards.

B.1.1. Cashless payments in the EU

Cashless payment transactions in the EU in 2004 amounted to 65.3 billion transactions[162]. Considering the importance of payments for all economic sectors, for customers and SMEs, effective competition between banks and between payments has an important role to play in improving the efficiency of services, reducing prices for customers and enhancing the competitiveness of the whole economy.

Figure 20:

[pic]

The number of electronic payments (card payments, credit transfers and direct debits) has constantly increased over the last years, progressively replacing payments traditionally made with paper instruments (cheques) and cash. The total number of cheques has declined from 8.9 billion in 2000 to 7.5 billion in 2004, whereas the total number of electronic payments has increased from 37.9 billion in 2000 to 57.9 billion in 2004. The most common electronic instrument was payment card, representing 41% of the total cashless transaction volume.

The recent evolution of cashless payment instruments shows a similar trend in all Member States, even though national differences concerning the relative importance of each type of payment instrument exist. For example, some countries still rely on paper-based payment instruments whereas in other countries electronic payment methods are already widely used.

Figure 21 shows the number of transactions per capita, distinguishing among payment instruments. In most Member States, the most common instruments are payment cards and credit transfers. Furthermore, some countries such as Austria, Germany, Netherlands and Spain show a quite high number of direct debits per capita. Concerning the number of cheques per capita, a few Member States reach values higher than 25 transactions per capita[163].

Figure 21:

[pic]

B.1.2. Creating a single market for payment services

The report presents an assessment of the state of competition as resulting from the Commission investigation into existing systems. The advent of SEPA (Single Euro Payment Area) will change significantly the landscape for card/non-card payments and related infrastructures in the EU. SEPA aims to create a single market for payments throughout the euro area by integrating national payments systems. This will permit economies of scale to be realised and make cross-border competition feasible. The end result should be more effective competition and several of the competition issues that are highlighted in this report can contribute to the establishment of a more pro-competitive SEPA.

However, although SEPA is clearly pro-competitive at the conceptual level, it is vital that SEPA is implemented in such a way that supports more effective competition and innovation thus enabling the realised cost savings to be passed on to businesses and consumers.

The Commission has adopted in December 2005 a proposal for a Directive on Payments Services in the Internal Market. [164] From a competition perspective, this proposal has a double aim. First, to enhance competition by establishing an appropriate prudential framework for payment service providers to allow new players to enter payments markets. Second, to facilitate cross-border competition for payment services by harmonising the necessary legal rules regarding the provision of payment services (e.g. who pays for payment transactions, transparency in pricing, execution times, liability in case of default, consumer information and rules on revocability of payment orders).

B.1.3. Organisation of POS Card Payment Systems

POS card payment systems enable consumers to use plastic cards for payment transactions at the point of sale (POS), which is — most often — a payment terminal in a merchant outlet. These POS systems are to be distinguished from ATM card payment systems, which enable consumers to use plastic cards for withdrawing money from automated teller machines (“ATM” or “cash machines”). In practice, POS and ATM systems may be combined so that consumers can use one plastic card both for payment at POS terminals and for withdrawing cash at ATMs, as well as for other ATM services such as printing of statements, balance reporting, credit transfers, etc. This sector inquiry concentrated on POS card payment systems and all subsequent observations therefore cover POS systems only. Where observations exceptionally relate to ATM systems as well, this will be spelled out explicitly.

POS card payment systems involve a wide range of services and service providers. Figure 22 below gives a structural overview of a de-integrated POS system where the roles of scheme ownership, network operation and financial services are attributed to different entities.

Figure 22:

[pic]

The above figure shows three main groups of players: (i) cardholders and merchants (ii) scheme owner (iii) issuers and acquirers. Cardholders and merchants engage in a payment transaction through the intermediary of banks and scheme owners. The cardholder receives payment services and credit services from the entity that issued the card (the issuer). The merchant receives payment services from the entity that deals with the merchant (the acquirer). Acquirers may also be issuers.

When the cardholder uses the card to buy from the merchant, the merchant receives from the acquirer the retail price less a merchant service charge. The issuer pays the acquirer the retail price minus or plus any interchange fee[165]. This interchange fee is determined by the card association members of, for instance, MasterCard and Visa. In addition to the interchange fee from the acquirer, the issuer receives from the customer the payment, any annual fee, any interest payment on debt outstanding, late payment fees, etc., and might conversely give the customer rebates, loyalty rewards and the like.

Issuers issue cards to cardholders and acquirers recruit merchants for payment card acceptance. Payment card issuing is the business of distributing payment cards to consumers on own account and risk while payment card acquiring is the business of contracting merchants for payment card acceptance on own account and risk. Both activities involve certain financial risks with regard to the settlement of a payment card transaction. An acquirer in particular risks losing money on chargeback claims of cardholders. An acquiring bank may be faced with chargeback claims up to several months after it has credited the merchant. If the merchant goes bankrupt in the meantime and if cardholders claim back their money, the acquiring bank may bear the financial costs of the chargeback claim vis-à-vis the issuing bank.

Typically, scheme owners reserve issuing and acquiring to credit institutions or entities controlled by credit institutions. Acquiring typically involves the marketing of card acceptance to merchants and therefore requires sales staff. Acquirers also provide customer service to merchants (e.g.: they defend them against chargeback claims of cardholders, check claims that money has not been transferred, etc.).

An issuing processor opens and manages the cardholder’s account on behalf of the issuer, books card transactions on these accounts, authorises card transactions on behalf of the issuing bank, sometimes arranges the clearing and settlement, provides cardholder statements and sometimes operates a cardholder call centre (for lost and stolen cards) and sometimes also handles chargeback claims of cardholders. An acquirer processor opens and manages the merchant’s account on behalf of the acquirer, forwards authorisation requests to a switch (or switches authorisation requests directly to the issuer or issuing processor) and sometimes also supplies voice authorisation centres, books transactions on merchant accounts, charges merchant discount rates to merchants and produces merchant statements. Many acquirers also rent out POS terminals to merchants.

Issuing and acquiring processing is often done by the issuers and acquirers themselves. The scheme owner is responsible for: (i) granting licenses (and membership status) to independent financial institutions for the use of a card logo and for performing issuing and acquiring services within the network; it may also (ii) certify non-financial institutions for performing technical activities such as clearing and processing within the system; it usually (iii) sets the network rules and the technical (message) standards; and it (iv) implements these network rules and standards by executing audits at member banks and certificate holders and by organising arbitration in the case of settlement disputes.

The figure shows other players who may or may not participate in a POS payment card system, depending on the degree of vertical integration of that system. In a largely decentralized system, there is scope for competition between non-bank entities that (i) produce payment terminals, (ii) rent out terminals, including maintenance services, (iii) switch transactions between banks, (iv) process transactions on behalf of the issuing bank and/or the acquiring bank, and (v) produce, personalise and destroy payment cards. In many European POS card payment systems, these technical services are concentrated in the hands of a single “network service provider”. In the Dutch debit card system PIN, for instance, the switching, processing and clearing is done by the inter-bank association Interpay, while in the Danish Dankort debit card system the inter-bank association PBS processes, clears and settles card transactions and even acquires merchants for Dankort acceptance. In some systems, these technical services are provided by the scheme owner, who then not only sets the rules and standards but is also involved in the operational aspects of the payment card system.

B.1.4. Fees paid in a POS card payment system

Figure 23 illustrates the flow of fees for a POS transaction in a four-party card system.

Figure 23: The flow of fees in a POS card transaction

[pic]

When a cardholder uses his/her card to buy at a merchant outlet, the merchant receives from the acquirer the retail price less the merchant discount rate or merchant service charge (MSC). The issuer pays the acquirer the retail price minus an interchange fee[166]. In addition, the issuer receives cardholder fees, interest payments on any debt outstanding and fees for late payment or other reasons. , However, the issuer may offer their customers rebates, loyalty rewards and other incentives.

Two customer groups may be charged the costs of services provided in a POS system: cardholders (mainly consumers) and merchants, the firm that accept payment cards. The academic literature on the payment cards industry describes this as ‘two-sided market’. There remains considerable debate on how issuing and acquiring banks should recoup their costs while ensuring the efficiency of the system. For example, it is widespread practice for issuing and acquiring banks to subsidise card usage by charging a zero (or even negative) fee to cardholders, while recouping the corresponding costs from merchants through interchange fees.

B.1.5. The inquiry’s data and methodology on payment cards[167]

DG Competition has collected information from two main sources. Firstly, information on acquiring and issuing was collected through a questionnaire sent out in July 2005 to a representative sample of 203 acquirers and issuers.

Secondly, data on payment card networks was obtained from a questionnaire sent out in August 2005 to 26 domestic and international payment card systems. Following this initial request, DG Competition sent out an additional questionnaire to payment card networks in December 2005.[168] The geographical scope of both questionnaires was the EU-25. In addition, DG Competition also had at its disposal surveys of consumers’ and merchants’ behaviour and a range of further market studies[169].

The questions put to acquirers and issuers addressed only debit and credit cards (deferred debit cards were treated as credit cards)[170]. Moreover, they focused only on transactions made at physical points of sale (POS) and did not cover ATM transactions. In contrast, the questionnaire sent to payment card systems covered a wide range of rules and activities developed by these institutions, including ATM and the relevant price and cost data. The market survey also gathered information on non-price variables affecting competition between payment card networks. The inquiry's data and findings on non-price competition variables are not reported here but are presented in Chapter 15 of Interim Report I.

Information was mainly collected on a yearly basis and over the period 2000-2004. Some data, however, were collected on a quarterly basis. A significant amount of the requested information concerned financial aspects (e.g. prices and costs). In order to harmonise the financial data, respondents were asked to convert their data into euro currency. Some of the requested data required an allocation of revenues and costs based on accounting data. This allocation was made by the respondents themselves.

The sample was performed on a list of banks submitted by the two largest international payment networks in Europe: MasterCard and Visa. Two different techniques were used to select institutions active in issuing and acquiring. These techniques were the same for both debit and credit cards. The data set comprised a list of issuing institutions sorted by countries. The sampling was done on a per-country basis, as a random selection could have led to the under-representation of some EU countries in the sample (particularly small ones). The technique had a number of statistical defects, which were considered prior to sampling: large institutions are intrinsically over-represented and there is a strong bias towards MasterCard and Visa network members.

Structure of Part B of the technical annex

The analysis of markets for payment cards and payment systems is structured as follows:

- chapter 2 examines concentration and integration in the payment cards market;

- chapter 3 discusses cardholder fees;

- chapter 4 discusses merchant fees;

- chapter 5 examines interchange fees;

- chapter 6 analyses profitability of the payment cards industry;

- chapter 7 considers non-price competition variables;

- chapter 8 examines membership rules and governance;

- chapter 9 discusses cross-border competition in acquiring;

- chapter 10 discusses payment infrastructures; and

- chapter 11 examines interchange fees for ATMs and non-card payments.

A Glossary of the terms used in this section can be found at Annex A.

B.2. Market concentration and integration

Market structures in payment cards markets vary considerably across the EU. Card acquiring markets in particular show evidence of highly concentration levels in some Member States and low concentration levels in others. Especially where market concentration is high, incumbents may have the opportunity to erect barriers to entry in either issuing or acquiring, or be able to exercise market power in the setting of card fees.

The chapter is structured as follows:

- section 1 examines concentration in acquiring and issuing markets;

- section 2 studies the relationship between market concentration and merchant fees;

- section 3 discusses integration of card payment systems;

- section 4 examines relevant competition issues; and

- section 5 concludes.

B.2.1. Concentration in acquiring and issuing markets

The analysis of the level of concentration on the issuing side of the payment card market yielded no evidence of excessive concentration across the EU-25 Member States. Generally, issuing is characterised by a high number of market players of varying size. No cases of a single issuer have been reported.

The remainder of this chapter examines market concentration and market structures in the card acquiring market. This section discusses:

- acquiring in international payment networks for credit cards;

- acquiring in international payment networks for debit cards; and

- acquiring in national payment networks.

Acquiring in international payment networks for credit cards

The business of acquiring credit cards in the international networks appears highly concentrated. Figure 24 shows the level of concentration in the acquiring of credit cards for one of the international networks, measured in terms of the Herfindahl-Hirschman Index (HHI), and the number of acquirers across the EU-25 Member States for 2004.

Figure 24:

[pic]

An HHI of up to 2000 (shown by the red line) is assumed to raise no substantial competition concerns. Spain is the only country where the level of concentration in the acquiring of credit cards for one of the international networks is below 2000, average HHI across the EU being at the level of about 5800. In five Member States (namely Austria, Cyprus, Denmark, Finland and Portugal) acquiring for MasterCard credit card transactions is performed by a single institution. Furthermore, three other Member States (Luxembourg, Netherlands and Germany[171]) have the HHI higher than 8000, meaning that acquiring is very concentrated, albeit not in hands of a sole acquirer.

The analysis shows that the high concentration goes hand in hand with a small number of acquirers and low concentration is usually combined with a larger number of acquirers. For example, Spain has a low concentration and the highest number of the banks (19) performing acquiring in the network. However, there are several countries (e.g. Slovakia, the UK and Hungary) where the level of concentration remains quite high despite of a large number of acquirers.

A simplified analysis of changing levels of HHI across the EU-25 revealed that falling and raising levels of concentration in many Member States from 2000 to 2004. A dynamic analysis of market concentration shows that in 10 EU Member States, the variation over the period in the HHI was above 10%. Seven of these countries are new Member States (the Czech Republic, Hungary, Latvia, Lithuania, Malta, Poland and Slovakia). Since these countries are characterised by fairly immature and unstable payment card markets, these changes may be explained by market adjustments. In three old Member States (Greece, Ireland and Luxembourg) variations in the level of concentration were caused by either increase of the total number of the acquirers (Greece), “reshuffling” of the market shares of the three biggest banks (Ireland) or expansion of its market share by the main acquirer while the share of the main competitor had been falling (Luxembourg, with an increase of the concentration index by astonishing 40%).

Looking at one of the international systems, where acquiring for credit cards has nearly doubled in the period from 2000 to 2004, it appears that the EU average concentration index — at the level of 5800 in 2004 — is quite high. This is almost three times the threshold of 2000 above which competition concerns start to arise.

Acquiring in international payment networks for debit cards

The business of acquiring debit cards in the international systems appears equally highly concentrated. Looking at one of the international systems, where this business expanded by more than 400% between years 2000 and 2004, the average HHI exceeded the threshold of 2000 in 16 Member States. Only in two Member States (Spain and Latvia) the level of an HHI is below or slightly above an HHI of 2000. The same group of countries as for credit cards (with the exception of Luxembourg and Finland) report a single acquirer for debit cards, namely Cyprus, Denmark, Luxembourg and Portugal.

Figure 25:[172]

[pic]

The analysis of the dynamics in the HHI in the time series (from 2000 to 2004) shows that 5 of the 18 countries analysed, which also happen to be countries with a single acquirer, show no changes in terms of their respective levels of concentration over the 2000-2004 period (i.e. Austria, Cyprus, Denmark, Luxembourg and Portugal).

As with credit cards, the Member States with significant changes (above 10%) in the concentration index are mainly new Member States with quite “immature” acquiring markets and therefore possibly unstable acquiring. As in case of credit cards, no particular pattern can be identified (both increasing and decreasing concentrations are equally reported).

Among the old Member States, a substantial variation in concentration levels was observed in four countries: Greece, Spain, Sweden and the UK. In Greece and Spain, the fall of the HHI concentration index by almost 60% was mostly due to the fact that the biggest acquirer (or 2 biggest acquirers as was the case in Spain) in 2001 subsequently lost a significant portion of its market share to its competitors in the following years and the acquiring market shares were redistributed among the existing acquiring institutions.

In the UK, a similar drop of almost 60% in the concentration index was caused primarily by the increase of the overall number of acquirers for debit card transactions in the network and the loss by the market share leader in 2001 of a substantial part of its market by 2004. On the other hand, two other acquirers had very much strengthened their position.- In Sweden the level of concentration had actually grown by 2004 due to the fact that the total number of acquirers fell over the period examined. Furthermore, the market leader had expanded its market share in the acquiring of debit card transactions in the network considered.

Acquiring in national payment networks

Ten national card payment systems were analysed. The results may be biased due to the small sample size and to the fact that many of the domestic debit cards issued in the national networks tend to be co-branded with international network brands, such as Maestro and Visa debit. Therefore, some of the volume reported for the national network may actually include turnover data reported by international card networks.

The joint acquiring volume of these ten analysed national systems grew by 50% in relative terms between 2000 and 2004. This cumulative turnover was more than four times higher than the turnover generated on one of the international debit networks in the same period. This result would be even more striking if the turnover generated in the Spanish national networks (ServiRed, Sistema 4B and Euro 6000) had been included. The level of concentration index in the acquiring in national systems is presented in Figure 26 below.

Figure 26:

[pic]

The majority of national networks are characterised by a very high level of concentration in the acquiring business. In fact, 4 of the 10 networks analysed, according to the reported data, had a single acquirer in 2004 (Belgian Bancontact, Danish Dankort, Maltese Cashlink and Dutch Currence (PIN)). The lowest level of concentration was found in the Italian national network Cogeban (Pagobancomat), with an HHI below 600. This low HHI was primarily due to an extremely high number of acquirers in the market (about 430) as well as the fact that no single acquirer has a large share of the market. Given the low level of concentration in Italy, its HHI has not been taken into account in calculating the average level of concentration and average number of acquirers in national networks across the EU-25. Thus, excluding Italy, the average level of concentration in the EU-25 in 2004 was about 7100, which exceeded by about 24% the corresponding average level of concentration in the international networks considered (both debit and credit).

B.2.2. The relationship between market concentration and merchant fees

High concentration does not always give rise to competition concerns. However, it may lead to do so where highly concentrated markets exhibit higher prices; e.g. where the level of concentration positively correlates with high merchant fees in a given country.

Such correlation analysis comparing HHI levels and the average levels of merchant service charges (MSCs) did not lead to conclusive results. Even though the correlation is positive for both the national systems and the one international system considered (meaning that higher concentrations on average lead to higher levels of MSC), the absolute level of the correlation coefficient does not exceed 50%. Moreover, this correlation analysis did not control other variables which may affect the level of the merchant rates.

Table 27: Degree of Correlation between level of HHI, number of acquirers and level of MSC across the EU-25, 2004, %

[pic]

B.2.3. Integration of card Payment Systems

This section discusses:

- degrees of vertical integration of card payment systems in the EU;

- a case study of partial de-integration of the Dutch Interpay scheme; and

- joint ventures for acquiring services.

Degrees of vertical integration of card payment systems in the EU

In the EU-25, a wide range of different card payment systems with a varying degree of vertical integration can be observed. The industry generally distinguishes between “ open ” or “ four-party ” card payment systems and “ closed ” or “ three-party ” card payment systems, where the scheme owner also engages in the financial aspects of the payment card business by issuing cards and acquiring merchants. This is the case for American Express, Citibank (Diners Club) and JCB, which (mainly) issue and acquire cards themselves (or under licences). These systems are also referred to as “proprietary” systems, as the scheme owner typically is the proprietor of part or all of the technical network used for routing, switching, clearing and processing the transactions. The industry also calls them “T&E card” systems as these systems predominantly target cardholders who use cards in the travel and entertainment (T&E) industry.

For the purposes of the sector inquiry, we compared the degree of vertical integration using a scale from 1 to 6 with 1 being the lowest degree. Systems where the entity owning the card brand essentially does not engage in any activity other than setting the technical standards and parameters for access to the network operate at level “1”. Here, scheme ownership is legally separated from network ownership and the financial business of issuing and acquiring.

Where a scheme owner engages in further — technical or financial — parts of the cards business, further integration levels are reached as follows:

- + 1 level : scheme owner switches authorisation requests itself

- + 1 level: scheme owner authorises and processes transactions

- + 1 level : scheme owner clears and/or settles transactions

- + 1 level: scheme owner acquires merchants

- + 1 level : scheme owner sells and/or rents POS equipment

This classification is only a starting point for a complex assessment, as the separation of scheme ownership from the technical/financial aspects of the business may not be sufficient alone to realise the full potential of competition in a card payment system. On a scale from 1 to 6, five national systems had the lowest degree of 1, one system was classified as degree 2, one reached a degree of 3 and three a degree of 4. Finally, one system had the highest integration of 6 (please see Table 28 below).

Table 28:

5 | 5 | 5 |

4 | 4 | 4 |

3 | 3 | 3 | 3 | 3 |

2 | 2 | 2 | 2 | 2 |

1 | 1 | 1 | 1 | 1 |

MC | VISA | JCB | Amex | Diners |

This comparison is only the first step in a complex analysis and no quick conclusions should be drawn from it. Even systems with an integration degree of “1” may be difficult to penetrate for foreign banks (and non-bank service providers), because the separation of the card scheme from the technical/financial aspects of the payment cards business alone may not be sufficient to allow for real competition. This may be the case because other barriers to competition exist or because a scheme’s principal members may be able to adopt measures to reduce or eliminate the scope for competition.

Despite these caveats, the separation of scheme ownership, network operation and the financial aspects of the payment cards business, i.e. issuing and acquiring, may be a first important step towards more competition within a POS card payment system.

The extent to which a card system is vertically integrated matters from the competition viewpoint, particularly with respect to the financial aspects of the cards business. These relate to the guarantees given by banks to both cardholders and merchants that a transaction will be settled if all formal requirements[173] are fulfilled. For issuing banks, the risk calls for, amongst other things, the careful assessment of a cardholder’s creditworthiness and for acquiring banks it entails, amongst other things, the evaluation of fraud risks at merchant outlets[174]. Where inter-bank associations issue cards and/or acquire merchants on behalf of shareholder banks, the price for these financial services is not subject to competition between these banks. Examples include Banksys in Belgium and many inter-bank associations for acquiring merchants in the MasterCard and Visa systems (see the next section).

Case study of partial de-integration of the Dutch Interpay scheme

In the Netherlands, the scheme owner Interpay previously licensed banks to issue debit cards, processed, switched and routed the transactions and also signed up merchants for debit card acceptance. Thus, the cooperation of the shareholder banks in the joint venture included both the technical and financial aspects of the cards business. In April 2004, the Dutch competition authority NMa adopted a decision declaring that the shareholder banks of Interpay infringed national competition law by extending their cooperation beyond the technical aspects of the cards business and by jointly selling acquiring services to Dutch merchants. After the initiation of competition proceedings and a recommendation of the Dutch Central Bank, the shareholders of Interpay decided to cease selling acquiring services collectively and took over merchant contracts from Interpay for a transitional period.

Thus, the system was separated at downstream level, where banks starting selling acquiring services to merchants in a competitive way. As a consequence, 12% of Dutch merchants entered into negotiations with an acquiring bank for better prices, resulting in average cost savings of 7.4% (according to a NMa study) for those contracts that were re-negotiated.

The Dutch Interpay case provides an interesting example that, where potential competition at the downstream level of a card scheme is foreclosed due to vertical integration, there may be scope for price reductions if the system is de-integrated.

Joint ventures for acquiring services

In eight EU Member States member banks of national and international payment card systems provide acquiring services through joint ventures. Merchants therefore face one single offer instead of many competing offers. Moreover, the existence of joint ventures in acquiring services can effectively prevent a foreign acquirer from making a competitive offer to local merchants. Central acquirers informally told the Commission that this was the situation in several EU Member States.

In EU Member States and networks where inter-bank associations (joint ventures) acquire merchants, market access for foreign banks may be particularly burdensome. Local issuing banks may agree on preferential (“on us”) interchange fees with the incumbent acquirer (an inter-bank association in which they have financial interests) but charge higher, multilaterally agreed interchange fees to any foreign acquirers attempting to compete with the incumbent.

Informal complaints against Europay Austria in Austria and UNICRE in Portugal suggest that preferential bilateral interchange fees within the MasterCard and Visa payment card systems may raise competition concerns. In Portugal, an informal complaint by an acquirer alleges that that the structure and level of domestic interchange fees in the Portuguese Visa system discriminated against foreign acquirers, as UNICRE had no problem agreeing preferential tariffs with its shareholder banks while foreign acquirers could not obtain equally low fees.

After the publication of Interim Report I in April 2006, the Commission held several meetings with banks in a number of Member States in order to analyse where self regulation is possible or whether antitrust enforcement is necessary. These talks were held in close cooperation with the local competition authorities who would then take the lead in addressing competition problems if the talks should fail.

So far, this approach has yielded promising results. Austria provides a concrete example: from a competition perspective, an immediate goal for this market is that the two dominant acquiring joint ventures (Europay Austria and Visa Austria) engage in genuine competition. Austrian banks pledged to instruct their joint ventures accordingly and will report to the Commission regularly on the implementation. The banking industry moreover pledged to review inter-bank arrangements on domestic interchange fees for MasterCard and Visa payment cards and announced that a reduction of these fees can be expected. A solution for the structural competition problems is also under discussion and – if implemented – would address a major issue identified in the inquiry's Interim Report on the payment cards market.

B.2.4. Competition issues

This section discusses:

- implications of clearing arrangements;

- membership requirements to buy processing services; and

- co-branding.

Implications of clearing arrangements

Some card payment systems do not have a multilateral clearing platform. Banks then have to arrange clearing (and chargebacks) bilaterally. This “ peer-to-peer clearing” may, however, raise the cost of market entry for foreign banks.

The Commission’s sector inquiry provided indications that multilateral clearing platforms may be competition-enhancing within domestic card payment systems, as they facilitate market entry for foreign banks. In systems with bilateral clearing arrangements, foreign banks may have difficulties in gaining access to clearing facilities as this depends on the willingness of all local banks to enter into bilateral clearing arrangements or on the goodwill of a “sponsor”. The existence of sponsorship alone may not be sufficient to allow market entry if local banks have no commercial interest in sponsoring a potential competitor. In order to promote cross-border competition, card payment systems should be invited to set objective and verifiable rules that grant new entrants a right of access to sponsorship by one of the incumbent banks, or — if this is technically feasible — set up a multilateral clearing platform. Similar concerns may arise where membership in a card payment scheme as such relies on being sponsored by a principal member, normally an incumbent in the market.

Membership requirements to buy processing services

In systems where scheme owners also provide processing services, member banks may be required to buy processing services in order to obtain a license for issuing a certain card or acquiring merchants for a certain card brand. One of the international schemes relies on an exclusivity arrangement with member banks regarding its clearing services for domestic debit card transactions. In one case, for seven years after migrating the national scheme to the international scheme, the member banks are obliged to use the processing facilities of the international network for domestic payment card transactions as well. Only after these seven years will members be free again to use third party providers for processing services with regard to that card brand. It remains to be seen whether banks agreed to this exclusivity arrangement in return for investment by the international scheme in the processing facilities of the former national scheme.

Co-branding

Definitions vary of ‘co-branding’. There appear to be three possible definitions:

- Co-branding with a non competitor such as a retailer or a non profit association (affinity cards)

- Co-branding with an indirect competitor: a national payment system with an international one.

- Co-branding with a direct competitor.

Most schemes use the term “co-branding” in a broad sense for the co-existence of their own logo together with another logo on the face of the card. One international payment scheme and some other networks further distinguish between “co-branding”, i.e. a cooperative agreement between an issuer and a non-member co-branding partner; and “co-badging”, which involves the application of the mark of a national payment scheme on the face of a card together with the logo of the international payment scheme.

The co-branded logo can be that of another network, run by a financial institution (or “bank”) and/or by a non-bank organisation, such as retailers. The co-branded network can further be national or international. It can cover debit and/or credit functions. There can also be more than one co-brand partner for one and the same card, e.g. the regulations of one of the international payment schemes allow a maximum of three. On the basis of the replies, most schemes seem to allow co-branding with networks that are not deemed competitors, at both national and international level.

In principle, a national debit payment system does not seem to be considered a competitor of an international credit payment system. Nor does a national debit payment seem to be considered a competitor of an international debit payment system. This is sometimes, but not always, because of the geographic coverage of the respective schemes. In many countries, the rules of the national scheme apply while a card is used in a domestic context, but once the card is used cross-border, the rules of the international network apply. In some countries, however, both the national and the international debit card scheme may cover national payments. Most national payment schemes offer international payment card functions, e.g. for MasterCard and/or Visa. Likewise, the regulations of one of the international payment schemes explicitly allow co-badging with any national scheme as long as the scheme is not deemed a competitor. However, most schemes seem to prohibit, either explicitly or implicitly, co-branding with networks deemed to be competitors, at both international and national level.

The rules of two of the international payment schemes explicitly prohibit the use of brands deemed to be competitors by the decision-making forum. At national level, the same principle seems to apply but often implicitly. In Spain, for instance, the logos of the three national schemes, i.e., Sistema 4B, ServiRed and Euro 6000, may not co-exist on the same card, although the systems are claimed to be fully interoperable. This appears to follow from the unique membership rule, according to which a financial institution cannot be a member of two systems at the same time.

Although exceptional, there are a couple of examples of co-branding between competing networks at both international and national level. In Italy, for instance, national networks allow co-branding with each other.

Co-branding with non-banks seems to be generally accepted by the international payment networks. As most national payment system co-brand their cards with an international payment function, they are bound by the rules of the international networks on co-branding with non-banks. In practice, most national payment schemes allow co-branding with non-banks. However, two national payment schemes prohibit co-branding with non-banks, both through the design rules for the cards.

The prohibition on co-branding with networks deemed to be competitors and with non-banks might limit not only actual but also potential competition between networks and between banks and non-banks, respectively. For instance, the risk of being deemed a competitor and thus loosing the right to offer international payment functions might hinder national debit schemes from entering into competition with MasterCard and Visa for the processing of cross-border debit card transactions. Similarly, the prohibition on co-branding in two national payment schemes might reduce the choice of cardholders and thus their possibility to put pressure on the member/licensed banks to compete with better prices and conditions. Finally, by prohibiting co-branding with non-banks, national payment schemes might prevent retailers from competing with banks in the market for card issuing. In countries where such co-branding is permitted (for instance, the United Kingdom and Germany), it can be observed that co-branded cards are used as a vehicle for market entry by new issuers or for the competitive expansion of card issuing by existing issuers. In systems where such co-branding is prohibited this route towards more intense competition with the incumbent card issuers may be foreclosed.

B.2.5. Conclusions

The business of acquiring credit cards and debit cards in the international networks appears highly concentrated. The majority of national networks are characterised by a very high level of concentration in card acquiring. Issuing, on the other hand, is much less concentrated.

The structure of point-of-sale card systems in the EU is heterogeneous. The inquiry compared the degrees of vertical integration in the national card systems in various Member States. Five of the national systems examined had the lowest degree of vertical integration, setting only technical standards and the parameters for network access. Meanwhile three systems surveyed had relatively high degrees of integration since they also conduct authorisation, processing and clearing of transactions.

The degree of vertical integration in the large international systems MasterCard and Visa differs from one Member State to another. In general these systems show moderate degrees of vertical integration. However in some Member States, transactions are routed through the network of a local network operator that also acquires all merchants in the market.

The sector inquiry has provided some indications that joint ventures in acquiring may be a structural issue leading to various entry problems for foreign acquirers. For example, local issuing banks may agree on preferential ('on us') interchange fees with the incumbent acquirer (an inter-bank association in which they have financial interests) but charge higher, multilaterally agreed interchange fees to any foreign acquirers attempting to compete with the incumbent.

Access to clearing facilities, as a pre-condition for banks to enter new markets, may be an obstacle where local banks have no commercial interest in sponsoring a potential competitor. In order to promote cross-border competition, card payment systems should be invited to set objective and verifiable rules to grant new entrants a right of access to sponsorship by one of the incumbent banks or – if technically feasible – set up a multilateral clearing platform.

The prohibition on cooperative agreements with competing networks or non-banks, i.e. co-branding, may hinder national debit card payment systems from entering into competition with MasterCard and Visa or impede retailers or other operators from entering into competition with the incumbent card issuer.

After the publication of Interim Report I in April 2006, the Commission held several meetings with banks in a number of Member States in order to analyse where self regulation is possible or whether antitrust enforcement is necessary. These talks were held in close cooperation with the local competition authorities who would then take the lead in addressing competition problems if the talks should fail. So far this approach has yielded promising results, notably in the Austrian payment cards market.

B.3. Cardholder fees

Cardholders have a contractual relationship with the card issuer, i.e. the bank whose name is on the card. By charging cardholders for card services, issuing banks can recoup the costs of services provided (e.g. transaction processing and billing) and earn a profit margin. Issuers usually charge several fees to cardholders, such as annual fees and transaction fees. Furthermore, issuers may use payment cards as a way to attract costumers to purchase other products, such as current accounts and credit, which may imply that cardholder fees are not determined in a fully autonomous manner. All these factors imply that different issuing institutions may have different pricing policies and, consequently, a comparison of cardholder fees across countries needs to be undertaken carefully.

This chapter is structured as follows:

- Section 1 examines cardholder fees for credit cards;

- Section 2 examines cardholder fees for debit cards;

- Section 3 analyses the correlation between cardholder and interchange fees; and

- Section 4 concludes.

This chapter compares four types of cardholder fees. These fees are: (1) the fee per card, which is an annual fee given in euros; (2) the card issuance fee, which is a fee charged only when the card is issued, also given in euros; (3) the fee per transaction, which is charged as a percentage or in a (euro) amount per transaction and (4) the account statement and billing information fee, which again is an annual fee (or the equivalent) in euros. These four fees were requested for a “typical cardholder” with standard/classical credit and debit cards, for each year over the period 2000-2004.

Among these four fees, the annual fee per card is the most widely used in the EU-25 Member States. In fact, our sample contains more than 600 positive observations for this fee. This compares with 320 positive observations for the issuance fee, 200 for the account statement and billing information fee and 100 for the fee per transaction. These patterns show the considerable heterogeneity in pricing policies in the card issuing business.

Figure 30: The flow of cardholder fees

[pic]

B.3.1. Credit cards

Fee per card

Looking at the simple average of the fee per card actually paid in 2004 by a typical cardholder holding a classical or standard credit card in 23 Member States[175], the results show that Diners Club is the network where cardholders pay the highest fee (57 euros). American Express charges on average 47 euros at EU level. Cardholders in the two most important international networks pay relatively the same amount of fee per card annually: 24 euros for MasterCard and 23 euros for Visa.

The inquiry compared the annual average fee across countries in 2004 between the two major networks. Figure 31 shows the country annual average for 23 EU Member States in 2004.

Figure 31:

[pic]

The figure shows that the average fee incurred by a typical cardholder for the two main credit card brands is relatively similar within the countries for which data are available. However, the average fee for both networks varies significantly across countries. The averages at EU level for MasterCard and Visa are 23 and 24 euros, although two countries report fees well above the EU average.[176]

Card issuance fee

The second most widely used fee charged to cardholders in the EU-25 Member States is the card issuance fee[177]. An analysis of the data shows that this fee is not applied in 5 Member States. In the remaining 20 countries, however, it is interesting to observe that this fee is, on average, higher for cardholders holding credit cards issued in the MasterCard and Visa networks (14 euros for both) than for those with cards issued by American Express and Diners Club (11 and 5 euros, respectively).

Account statement and billing information fee

The third annual fee analysed is the fee for account statements and billing information. An analysis of the data shows that, in contrast with the first two fees, no account statement and billing information fee is charged in the majority of Member States.

Fee per transaction

Finally, we have analysed the fee per transaction, defined either in euros or as a percentage. In 19 of the 25 countries, respondents claim that they do not charge their cardholders for each transaction they make, irrespective of the network. For the remaining 6 countries, the fee per transaction varies considerably, including between networks[178]. Where a fee per transaction is charged as a percentage of the transaction volume, this fee varies from 0.1% to 0.7% in the MasterCard network and from 0.5% to 0.7% in the Visa network.

B.3.2. Debit cards

Fee per card

As with the analysis carried out for credit cards, the fees paid by cardholders for debit cards in the EU Member states are compared here for 2004. For the purpose of this analysis, the fees charged for debit cards in the MasterCard, Visa and national debit networks in 20 EU countries are compared.[179]

Simple comparisons show that, on average, the fee per card is significantly lower for debit than for credit cards. Indeed, the fee per debit card is on average 10 euros for MasterCard (Maestro) (as against 24 euros for credit cards) and 11 euros under the Visa brand (Visa Electron) (as against 23 euros for credit cards).

Another interesting finding emerges from a comparison between the fees in international networks (MasterCard and Visa) and those for national debit networks. The average fee per card in the national schemes amounts to 9 euros (as against 10 and 11 euros for MasterCard and Visa, respectively). On average, the fee per card is significantly lower in national debit schemes than in the international debit networks.

Card issuance fee

An analysis of the data shows that this fee is not applied in most EU Member States. In the remaining 13 countries where this fee is applied, it is interesting to note that cardholders pay an issuance fee of more than 6 euros on average.

Account statement and billing information fee

An analysis of the data shows that no fee for account statements and billing information is charged to cardholders in most Member States.

Fee per transaction

In 17 of the 25 countries, respondents claimed that they do not charge their cardholders for each transaction they make, irrespective of the network. For the remaining 8 countries, results show that, for those where a transaction fee is charged as a percentage of the transaction volume, the fees vary from 0.1 % to 0.75%.

B.3.3. Correlation between cardholder and interchange fees

As indicated once again by comments received, Industry participants and mainstream economic theory suggest that increasing the level of the interchange fee would, ceteris paribus, raise merchant service charges but would lower cardholder fees through the interchange fee mechanism. That is, in the absence of interchange fees paid by acquirers to issuers, issuers would have to recoup all their costs from cardholders, with the result that cardholder fees are higher if the interchange fee decreases and lower if it increases. Such a hypothesis would be supported if a strong negative correlation could be shown between the average fee per card and the level of interchange fee for a given country and network. In order to test this hypothesis, simple correlation coefficients have been estimated between the country-average fee per card and the level of the interchange fee for the MasterCard and Visa networks over the period 2004-2004.

Table 32: Correlation coefficients between “fee per card” and “interchange fee”

Years | Visa | MasterCard |

2000 | 0.11 | -0.27 |

2001 | 0.15 | 0.20 |

2002 | 0.18 | -0. 05 |

2003 | -0.13 | 0.13 |

2004 | 0.11 | 0.05 |

Since in only 3 out of the above 10 cases, the correlation coefficient is negative, these results suggest that there is not a strong negative relationship between the level of the cardholder fee and the level of the interchange fee. This pattern is common to both networks and relatively consistent over time. The fact that the low correlation values remained relatively unchanged over time may imply that a possible increase (decrease) in the interchange fee during this period does not seem to have been passed on in lower (higher) cardholder fees.

These simple correlation coefficients do not control for other factors that may affect the fee per card level. However, an econometric estimation controlling for other variables that may affect the fee per card level shows that if the interchange fee increases by 1 Euro only 25 cents are passed on to consumers in lower fees[180]. While the exact percentage of this pass-through is, of course, difficult to estimate with complete confidence, it seems fair to conclude that this result challenges the hypothesis advanced by some industry participants and the economic literature that an increase in interchange fees is fully offset by reductions in cardholder fees. These results confirm the findings described in the chapter on profitability and may cast doubt on the relevance of the arguments put forward by industry participants and some of the economic literature on two sided markets as regards the role played by the interchange fee in this industry. Indeed, if issuers do not pass return the additional interchange fee revenues back to cardholders this implies that interchange fees are a way to transfer profits to the side of the scheme where they are least likely to be competed away.

B.3.4. Conclusions

This chapter has examined four kinds of fees charged to credit and debit cardholders:

(i) annual fees per card;

(ii) card issuance fees;

(iii) fees per transaction; and

(iv) account statement and billing information fees.

Of these fees, annual fees per card are the most important component of cardholder revenues, for both debit and credit cards. The average levels of cardholder fees charged by Visa and MasterCard in a particular Member State tend to be similar, though levels can vary substantially across Member States.

Simple correlation analysis of sector inquiry data during the period 2000 to 2004 suggest that there is no strong negative relationship between the level of the cardholder fee and the level of the interchange fee. This pattern is common to both networks and relatively consistent over time. Moreover, an econometric estimation controlling for other variables that may affect the fee per card level shows that if the interchange fee increases by 1 Euro, typically only 25 cents are passed on to consumers in lower fees.

These findings challenge the hypothesis advanced by some industry participants and some of the economic literature that an increase in interchange fees is fully offset by reductions in cardholder fees. These results are consistent with the findings of the inquiry's analysis on profitability and may cast doubt on the relevance of the arguments put forward by industry participants and the economic literature concerning the role played by the interchange fee in the payment cards industry. Indeed, if issuers do not pass return the additional interchange fee revenues back to cardholders this implies that interchange fees are a way to transfer profits to the side of the scheme where they are least likely to be competed away.

B.4. Merchant fees

The merchant service charge (MSC) is the price that a retailer (merchant) has to pay per transaction to the card acquirer, which processes the merchant’s transaction through the network and obtains the funds from the cardholder’s bank (the issuing bank).[181] Thus the MSC represents the share of retailers’ revenue from a payment card transaction that is paid directly to the card acquirer. A significant proportion of the level of the MSC derives from interchange fees (discussed further in Chapter B.5.), which is by the acquiring to the issuing bank. Other elements of the MSC cover other acquiring costs as well as a profit margin. The flow of the MSC is illustrated below in Figure 33.

Figure 33: The flow of merchant service charges

[pic]

Different acquiring institutions may have different business or pricing policies. For example, some acquirers claim that they do not extract significant (if any) profit from their acquiring business and offer it as a supplementary service to existing clients. Others, on the contrary, see acquiring as a profitable activity and by combining it with issuing, enjoy synergies and substantial profits (see Chapter B.6 on profitability analysis of issuing and acquiring). All this inevitably affects the way the MSC is negotiated between acquirers and merchants.

Some acquiring costs are not normally included in the MSC fee. For example, the majority of acquirers stated that, when leasing a terminal to the merchant, the charge for the lease typically does not constitute part of the MSC fee. A few acquirers claimed to charge terminal fees as a component of the MSC. Therefore MSC levels and the services they cover are not totally identical.

This chapter is structured as follows:

- section 1 examines MSCs for credit cards;

- section 2 examines MSCs for debit cards;

- section 3 discusses countries with highly regulated or zero MSCs;

- section 4 discusses the effect of blending and surcharges on setting of MSCs; and

- section 5 concludes.

B.4.1. MSCs for Visa and MasterCard credit cards

This section discusses:

- weighted average MSCs in international credit card networks; and

- MSCs for different categories of merchants.

Weighted average MSCs in international credit card networks

The level of the MasterCard and Visa weighted average MSC charged on credit cards was much lower than the corresponding level charged in Amex, Diners Club and JCB. Over the period 2000-2004, the lowest weighted average MSC fee was charged in the Visa network (average of 1.8%), while the highest was seen with American Express (average of 3.14%).

The country-specific analysis of the weighted average MSC credit card rates revealed a substantial variation in MSC rates across the Member States. In the Visa network, the fees varied from a low of 0.77% up to 3.10%, the difference being around 300%. In the MasterCard network, the lowest weighted average MSC was reported to be 0.95%, while the highest was 2.98%, a difference of more than 200%.

Despite some exceptions, most of the countries with relatively high MSC rates for one of the two main international networks had quite high rates for the other one. In general, the MSC levels in the two networks were quite similar within one country for most EU-25 Member States.

MSCs for different categories of merchants

As the replies of the respondents indicate (see Figure 34), in 2004 all payment card networks charged on average much higher MSC rates for credit cards to smaller merchants (as compared to larger merchants). Whereas absolute MSC levels were higher in the Amex, Diners Club and JCB networks, the relative difference in MSC rates between the two groups of merchants ( “smaller” and “larger” merchants) was considerably higher in the MasterCard and Visa networks (around 70% in both networks).

Figure 34:

[pic]

B.4.2. MSCs for debit cards

This section discusses:

- weighted average MSCs for national and international debit card networks; and

- MSC for different categories of merchants ; and

- c ountries with highly regulated or zero MSCs.

Weighted average MSCs for national and international debit card networks

As with the trend in the weighted average MSC charged to merchants on credit card transactions, the weighted average MSC levels across all debit card payment networks decreased over the 2000-2004 period (see Figure 35).The lowest average MSC level was reported in national debit networks, the highest for Maestro cards (MasterCard network) (1.17% vs 1.60%).

Across the EU-25 and over the 2000-2004 period, the weighted average MSC rate charged in national debit networks was on average 30% lower than the corresponding Visa debit MSC rate, and almost 40% lower than the corresponding Maestro MSC rate. In contrast, the average difference between the weighted average MSC rates charged on Maestro and Visa debit transactions was quite limited and amounted to only about 6%.

As with weighted average MSC levels on credit cards, the weighted average MSC charged on debit cards showed considerable variation across the Member States. According to the data, the weighted average MSC fees in the Visa network varied from a low of 0.32% up to roughly 1.9%, the difference being around 500%. In the MasterCard network, the lowest weighted average MSC was reported to be 0.36%[182], while the highest was above 2%, which is 6.5 times higher.

Figure 35:

[pic]

Maestro MSC rates tended to exceed (in some cases – significantly) those for national debit cards. For example, in the case of one country, the weighted average MSC rate in the domestic debit network in 2004 was less than half the corresponding rate in the Maestro network.

MSC for different categories of merchants

When the MSC levels charged to smaller and larger merchants for debit card transactions are analysed, it is clear that smaller merchants pay far higher MSC rates to the large international networks (MasterCard and Visa). In the national debit card networks, which carry the bulk of payment card volumes (national debit transactions sometimes account for about 80% of the total domestic card transaction volume in a given country), smaller merchants on average have to pay 7% higher MSC rates as compared to larger merchants. The difference in the fees charged to small and large merchants is about 10 times smaller in national debit networks than in the MasterCard and Visa networks (both for debit and credit cards)[183]. If the explanation for the different treatment of larger versus smaller merchants is only the scale factor, it should also apply to the national debit networks. As the difference in national debit networks is much smaller, however, it may be asked whether small merchants pay a premium for the use of the MasterCard and Visa networks and what justification there is for this.

Figure 36:

[pic]

Countries with highly regulated or zero MSCs

As argued above, acquiring institutions are believed to use MSCs mainly in order to pass on to merchants the cost of the interchange fee they pay to issuing institutions for each card transaction. It is also alleged by most market players that interchange fee payments account for the substantial bulk of total acquiring costs. Some domestic debit networks can operate without an interchange fee mechanism, but with an MSC, while nonetheless securing sufficient transaction volumes and wide card acceptance. The Netherlands, Finland are examples of such systems. Denmark is a particular case where there is no interchange and no MSC, as a result of legislation passed in March 2005[184].

B.4.3. The effect of blending and surcharges on the setting of MSCs

This section discusses:

- blended merchant service charges; and

- networks' ban on surcharge

Blended merchant service charges

‘Blending’ generally refers to the situation where the same MSC rate is offered to merchants accepting cards issued in two or more different networks, irrespective of any differences in their level of interchange fees. The analysis of the replies from acquirers revealed that blending is a frequent phenomenon across the EU-25.

According to respondents, blending is common between MasterCard and Visa card products. In other words, blending occurs between networks with similar levels of interchange fees, and therefore with similar cost components for the MSC. Due to substantial differences in interchange fees, blending between MSC rates charged on debit and credit cards is less frequent.

The data analysis shows that nearly three quarters of Member States surveyed have some form of blending of MSC rates. Furthermore, one fifth (Belgium, Denmark, Hungary, Ireland, etc.) seem to have full blending in the market: all of the responding acquirers indicated that they blend MasterCard and Visa MSC rates to 100% of their client base.

Networks' ban on surcharge

A merchant can pass the cost of accepting cards as a method of payment to the customers either by charging a fee for the use of the card, surcharging, or by including the fees in the product/service prices but granting a discount to customers paying in cash, cash discounts. Most networks refer to a clause prohibiting such surcharges and/or cash discounts as a “ no discrimination clause ”, i.e. the merchants are prohibited from applying higher prices and/or less favourable conditions to card transactions than to cash transactions. Some networks also refer to the practice of charging different prices depending on the method of payment as “ dual pricing ”.

About half of the 25 addressees of the inquiry explicitly allow surcharging and/or discounts for cash, or claim not to have any rules regulating and/or limiting such practices. National payment networks with no rules on surcharging may nevertheless be bound by those of the international payment networks when co-branding their cards with international payment card functions.

For instance, the Switch/MasterCard network in the UK allows surcharging provided the charge is advertised to the cardholder in advance and bears a reasonable relationship to the retailer’s cost in accepting cards. Likewise, retailers in Finland may charge their customers a processing fee for the use of the national debit card, Pankkikortti. As from 1 January 2005, merchants accepting the cards of MasterCard in the European Economic Area have the option of surcharging. Two of the other international networks also permit surcharging. According to the membership rules of both the latter networks, however, merchants may not discriminate between cards of these networks and other cards.

The remaining addressees of the inquiry explicitly prohibit or discourage surcharging and/or cash discounts. Although not party to the contracts between acquirers and retailers, most networks prohibiting surcharging and/or cash discounts in their network assume that a provision to this effect is included in these contracts. In Germany, the prohibition against surcharging stems from the framework agreement between issuers and merchants governing access to the electronic cash card system.

Possible sanctions for breach of the surcharge prohibition range from commercial pressure on the merchant to comply to pecuniary penalties and/or warnings with possible termination of contract. In Belgium, for instance, surcharging merchants are subject to higher merchant service charges under an oral agreement concluded in 1998 between Banksys, the banks and the Belgian Ministry of Economic Affairs. The rules of both one of the international payment schemes and one national scheme contain a “ dispute resolution mechanism ”, via which consumers can have their issuer seek, directly or via the networks, a refund of the value of the surcharge from the retail.

From a competition perspective, the surcharge prohibition restricts the freedom of merchants to pass on to cardholders the transaction costs of accepting payment cards. In its decision of 7 August 2001 (Commission’s Visa Decision)[185], the Commission gave a “negative clearance” to the Visa surcharge prohibition (i.e. the ‘no discrimination rule’) on the grounds of lack of appreciable effects[186]. This conclusion was based on studies commissioned by the Commission on the effects of lifting the surcharge prohibition in Sweden and in the Netherlands. These studies found that relatively few merchants made use of their possibility to surcharge.

In some of the literature in this field, a surcharge prohibition has been considered necessary to prevents merchants from passing on the interchange fee to cardholders, thereby stimulating the diffusion of cards[187]. It has been argued that if the prohibition is lifted and merchants can price discriminate freely, the interchange fee will no longer influence the level of payment card usage[188]. Other authors have argued that there is less need to use the interchange fee for this purpose in mature systems[189]. Concerns have also been raised that if surcharges are allowed merchants may overcharge for the use of cards, which will be difficult for consumers to verify.

B.4.4. Conclusions

Small merchants on average pay 70% more for payment card acceptance than large merchants. In theory, this could be explained by the lower average costs of acquiring merchants with higher transaction volumes. However, a comparison of price differentials between large and small merchants in the international schemes (MC/Visa: 70%, Amex 50%, JCB 40%, Diners 35%) with those in domestic systems (7% on average) suggests that scale may not be the decisive factor. It could be that smaller merchants pay a premium for accepting MasterCard and Visa cards. If that were true, the differentiation of prices according to the size of the merchant could be a measure for the exercise of market power by banks within a given system.

Merchants paying the highest average rates for MasterCard and Visa card acceptance (florists, restaurants, professional services, car rental, hotels) are typically those active in the T&E sector, where travellers expect to pay with cards, while merchants paying lower fees are typically to be found in segments with low profit margins .

Blending of merchant fees by acquirers can have direct implications for inter-network competition, as it removes an important parameter of price competition; differential MSC levels. The potential outcome of blending may be higher rates than the merchant needs to pay for acquiring services, since there is no pressure to drive down these charges through inter-network competition. Where the price difference charged in two networks is negligible, blending may be done on efficiency grounds (since it could reduce accounting and reporting requirements). However, blending would tend to nullify the effect of “multi-homing”, which, according to some claims in the academic literature, is an important countervailing force to the market power of card payment networks.

The ban on surc harging appears to restrict inter-network competition, notably by concealing the true cost of payment cards for consumers via cross-subsidisation, and may result in the use of non-optimal payment instruments. It has been suggested that the surcharging prohibition may constitute a barrier to entry for alternative non-cash payment instruments, such as mobile phones or e-money. Retailers surveyed in the inquiry were strongly in favour of the possibility to surcharge, since it would strengthen the incentive for consumers to use cheaper payment instruments.

B.5. Interchange Fees

Interchange fees are inter-bank fees paid between the payers’ and the payees’ banks for the conclusion of a payment transaction. These inter-bank fees can be agreed bilaterally between the banks or can be the subject of multilateral agreements among banks participating to a certain payment network/scheme. The setting of interchange fees and fee levels are important issues for competition in the European payment industry. The flow of these interchange fees is illustrated below in Figure 37.

Figure 37: The flow of multilateral interchange fees

[pic]

This chapter is structured as follows:

- section 1 describes the types of interchange fee applied in international networks;

- section 2 examines credit card interchange fees in international networks;

- section 3 discusses debit card interchange fees;

- section 4 summarises industry and academic views on the role of interchange fees;

- section 5 outlines the competition issues arising from interchange fees; and

- section 6 concludes.

B.5.1. Types of interchange fee applied in international networks

As already noted in the Commission Decision on Visa cross-border interchange fees[190], the international systems distinguish between three types of interchange fees: national; intra-regional; and inter-regional fees. National interchange fees apply to transactions in the same country in which the card is issued. Intra-regional interchange fees (hereafter referred to as “cross-border interchange fees”) apply to transactions that are completed at a merchant outlet outside the country but within the geographical region in which the card is issued. Inter-regional interchange fees apply to transactions between Europe and Asia or the US. These fees are not discussed in this chapter.

Interchange fees may also differ according to the method of processing (on-line, off-line, card present/not present etc.) and the type of card used (whether consumer or corporate cards; magnetic stripe card or chip card, etc).

B.5.2 . Credit card interchange fees in international networks

This section discusses:

- nominal interchange fee levels for national transactions;

- weighted interchange fee levels for national transactions; and

- weighted interchange fees for cross-border transactions.

Nominal interchange fee levels for national transactions

Country divergences as regards national interchange fees (nominal rates) in the MasterCard and Visa systems are considerable. The level of national interchange fees for Visa cards diverged by as much as 323% across the EU and by 329% for MasterCard cards. In most Central European countries, the nominal rates for MasterCard and Visa credit and debit cards are set at identical levels.

Three important conclusions can be drawn from the comparison of MasterCard and Visa national interchange fee rates. First, there are considerable differences between interchange fees from one merchant segment to another. Acquirers in the same country may pay roughly half the interchange fee for credit card payments at a petrol station than for a credit card payment to an airline. Second, country-specific differences in a given merchant segment are considerable as well. For instance, Portuguese acquirers bear roughly 165% more interchange fee cost for a credit card transaction at a petrol station than their German counterparts. Third, merchant-specific fees within the same country and the same merchant segment also differ to some extent between MasterCard and Visa.

Weighted interchange fee levels for national transactions

Comparing only nominal fee levels would not take into account the frequency with which different rates are applied in practice. Therefore, a weighting exercise was carried out. Weights were calculated according to the respective turnovers for each respective interchange fee per EU-25 Member State (see methodological details in Interim Report I Annex 1). The inquiry revealed significant variations in the weighted average of national interchange fees across the EU-25 Member States (see Figure 38[191]). The difference between the highest (above 1.5%) and the lowest weighted average fees for credit cards in 2004 was about 250%[192].

Figure 38:

[pic]

Since 2000, Visa national weighted average interchange fees have fallen gradually, while MasterCard interchange fees show no distinct trend, making it difficult to draw any precise conclusions.

Weighted interchange fees for cross-border transactions

As is apparent from the Visa website[193], Visa Europe’s cross-border (intra-regional) interchange fee scale has several categories for consumer credit cards. Meanwhile, MasterCard’s intra-European interchange fee[194] scale has ten different categories for consumer credit cards. Unlike Visa, MasterCard also publishes its cross-border interchange fees for commercial cards.

Before 2001, Visa acquirers paid on average somewhat higher interchange fees for cross-border transactions than MasterCard acquirers. As of 2001, MasterCard interchange fees started to exceed Visa interchange fees and continued to do so up to 2004, the end point of the analysis. In general terms, over 2000-2004 MasterCard acquirers seemed to pay on average 6% higher interchange fees on credit card transactions across the EU-25. Throughout the whole period from 2000 to 2004, MasterCard cross-border interchange fees kept rising (up to 2004, when they fell only insignificantly), while Visa rates followed a steady falling trend.

One of the most likely explanations for falling Visa rates is the adoption by the European Commission in 2002 of the Visa Decision[195]. This appears to have had the effect of reducing Visa cross-border interchange rates. MasterCard cross-border rates remained unregulated, which allowed the network to keep interchange fees significantly above the rates of Visa.

B.5.3. Debit card interchange fees

This section discusses:

- patterns of debit card use in Europe

- nominal interchange fee levels for national transactions; and

- cross-border interchange fee levels for debit cards.

Patterns of debit card use in Europe

National debit card transactions account for the largest and economically most significant share of card payments in Europe (see Figure 39). Historically, debit cards were often the first type of payment card to be introduced and hence customised debit card networks have evolved in a number of Member States. These networks still carry the majority of card transactions in many countries. In addition, MasterCard and Visa also offer national debit cards. These products do not have large market shares in all countries, but have been introduced as standard debit cards mainly in the new Member States of central Europe, which had historically not built up national card networks.

Figure 39:

[pic]

Source: RBR Report, 2006 (based on overall number of transaction volume).

The bulk of national debit card transactions run on national debit networks. Most national card networks offer debit cards only, with a few exceptions. These networks are country-specific, i.e. operate only within a single country, and for the most part lack interoperability with each other. In many countries, the networks were historically run by a consortium or an association of local banks, which sometimes jointly owned the network.

The interchange fee patterns in these national debit networks are highly varied, both in terms of fee structure and level of fees. Some systems use flat-rate interchange fees while others use a percentage and some use a combination of both. However, in contrast to MasterCard and Visa, most national debit card systems do not use different interchange fee “tiers” distinguishing between types of card or types of transaction.

It should be noted that POS interchange fee agreements between banks in debit card systems are not always an intrinsic feature of these systems. Table 40 below shows the EU countries where banks cooperate in payment card systems without charging one another interchange fees for POS transactions[196].

Table 40:

FI | LU | DK | NL |

Name | Pankkikortti | Bancomat | Dankort | PIN |

Nominal interchange fee levels for national transactions

It should first be noted that in some countries where Maestro and Visa branded debit cards are issued, these cards may not be relevant for national payments. Thus, national interchange fees may not be set for such cards. In some Member States[197] with established national debit systems, national debit cards are co-branded with an international debit card logo (e.g. Maestro or Visa debit) to allow mostly for cross-border operability.

While interchange fees for credit card transactions are denoted for both networks in terms of a percentage of the transaction value, interchange fees for debit card transactions are denoted differently for each network. Visa applies a fixed fee per transaction, whereas MasterCard opts for a percentage of the transaction value. A fee structure analysis shows that for an average transaction value (ATV) below 49 euros, the Visa fee mechanism generates higher interchange revenues than the MasterCard fee mechanism, while the opposite is true for an ATV above 49 euros).[198]

A comparison of nominal flat-rate fees, ad valorem fees and combined (i.e.: ad valorem and flat rate) fees necessitates a simulation. Two simulations have been conducted for a small (€5) and a medium-sized (€50) transaction value. Figure 41 below provides the result of the simulation for the value of €50, which is a good proxy for an average debit card transaction in Europe.

Figure 41:

[pic]

It appears that the level of interchange fees for eight national debit cards in the EU-25 is below those for Maestro and Visa debit cards while four national debit cards are above (though three of these four cards are insignificant in terms of transaction volumes). Similar results emerge when a €5 transaction is used as the basis for comparison. However, it is notable that for small transactions payment cards with a flat fee are more expensive than cards with an ad valorem fee.

The analysis of the weighted average national interchange fee for debit card transactions across the EU-25 reveals that up to 2002 Visa had on average a higher interchange fee for debit cards than MasterCard. As of 2002, however, the weighted national interchange fee on Maestro transactions started to exceed the fee for Visa debit, the average difference over the following three years being about 11%. Interestingly, there were almost no changes in either the Maestro or the Visa debit weighted average national fee levels up to 2002, coincidentally the year of the Visa Decision, when suddenly the Visa debit average interchange fee fell sharply, thus leading to a difference of more than 13% between the weighted average debit interchange fees in the two networks. In contrast, the Maestro weighted average interchange fee started falling only as of 2003 and at a much more moderate rate.

The results of the inquiry revealed significant variation between weighted average national interchange fees on debit cards of MasterCard and Visa. The highest weighted average interchange fees on debit card transactions are observed in some of the new Member States. If the lowest interchange fee value is excluded from the sample to avoid possible bias, the absolute difference becomes 300% in Maestro and 400% in Visa debit.

Cross-border interchange fee levels for debit cards

The patterns observed in cross-border interchange fees for debit cards to some extent mimic the patterns observed in cross-border fees for credit cards. Over the whole period, MasterCard acquirers paid on average 12% higher interchange fees on debit cards than Visa acquirers.

B.5.4. Industry and academic views on the role of interchange fees

From the responses of payment card providers, there appear to be four different means of setting interchange fees (where interchange fees are set at all). First, the scheme’s management sets interchange fees without the intervention of member banks. Second, member banks bilaterally agree on interchange fees. Third, member banks multilaterally agree on interchange fees. Fourth, member banks multilaterally agree on a fee paid by merchants to processors, who collect this fee and then transfer it to the appropriate issuing bank without the involvement of an acquiring bank. The last system is unique to one Member State (Germany).

Academic views on the role of interchange fees

Most economic analyses of payment card systems have focused on the incentives for payment card systems when they set interchange fees. Several studies have considered whether such fees promote a socially optimal choice of payment instruments[199]. Rochet and Tirole[200] (2002) compare privately and socially optimal interchange fee levels. They consider the full welfare effects of different interchange fees, also allowing for the effects on cash-paying consumers. Wright (2004) finds that privately optimal interchange fees may be too high, notably if merchant fees increase with interchange fees but issuers do not pass the additional interchange fee revenue back to cardholders. Where this rebate is not provided, high interchange fees may have the effect of transferring profits to the side of the scheme where they are least likely to be competed away, resulting in a restriction on output.

To summarise therefore, two competing assessments can be distilled from the economic literature on interchange fees in payment card systems: either that their effect is neutral and provides efficient incentives for card issuers to expand output; or that high interchange fees offer a means of transferring rent (which cannot be competed away) from acquiring to issuing banks.

Industry views on the role of interchange fees

The Commission asked international and national payment card networks to explain the economic function that interchange fees fulfil in their networks. The networks were also asked to provide the market context (for instance, in terms of the mix of different payment means, the maturity of the payment card segment, or regulation) explaining why interchange fees were used or not.[201]

One of the international payment card systems believes that in the absence of POS interchange fees paid by acquirers to issuers, issuers would have to recoup all of their costs from cardholders and this would lead to a sub-optimal level of card issuing for the system as a whole. The other system identifies an imbalance between issuing and acquiring costs necessitating a transfer of revenues from acquirers to issuers. The common feature of the MasterCard and Visa replies seems to be that both networks assume that issuing banks would not gain sufficient revenue from issuing MasterCard/Visa cards in the absence of POS interchange fees. (This assumption is challenged empirically in this report.) On the basis of this assumption, MasterCard and Visa assert that total system output would fall if card issuers were not subsidised through a transfer of revenues from acquirers.

National payment card networks often did not explain why POS interchange fees are used in their system, but referred to a declaration of banks represented on the European Payments Council. This declaration stated that interchange fees “have proven to be necessary enablers for the operation and development of the cards business and for sound cooperation between competing banks” . Amongst those national networks that commented on the purpose of POS interchange fees, opinions diverge as to the character of such interchange fees. Some view interchange fees as “ remuneration ” for services provided by issuing banks to acquiring banks, which appears similar to the purpose of interchange fees in an ATM system. Accordingly such networks often set interchange fees on the basis of the costs of the services that the participating banks provide to each other. Other networks, however, reject the idea that POS interchange fees could be “a price” and argued that POS interchange fees are “ a tool” to shift costs and revenues in a way that is neutral in terms of the overall costs and revenues incurred/charged by the banks in the system.

To summarise, a substantial number of networks maintain the “traditional” view that POS interchange fees remunerate services that banks provide to each other within the network and compensate for corresponding costs. This is reflected in the fee-setting practice of a number of networks. Other networks have adopted views which have been recently advanced by academic authors.

B.5.5. Competition issues arising from interchange fees

In a POS system, agreements on interchange fees lead to a transfer of revenues from acquirers to issuers and thereby distort price competition between acquiring banks. The Commission has in the past considered that multilaterally set interchange fees in the Visa system restrict competition between banks for providing services to cardholders and to merchants, as they largely determine the fees charged to both consumer groups. Visa interchange fees were allowed only after Visa committed itself to set interchange fees on the basis of objective costs incurred by issuers for providing concrete services to merchants; and to allow member banks to disclose these fees to merchants (cf. the Commission’s Visa Decision of 24 July 2002, OJ L 318/17 of 22 November 2002). In subsequent years, national competition authorities such as the UK Office of Fair Trading, the Spanish Tribunal for the Defence of Competition and the Italian Central Bank have concluded that interchange fee agreements infringe competition law, but that they could be allowed if the fees were set on the basis of costs incurred by issuing banks for providing card-related services.

Moreover, there are indications that the setting of interchange fees in the international systems may possibly have the object and/or effect of creating market entry barriers to competition between local and foreign member banks. Both MasterCard and Visa allow the parallel existence of multilaterally set (“fallback”) and bilaterally set (“on us”) interchange fees. While multilateral fees apply to all national payments in a given country (irrespective of the bank’s identity), bilaterally agreed fees only apply between the parties to the agreement. Therefore, in countries where local banks wish to set low interchange fees specifically for certain merchant segments (e.g.: food retail sector, petrol sector), they have a basic choice. They can either set these rates by multilateral decision in a local board or they can go through the more burdensome route of setting the same rates in several bilateral agreements between each issuer and each acquirer in a given country. Under the network rules of MasterCard and Visa, only in EU Member States where local banks set merchant-specific rates multilaterally in a local board are foreign banks able to benefit from such preferential rates. If, on the contrary, the same rates are set in a bundle of identical bilateral interchange fee agreements, the foreign bank pays higher fallback rates

A comparison of the absolute levels of MasterCard and Visa national interchange fees suggests that the relatively high level of some merchant-specific rates as opposed to others may have historical reasons and/or may be a question of market power.

Turning to the analysis of cross-border interchange fees , the evolution of MasterCard and Visa fees between 2001 and 2004 raises the question why the weighted average of MasterCard cross-border interchange fees for credit cards increased from 2002 even though Visa’s weighted average interchange fees for cross-border payments decreased from that year onwards. In other words, does inter-system competition between MasterCard and Visa act as a disciplining market force on bodies setting interchange fees in these networks? The development of MasterCard cross-border interchange fees would rather suggest that inter-system competition did not restrain MasterCard from maintaining higher cross-border interchange fees than those of Visa over more than three years (2002 to 2004). Market forces may therefore be insufficient to 'penalise' card systems with relatively high interchange fees, at least as far as fees for cross-border payments are concerned

Finally, a specific issue relevant for competition within the MasterCard and Visa systems is the co-existence of bilaterally and multilaterally agreed interchange fees. The former are often referred to in the industry as “on-us” fees. Strictly speaking, “on-us” transactions are transactions where one bank is both the issuer and the acquirer. However, in countries where an inter-bank association acquires, for instance, Maestro or MasterCard transactions, local banks that are co-shareholders of this inter-bank association may be able to offer lower fees to the association. This means that parties to these “on us” agreements can offer lower merchant fees and thereby prevent new competitors form entering a market.

B.5.6. Conclusions

The Commission’s sector inquiry provides indications that interchange fees are not intrinsic to the operation of card payment systems. Several national systems operate without an interchange fee mechanism, resulting in generally lower merchant fees.

In the international networks, Visa and MasterCard, the inquiry revealed significant variations in the weighted average of national credit card interchange fees across the Member States. In 2004 the level of the highest fees (over 1.5% of transaction value) was two-and-a-half times greater than the lowest weighted average fees. For Visa and MasterCard debit cards, the highest fees were observed in some of the new Member States. For Maestro cards weighted average interchange fees were more than three times higher in some Member States than others, and more than four times higher in some Member States than others for Visa debit cards.

The use of interchange fees may serve several purposes. Card payment networks argue that, given the typical set-up of card payment mechanisms, the card issuers typically bear the main costs of the payment system, while most of the revenues are collected on the acquiring side as merchant fees. Therefore, they claim that there is a need to redress cost imbalances by an interchange fee mechanism, i.e. a fee paid by the acquirers to the issuers. Other systems argued that interchange fees are a co-ordinating mechanism necessary to optimise the operation of four-party payment card systems.

Two competing assessments can be distilled from the economic literature on interchange fees in payment card systems: either that their effect is neutral and provides efficient incentives for card issuers to expand output; or that high interchange fees offer a means of transferring rent (which cannot be competed away) from acquiring to issuing banks. From a competition viewpoint, it is important to assess whether interchange fees are used to extract rents from merchants. Some of the inquiry's findings – in particular concerning large divergences in interchange fees between countries and between merchant segments – may provide indications that the setting of interchange fees could be subject to the exercise of market power in some Member States.

By concluding and acting on a basis of preferential interchange agreements, incumbent players, involved in both issuing and acquiring activities, may indirectly obstruct new entry to the acquiring by not extending the same favourable conditions to newcomers.

B.6. Profitability of the payment cards industry

The profitability of a business may provide important information for a competition analysis. On one hand, the existence of significant economic profits may be the reward for taking risks and for innovating and/or it may be the reward for superior efficiency and better management. On the other hand, high profits may also be the result of having and exerting market power, in particular if profit margins remain high over a long time period in a relatively mature market. The sector inquiry therefore analysed to what extent issuing and acquiring are profitable and how profits developed during the period from 2000 to 2004. The second purpose of the inquiry's profitability analysis was to assess to what extent the profitability of the issuing business depends on revenues generated by interchange fees.

This chapter is structured as follows:

- Section 1 outlines the methodology for the inquiry’s profitability analysis;

- Section 2 examines the profitability of credit card business;

- Section 3 examines the profitability of debit card business;

- Section 4 discusses the findings on profitability in the payment cards industry; and

- Section 5 concludes.

B.6.1. Methodology for the inquiry’s profitability analysis

This chapter provides a descriptive comparison of profitability trends for issuing and acquiring businesses in credit and debit cards for all EU-25 Member States. This analysis covers the period 2000-2004. Looking at profitability may yield important information for competition analysis. In fact, while the existence of significant rents may be the reward for taking risks and innovating, superior efficiency or better management, it could also be the result of having and exerting market power. High and persistent rents in relatively mature markets where some prices, such as interchange fees, are determined collectively may suggest the latter. These findings, together with other evidence obtained by this inquiry, may reveal whether a firm or a group of firms is exercising market power to the detriment of consumers in a particular market. Taking advantage of the detailed data set available, this chapter also aims at examining further the role played by the interchange fee in a “two-sided” industry.

The measurement of the profitability of a specific activity is typically subject to problems related to the allocation of costs that are common to other activities. This could be also relevant in our case, because acquirers and issuers (which may be multi-product firms) may also carry out other activities[202]. However, it is worth noting that the allocation of revenues and costs, based on accounting data, was made by the respondents. Consequently, the computation of key cost and revenue parameters by the respondents themselves reduces significantly the degree of uncertainty as to their true level. Moreover, the revenues and costs are not separated by the different payment systems in which acquirers and issuers participate, further decreasing this uncertainty.

For the purpose of the inquiry, both issuing and acquiring institutions were requested to report their total revenues and total costs associated with the issuing and acquiring of credit and debit cards. The questionnaire provided a breakdown of the most relevant parameters for total revenues and total costs. In the acquiring business, total revenues comprise: merchant service charges; terminal processing fees; currency conversion fees; and “other type of incomes”. Total acquiring costs comprise interchange fees; transaction processing costs; and “other type of costs”. In the issuing business, total revenues comprise: interest charged; interchange fees; cardholder fees; currency conversion fees; income from co-branding; and “other type of incomes”. Total issuing costs comprise: costs for the provision of a free funding period; card production and transaction processing costs; billing; fraud; credit losses; costs related to rebates; staff costs; and “other type of costs”. The parameter “other type of income/cost” aims to capture any other relevant type of income or cost in the acquiring and issuing of cards, as perceived by the respondents, which does not fall under the other categories. Costs related to the depreciation of assets, for instance, could be included in this category.

In order to investigate the magnitude and evolution of profitability payment card business in the Member States over the period 2000-2004, we make use of a simple profit-to-cost ratio. The operational profit-to-cost ratio before taxes (hereafter “profit ratio”) in the acquiring of credit cards by acquirer B in country A at time t is given by:

[pic]

This measure of profitability is used throughout this chapter.

B.6.2. The profitability of credit card business

This section discusses:

- the profitability of credit card acquiring;

- the profitability of credit card issuing;

- the profitability of credit card issuing and acquiring together; and

- the relationship between credit card issuing profitability and interchange fees.

Acquiring business for credit cards

Of the 83 respondents that reported figures for the acquiring of credit cards in 2004, 52 reported a positive and 31 a negative profit ratio. In order to avoid giving equal weight to both small and large acquirers in the determination of the overall country profit ratio, this ratio is a weighted average of all the acquirers in the country in question (the weight is given by the total income of acquirers). Figure 42 below shows the weighted profit ratio for credit card acquiring in 2004.

Figure 42:

[pic]

The figure shows that credit card acquiring was profitable in 19 of the 25 EU countries in 2004. The profit ratios vary from -16% to 62%.[203] It can also be observed that 9 countries are above the EU-25 weighted profit ratio (15.9%).

Issuing business for credit cards

Of the 136 respondents reporting figures for the issuing of credit cards in 2004, only 20 issuer institutions reported a negative profit ratio. The weighted mean profit ratio was 65%, with a median profit ratio of 61.4%.The third quartile of issuers reported profitability between 61.4% and 131.8%, while the fourth (highest) quartile reported figures above 131.8%. Of the issuers whose profitability was below the median value, the first (lowest) quartile reported profit ratios between -50% and 14.7% while the issuers in the second quartile reported profitability between 14.7% and 61. 4%.

As with the acquiring business, we have also carried out an analysis of the profitability of the issuing of credit cards at country level for 2004. Again, in order to avoid giving equal weight to small and large issuers in the determination of the overall profit ratio, the country profit ratio is a weighted average of all the issuers in the country in question (the weight is given by the total income of issuers). Figure 43 shows this weighted average profit ratio for all Member States for 2004.

Figure 43:

[pic]

The figure shows that the income generated by the issuing of credit cards is higher than the associated costs in all 25 Member States. The weighted average profit ratios vary from 3% to 147%. The EU-25 weighted average is 65%, with 11 countries above this figure.

As with the data on acquiring, some of these profit ratios are based on only a limited number of observations, implying that results may not be entirely representative of profitability for a given country. However, with the exception of two countries, the number of observations per country is clearly higher for issuing than for acquiring.

It is interesting to assess the degree of variability of profit ratios within a country, i.e. whether profits in each country are evenly distributed among issuers. Taking as examples the four countries with a high number of observations and displaying a weighted average profit ratio above the EU-25 average, we can observe a fairly similar pattern as regards profit ratios: while almost all issuers reported positive profit ratios, there are some discrepancies in the profit ratios of some top issuing institutions. Looking in detail at the data, it is possible to conclude that the differences in cost structures may explain to a large extent the discrepancies in profit ratios among top issuing institutions. Therefore, the differences observed in the profit ratios of some top issuing institutions in the same country seem mainly to reflect a different level of efficiency and not a fierce competition on prices.

As with the acquiring of credit cards, it is necessary to analyse further the dynamics of the profit ratios at country level over the period 2000-2004 in order to detect the influence of different stages of the business cycle in each country. It is apparent that profit patterns were relatively consistent over this period in almost all countries[204]. More importantly, these results suggest that, as for the acquiring business, the magnitude of profit ratios is not related to the different stages of the business cycle in each market but rather follows a medium-term trend.

In summary, credit card issuing is highly profitable in the large majority of the EU-25 Member States. The persistence of high profit ratios over a sustained period suggests that this might be the result of having and exerting market power.

Profitability of credit card issuing and acquiring together

According to the mainstream theory on two-sided markets, where benefits arise on two sides of the industry, there may be no meaningful economic relationship between benefits and costs on either side of the market considered by itself. Thus, it is relevant to analyse the profitability patterns of both acquiring and issuing together, and to explore the findings in the light of this theory. Figure 44 shows the weighted average profit ratios of acquiring and issuing credit card businesses for all EU Member States for 2004. An important finding is that issuing business is clearly more profitable than the acquiring business in almost all countries.

Figure 44:

[pic]

The relationship between credit card issuing profitability and interchange fees

Assessing the extent to which the profitability of issuing business depends on the revenues generated by the current level of interchange fees may provide further insight into the accuracy of the two different theoretical predictions and also the pertinence of the arguments put forward by the payment card systems. Another strand of economic literature suggests that privately optimal interchange fees may be too high, notably if merchant fees increase with interchange fees but issuers do not pass the additional interchange fee revenue back to cardholders. In this case, high interchange fees are a way to transfer rents to the side of the scheme where they are least likely to be competed away (see Wright, 2004, op. cit. and Bergman, 2005, op. cit.).

Of the 136 institutions reporting data for the credit card issuing business for 2004, 118 also reported data on the total revenue obtained through interchange fees. It should be noted that 100 of these 118 reported positive profit ratios. In order to quantify the importance of the interchange fee in the total income of the issuing institutions, we have carried out an additional exercise. We compared the total income of these 100 issuers, i.e. including the part generated by interchange fees, with the income that the same issuers would obtain if the interchange fees were taken out from their total revenue (which is equivalent to at par clearing ).

This exercise reveals that if that part of total income due to interchange fees were to be taken out, 62 of the 100 institutions reporting positive ratio profits would nevertheless remain profitable[205]. These findings may partly be explained by the likelihood that the income from cardholder fees and interest may make issuing profitable anyway. Figure 45 shows the country weighted average profit ratio for the 118 issuers, both when that part of revenue due to interchange fees is included in total revenue and when it is excluded.

Figure 45:

[pic]

The fact that a high number of issuing institutions remain profitable in the extreme situation of a “zero” interchange fee is relevant. From our exercise, it can be concluded that in 20 of the 25 countries, the interchange fee significantly adds to the positive level of profits in the credit card issuing business that would be obtained anyway with zero interchange fees.

This exercise seems to partially invalidate one of the main results of the theoretical models described in the chapter on the economic literature, which suggest that a positive “optimum” level of interchange fee is needed because price market mechanisms fail to internalise the existing externalities, with the result that total system output would suffer if issuing were not subsidised through the transfer of revenues from acquirers. The aim of this analysis is not to argue in favour of a zero interchange fee. However, in the light of the results, it is legitimate to question the “optimality” of the current level of interchange fees in several countries. Our findings seem to confirm some recent theoretical predictions of the two-sided market literature, which suggest that privately optimal interchange fees may be too high, notably if merchant fees increase with interchange fees but issuers do not pass the additional interchange fee revenue back to cardholders.

Similarly, these results also seem to cast doubt on the justifications for the existence of interchange fees put forward by the payment card systems. For instance, one international network believes that in the absence of POS interchange fees paid by acquirers to issuers, issuers would have to recoup all of their costs from cardholders and this would lead to a level of card issuing that is “not optimal” for the system as a whole. In addition to the fact that system optimality does not mean welfare optimality, this statement seems to be largely refuted by our results. The justification put forward by another international network, which considers that the interchange fee provides for a transfer of revenue between issuers and acquirers to achieve the optimal delivery of services by both acquirers and issuers to merchants and cardholders, is also not supported by our results. For instance, looking at country U in Figure 44, it can be seen that the issuing of credit cards is much more profitable than acquiring (which is even negative). Moreover, Figure 45 shows that issuing credit cards in country U would still be profitable even with no interchange fee.

B.6.3. Profitability of debit card business

This section discusses:

- the profitability of debit card acquiring;

- the profitability of debit card issuing; and

- the profitability of debit card issuing and acquiring together.

Acquiring business for debit cards

Of the 53 respondents that reported figures for the acquiring of debit cards in 2004, 30 reported a positive and 21 a negative profit ratio. The overall weighted average profit ratio is 5% (weights are given by the total income of acquirers), which suggests that the acquiring of debit cards is on average significantly less profitable than the acquiring of credit cards at EU level. Turning to the analysis of profitability at country level, Figure 46 shows the country weighted profit ratios in the acquiring of debit cards for 2004.

Figure 46:

[pic]

The figure shows a strong variation in the weighted profit ratio across countries. The weighted country profit ratio varies from -32% to 35%. Debit card acquiring is only profitable in 12 out of 19 countries in the EU in 2004.

As with the analysis of credit cards, some of these profit ratios are based on only a limited number of observations, which again means that results may not be entirely representative of profitability for a given country. In certain cases, the sole observations in the sample are representative of the country, given that they are from large and specialised acquiring institutions. However, some caution is necessary in cases where the country figure is based only on a small acquirer.

Issuing business for debit cards

Four countries were excluded from the analysis due to data unavailability. Of the 71 respondents reporting figures for the issuing of debit cards in 2004, 21 reported a negative profit ratio. Seventeen issuer institutions reported a profit ratio below -10%, 18 reported a profit ratio between -10% and 33%, 18 between 33% and 120% and 18 above 120%.

If one also takes into consideration that the weighted profit ratio of all respondents for 2004 is 47% (weights are given by the total income of debit card issuers), it seems reasonable to conclude that the debit card issuing business is profitable. As for credit cards, a simple comparison shows that issuing for debit cards is significantly more profitable than acquiring.

As with acquiring, we also carried out an analysis of the profitability of issuing at country level for 2004. Again, in order to avoid giving equal weight to small and big acquirers in the determination of the overall country profit ratio, the country profit ratio is a weighted average for all issuers of debit cards in the country in question (the weight is given by the total income of debit card issuers). Figure 47 shows this weighted average profit ratio for 21 Member States for 2004.

Figure 47:

[pic]

As shown in Figure 47, the income generated by the issuing of debit cards is higher than the associated costs in 19 out of 21 Member States. The country weighted average profit ratios vary from -20% to more than 140%. It seems unquestionable that for countries S and T , the existence and persistence of a very high profit ratio over a relatively long period of time seems to follow a medium-term trend. Unfortunately, it is not possible to draw any conclusion regarding country U because the sample does not contain observations for the entire period.

Profitability of debit card issuing and acquiring together

An analysis of Figure 48, which gives a comparison of weighted average profit ratios for the issuing and acquiring of debit cards across countries, shows that, as for the credit card analysis, issuing is clearly more profitable than acquiring in almost all countries.

Figure 48:

[pic]

The reason why issuing is several times more profitable than acquiring in some countries, e.g. country S , is again because of the role played by the interchange and cardholder fees. In fact, while the level of the MSC charged by an acquirer in country S to merchants is very high, the amount of interchange fee per transaction, which is transferred to the issuing side, erodes its profit margin. This also confirms the results in the chapter on merchant service charges, where it is shown that this country had the highest weighted average interchange fee as a proportion of the weighted average merchant service charge in 2004. As regards the debit card issuing industry in country S , one can observe that the two largest issuers receive revenue from the interchange fee that is 5 times higher than the cost associated with the transactions carried out. Additionally, these issuers also receive considerable revenues from cardholder fees. Consequently, the issuing of debit cards is very profitable in this country.

B.6.4. Findings on profitability in the payment cards industry

The analysis of profit ratios in POS payment card systems may provide valuable information for a competition analysis. In this respect, the inquiry has made several important findings. Firstly, the issuing of credit cards is very profitable. On a pan-EU scale, credit card issuers had a weighted average profit-to-cost ratio of 65% in 2004 while debit card issuers had a weighted average profit ratio of 47%. In most EU Member States, the weighted average profit ratios remained fairly stable over the period 2000 to 2004. Secondly, interchange fees appear to magnify these profits. It appears that 62% of all banks surveyed would still make profits with credit card issuing even if they received no interchange fee revenues. In 23 EU Member States, at least one bank participating in the survey was able to make a profit from issuing credit cards without interchange fees. Thirdly, the reason why issuing is considerably more profitable than acquiring in the majority of countries is the role played by the interchange fee as a cost and revenue element in the payment card system.

Industry comments on the inquiry's profitability analysis

In their responses to the public consultation on Interim Report I, banks and networks criticised the methodology used to assess profits. The argued firstly that the inquiry's methodology failed to consider the opportunity cost of capital and risks, thus overestimating the extent of profitability. Secondly, it was argued that that the profit ratio produced biased profitability estimates because of the absence of a uniform cost-allocation methodology. Thirdly, it was argued that the inquiry's profitability methodology may not have accurately reflected the relative profitability of issuing and acquiring. Finally, it was argued that the analysis of levels of profitability failed to compare the observed profit ratio with an appropriate benchmark.

The calculation of revenues and costs was based on the figures provided by the respondents, using a common standard profit and loss statement. This profit and loss statement included an item called “other types of costs” that allowed banks to account for any other relevant type of cost that does not fall within one of the other categories. All types of relevant business costs including capital costs could therefore have been included in this category. Moreover, the profitability analysis did not aim at providing a complex accounting exercise, or to define “reasonable” and “excessive” profits.

Also, it is worth noting that the same measure of profits was consistently applied across all respondents in order to compare country disparities. Firstly, it is not clear that possible inaccuracies in the methodology should bias the profitability results of particular Member States and not others. Secondly, the extent of variation in profitability across the Member States is unmistakable: the results show that the profitability of issuing activity may be more than 10 times higher in some countries than in others. Moreover, the results show that high profits not only have been sustained over time but that they are also correlated with high fees and with some specific market structures such as acquiring joint ventures.

Comments on possible overestimations of profitability arising from the inquiry's methodology appear most relevant to card issuing. The inquiry's estimates show generally modest profitability in acquiring of both credit and debit cards (showing an EU average profit ratio of 15.9% and 5% respectively in 2004). The implication of overestimation in acquiring profitability would be that 'true' ratios would on average be close to zero and negative in a considerable number of Member States. It was argued that capital costs would be likely to be higher in issuing than on the acquiring side, as a result of banks' need to provision against default by cardholders. Thus, to the extent that issuers failed to accurately report their capital costs, the inquiry's estimates may underestimate total costs and overestimate profitability of card issuing. However this would not significantly alter the inquiry's findings on the relative profitability of issuing and acquiring activity; nor on the impact of interchange fees on the profitability of both sides of the market.

Some banks also argued that in order to evaluate the effects of abolishing interchange fees, complex simulations and model calculations were necessary. It should be noted, however, that a complex, dynamic simulation exercise clearly goes beyond the scope of a sector inquiry. The findings on the profitability of payment card issuing clearly cast doubt on the assumption that in the absence of interchange fees, issuers could not recoup their costs from cardholders. These findings are sufficiently robust, even if not obtained with model calculations. Nonetheless, the sector inquiry does not exclude that systems may be more efficient with positive interchange fees.

The question whether card issuers can offer payment cards at affordable prices to consumers in the absence of interchange revenues is of relevance for a competition analysis of interchange fee agreements. If the multilateral transfer of revenues were necessary for the operation of a payment card system, then multilateral interchange fee agreements may not be caught by Article 81(1) EC, even if the fees determine the prices charged by an acquirer to merchants. However, the above findings on the profitability of payment card issuing cast doubt on the assumption that in the absence of interchange fees, issuers could not efficiently recoup their costs from cardholders.

This observation does not exclude, that the use of interchange fees may lead to certain efficiencies in the operation of a POS system. However, it does seems to confirm some recent theoretical predictions in the literature on two-sided markets suggesting that privately optimal interchange fees may be too high; notably if merchant fees increase with interchange fees but issuers do not return the additional interchange fee revenue to cardholders. In this case, high interchange fees may provide the means to transfer rents to the issuing side of the scheme, where they are least likely to be competed away.

B.6.5. Conclusions

Credit cards issuing is highly profitable. On a pan-EU scale, the inquiry estimates that credit card issuers had a weighted average profit-to-cost ratio of 65% in 2004 while debit card issuers had a weighted average profit ratio of 47%. In most EU Member States, weighted average profit ratios remained fairly stable over the period 2000 to 2004.

Interchange fees appear to magnify the profits of card issuers. It appears that 62% of all banks surveyed would still make profits with credit card issuing even if they did not receive any interchange fee revenues at all. In 23 EU Member States, at least one bank participating in the survey was able to make a profit from issuing credit cards without interchange fees. This exercise seems to partially invalidate explanations put forward by the industry that total system output would suffer if issuing were not subsidised through the transfer of revenues from acquirers. The aim of this analysis is not to argue in favour of a zero interchange fee. However, in the light of the results, it is legitimate to question the optimality of the current level of interchange fees in several countries.

Industry comments on possible overestimations of profitability arising from the inquiry's methodology appear most relevant to card issuing. It was argued that capital costs would be likely to be higher in issuing than on the acquiring side (e.g. as a result of banks' need to provision against default by cardholders). Thus, to the extent that issuers failed to accurately report their capital costs, the inquiry's estimates may underestimate total costs and overestimate profitability of card issuing. However this would not significantly alter the inquiry's findings on the relative profitability of issuing and acquiring activity; nor on the impact of interchange fees on the profitability of both sides of the market.

The inquiry has found high and persistent profit ratios in relatively mature markets, together with other evidence collected on entry barriers, suggesting the existence and exercise of market power in these markets. The question whether card issuers can offer payment cards at affordable prices to consumers in the absence of interchange fee revenues is also relevant for a competition analysis of interchange fee agreements. If the multilateral transfer of revenues were necessary for the operation of a payment card system, then multilateral interchange fee agreements may not be caught by Article 81(1) EC, even if the fees determine the prices charged by an acquirer to merchants. However, the above findings on the profitability of payment card issuing cast doubt on the assumption that in the absence of interchange fees, issuers could not recoup their costs from cardholders.

These observations do not exclude that the use of interchange fees may lead to certain efficiencies in the operation of a POS system. However, they seem to confirm some recent theoretical predictions in the literature on two-sided markets suggesting that privately optimal interchange fees may be too high from a welfare perspective; notably if merchant fees increase with interchange fees but issuers do not return the additional interchange fee revenues to cardholders.

B.7. Free funding periods for payment card transactions

This chapter discusses funding periods for payment card transactions in different networks throughout Europe. Banks may delay the settlement of a card transaction on the current account of a customer by days or even weeks. To the extent that cardholders are not charged interest for this time period (“free funding period”), an issuing bank incurs costs. It has been argued that these costs should be co-financed by merchants through an interchange fee as cardholders make greater use of cards in their shops if they benefit from a free funding period. From a competition viewpoint it is therefore interesting to examine the extent to which banks indeed incur costs by delaying the settlement of a card transaction.

This chapter is structured as follows:

- Section 1 defines the main parameters affecting funding periods for card transactions;

- Section 2 analyses the free funding period and net float per card brand;

- Section 3 analyses the free funding period and net float in the Member States; and

- Section 4 concludes.

B.7.1. Main parameters affecting funding periods for card transactions

This chapter analyses:

- the average number of days between authorisation of a POS card transaction and deduction of the money from a cardholder’s bank account; and

- the average number of days that card issuing banks delay the transfer of funds to the acquiring bank.[206]

Data presented here are the based on the responses of 114 banking groups across the entire European Union. Data from VISA and VISA Electron and MasterCard and Maestro have been averaged together. The averages given are simple arithmetical averages.

We will use the following definitions:

- Free funding period: the time delay (measured in days) between the time a POS transaction is authorised and the time the issuing bank debits the cardholder’s bank current account;

- Transfer period: the time delay (measured in days) between the time a POS transaction is authorised and the time the issuing bank transfers the corresponding funds to the acquiring bank; and

- Net float: the sum of the ‘transfer period’ minus the ‘free funding period’ (measured in days).

From the perspective of the issuing bank, net float occurs if there is a divergence between the time it debits the cardholder and the time it transfers money to the acquirer. Net float may be positive, zero or negative. It is:

1. negative if the bank debits the cardholder after the transfer

2. positive if the bank debits the cardholder before the transfer

3. zero if the bank debits the cardholder the same day the transfer occurs.

Where the net float duration is positive, the issuing bank has the opportunity to earn a return on the transfer amount. Where the net float duration is negative, the issuing bank is required to advance the transfer amount, which will create a cost for the issuing bank.

B.7.2. Analysis of free funding period and net float per card brand

The figure below sets out the average free funding periods (also referred to as “grace periods”) for payment card transactions with the main card brands in the EU: VISA, MasterCard, VISA Electron, as well as some selected national debit cards such as EC Cash (Germany), Multibanco (Portugal), Karanta (Slovenia) and Laser (Ireland).

Figure 49:

[pic]

Figure 50:

[pic]

Together Figures 49 and 50 show that:

4. On a pan-European scale, VISA cards have an average free funding period of 14.87 days and the net float financed by issuing banks is -13.62 days on average. This compares to an average free funding period of 8.74 days for MasterCard branded cards, where the net float is -6.68 days.

5. Banks finance a small part of the costs of issuing debit cards through positive float.

6. Exceptionally, banks may achieve positive net floats even with credit cards.

B.7.3. Analysis of free funding period and net float in the Member States

This section discusses:

- average transfer periods;

- average free funding period; and

- average net float.

Average transfer periods

On a pan-EU scale, there are some divergences between the transfer periods for international cards. These periods range from a minimum of 0.76 days (Portugal) to 3.07 days (Germany) and the average is closed to 1.8 days. In most EU countries, transactions with one of the international cards are debited on the second day after the transaction. A comparison with national cards shows a similar picture. Here as well, there are differences on a pan-EU scale ranging from zero days (NL) to 8.9 days (Slovenia) and the average is also closed to 1.8 days. In most instances, however, transactions with national cards are settled the day after the transaction or on the subsequent day.[207] Figures 51 and 52 compare the average transfer periods in the international networks and domestic debit networks respectively across the EU.

Figure 51:

[pic]

Figure 52:

[pic]

Average free funding period

The differences between EU Member States are much more pronounced for free funding periods than for transfers periods. A comparison of all international cards (taken together) shows periods ranging between 1.36 days (SK) and 20.33 days (LU). It appears that free funding periods for international cards are particularly short in the Central European countries while they are long in Western Europe (the Netherlands: 18.17, Germany: 15.93, Italy: 12.82 days).

The average pan-EU free funding period on national cards (3.90 days) is nearly half the length of the average pan-EU free funding period for international cards (7.39 days). On a country by country basis, divergences between national systems are marked, with free funding periods ranging from 0.6 days (BE) to 8.13 (SI).

Average net float

The net float is the difference between the length of the free funding period and the transfer period. Owing to the considerable divergences in free funding periods across Member States, there is a significant variation in the average net float for international cards. Latvia has the highest positive figure with 0.34; whereas Belgium has the largest negative float with -19.65 days. On average, EU banks bear -5.57 days net float for each transaction with an international card in the EU. Figures 53 and 54 compare the net float durations in the international networks and domestic debit networks respectively across the EU.

Figure 53:

[pic]

Figure 54:

[pic]

It would appear that many banks in Central and Eastern European Member States treat MasterCard, VISA, Visa Electron and Maestro equally with regard to the transfer period, the free funding period and net float. In other words, there often appear to be no appreciable differences in the technical product characteristics of various card brands regarding these features. The data show that banks in eight Member States do not distinguish between card brands in terms of the free funding period. This group includes five Central and Eastern European countries along with Cyprus, Greece and the UK.

B.7.4. Conclusions

On average, acquirers in the EU receive the funds of the second day after the transaction. On a pan-EU scale, the free funding period for international cards varies according to the card brand. MasterCard branded (credit and charge) cards carry roughly half of the free funding period of VISA (credit and charge) cards. VISA and MasterCard cards in turn both typically carry longer free funding periods than VISA Electron and Maestro branded cards. Contrary to this general trend, banks in eight EU Member States (in particular Central and Eastern European countries) appear to treat all international card brands equally in terms of free funding periods.

The average free funding period for national payment cards (Bancontact/Mr. Cash, PIN, Bancomat, Pagobancomat etc.) is only half as long as that of international payment cards (VISA, VISA electron, MasterCard, Maestro)

Hardly any bank surveyed delayed the transfer of money to the merchant’s bank long enough to recoup the entirety of its costs for funding delayed payment. In national schemes, banks delayed the transfer of funds to the merchant’s bank by between zero days (NL) up to 8.9 days (SI) and in the international schemes (all brands) they delayed the transfer by between 0.76 days (PT) and 3.07 days (DE). This compares to average free funding periods ranging, for national cards, between 0.6 days (BE) and 8.13 days (SI) and, for international cards (all brands), between 1.36 days (SK) and 20.33 days (LU).

B.8. Membership and governance rules

Membership conditions and joining fees in payment card networks may under certain circumstances impede new entrants from joining the network. The analysis of the membership conditions revealed that the requirements such as those related to being a financial institution and having a local establishment may create market barriers. Similarly, joining fees – even though as such they do not raise competition concerns – under certain circumstances may hamper or even hinder effective intra-system competition by dissuading entrants or raising their costs significantly.

This chapter is structured as follows:

- Section 1 examines membership conditions in payment card networks

- Section 2 examines network joining fees;

- Section 3 discusses governance in payment card networks; and

- Section 4 concludes.

B.8.1. Membership conditions in payment card networks

This section considers two types of membership conditions in payment card networks which may reduce market entry and thereby restrict competition. These conditions are:

- the financial institution requirement; and

- the local establishment requirement.

Financial institution requirement

Both international and domestic card payment systems reserve the issuing of cards and acquiring of merchants, to credit institutions or entities controlled by credit institutions.

International systems

One of the international systems restricts membership to financial institutions that are organised under the commercial banking laws of its own country and licensed to accept demand deposits or which are controlled by another such organisation.

The other system likewise reserves membership to financial institutions, which are defined as entities authorised to engage in financial transactions under the laws of the country where they principally engage in business. The concept of “financial institution” in the latter case is somewhat wider than that of “credit institution” within the meaning of Article 1 of Directive 2001/12 of 20 March 2000, as it also includes entities that do not take deposits, but which substantially conduct all of their business by executing “financial transactions”. It likewise allows non-credit institutions to apply for membership if banks are “directly or indirectly” controlling them.

Domestic systems

With the exception of one domestic system, all card systems surveyed stated that they reserve merchant acquiring to credit institutions. In Germany, however, network service providers (“Netzbetreiber”) may in practice act as acquirers in the domestic debit card system “EC Karte”. By providing collection services (i.e.: the payment does not go via the merchant’s bank), the processor takes on the settlement risk vis-à-vis the merchant. Thus, there is no network rule preventing Netzbetreiber from handling the financial aspects of the payment business and merchant service providers offer acquiring services to merchants without de-stabilising effects on the system.

If the proposed Directive on the payment services in the internal market[208] enters into force, payment service providers like the German network service providers should gain access to the acquiring business in other POS systems across Europe.

Local establishment requirement

International card payment systems allow banks to operate cross-border without establishing a physical presence in the country where they issue cards and/or acquire merchants. Physical presence may, however, be a prerequisite to operate in domestic card payment systems, as some of the domestic schemes and seemingly also some central banks legally or de facto (e.g. under the requirement of supervision by the national central bank) require financial institutions to have a physical presence in order to participate in the domestic payment card systems of their country. To the extent that these rules are based on legislation or decisions by the national central bank, the requirement may moreover be in contradiction with the freedom of services and internal market banking directives.

International systems

In one of the international systems, both principal and associate members may apply for central and cross-border issuing and/or acquiring licenses, which enable these banks to offer their services without having a physical presence in the “host” Member State. Member banks in principle can issue cards outside their home country provided that the associated activities are carried out in the member’s home country. The principle members can moreover apply to obtain a “branch license” for their branch in another EU Member State. However, a subsidiary of a principle member needs to join the system as a separate entity.

In the other international system, banks may not acquire merchants outside the area of use of their license, but may either ask for an extension of this area of use or apply for a so-called “Central Acquiring license”, which then enables them to acquire merchants centrally or cross-border in countries outside the country where they are established.

Domestic systems

Among the responding domestic systems, three confirmed that admission to the scheme is conditioned on the physical presence in the country where the banks issue cards and/or acquire merchants. This requirement is usually a result of other conditions imposed either by a scheme itself or the central banks:

- admitting only credit institutions that “carry out retail banking in respect of accounts domiciled in [country A] to participate in the [..] Scheme” (requirement of the scheme); or

- participation reserved to credit institutions that are registered with and/or supervised by the national central bank (requirement of scheme or national central bank).

These requirements would need to be further investigated as it may imply a local establishment requirement contrary to internal market rules and impede competition by excluding foreign banks from issuing cards and acquiring card.

Other systems do not require banks to establish a physical presence prior to joining their system. One national system reported that it admits any banks certified by any of the central banks of the European Union. Another national system also reported that it did not require the physical presence of a financial institution in their country and that some members indeed were foreign banks.

B.8.2. Network joining fees

As explained above, networks can be either open for membership to independent financial institutions (“open systems”) or not (“closed systems”). In most open systems, either the scheme owners and/or the network operators charge some sort of fee for an institution to join the system (“joining fee”), which can be one-off or recurring.

This section first discusses the factors determining joining fees in payment card networks. Secondly, it examines the level of joining fees in payment card networks across the EU.

Factors determining joining fees

On the basis of the replies of the open systems, the joining fees seem to vary depending on a number of factors such as:

- type of a membership,

- type of service used by members,

- activities of the members (i.e. issuers and/or acquirers),

- assets of the member or a proportion of the market participation,

- shareholder capital (where the level of joining fee is linked to the shareholder value or subscription and payment of certain amount of shares),

- number of ATM and/or POS terminals in the network (the sole example being Bancontact scheme in Belgium),

- number of cards that the new member will issue over the first five years or the number of cards issued by all members at the time of member entry, depending on whether the new member is an issuer or a collector,

- projected number of cards to be issued during a certain period.

In Denmark and Finland additional fee has to be paid in order to join the inter-bank agreements administered by the respective bankers’ associations. In Denmark joining institutions have to pay a licensing fee and a fee to join PBS A/S for processing and clearing of Dankort-transactions. In France, for the use inter-bank service, the e-rsb (“réseau des services aux banques”), every new member of the CB system has to pay a joining fee.

Finally, some payment systems also require an international licence. For instance, in addition to a flat-rate joining fee, the UK network Switch/Maestro requires a Maestro licence. Similarly, members of one of the Italian networks must de facto be Visa Participant/MasterCard Affiliate Members.

Level of joining fees

The level of joining fees varies considerably across the Member States as regards the domestic payment systems, whereas the joining fee for the two international payment systems increases with the assets of the member.

International systems

Compared to the level of joining fees in the open domestic payment systems those for the open international ones are in the middle range. The joining fee for two of the international payment systems increases with the assets of the member. For instance, one of the international payment systems charges their Principal Members €7 for every million euros of assets, with a minimum fee of €108 500 and a maximum of €542 300. Participant Members pay a flat rate fee of €10 850.

The joining fee for the other international payment system increases in three steps for both Principal Members (PM) and Affiliate Members (AM) depending on whether their total assets amount to less than €50 billion, between €50 billion and €100 billion, or more than €100 billion. This results in a joining fee for PMs of €30 000, €90 000 and €150 000, respectively, i.e. between €0.60 and €1.50 per million euros of assets. Similarly, the joining fee for AMs amounts to €15 000, €45 000 and €75 000, respectively, i.e. between €0.3 and €0.75 per million euros of assets. In addition, both PMs and AMs pay a one-time application fee of €10 000 and €20 000, respectively.

It is interesting to note that the joining fee for the Principal Members of one international payment system is approximately three times as high as that for one of the other international payment systems, whereas Affiliate Members of the latter pay almost seven times as much, relatively, compared with the Participant Members of the former.

Domestic systems

Depending on the level of the joining fee, the open domestic payment systems can in principle be divided into three:

- No joining fee (in Germany and some systems in Italy) or the fee is less than €15 000;

- The joining fee ranges from €30 000 to €150 000, with most systems charging about €50 000, such as the GCB in France, the PIN scheme in the Netherlands and one system in Spain;

- The joining fee amounts to between €1.1 million and €1.9 million, e.g. Danish scheme Dankort and Belgian Bancontact scheme, as well as access fee for entering the bank card system Pankkikortti in Finland together with fee for the PMJ and POPS.

- One national system constitutes a category of its own, as a member participating as an acquirer and issuer may pay up to €6.7 million in joining fees. The joining fee for an issuer increases in five steps depending on the number of cards projected to be issued during the first three years. Calculated per card projected to be issued, the joining fee of this scheme appears to be set so as to discourage issuers from increasing their volume and constitute a barrier to entry.

B.8.3. Governance in Card Payment Systems

Various schemes (though not all) distinguish between different classes of membership, although to varying degrees and following different criteria of such division. From a competition point of view, the only arrangements relevant are those that create some risk of distorting the conditions under which individual member institutions compete with each other or under which potential new members can compete with the incumbent ones.

This section considers first the various classes of membership available in payment card networks, and the reasons networks distinguish between members. Secondly, it considers the implications and possible competition restraints arising from membership rules.

Classes of membership and reasons for distinguishing

There are three types of distinction between different classes of membership that could be observed on the basis of networks’ responses:

7. distinction between the principal and secondary members (“associate”, “affiliate” or “regular”) members – common for two international systems Visa and MasterCard, and present also in several national systems (where, however, principle members are limited in number to the incumbents in certain national market);

8. distinction according to functional role played by the members in the system (e.g. division between (i) card issuers, (ii) ATM acquirers, (iii) POS acquirers and (iv) collectors in one domestic system, and division between (i) member banks with access to clearing facilities and (ii) those with access to the clearing infrastructure through a principal members in another domestic system); and

9. distinction according to the ownership status, where shareholder banks may vote on the shareholder board and have certain control rights as opposed by the banks use the system’s services without becoming shareholders.

Various reasons for such distinguishing between classes were quoted by the networks:

- affiliate status as an encouragement for participation by small or new financial institutions in the scheme in the way that does no involve excessive undertakings and risks (MasterCard),

- greater flexibility in order to accommodate the different ways of providing card payment services in different regional and local markets (Visa),

- simplifying the decision-making process by reducing the number of entities participating in it,

- historical reasons quoted especially with regards to the distinction according to functional role played by the members and the ownership-based distinction,

- increase accessibility of clearing and settlement system to the smaller credit institutions without the need for complicated and costly bilateral settlement arrangements – quoted as a reason valid for those systems where the membership of smaller bank depends on ‘sponsorship’ of principle member,

- flexible access to institutions with an ‘asymmetric profile’ i.e. solely issuers or acquirers.

Implications and possible competition restraints arising from membership rules

Various classes of membership confer different types of rights on members. Generally these rights can be divided into four categories:

10. collection of business-sensitive data; in one international system, information on the business activity of affiliate members has to be supplied to principal members and certain activities of the secondary members – such as introduction of a new card programme or if an Associate itself wished to sponsor a Participant Member - may even require a prior approval of the principal members. Also in some domestic systems obligation to supply information applies only to secondary members. Similarly, also the scope of information demanded varies: from statistical data regarding volumes of POS transactions and ATM withdrawals to all data on the business activity;

11. decision making on issues affecting intra-system competition; in one particular national system, there are differences in the role played by the two classes of members (Principal and Affiliate) in decision-making, including decisions that may affect the way in which members compete in carrying out their activities. All members participate with the right of discussion and vote in the General Assembly. However, only the Principal Members sit on the Board of Directors, which is the body that defines the general policy of the system and takes all the important decisions (including those regarding admission or exclusion of members, adoption of sanctions against members, adoption of different categories and levels of fees in the system, decisions regarding using brands in the system, compulsory rules regarding issuing and acquiring, security measures, technical rules and specifications, and other rules for the functioning of the system, etc.);

12. supervisory and sanctioning powers; within one particular national system, the Principal Members have a number of supervisory and sanctioning powers vis-à-vis regular members. In another system the Board of Directors is the competent body to impose sanctions, extending to the expulsion of a member from the system, for non-respect of the rules. In the system, agreements by members with other networks regarding cards or access to the system, such as the opening of ATMs or POS to non-system cards, have to be submitted and validated by the Board of Directors, which will ensure that all necessary measures to protect the brand and the security and the integrity of the system have been taken;

13. intermediating in membership applications; in one particular national system, the Principal Members transmit to the scheme owner membership applications on behalf of new applicant members. New membership is moreover limited to the class of “regular member”, i.e. there is no possibility of becoming a Principal Member with the associated rights and powers.

Most of the membership arrangements reviewed do not seem to raise concerns from the competition angle, in particular where these arrangements concern the functional role played by different members. However, the distinction between principal and secondary (“associate”, “affiliate” or “regular”) members, if combined with materially different co-decision and participation rights, requires further assessment. This distinction is made by the large international networks (MasterCard and Visa) and may find some explanation in the very large number of members that make up both networks (even though it could be imagined that efficient decision-making processes could be organised even in such large schemes without distinguishing between principal and secondary members). However, only one domestic card network makes a distinction of this kind.

Networks naturally appear keen to collect data on transaction volumes from members for statistical purposes and in order to collect transaction-related fees. However, the collection of business-sensitive data through principal member banks as “intermediaries” leads to a one-sided information exchange, as secondary members have to share business-sensitive information with principal members. The information-collecting bank may therefore gain a competitive advantage over the reporting bank. As the information-collecting bank is typically a bank with voting rights on the scheme’s board, such a one-sided exchange of information may reinforce the concern that decisions that might be taken limit competition. Again, the quality and amount of data that principal members collect from the secondary members attached to them appears more significant within one particular network than in other networks.

On the other hand, the collection of business-sensitive data by scheme owners through member banks is necessary neither for statistical purposes nor for the calculation of transaction fees. As exemplified by other systems (e.g. PIN, SIBS-Multibanco), member banks can provide this information directly to a scheme owner. Also, both MasterCard and Visa collect data on transaction volumes directly from member banks, not through principal members. The possibility of collecting business-sensitive data directly from member banks raises the question whether the exchange of such data between principal and secondary member banks serves purposes other than the calculation of fees by the scheme owner or the verification of the financial soundness of affiliates. Considering that all banks within the EU are under the supervision of a financial supervisor, it is also questionable whether the duty of affiliates to report to principals provides stronger guarantees for the financial stability of an affiliate than already provided by the yearly screening of a bank by a financial market supervisor.

While there may be reasons for differentiating between primary and secondary members, at least in large international networks, it may lead to restrictions on competition where domestic card networks reserve far-reaching decision-making powers exclusively to a limited group of local incumbent banks. In one country, the local card network has reserved to the incumbent banks the power to determine many essential parameters of competition. Here, the risk of competitive distortions and restrictions can arise.

B.8.4. Conclusions

Most card payment systems reserve card issuing and merchant acquiring to credit institutions or entities controlled by credit institutions, which may inhibit processors from entering the business and from competing with the banks. The financial institution requirement may, however, no longer apply when the proposed Payment Services Directive is implemented.

Card systems in some Member States require that credit institutions are registered with the national central bank in order to participate in the domestic payment system, which may inhibit cross-border competition. To the extent that these rules are based on legislation or decisions by national central banks, they may also be in contradiction with the freedom of services and internal market directives.

The high level of joining fees and their structure may hinder effective intra-system competition. It is remarkable that the joining fee in open domestic payment systems varies from no fee in Germany and in some systems in Italy to fees in millions of Euro in particular Member States.

To the extent that scheme owners enter into exclusivity agreements with member banks for the sale of processing services, potential competition with other processors may be inhibited.

B.9. Cross-border Competition in Acquiring

The provision of cross-border services to merchants is not developing as fast as it could. Only around 10% of cross-border acquirers have opted for opening a cross-border branch or buying a foreign bank. The cross-border acquiring of merchants appears to be currently limited almost solely to the international networks of MasterCard and Visa; 86% of foreign entrants have used a cross-border acquiring licence from an international network.

Within the two large international networks, very few banks acquire merchants cross-border. 9% of banks surveyed attempted cross-border entry, with the UK banks being the most active. While Italian banks were able to secure more contracts cross-border than their UK counterparts, UK providers lead in terms of turnover. Cross-border acquiring is most often offered by EU-15 banks, whereas acquirers from new Member States seem to refrain from cross-border acquiring owing to incomplete saturation of the local credit card market, the generally relatively small size of such acquirers, and their lack of technical expertise.

It should be noted that the purchase of a local acquirer does not necessarily mean that the entrant will focus on merchant acquiring in that country. Merchant acquiring is more often perceived as part of a general strategy to establish a commercial presence in a foreign country. Banks may see acquiring as part of a “full service package”; however this may not be their main product offer to their corporate clients[209]. The option of offering services through a cross-border acquiring licence (a.k.a. central acquiring licence) is, on the contrary, taken up solely where the bank is keen to expand its acquiring business.

B.9.1. Merchants acquired cross-border

Merchants acquired cross-border tend to be large multinational companies. In 2004, the share of such merchants came to almost 90% of total turnover generated in cross-border acquiring.

The share of purely national merchants in cross-border acquiring seems to be rising. However, despite their rising share in the absolute numbers of the contracts signed, the corresponding card turnover still remains quite limited. Such a strong contrast between the number of contracts and card turnover only confirms the supposition that national merchants tend to be of smaller size than multinational companies.

Factors impeding the development of cross-border acquiring

Notwithstanding recent growth in cross-border acquiring[210], resulting mainly from the rapid development of the e-commerce sector and the expansion of gasoline companies, several respondents to the Commission’s inquiry indicated that cross-border acquiring is not developing as fast as it could. The main reason appears to be the existence of barriers to entry into domestic card acquiring markets. Ireland, Spain and France were mentioned as the acquiring markets that were most difficult to enter.

Statements by acquirers suggest that many merchants often prefer to have an acquiring relationship with a single bank. Owing to the fact that debit card transactions in Western Europe[211] on average constitute about 60% of the total card transaction volume, while credit and charge card transactions account for only 40% (with strong regional differences: in many countries the debit card transaction volume is significantly higher), merchants are particularly motivated to accept debit cards. However domestic card networks may be particularly difficult to access for foreign entrants, due to a number of technical, administrative and financial reasons.

Technical standards: communication protocols, security standards and certification

Domestic systems tend to be technically closed networks and for the most part lack interoperability. The co-existence of different technical rules and standards within the major card payment systems and between the national debit systems may inhibit the cross-border competition of merchant acquirers and processors in the EU. In addition, there are also different communication protocols for domestic credit card transactions within the Master Card and Visa systems.

While the harmonisation of all national communication protocols, including the protocols for transmitting transaction batches to the banks/merchant acquirers and the security concepts, would remove such barriers, a number of alternative steps could be considered as an intermediate solution.

Interchange fee arrangements

A further practice that can substantially inhibit or even prevent cross-border acquiring may be the obligation on foreign acquirers to pay the fallback interchange fee in the target country. Such fallback interchange rates create an obstacle to entry where local incumbent acquirers (often joint ventures created by domestic banks and sometimes the sole providers of acquiring services in a network) are able to agree favourable “on us” interchange rates with domestic card issuers. These “on us” interchange rates are presumably considered sufficient by domestic banks; however, the fact that a higher fallback interchange rate is imposed on a foreign acquirer may substantially raise the latter’s cost, compared with that of incumbents, and limit its ability to offer competitive merchant fees. Acquirers have informally complained to the Commission that this situation exists in at least two countries and impedes their access to the market for acquiring MasterCard/Visa transactions.

Structural barriers

Vertical integration of card payment systems gives rise to structural barriers that may impede new entrants, in particular non-banks, from competing with the incumbent in one segment of the market (for details, see Chapter B.2). Furthermore, the lack of multilateral clearing platforms may create entry barriers for foreign banks seeking access to clearing facilities. In systems with bilateral clearing arrangements, foreign banks depend on the goodwill of a local bank to “sponsor” its participation in the clearing of card transactions. New entrants thereby depend on incumbent banks for market access. It would appear that the absence of a multilateral clearing platform has impeded market access for foreign banks in at least one EU Member State.

Membership requirements and joining fees

Specific requirements, such as local establishment requirements[212] in some countries may inhibit market entry for the cross-border provision of services. It appears that this practice excludes the provision of cross-border payment services without a local presence, which in turn may raise entry costs. In contrast, the large international systems, MasterCard and Visa, allow for cross-border issuing and acquiring services to be provided under pan-European licenses. Furthermore, high joining fees may also dissuade membership in some domestic debit card networks and make entry unprofitable, particularly in small markets.[213]

Governance

Some governance arrangements within card payment systems risk distorting the conditions for competition between member banks , in particular between new entrants and incumbent banks (for details, see the chapter on governance). For instance, in some networks associate members have to communicate business-sensitive information to the principal members without reciprocal information-sharing. In other systems, decision-making on issues affecting intra-system competition, such as fees, membership rules and technical specifications, is reserved to the principal members.

B.9.2. Conclusions

The provision of cross-border services to merchants is developing very slowly and is limited almost solely to the international networks, MasterCard and Visa. The sector inquiry revealed a number of competition barriers on the market, which were confirmed by several market participants.

Technical barriers

Diverging technical standards for message protocols and security requirements in national and international schemes hinder processors and terminal vendors from operating on a pan-European scale. This in turn inflates input costs for banks and ultimately for merchants, and in the same time serves as a barrier to entry for cross-border acquirers.

Interchange fee arrangements

The obligation on foreign acquirers to pay the fallback interchange fee where local incumbent acquirers are able to agree a favourable "on us" interchange rate with domestic issuers, may create obstacle to new foreign entry.

Structural barriers

Vertical integration of card payment systems and lack of multilateral clearing platforms may impede new entrants from competing with the incumbent in one segment of the market.

Membership conditions and joining fees

Membership requirements such as those relating to registration with the local central bank or to being a credit institutions may exclude the provision of cross-border payment services. Also high joining fees for card payment systems may result in discouraging new entry.

Governance

Certain governance arrangements, e.g. such as obliging some members of the system to provide business-sensitive information to principle members, without reciprocal information sharing, may distort conditions for competition between the member banks.

B.10. Payment infrastructures

Access to payment systems is necessary for any bank considering entering a retail banking market and intending to offer customers core banking services, such as current accounts or payment cards. The analysis in this chapter is therefore relevant not only to the payment service markets themselves, but also to wider issues of competition between banks.

This chapter is structured as follows:

- Section 1 describes the typical operation of payment infrastructures;

- Section 2 discusses the scope and methodology of the inquiry into payment infrastructures;

- Section 3 outlines the main features of the surveyed payment systems;

- Section 4 describes the ownership and management of payment infrastructures;

- Section 5 discusses access issues;

- Section 6 examines fees charged to users;

- Section 7 provides a competition analysis;

- Section 8 discusses the transition to the Single Euro Payments Area; and

- Section 9 concludes.

B.10.1. The typical operation of payment infrastructures

In principle, each cashless form of payment involves five different parties: the payer; the payee; two intermediaries (banks or other payment service providers) offering customers transaction facilities; and an inter-bank payment arrangement for executing funds transfers between the two intermediaries. Figure 55 illustrates this structure.

Figure 55:

[pic]

To make payment services available to consumers, banks and other payment service providers need access to facilities to conclude the payment transaction. The way a payment is executed between two banks (“Payment arrangements”) requires a number of supporting activities. Two activities are particularly important:

- payment transmission, clearing and settlement of payments (normally referred to as the 'payment infrastructure'); and

- agreements to fix standards covering technical, operational and sometimes commercial aspects as well as financial aspects of the inter-bank relations (normally referred to as the 'payment scheme').

When the payer and the payee have an account at the same bank – or the same group - the exchange of information and balance calculation occur normally within the institution. These transactions are referred to in this report as “on us” payments.

When the payer and payee are customers of a different bank, some kind of inter-bank arrangement is required. These arrangements may take different forms: they may be bilateral (such as in Austria) or multilateral.[214] In this latter case, financial institutions present and exchange data and/or documents relating to funds transfer to other financial institutions under a common set of rules and procedures established at central level and compelling for all participants to the arrangement. These payment arrangements can have targeted membership, i.e. they are opened to institutions belonging to special categories[215], or they can be open systems. In this category of payment arrangements, two main models have developed in EU: either a centralized payment system, generally organised in an ACH (Automated Clearing House) or a decentralized payment system: the clearing system is not based on a centralized infrastructure in common ownership but rather is based on bilateral infrastructures provided by each member. The company managing the network is a coordinating body for collective governance purpose only. This model is in place in Finland and Ireland.

B.10.2. Methodology and scope of the inquiry into payment infrastructures

The scope of the analysis is limited to multilateral clearing arrangements with open membership.[216] Infrastructures that clear only payment card transactions are also excluded. Therefore the inquiry does not consider all types of payment infrastructure.[217]

A questionnaire was sent to clearing infrastructures that operate in the EU-25 Member States; the addressees were selected from the list of payment system infrastructures provided to the Commission by National Central Banks (NCBs). Table 56 below provides the list of the surveyed retail payment systems.

Table 56: Payment systems surveyed in the sector inquiry

Belgium | CEC (CEC) |

Cyprus | JCC TRANSFER (JCC Payment System Ltd) |

Czech Republic | CERTIS (Czech National Bank) |

Denmark | SUMCLEARINGEN |

Estonia | ESTA (Eesti Pank) |

Finland | PMJ (Finnish Banking Association) |

France | SIT (GSIT) |

Germany | RPS (Deutsche Bundensbank) |

Greece | DIAS (DIAS S.A.) and ACO (ACO) |

Hungary | ICS (GIRO Zrt.) |

Ireland | IRECC (IRECC Ltd.) and IPCC (IPCC Ltd.) |

Italy | BI-COMP (Bank of Italy) |

Latvia | EKS (Bank of Latvia) |

Lithuania | LITAS (Bank of Lithuania) and KUBAS (LCKU) |

Luxembourg | LIPS_NET (Sypal Gie) and DOM-ELECTRONIQUES |

Malta | MARIS and CHM (Central Bank of Malta) |

Netherlands | CSS (Interpay Nederland BV) |

Poland | ELIXIR (KIR) |

Portugal | SICOI (Bank of Portugal) |

Slovakia | SIPS (National Bank of Slovakia) |

Slovenia | GC (Banka Slovenije) |

Spain | SNCE (SESP) |

Sweden | BANKGIROT(BGC) |

UK | BACS (BPSL) and C&CCC (C&CCC Ltd) |

Europe | STEP2 (EBA Clearing) |

B.10.3. Main features of surveyed payment systems

In most EU Member States there is only one open clearing infrastructure for domestic retail payments.[218] Where there is more than one, such as in Greece, Ireland, Lithuania, Luxembourg, Malta, or the UK, they are usually not in direct competition; rather they are complementary.[219] Payment systems in the Member States have generally evolved independently of each other and within the scope of their national boundaries. Payment systems have selected and implemented at a national level their own technologies, formats and service levels, as well as governance models. The outcomes are nationally-based and fragmented payment systems across EU.

Even though funds transfers between two banks can be performed via a number of alternative channels (e.g. through bilateral agreements or through a third financial institution, known as correspondent bank[220]), the share of transactions processed and cleared by the surveyed systems is generally significant. On the basis of data from the ECB Blue Book, nearly half of the surveyed retail payment systems handle more than 50% of the total domestic transaction volume. In two countries the percentage is nearly close to 100%. Most surveyed systems service the majority of their country’s domestic credit transfers and direct debits, with the exception of intra-bank transactions (on-us transactions).[221]

With respect to card payments, only six systems offer this type of service and, among these, four of them process more than 50% of the total number of card payments made by national customers. In other countries this service is generally provided by other private operators specialized on payment card operations (among these, the most important at EU level are MasterCard Europe and Visa Europe).

In order to better assess the economies of scale and operating efficiency of clearing infrastructures, respondents have been asked to provide data on operating costs. The ratio of operating costs[222] to total number of transactions by payment system has been calculated for each system that was able to provide operating cost data. The cost data show significant variability across schemes, taking values ranging from 0.1 cent to 28 cents. The variability across countries could be explained at least partially by the fact that each scheme is organized differently and handles different types of transactions, implying different costs. Countries which still have both paper and non-paper clearing systems, purely paper-based clearing houses report operational costs per transaction more than three times higher than that of non-paper clearing systems.

Furthermore, the variability of the cost indicator across countries could also be explained in terms of economies of scale: indeed systems with the highest operating cost ratio generally handle a lower number of transactions than systems with the lowest operating cost ratio.

B.10.4. Ownership and management of payment infrastructures

Concerning the legal-status of clearing and settlement infrastructures, they were traditionally created on a non-profit basis. This arrangement has evolved over time and, currently, some systems are moving towards a profit-oriented organisational structure. Currently nine of the surveyed domestic payment systems and the European retail payment system STEP2 are operated on a for-profit basis.

In the past, ownership of infrastructure often involved National Central Banks (NCBs): this ownership arrangement had the explicit objective of fostering financial stability and promoting the soundness of payment and settlement systems. Currently only ten Member States have payment systems that are owned and managed by the NCB[223]. The remaining systems have opted for a so called “mutual governance model”, i.e. a system owned and or managed by all (or more often the largest) users of the system.

Where the infrastructures are directly operated by the NCB, the decision making bodies of the central bank takes the most important decisions in terms of pricing and access issues to the clearing system. Where the system is operated by a joint venture of banks, factors such as the operating volume play an important role and may determine the number of votes for decisions concerning fees and access rules. In one Member State only the founding members are involved in the decision-making process.

B.10.5. Access issues

The concept of membership in payment systems is typically applied to open clearing systems, where banks that wish to use a clearing infrastructure have to obtain membership of the organisation that provides the clearing services.

Nine of the payment systems that were surveyed have only one membership class. However, many other systems choose to distinguish members according to classes of membership. The most common distinction is between direct[224] and indirect participants. Direct members have the benefit of being directly in contact with the clearing operator and of settling their operations in their own account held at the NCB. The notion of “indirect membership”, instead, is quite wide. It ranges from systems where indirect member simply settles its positions in the Real time gross settlement system (RTGS) account of a direct member, to systems where it is not recognized as member of the network. In the latter event, indirect participants access the system via an agency agreement negotiated bilaterally with a direct member. In the view of the various shapes indirect membership can concretely assume, indirect participants’ duties and rights should be analysed on a case-by-case basis. Normally, the indirect member does not get involved in the collective decision-making process of the system, which exclusively involves direct members.

A number of systems define participation in terms of scheme ownership, i.e., shareholder banks that vote on the shareholder board and have certain control rights, as opposed to user banks which simply use the system’s services without becoming a shareholder. In other systems, instead, the concept of membership has evolved into the notion of client-relationship.

In all surveyed clearing systems participation is limited to credit institutions or, in some systems, to non-bank financial institutions (for example to payment card companies). The requirements that only banks are allowed to be direct participants in certain clearing infrastructures (in particular reference was made to the European STEP 2 clearing system) was signalled as possibly “ impeding competition ” in the infrastructure market. In particular, domestic clearing houses, that do not hold a banking license, are not admitted as direct participants to STEP2.

Additional requirements are imposed on direct participants by the rules governing the infrastructure. One is the obligation to have an account with the Central Banks and/or being a member of the RTGS system. These requirements are generally linked to the settlement of transactions. In this case, the direct member shall also comply with the specific requirements to participate in the RTGS system[225].

Some of the eligibility criteria applied by payment schemes may, in certain cases, make it more difficult for a new entrant to join the system as a direct member. Examples of such criteria include:

- the need to have minimum level of activity, expressed either as share of individual transaction volume on the total number of transactions;

- the need to become a shareholder of the owner of the infrastructure;

- then need to be member of the national banking association[226]; and

- the need to be operating in the country for a certain period or to have physical presence/branch/subsidiary in the country.

B.10.6. Fees charged to users

Fees charged to banks for the use of the payment infrastructure could be an important determinant of the overall cost of certain retail financial services. The fee structure applied in various payment systems may imply significant cost differentials depending on individual bank characteristics. Fees charged can generally be divided into two categories: joining fees and clearing fees. Two of the surveyed payment systems also apply an exit charge: the UK C&CCC and EBA STEP2. According to respondents, exit fees are charged to cover any direct costs that the company reasonably incurs as a result of the withdrawal or exclusion.

This section discusses:

- j oining fees;

- clearing fees;

- settlement fees; and

- economies of scale arising from fee structures

Joining fees

Depending on the level of the joining fee, the surveyed systems can be divided into three categories, charging: no joining fee; a one-off joining fee equal for all members; or different joining fees for direct and indirect members.[227] In most EU systems indirect participants do not pay to join the scheme, or pay a lower fee. In two systems, direct and indirect members have to pay the same joining fee.

In systems where there is a joining fee, it is typically a one-off payment, whose amount can be either fixed or linked to parameters such as: the characteristics of the bank’s technical infrastructure; costs incurred by existing members to accommodate the new entry; individual transactions volume; or the package of services bought.

In one system the joining fee payable by new participants is calculated on the basis of a formula which is linked to a scheme that provides for the reimbursement of costs incurred by the founding members at the time the system was created. The highest joining fee charged by surveyed systems to new members amounts €1.2 million. The second highest fee is €250 000.

Clearing fees

According to the results of the survey, payment infrastructures charge a periodical (annual or monthly) fee and/or a fee per transaction. These fees can be either fixed or linked to parameters such as: the bank’s transaction volume; the bank’s type of technical connection to the network; the time at which the transaction enters the system; or the type of transaction.

Periodical fees charged by surveyed clearing systems range from zero to €216 300 per year. Per transaction fees range from zero to a maximum of €0.23.

Settlement fees

In most of the surveyed payment systems, fees charged by the clearing operator cover both clearing and settlement services. Nevertheless, in some systems (e.g. Cyprus, Finland, Greece, Italy, Poland and Portugal), direct members are charged an additional fee for settlement. It should also be noted that charges reported by replies may include a different range of services provided (e.g. in some cases, fees do not include costs of data transmission and/or costs for holding collateral with the NCB).

Economies of scale arising from fee structures

With regard to the economies of scale created by the fee structure, all but five systems surveyed show some type of 'regressive' fee structure. Fee structures could be regressive in several ways:

14. a fixed membership fee

15. a periodical (annual or monthly) fee

16. volume-discounts, offered through regressive fees per transaction.

The effects of these regressive fee structures are illustrated in a simulation exercise conducted as part of the inquiry.[228]. The exercise has been conducted for a standard domestic credit transfer. Accordingly, all domestic systems that do not process credit transfers have been excluded from the simulation[229]. Additionally, the analysis has been limited to prices charged to direct participants, as those charged to indirect members would also have required additional data for the payments from indirect members to direct members, which are the result of private bilateral agreements.

The price per transaction in nearly all systems[230] decreases with bank size. These cost savings arise from a fixed membership fee, a periodical (annual or monthly) fee and or volume-discounts, offered through regressive fees per transaction. This price structure implies that cost per transaction for a small bank may vary significantly across countries. For example, if a bank with up to 50,000 monthly transactions is considered, total fees per transaction vary from 0.15 euro cents to 0.69 euro cents. The simulation shows that for certain systems, banks with low transaction volumes (such as new entrants or niche players) could incur much higher unit costs than incumbents.

B.10.7. Competition analysis

In general terms, retail payment markets should achieve an adequate balance between competition and cooperation to benefit market users; transparent market should promote competition and contestability; and the pricing structure should encourage an efficient allocation of resources and payment risks.[231]If markets are insufficiently competitive or contestable, efficiency benefits from innovation, consolidation, exploitation of economies of scope and scale may fail to be realized or to be passed on to consumers. In particular, established networks are potentially in a position to create entry barriers that impede competition and innovation. Entry barriers can be created either directly by imposing access restrictions or by more indirect means, for example, by a choice of standards and rules that are inappropriate, or difficult to adopt.[232]

This section discusses:

- the operation of clearing infrastructures and lack of inter-system competition;

- the need to adapt to different national standards;

- different classes of membership and special requirements for direct members;

- the “need to be a bank” requirement; and

- fees and fees structure.

Operation of clearing infrastructures and lack of inter-system competition

In most countries there is only one retail clearing infrastructure that is operated either by the National Central Bank or by a membership associations controlled by (the main) banks operating in the country. The existence of one 'dominant' system can be explained by the specific structural characteristics of the market, i.e. economies of scale and network effects. Even though consolidation process in payment systems market may be justified by economic reasons, competition may still help provide lower prices and a greater range of services. When a joint-venture of banks owns and manages the infrastructure, the decision-making body is normally composed of the largest participants. This factor may raise anti-competitive concerns, as discussed further below. [233]

The same banks are often members of different payment systems (e.g. card and non-card systems, national and cross border systems). This can create conflict of interests for members and reduce incentives for inter-system competition, as a bank that is a member of a payment system may have less incentive to promote a strategy of intense competition with another network it participates in. For example, major payment card schemes have a clearing system which does not compete with other clearing systems.

In August 2005 the European Central Bank published a policy statement[234] regarding central banks’ provision of retail payment services to credit institutions in euros. This statement mentions possible competition problems and recognises the importance of avoiding competitive distortions or crowding-out of market initiatives when NCBs provide retail payment services to credit institutions.

Need to adapt to different national standards

Member banks normally have to respect certain technical specifications, and a testing and certification procedure. In some systems this can take between 6 and 12 months. Banks that operate in different Member States need to adapt to some 25 different procedures and technical standards.

Different classes of membership and special requirements for direct members

Various clearing systems distinguish between different classes of membership, although to varying degrees. From a competition point of view, arrangements are only relevant where they pose some risk of distorting the conditions under which the individual member institutions concerned compete with each other or under which potential new members can compete with the incumbent ones.

The distinction between “direct” ("principal") and “indirect” (“ancillary”, “affiliate” or others) members, in combination with a different participation in decision making and participation rights, requires further assessment. Firstly, an indirect member will depend on the “good will” of a direct member (a competitor in the downstream market) with whom the indirect member will have to negotiate an agency contract. This also adds an additional layer of intermediation to the system and possibly leads to an increase in total costs and/or a lengthening in the clearing cycles and/or an imposition of unfair requirements. At the same time, the possibility of joining a clearing system as indirect member could be seen as increasing the choice for smaller banks and niche players, who can benefit from not having to comply with the requirements linked to settlement and direct membership. Conditions of these contracts are not included in this report.

Secondly, indirect members normally do not fully participate in the decision making process (determining prices, deciding on membership application, on technical standards and other rules). In practice, direct members might in some way decide the costs that all banks will have to bear to use the infrastructure.

Thirdly, direct members get better information than indirect members, both concerning the systems as a whole and the data they receive from indirect members. On this point, the collection of business-sensitive data through direct member banks as “agent” leads to a one-sided information exchange, as indirect members have to share their list of payments with principal members. The information collecting bank may therefore gain a competitive advantage over the indirect member. As the information collecting bank typically is a bank with voting rights on the scheme’s Board, such one-sided exchange of information may reinforce the concern that decisions might be taken that limit competition.

Concerning the possibility of becoming direct member, the investigation has shown that there are a number of requirements that banks have to meet in certain systems: such as a minimum level of activity, the need to become a shareholder of the owner of the infrastructure, the need to be member of the national banking association or the need to be operating in the country for a certain period of time. These rules are fixed by incumbents and may be difficult to meet by new entrants.

The “need to be a bank” requirement

All surveyed systems require members to be regulated financial institutions. Some of these schemes also require banks to be supervised by the NCB, or require a physical presence in the Member State. According to respondents, these restrictions address the need to ensure that a system is financially secure, minimise systemic risk and ensure that new members are able to interact properly from a technical and operational perspective. However, while the oversight by NCBs may be an efficient tool to guarantee the financial reliability of players acting in payment systems, it could be worthwhile to explore other ways to achieve financial stability within these systems. The proposal for a Directive for a New Legal Framework (NLF)[235] for Payments in the Internal Market is also meant to open up EU payment systems to non-banks. Many replies to the public consultation highlighted the concern shared by credit institutions in relation to the openness of clearing infrastructure to the so-called “payment institutions”. They argue these latter should be subject to the same prudential requirements in order to ensure a level-playing field among supervised and non supervised institutions.

The exclusion of non-banks means that non-bank enterprises cannot be direct members of the clearing system. This also means that non financial institutions (such as some processors or customers) are not involved in a network’s decision making and thus a network may develop in ways that do not meet the needs of a significant sector of users. Concern has also been expressed that the inability of corporate clients to access clearing systems directly might tie them unduly into their current banking arrangements. For example, in their responses to the public consultation, some credit unions’ organizations complained of their members' inability to directly access retail payment systems some Member States. It was argued that this would leave credit unions at a competitive disadvantage in supplying retail banking services.

It is worth noting that in STEP 2, currently the sole pan-European clearing system, only banks can be direct members. This requirement has been criticized by one clearing infrastructure as restrictive. Linkages between clearing houses could possibly expand the availability of their services to a wider group of financial institutions and their customers.

Fees and fees structure

In certain Member States the way in which the fee system is structured could potentially be considered a barrier to entry for new or small players. The joining, annual and transaction fees of the multilateral inter-bank networks are generally set by boards made up of representatives of their shareholders, who are also (some of) the network’s members. Fees paid by the new members in some cases cover initial members’ costs for developing the scheme.

However, the question arises as to whether joining fees charged hamper effective competition by dissuading entrants or raising their cost significantly. In one country, one bank withdrew its request to participate in the system, allegedly because of the high entry fee. One bank pointed out that the fee system in one clearing house, by offering large volume discounts, creates a competitive advantage for the largest entities. Regressive fees on the basis of volume also apply in other systems. As was shown in the simulation exercise in Interim Report II, volume discounts provided through regressive fees per transaction or fixed fees (one-off and periodical) may influence banks with low transaction volume, typically new entrants or niche players, in their choice to enter a market. Nevertheless, the welfare effects of such pricing policies in terms of welfare are ambiguous. Consequently, the effects that a specific pricing structure can produce should be assessed on case-by-case basis.

B.10.8. The European context: towards a Single Euro Payments Area

The advent of a Single European Payment Area (SEPA) will change the basic infrastructure for clearing and settlement systems. SEPA aims to create a single market for payments throughout the euro area, integrating payments systems and increasing efficiency. Therefore it is vital that the SEPA framework is conceived in a way that supports competition and innovation and enables cost savings to be passed on to businesses and consumers. Several of the competition barriers that the sector inquiry has highlighted in payment systems can be remedied through the establishment of a pro-competitive SEPA.

In this direction, the Commission has adopted in December 2005 a proposal for a Directive for a New Legal Framework (NLF) for Payments in the Internal Market, known under the name "Payment Services Directive".[236] The aim of this proposal is to establish rules for being a payment service provider and to harmonise legal rules regarding the provision of payment services (e.g. who pays for payment transactions, transparency in pricing, execution times, liability in case of default, customer information and rules on revocability of payment orders). The proposal also contains provisions (Article. 23 of the Draft Directive) on non-discriminatory access to payment infrastructures.

In the field of clearing and settlement infrastructures, the objective of SEPA is for retail payment systems to be able to process “SEPA compliant” payments and to be fully interoperable for basic services. The prospect of the transformation from domestic clearers to one or several pan-European automated clearing houses - (PE-ACH) is seen by some as offering new growth opportunities, while for others it threatens their longstanding business model[237]. It is also expected that existing market infrastructures will consolidate in order to exploit economies of scale: therefore both the number of retail payment clearing and settlement infrastructures and the costs related to their services are expected to decrease[238].

B.10.9. Conclusions

Retail payment systems in the EU are not yet integrated and their organisation and structures remain highly varied. This means that a bank operating in different Member States has to join the various national systems, adapt to different standards and face different costs.

In most EU Member States there is one national clearing infrastructure, which is operated either by the central bank or by a membership association controlled by (the main) banks in one country. Some Member States may have two or more payment systems, which are not in competition but rather are complementary, clearing different payment instruments.

Access to clearing and settlement systems is necessary for any bank considering entering a retail banking market. Operators of the established infrastructures are potentially in a position to create entry barriers which may take a variety of forms. In particular some membership rules and the way the fee system is structured may raise barriers to entry for new or small players.

The advent of the Single Euro Payments Area will change the competitive landscape. SEPA aims to create a single market for payments throughout the Euro area by integrating national payments systems. This will permit economies of scale to be realised and make cross-border competition feasible. The end result should be more effective competition in the market for payment services.

B.11. Multilaterally agreed interchange fees for ATMs and for non-card payments

This chapter considers the purpose and effect of interchange fees for means of payment other than payment cards. The chapter is structured as follows:

- Section 1 examines interchange fees for ATM withdrawals;

- Section 2 examines interchange fees for non-card payments (specifically direct debits, credit transfers and cheques); and

- Section 3 concludes.

B.11.1. Interchange fees for ATM withdrawals

Interchange fees for ATM withdrawals in the examined schemes are inter-bank fees paid by the issuing banks to the acquiring banks for the cash withdrawal carried by the cardholder using the ATM of an acquiring bank. Interchange fees for the ATM withdrawals are applicable under international and national schemes and with respect to both inter-European and domestic withdrawals.

This section discusses:

- the level of ATM interchange fees;

- the relationship between ATM interchange fee levels and cardholder fees; and

- competition issues.

The level of ATM interchange fees

During the course of its inquiry the Commission gathered data from several national schemes and two international card schemes. The level of data detail related to national schemes in some Member States, however, varied significantly. Therefore, not all data could be taken into account for the purposes of this analysis. In all schemes surveyed the interchange fees for ATM withdrawals are paid by the issuing bank to the acquirers. This contrasts with POS transactions where the interchange fee is usually paid by the acquiring bank to the issuing bank.

The interchange fee for ATM withdrawals can be agreed bilaterally between the banks or can be subject of multilateral agreements among banks participating in a certain payment scheme. Regarding national schemes, the interchange fee for ATM withdrawals is in some cases agreed bilaterally between the issuing and acquiring banks. The multilaterally agreed interchange fee is applicable in at least 6 of the surveyed schemes operating in different Member States (e.g. Denmark, Hungary, Italy, Luxembourg, Portugal and United Kingdom). The level of the interchange fee under two of the surveyed national schemes is related to the balance between the issuing and acquiring activity of the bank (e.g. the fee is higher if the issuing bank is not an acquirer).

Under the international card schemes, specific interchange fees have been established for the intra-European ATM withdrawals. These fees apply by default also to domestic withdrawals, i.e. where there is no bilateral arrangement between the issuing and acquiring bank and/or where no nationally established interchange fee applies.

The Commission carried out a simulation to analyse the level of interchange fee applicable at an ATM withdrawal of €100. The simulation concerned 27 card schemes (both national and international) within 11 Member States. The results of this exercise are presented in Figure 53 below.

Figure 53

[pic]

Relationship between ATM interchange fee levels and cardholder fees

The Commission examined interchange fees for ATM withdrawals and fees charged to cardholders. The level of interchange fee may affect the fees paid by the cardholder: a comparison of these two types of fees showed that in most of the cases the ATM fees charged to the cardholder were higher than the interchange fees or at least equal.

Competition issues

In several Member States banks multilaterally agree upon interchange fee for ATM withdrawals. These fees are fixed by banks or bank associations; i.e. by likely competitors in the downstream retail banking market. Multilaterally agreed interchange fees may distort competition between different means of payment as well as between banks to provide services to their customers. Competition between means of payment may be distorted in so far as banks have an incentive to promote the use of payment means that have high interchange fees. Competition between banks may be affected in so far as the fee is artificially kept at a level that penalise new entrants to the advantage of incumbent that normally are the owner of major ATM networks.

B.11.2. Multilateral interchange fees for other non-card payments

This section discusses:

- the setting of multilateral interchange fees for non-card payments;

- the level of multilateral interchange fees;

- economic justifications provided by industry for interchange fees; and

- competition issues.

The setting of multilateral interchange fees

Interchange fees are inter-bank fees paid between the payers’ and the payees’ banks for the conclusion of a payment transaction and/or for the provision of services in relation to a given payment. These inter-bank fees can be agreed bilaterally between the banks or can be the subject of multilateral agreements among banks participating in a certain payment scheme. Multilaterally agreed interchange fees have been agreed upon, in some Member States, also in relation to credit transfers, direct debits and cheques. The Commission’s inquiry has collected data on interchange fees for non-card payments.

Multilateral interchange fees are normally fixed by the banking communities. Clearing and settlement infrastructures are not part of the agreement and normally only have the function of settling the fees. However, in some countries, where the management of the infrastructure is not clearly separate from the management of the scheme, it is the decision making body of the infrastructure that fixes the fee. In all but one of the Member States surveyed, multilaterally agreed interchange fees function by default, and bilateral arrangements are possible.

The level of multilateral interchange fees

The interchange fee patterns in the national schemes are quite different, both in terms of fee structure and level of fees. According to data collected from clearing systems and National Banking Associations, a multilateral interchange fee has been agreed upon in eight countries for direct debit transactions and in six countries for credit transfers. The fee is always applied per transaction but the type and level of fees differs between systems. None of the surveyed systems in the new Member States reports the existence of multilaterally agreed interchange fees for credit transfers or direct debits.

For credit transfers, the fee charging option (Shared, Our, Beneficiary)[239] seems to influence the fee level in some countries. The fee charging option determines the way bank’s costs are charged between the payer and the payee. For example, the cross border payment “CREDEURO” is a “Share” payment and no interchange fee is paid between the payer’s and payee’s bank. In addition, types of payment, i.e. paper form, electronic form or the classification of transaction in STP (straight through processing) and non-STP payments, determine the level of fees in some systems.

The following fees structures were reported for the various payment instruments:

For direct debit transactions: there is currently an agreement on a multilateral interchange fee in 8 Member States: the payee's bank has to pay the payer's bank a fixed fee for every transaction. The fees vary from €0.02 to more than €2. In some countries (e.g. France and Portugal) different fees apply to different transaction types. Additional fees, often higher than the above mentioned fees, are charged for returned payments.

For credit transfers: in 6 Member States there are multilateral interchange fees for specific transactions, namely for paper transactions or for non STP transactions; for transactions with specific fee-payment options (e.g. for “our” payments); for specific services rendered by the payer’s bank to the payee’s bank. In one Member State, interchange fees up to €15 apply for certain payments and additional €5 are charged for non STP transactions. For credit transfers, the fees can be transferred from the payer’s bank to the payee’s bank, but also vice versa.

For cheques: there is a multilaterally agreed interchange fee in 5 Member States: These fees cover both “normal” transactions and additional services provided between banks.

Except for Italy and the Netherlands, where the fees have been scrutinised by the National Competition Authorities and decreased, fees in most systems have remained fairly stable over the last five years.

Economic justifications provided by industry for interchange fees

The Commission asked respondents to explain the economic function that interchange fees fulfilled in their systems. The main justifications provided were:

the fees are a mechanism to have the maximum number of participants on board and to maximise network externalities;

the fees are an “incentive” to banks to switch to more efficient payment instruments, in particular paper less and STP transactions (e.g. by introducing an interchange fee payment for banks starting a paper transaction);

the fee is a cost recovery mechanism to compensate for costs incurred by the bank of the beneficiary that are not recuperated via charging the customers, either because of a bank agreement on cost sharing (e.g. in case of “OUR” transfer) or because the bank of the payer chooses not to charge or to charge only to a limited extent (e.g. for collection of direct debits).

However, industry respondents did not provide economic evidence to support these justifications.

Competition issues

In several Member States banks have agreed upon an inter-bank payment for direct debit and for certain types of credit transfer and cheques. According to respondents, inter-bank fees have multiple purposes, including maximising network externalities; promoting certain types of payments (e.g., electronic payments); or covering costs for services provided from one bank to another. However, the scale and nature of these costs were not clearly justified by the surveyed institutions. Furthermore, interchange fees for credit transfers can go from payer’s bank to payee’s bank, or in the opposite direction.

In practice, the interchange fee creates a multilateral transfer between banks, which by being multilaterally fixed, does not take into account bank specificities and the relationship of a bank with its customers. As said above, the existence of interchange fees may distort competition between means of payment as well as competition between banks to provide payment services to the customers. Competition between means of payment may be artificially distorted in so far as banks have an incentive to promote the use of payment means that have high interchange fees. The existence of interchange fee may also render the cost for providing the service in-transparent to consumers. For example, customers who chose to use direct debits for recurrent payments to utilities may consider that these are offered free of charge. De facto, however, costs are simply shifted from the payer to the payee.

Competition between banks in the downstream market could also be affected. In those countries where interchange fees for certain types of credit transfers are applied, this may affect customer prices. For example, in one country, banks indicated explicitly that the price charged to consumers for a credit transfer is a function of the interchange fee to be paid.

Market participants expressed mixed views during the consultation process in relation to the necessity of a multilateral interchange fee within the direct debit scheme. Only in those countries where currently a MIF exists, was it argued that such a fee was indispensable for the existence of the scheme.

B.11.3. Conclusions

Multilaterally agreed interchange fees have been agreed upon in some systems in relation to ATM withdrawals, credit transfers, direct debits and cheques. These fees are agreed upon by the banking community and may distort competition between different means of payment as well as competition to provide payment services to customers. The level of these fees varies significantly across the EU and may, in some cases, raise competition concerns, which should be assessed on the basis of a full examination of the specifics of each case.

Part C: Conclusions and possible next steps

C.1. Conclusions and possible next steps

This chapter summarises the main findings and recommendations of the sector inquiry into retail banking. The chapter is structured as follows:

- Section 1 outlines the findings on retail banking market structure and performance;

- Section 2 summarises the findings and recommendations on current accounts and related services; and

- Section 3 summarises the findings and recommendations on payment cards and payment systems.

C.1.1. Findings on retail banking market structure and performance

The inquiry has identified important characteristics in the operation of the supply side in retail banking markets. Firstly, market infrastructures such as payment systems and credit registers are generally fragmented along national lines. Secondly, retail banking typically displays high levels of cooperation among industry players, who encounter each other in several product markets. Thirdly, barriers to entry remain in retail banking, arising from several sources. Such barriers may be natural barriers as well as those arising from regulation or anticompetitive behaviour.

On the demand side of retail banking, two factors may weaken the operation of a competitive market. Firstly, information asymmetry – where banking consumers lack or are unable to act on full information – reduces the intensity of price competition. Secondly, switching costs – notably the informational and transactional costs of changing some banking products – discourage consumers from leaving their current provider.

The inquiry’s analysis of market concentration suggests that European retail banking markets in general are 'mildly' concentrated, at least at national level. They also confirm the perception that Belgium and the Netherlands on the one hand and the Nordic countries on the other, have more concentrated retail markets than the European average. Some other Member States, most notably Germany, but also others such as Spain and Italy, show comparatively low concentration ratios.

Integration of European retail banking markets remains low. With the exception of the Benelux and Nordic countries, few players are among the leading five banks (measured by market share) in two or more countries. In general, very few foreign banks are among the top five in each Member State, though foreign banks hold strong positions in several New Member States.

Financial performance in retail banking varies considerably across the Member States. During the period 2002-2004 most Member States show average profitability ratios close to the EU average of 20 to 30%. Ireland, Spain and the Nordic countries reported sustained pre-tax profitability ratios of about 40%; well above the EU average. Several other Member States, including Germany, Austria, and Belgium reported low profitability throughout. The distribution of cost-income ratios naturally followed a similar pattern, with banks reporting lower cost-income ratios in Member States where profitability was higher.

The inquiry found wide national variations in banks’ income for specific product lines. Comparisons across a range of retail products show that banks’ income per customer is typically twice as high in the EU15 as in the new Member States. Overall, mortgages generate the highest share of banks’ gross income and high degrees of cross-selling.

Based on comparative OECD data the Commission concludes that the long-term trend of profitability is upwards in the EU banking sector as a whole. Moreover, from the clear overall trend of rising pre-tax profitability, it can also be inferred that retail banking profitability has risen over the long-term.

C.1.2. Findings and recommendations on current accounts and related services

The inquiry's main findings and recommendations on the market for current accounts and related services concern:

- credit registers;

- cooperation among banks;

- setting of prices and policies; and

- customer choice and mobility.

Credit registers

Banks and credit providers require access to good quality credit data in order to overcome information asymmetry when they set prices for new or potential borrowers. Thus credit registers are an important element of retail banking market infrastructure. To ensure strong competition among credit providers in retail banking markets it is vital that credit registers enable open and non-discriminatory access to credit data.

The inquiry has found that in several Member States, coverage of credit information markets is limited. Credit information markets remain also fragmented along national lines. Only a few credit bureaus conduct cross-border reporting, albeit for low volumes of data. While this may be largely explained by low demand for cross-border credit, regulatory barriers in some Member States further limit the development of cross-border data sharing.

The inquiry has highlighted three sets of issues in relation to credit registers which can weaken competition in retail banking markets: unfair access conditions; partial data sharing; and regulatory barriers. Some of these issues may be addressed through competition law. Other measures may also be appropriate in some cases, including full enforcement of data protection rules. In addition, a future assessment of access to credit registers may be warranted to ensure they comply fully with non-discriminatory principles.

Cooperation among banks

Retail banks co-operate in a variety of areas such as the setting of standards and infrastructures or the operation of payment systems. Savings and co-operative banks traditionally have even closer co-operative ties. These specific types of banks cover a significant proportion of the retail banking activities in Europe and play an important role in several Member States such as Germany, France, Austria, Italy or and Spain. Insofar as savings and co-operative banks remain legally independent, they tend to co-operate in a variety of fields. They often run their own payment infrastructures, have a joint risk management and protection scheme for deposits or may even have a common business and marketing strategy including a common brand. Moreover, some savings banks and/or co-operative banks apply territorial restrictions – the ‘regional principle – reserving a defined geographic area for the activities of an individual retail bank.

Certain forms and areas of cooperation are indispensable for bringing about efficiencies and consumer benefits. It usually does so where the banks involved are SMEs and/or jointly do not possess a significant market share. Cooperation is also necessary to agree on common standards and infrastructures for the operation of networks such as payment systems. On the other hand, benefits resulting from certain areas and forms of cooperation cannot justify all potential competition restrictions. In particular, severe competition restrictions such as market sharing or price fixing are unlikely to be outweighed by economic benefits. Even if individual cooperation agreements bring about economic benefits, the effects on market competition have to be thoroughly analysed on a case-by-case basis.

The Commission intends to further evaluate certain competition issues arising in the context of close banking cooperations. The Commission intends to investigate:

- company structures and areas of cooperation among those savings banks and co-operative banks that play a substantial role in one or more retail banking market(s);

- behaviour that results in substantial competition restrictions among the participants and on the market;

- economic benefits arising from these types of cooperation;

- State measures potentially requiring, leading to or reinforcing anticompetitive behaviour; and

- regulation and state intervention potentially infringing other Treaty provisions (e.g. free movement of capital, freedom of establishment) and/or potentially distorting competition by State aid favouring certain companies.

This evaluation has to be carried out by means of a thorough analysis on a case by case basis. Should it turn out that one or more of these cooperations raise antitrust issues, the Commission would take up those cases with a Community dimension. The same applies if regulatory aspects or other forms of state measures are involved. Depending on the issues in question, national authorities may be well placed to deal with certain cases.

Setting of prices and policies

The characteristics of the retail banking industry make it difficult to compare similar products and construct reliable indicators that will enable an evaluation of competitive structure. Nevertheless, the pricing behaviour of banks provides some initial indications on the degree of competition in the market. Based on the inquiry’s market survey, a commonly observed that banks tend to compete less aggressively for switchers than for new-to-market customers, though the intensity varies across Member States.

In relation to current accounts the inquiry has examined a range of fees and found evidence of significant variation in prices within and across Member States. Such patterns are evident in relation to the pricing of several parameters including account management fees; closing charges; excess borrowing fees; fees for ATM withdrawals; and fees for credit transfers.

Notwithstanding significant differences between Member States for some types of current account fee, convergence can also be observed in the pricing behaviour of banks in several Member States. For example, banks in Italy and Luxembourg reported the highest levels of fees for both account management and closing an account. Similarly, excess borrowing fees typically generate a small share of banks' total fee income on current accounts; less than 10 per cent in more than half of the Member States. However such fees generate over a quarter of the current fee income for banks in France, Spain, the UK and Cyprus. For credit transfers banks in around half of the Member States surveyed do not charge fees for domestic transactions, while banks in Greece typically charge fees well in excess of the euro area average.

The inquiry’s market survey suggests that in most Member States the majority of banks tie a current account to mortgages, personal loans and SME loans. Moreover, where the largest bank in a Member State ties its products, the inquiry’s data suggests that the majority of its competitors, including foreign entrants, choose to follow suit. From a competition view point, product tying in retail banking may weaken competition in three ways. Firstly, tying raises switching costs and therefore is likely to reduce customer mobility. Secondly, by binding customers into buying several products from the same bank, tying is likely to discourage the entry of new players and growth of smaller players. Thirdly, by introducing additional – perhaps unnecessary – products into the transaction, tying reduces price transparency and comparability among providers.

The Commission is concerned that possible anticompetitive effects will be strongest in markets where one or more large banks tie products. Product tying by one or more undertakings in a particular Member State may constitute an exclusionary abuse of dominance under Article 82 EC, where such undertakings have a dominant position. Clearly the assessment of a particular tying practice would depend on the specifics of the case.

Customer choice and mobility

It is likely that a large proportion of banking customers – probably the majority in most Member States – would describe themselves as satisfied with their current bank. For these customers the question of switching bank (and its related costs) does not arise. Thus the scope of the inquiry’s analysis of customer mobility is consumers and SMEs who are not fully satisfied with their current provider or are seeking to change bank for other reasons.

The inquiry’s analysis suggests that typically between 5.4% and 6.6% of current account customers in the EU will change provider per year. However, industry surveys suggest that the proportion of unsatisfied customers is typically much higher. For this group, the level of switching costs will be an important consideration. The sector inquiry has identified four sources of switching costs that are likely to reduce the ability of consumers to switch bank: administrative burden; information asymmetry and low price transparency; bundling and tying; and closing charges.

The evidence gathered by the sector inquiry suggests that high levels of switching costs in the retail banking industry may weaken competition in two ways. Firstly, switching costs may increase banks’ market power, enabling them to set higher prices for established customers who appear locked in to a banking relationship. Secondly, high switching costs and low customer mobility may limit prospects for market entry in full service retail banking, notably through greenfield operations.

Building on the preliminary findings of Interim Report II, the inquiry has conducted a multivariate analysis of the relationship between customer mobility and market performance. Owing to measurement problems, variables such as customer satisfaction, switching costs and financial stability could not be incorporated directly into the quantitative analysis. However, the inquiry has estimated the impact of customer mobility in the current account market on banks’ ability to exercise market power (using total retail banking profitability as a proxy). Multivariate analysis at bank level suggests that a one percentage point increase in the level of market churn corresponds to a similar reduction in banks' pre-tax profitability ratio. This effect is robust and statistically significant, suggesting that banks face greater pressure on profit margins where customers are more mobile. Therefore one conclusion from the inquiry is that simple, proportionate steps to reduce switching costs will enhance competition in retail banking.

C.1.3. Findings and recommendations on payment cards and payment systems

The inquiry's main findings and recommendations on the market for payment cards and payment systems concern:

- concentration and integration in the payment cards market;

- cardholder fees;

- merchant fees;

- interchange fees;

- profitability of the payment cards industry;

- membership and governance rules;

- cross-border competition in acquiring; and

- payment infrastructures.

The Commission welcomes that the publication of the preliminary findings of its sector inquiry fostered a constructive dialogue with the industry and led to self regulation in some Member States. The Commission invites banks in other Member States to seek a similar constructive dialogue with the Commission and the national competition authorities. Where no such initiatives are taken, the Commission will seek to bring more competition to the market through antitrust enforcement.

Concentration and integration in the payment cards market

The business of acquiring credit cards and debit cards in the international networks appears highly concentrated. The majority of national networks are characterised by a very high level of concentration in card acquiring. Issuing, on the other hand, is much less concentrated.

The inquiry compared the degrees of vertical integration in the national card systems in various Member States. Five of the national systems examined had the lowest degree of vertical integration, setting only technical standards and the parameters for network access. Meanwhile three systems surveyed had relatively high degrees of integration since they also conduct authorisation, processing and clearing of transactions.

The degree of vertical integration in the large international systems MasterCard and Visa differs from one Member State to another. In general these systems show moderate degrees of vertical integration. However in some Member States, transactions are routed through the network of a local network operator that also acquires all merchants in the market.

The sector inquiry has provided some indications that joint ventures in acquiring may be a structural issue leading to various entry problems for foreign acquirers. For example, local issuing banks may agree on preferential ('on us') interchange fees with the incumbent acquirer (an inter-bank association in which they have financial interests) but charge higher, multilaterally agreed interchange fees to any foreign acquirers attempting to compete with the incumbent.

Access to clearing facilities, as a pre-condition for banks to enter new markets, may be an obstacle where local banks have no commercial interest in sponsoring a potential competitor. In order to promote cross-border competition, card payment systems should be invited to set objective and verifiable rules to grant new entrants a right of access to sponsorship by one of the incumbent banks or – if technically feasible – set up a multilateral clearing platform.

The prohibition on cooperative agreements with competing networks or non-banks, i.e. co-branding, may hinder national debit card payment systems from entering into competition with MasterCard and Visa or impede retailers or other operators from entering into competition with the incumbent card issuer.

Cardholder fees

This inquiry has examined four kinds of fees charged to credit and debit cardholders: (i) annual fees per card; (ii) card issuance fees; (iii) fees per transaction; and (iv) account statement and billing information fees. Of these fees, annual fees per card are the most important component of cardholder revenues, for both debit and credit cards. The average levels of cardholder fees charged by Visa and MasterCard in a particular Member State tend to be similar, though levels can vary substantially across Member States.

Simple correlation analysis of sector inquiry data during the period 2000 to 2004 suggest that there is no strong negative relationship between the level of the cardholder fee and the level of the interchange fee. This pattern is common to both networks and relatively consistent over time. Moreover, an econometric estimation controlling for other variables that may affect the fee per card level shows that if the interchange fee increases by 1 Euro, typically only 25 cents are passed on to consumers in lower fees.

These findings challenge the hypothesis advanced by some industry participants and some of the economic literature that an increase in interchange fees is fully offset by reductions in cardholder fees. These results are consistent with the findings of the inquiry's analysis on profitability and may cast doubt on the relevance of the arguments put forward by industry participants and the economic literature concerning the role played by the interchange fee in the payment cards industry. Indeed, if issuers do not pass return the additional interchange fee revenues back to cardholders this implies that interchange fees are a way to transfer profits to the side of the scheme where they are least likely to be competed away.

Merchant fees

Small merchants on average pay 70% more for payment card acceptance than large merchants. In theory, this could be explained by the lower average costs of acquiring merchants with higher transaction volumes. However, a comparison of price differentials between large and small merchants in the international schemes (MC/Visa: 70%, Amex 50%, JCB 40%, Diners 35%) with those in domestic systems (7% on average) suggests that scale may not be the decisive factor. It could be that smaller merchants pay a premium for accepting MasterCard and Visa cards. If that were true, the differentiation of prices according to the size of the merchant could be a measure for the exercise of market power by banks within a given system.

Blending of merchant fees by acquirers can have direct implications for inter-network competition, as it removes an important parameter of price competition; differential MSC levels. The potential outcome of blending may be higher rates than the merchant needs to pay for acquiring services, since there is no pressure to drive down these charges through inter-network competition.

The ban on surc harging appears to restrict inter-network competition, notably by concealing the true cost of payment cards for consumers via cross-subsidisation, and may result in the use of non-optimal payment instruments. It has been suggested that the surcharging prohibition may constitute a barrier to entry for alternative non-cash payment instruments, such as mobile phones or e-money. Retailers surveyed in the inquiry were strongly in favour of the possibility to surcharge, since it would strengthen the incentive for consumers to use cheaper payment instruments.

Interchange fees

The Commission’s sector inquiry provides indications that interchange fees are not intrinsic to the operation of card payment systems. Several national systems operate without an interchange fee mechanism, resulting in generally lower merchant fees.

In the international networks, Visa and MasterCard, the inquiry revealed significant variations in the weighted average of national credit card interchange fees across the Member States. In 2004 the level of the highest fees (over 1.5% of transaction value) was two-and-a-half times greater than the lowest weighted average fees. For Visa and MasterCard debit cards, the highest fees were observed in some of the new Member States. For Maestro cards weighted average interchange fees were more than three times higher in some Member States than others, and more than four times higher for Visa debit cards.

The use of interchange fees may serve several purposes. Card payment networks argue that, given the typical set-up of card payment mechanisms, the card issuers typically bear the main costs of the payment system, while most of the revenues are collected on the acquiring side as merchant fees. Therefore, they claim that there is a need to redress cost imbalances by an interchange fee mechanism, i.e. a fee paid by the acquirers to the issuers. Other systems argued that interchange fees are a co-ordinating mechanism necessary to optimise the operation of four-party payment card systems.

Two competing assessments can be distilled from the economic literature on interchange fees in payment card systems: either that their effect is neutral and provides efficient incentives for card issuers to expand output; or that high interchange fees offer a means of transferring rent (which cannot be competed away) from acquiring to issuing banks. From a competition viewpoint, it is important to assess whether interchange fees are used to extract rents from merchants. Some of the inquiry's findings – in particular concerning large divergences in interchange fees between countries and between merchant segments – may provide indications that the setting of interchange fees could be subject to the exercise of market power in some Member States.

By concluding and acting on a basis of preferential interchange agreements, incumbent players, involved in both issuing and acquiring activities, may indirectly obstruct new entry to the acquiring by not extending the same favourable conditions to newcomers.

Profitability of the payment cards industry

Credit cards issuing is highly profitable. On a pan-EU scale, the inquiry estimates that credit card issuers had a weighted average profit-to-cost ratio of 65% in 2004 while debit card issuers had a weighted average profit ratio of 47%. In most EU Member States, weighted average profit ratios remained fairly stable over the period 2000 to 2004.

Interchange fees appear to magnify the profits of card issuers. It appears that 62% of all banks surveyed would still make profits with credit card issuing even if they did not receive any interchange fee revenues at all. In 23 EU Member States, at least one bank participating in the survey was able to make a profit from issuing credit cards without interchange fees. This exercise seems to partially invalidate explanations put forward by the industry that total system output would suffer if issuing were not subsidised through the transfer of revenues from acquirers. The aim of this analysis is not to argue in favour of a zero interchange fee. However, in the light of the results, it is legitimate to question the optimality of the current level of interchange fees in several countries.

Industry comments on possible overestimations of profitability arising from the inquiry's methodology appear most relevant to card issuing. It was argued that capital costs would be likely to be higher in issuing than on the acquiring side (e.g. as a result of banks' need to provision against default by cardholders). Thus, to the extent that issuers failed to accurately report their capital costs, the inquiry may underestimate total costs and overestimate profitability of card issuing. However this would not significantly alter the inquiry's findings on the relative profitability of issuing and acquiring activity; nor on the impact of interchange fees on the profitability of both sides of the market.

The inquiry has found high and persistent profit ratios in relatively mature markets, together with other evidence collected on entry barriers, suggesting the existence and exercise of market power in these markets. The question whether card issuers can offer payment cards at affordable prices to consumers in the absence of interchange fee revenues is also relevant for a competition analysis of interchange fee agreements. If the multilateral transfer of revenues were necessary for the operation of a payment card system, then multilateral interchange fee agreements may not be caught by Article 81(1) EC, even if the fees determine the prices charged by an acquirer to merchants. However, the above findings on the profitability of payment card issuing cast doubt on the assumption that in the absence of interchange fees, issuers could not recoup their costs from cardholders.

These observations do not exclude that the use of interchange fees may lead to certain efficiencies in the operation of a POS system. However, they seem to confirm some recent theoretical predictions in the literature on two-sided markets suggesting that privately optimal interchange fees may be too high from a welfare perspective; notably if merchant fees increase with interchange fees but issuers do not return the additional interchange fee revenues to cardholders.

Membership and governance rules

Most card payment systems reserve card issuing and merchant acquiring to credit institutions or entities controlled by credit institutions, which may inhibit processors from entering the business and from competing with the banks. The financial institution requirement may, however, no longer apply when the proposed Payment Services Directive is implemented.

Card systems in some Member States require that credit institutions are registered with the national central bank in order to participate in the domestic payment system, which may inhibit cross-border competition. To the extent that these rules are based on legislation or decisions by national central banks, they may also be in contradiction with the freedom of services and internal market directives.

The high level of joining fees and their structure may hinder effective intra-system competition. The joining fee in open domestic payment systems varies from no fee in Germany and in some systems in Italy to fees in millions of Euro in particular Member States.

Cross-border competition in acquiring

The provision of cross-border services to merchants is developing very slowly and is limited almost solely to the international networks, MasterCard and Visa. The sector inquiry revealed a number of competition barriers on the market, which were confirmed by several market participants:

- the obligation on foreign acquirers to pay the fallback interchange fee where local incumbent acquirers are able to agree a favourable "on us" interchange rate with domestic issuers, may create obstacle to new foreign entry;

- vertical integration of card payment systems and lack of multilateral clearing platforms may impede new entrants from competing with the incumbent in one segment of the market;

- membership requirements, such as those relating to registration with the local central bank or to being a credit institutions may exclude the provision of cross-border payment services. Also high joining fees for card payment systems may result in discouraging new entry;

- certain governance arrangements, such as obliging some members of the system to provide business-sensitive information to principle members, without reciprocal information sharing, may distort conditions for competition between the member banks; and

- diverging technical standards for message protocols and security requirements in national and international schemes hinder processors and terminal vendors from operating on a pan-European scale. This in turn inflates input costs for banks and ultimately for merchants, and in the same time serves as a barrier to entry for cross-border acquirers.

Payment infrastructures

Retail payment systems in the EU are not yet integrated and their organisation and structures remain highly varied. This means that a bank operating in different Member States has to join the various national systems, adapt to different standards and face different costs.

In most EU Member States there is one national clearing infrastructure, which is operated either by the central bank or by a membership association controlled by (the main) banks in one country. Some Member States may have two or more payment systems, which are not in competition but rather are complementary, clearing different payment instruments.

Access to clearing and settlement systems is necessary for any bank considering entering a retail banking market. Operators of the established infrastructures are potentially in a position to create entry barriers which may take a variety of forms. In particular some membership rules and the way the fee system is structured may raise barriers to entry for new or small players.

The advent of the Single Euro Payments Area will change the competitive landscape. SEPA aims to create a single market for payments throughout the Euro area by integrating national payments systems. This will permit economies of scale to be realised and make cross-border competition feasible. The end result should be more effective competition in the market for payment services.

Glossary

Automated Clearing House (ACH): electronic clearing system in which payment orders are exchanged among financial institutions primarily via magnetic media or telecommunication networks and handled by a data processing centre.

Automated teller machine (ATM): point where consumers can use plastic cards for withdrawing money.

Beneficiary fee charging option (BEN): charging option according to which the bank fees shall be paid by the payee.

Cardholder: the holder of the card, who uses it as a payment instrument.

Card acquirer (or acquiring institution): credit institution or other undertaking, and member of a card scheme that has a contractual relation with a merchant.

Card brand: the logo of a particular payment card that has been licensed for use in a given territory.

Cardholder fee: the one-off or recurrent fee (or a set of fees) paid by a typical cardholder for the ownership and/or use of a classic/standard debit and/or credit payment card (where no special conditions apply), as well for other ancillary services (e.g. account statement information).

Card issuer (or issuing institution): credit institution, and member of a card scheme, that has a contractual relation with a cardholder for the provision and use of a card of that card scheme. In a closed system, the card issuer is the scheme owner, while in open systems several credit institutions act as card issuers.

Card scheme owner: defines standards, rules, specifications and access policies and governs the card scheme.

Cheque: debit instrument in the form of written order from one party (the drawer) to another (the drawee; normally a bank) requiring the drawee to pay a specified sum on demand to the drawer or to a third party specified by the drawer when the instrument is presented to the payer’s bank.

Clearing: process of transmitting, reconciling and, in some cases, confirming payment orders between financial institutions prior to settlement, possibly including the netting of instructions and the establishment of final positions for settlement.

CredEuro Convention: it was established in November 2002 as a standard for the execution of a “basic” (meaning no added-value services) bank-to-bank pan-European credit transfer. The CredEuro Convention considers as basic EU payment a transfer in euro up to €50 000, with indication of IBAN/BIC and with charges allocated as SHARE.

Credit transfer: payment order (or sometimes a sequence of payment orders, which is referred to as standing orders) made for the purpose of placing funds at the disposal of the beneficiary.

Direct Debit: pre-authorised debit on the payer’s bank account initiated by the payee.

Direct Member: participant in a payment system which is responsible to the settlement agent for the settlement of its own payments, those of its customers and those of the indirect participants on whose behalf it is settling.

Four-party system (or open card payment system): the stakeholders involved are 1) the issuer, 2) the acquirer (may be the same as or different from the issuer), 3) the cardholder and 4) the merchant (in the case of ATM transactions it is usually the acquirer that offers its services via the ATM). Simply put, it can be said that “the parties involved are the cardholder, the merchant and their banks”. Examples are Visa, MasterCard, and several national schemes.

Indirect member: type of participant in a payment system that distinguishes from a direct member for its inability to directly perform some of the system activities (e.g. inputting of transfer orders, settlement). It thus requires the services of a direct member to perform those activities on its behalf, via a bilaterally negotiated agency agreement.

Interchange fee: fee paid by an acquiring institution to an issuing institution for each payment card transaction at the point of sale of a merchant. In certain networks, this may be positive in others it is zero.

International card system: has an international presence (issuers and acquirers operating in several countries). The fact that the cards issued in one country can be used in another country makes these systems international. Examples are Visa, MasterCard, American Express and Diners.

Large-value payment: a payment, generally involving a very large amount, which is mainly exchanged between banks or between participants in the financial markets and usually requires urgent and timely settlement.

Merchant: the entity that accepts payments by means of cards.

Merchant service charge (MSC) (or merchant fee or merchant discount rate): fee paid for each transaction by a merchant to an acquirer, who processes the merchant’s transaction through the network and obtains the funds from the cardholder’s bank (issuing institution). The transaction is considered to be executed when the corresponding funds, equal to the price of the sold item, are debited from the consumer’s account and, after deducting the merchant service charge, are credited to the merchant’s account.

National payment card system (or national/domestic payment card network/scheme): usually operates within a single country; i.e. the issuer and the acquirer are within the same country.

“On-us” transactions (as opposed to “off-us” transactions): in a narrow sense, on-us transactions are payment card transactions where the issuing bank and the acquiring bank are identical. This situation is prevalent in closed payment card systems. In a wider sense on-us transactions occur where the issuing bank and the acquiring bank are separate entities but pertain to a common group of banks. This situation typically arises where issuing banks set up a joint venture which acquires merchants. Transactions between this acquirer and its shareholders are often labelled "on-us" transactions, although strictly speaking issuing and acquiring banks are separate entities.

Our fee charging option (OUR): charging option according to which all bank fees shall be paid by the payer.

Payment card: card that allows the cardholder to make payments for goods and services at POS (point of sale) terminals or remotely (mail order, telephone order, internet) —card-not-present transactions, respectively. It may be one of the following:

- Debit card: a card that allows the cardholder to charge purchases directly and individually to a current account at a deposit-taking institution (serves as an access device to funds stored in bank accounts). It is recognized that debit cards may also be closely linked to other products offered by banks.

- Credit card: a card that allows the cardholder to make purchases up to a certain credit amount, which can then be settled in full by the end of a specified period or only in part, with the remaining balance taken as extended credit and being charged interest; credit cards may be linked to a current account at a deposit-taking bank, but also may be linked to an account that has been set up specifically for the use of the credit card.

- In this report deferred debit card , which is defined as card that allows the cardholder to make purchases but does not offer extended credit (the full amount of the debt incurred has to be settled by end of a specified period), is treated as a credit card.

Payment card system (or payment card scheme or payment card network) : technical and commercial infrastructure set up to serve one or more particular card brands and which provides the organisation, framework and rules necessary for the brand to function.

Point of sale (POS): point where consumers can use plastic cards for payment transactions at a merchant outlet (often a payment terminal).

Retail payment: payment between various consumers, businesses and governments of relatively low value and urgency. It is a payment which is not included in the definition of large-value payments.

Retail payment system : set of instruments, banking procedures and inter-bank funds transfer systems, which handle a large volume of retail payments.

Shared fee charging option (SHA): charging option according to which the bank fees shall be shared and paid both by the payer and the payee.

Settlement: an act which discharges obligations in respect of funds transfers between two or more parties. A settlement may be final or provisional.

Straight Through Processing (STP) transaction: automated end-to-end processing of payment transfers that can be processed without manual intervention and that includes the automated completion of confirmation, matching, generation, clearing and settlement of instructions.

Three-party system (or closed card payment system): the stakeholders involved are: 1) the card issuer and acquirer (it is the card scheme itself that fulfils both functions), 2) the cardholder and 3) the merchant. Examples are Diners, American Express and some national schemes.

[1] Available at http://europa.eu.int/comm/internal_market/finances/docs/white_paper/white_paper_en.pdf

[2] See the joint Commission/ECB statement on the vision for SEPA: http://europa.eu.int/rapid/pressReleasesAction.do?reference=IP/06/577&format=HTML&aged=1&language=EN&guiLanguage=en

[3] In terms of gross income.

[4] The terms 'banks' and 'retail banks' are used throughout this report to refer to credit institutions providing retail banking services.

[5] See “Time to Move Up A Gear" The European Commission's 2006 Annual Progress Report on Growth and Jobs at: http://europa.eu.int/growthandjobs/pdf/2006_annual_report_full_en.pdf

[6] http://ec.europa.eu/comm/competition/antitrust/others/sector_inquiries/financial_services/#key

[7] See, for instance Marc Ivaldi, Bruno Jullien, Patrick Rey, Paul Seabright, and Jean Tirole, The Economics of Tacit Collusion , p. 48: http://ec.europa.eu/comm/competition/mergers/studies_reports/the_economics_of_tacit_collusion_en.pdf

[8] See: OECD paper (2000): Mergers in Financial Services DAFFE/CLP (2000)17, p. 23

See also: Compecon Limited (2004): Study of Economic Impact of Increased Competition in Irish Banking Services , Report Prepared for the Bank of Scotland (Ireland), p. 23.

[9] The conduct of the Italian banking supervisor during the attempted takeover of Antonveneta in 2005 is a prominent example of such barriers.

[10] Switching rates tend to be higher in the New Member States than in the EU15. The inquiry's data suggest than between 5.4 and 6.6% of EU consumers changed their current account in 2005.

[11] ‘Full service’ banking provides a complete range of retail products to consumers and small businesses. In the case of consumers such products might start with current accounts and deposit accounts and extend to credit products (e.g. mortgages or loans), savings products (e.g. investment funds) and insurance.

[12] This factor was particularly emphasised by foreign banks seeking to expand into other Member States.

[13] In France and Spain the Economy Ministry still retains overall responsibility.

[14] See Interim Report II, Chapter 3 for a survey of the relevant literature.

[15] Since the 1980s there have been significant advances in supervisory tools and in the quality of banks’ own risk monitoring, notably the risk-based capital requirements framework developed by the Basel Committee. In the EU this framework is being enhanced and expanded through the Capital Requirements Directive. It is likely that these supervisory instruments have increased the stability of the banking system.

[16] The Commission has recently adopted two decisions concerning the so-called Livrets A and bleu in France. Under state aid rules, there is an extension of the formal investigation procedure into the fees paid by the State to Crédit Mutuel for distributing the ‘livret bleu’, to establish whether there has been overcompensation. The second decision is a letter of formal notice asking the French authorities to justify the necessity of the special rights granted to La Poste, the Caisses d’Épargne and Crédit Mutuel to distribute Livrets A and bleu. The Commission fears that these special rights may infringe the Treaty by raising obstacles to the freedom of establishment and freedom to provide services (Articles 43 and 49). Both decisions intend remove advantages to specific banks so to suppress barriers to entry on the French savings market and to widen the consumer’s choice in the field of financial services.

[17] This corresponds to around €550-600 per person across the EU.

[18] As described in the interim report, chapter 4, these estimates for illustration purposes did not imply a delineation of relevant markets in the sense of antitrust or merger control. Factors such as different supply and demand features for different retail products as well as the general preference of banking customers for local suppliers, the significance of a dense branch network and the need for the bank to be physically close to its customers tend to support the definition of narrower relevant product and geographic markets.

[19] See also page 48 of the Interim report II; as already described in this report some, in particular smaller Member States, were excluded from this evaluation due to variances and partial data distortion.

[20] Here 'domestic' banks as those where a controlling interest is held by a domestic institution. 'Foreign' banks have a controlling interest held by an institution domiciled in another Member State.

[21] For some Member States only three or four banks are described, depending on the number of banks sampled in the inquiry's market survey.

[22] For the purposes of the sector inquiry, the Commission defined SME banking as services (including current accounts, term loans, credit lines and leasing) for enterprises up to an annual turnover of €10 million.

[23] The choice of profitability measure used in the sector inquiry was criticised by some banks. It was alleged that the Commission’s methodology did not consider differing levels of risk. Some account was taken of this factor because banks were asked to specify risk provisions separately from costs and pre-tax profits. It is however true that the inquiry's estimates cannot account fully for differing levels of profitability arising from different risk conditions. Several banks argued for typical industry measures of profitability to be used, such as internal rate of return (IRR). However, IRR measures in retail banking would face methodological weaknesses, most notably in the consistent and accurate reporting of ‘retail banking’ assets by banks. For example, it would be very difficult for banks to apply common definitions of the labour, capital and IT employed in their retail division. Overall the Commission’s profitability ratio appears preferable since it relies on the relation between a reliable denominator (gross retail income) and a relatively more reliable numerator (pre-tax retail profits).

[24] Examination of the main component of the cost-income ratio – staffing costs – also supports the negative relationship between profitability and the cost ratio. At the country level, we observe that richer countries and some very small Member States have higher staffing costs, while in general the New Member States have among the lowest cost ratios in the Union.

[25] This point also applies over the long-term. Where two Member States at similar levels of economic development display markedly divergent rates of economic growth over several cycles, one would expect to see similar divergence in banking sector performance and profitability. Nonetheless, banks’ profitability in a fast-growing economy need not remain high or continue growing where market entry and competition exert countervailing pressure.

[26] Of course long-term economic growth and the development of credit markets are interlinked. Strong, sustained economic growth may be fuelled by continued expansion of credit and rising private debt.

[27] For example, in Ireland the Competition Authority produced a detailed report in September 2005 outlining extensive competition barriers (notably in the payment system), many of which are being addressed. Meanwhile in Denmark the Commission is aware of obstacles preventing some market participants enjoying a level playing field in all retail banking product markets.

[28] These factors may help explain the relatively low recent profitability of retail banking observed in Germany and Austria.

[29] This subject is discussed further in Chapter A.4 of this report.

[30] The OECD's methodology and definitions for estimating bank profitability can be found here: http://www.oecd.org/dataoecd/50/44/2373422.pdf

[31] The six non-OECD EU Member States are Cyprus, Estonia, Latvia, Lithuania, Malta and Slovenia.

[32] An additional finding was the general trend of falling tax rates as a share of banks’ operating profits for the clear majority of Member States. In some cases effective rates have fallen dramatically (e.g. in Sweden, Finland, Belgium and Luxembourg). The conjunction of rising pre-tax profits and falling effective tax rates implies that on average the post-tax profitability of European banks has increased significantly since the 1980s. This trend would be expected to strengthen profit incentives for European banks.

[33] Fixed costs are high because of the need to develop technology and set-up a network. Marginal costs tend to be low because credit information, once collected into a register, can simply be reproduced to clients.

[34] A survey of the economics of credit reporting systems is provided by JENTZSCH, N. (2006): The Economics and Regulation of Financial Privacy (Springer-Physica, Heidelberg/New York).

[35] BARRON, J.M. and STATEN, M. (2003) The Value of Comprehensive Credit Reports: Lessons from the U.S. Experience. In M. Miller (ed), Credit Reporting Systems and the International Economy (MIT Press, Cambridge): 273 – 311.

[36] Adverse selection problems arise when lenders cannot distinguish 'good' from 'bad' borrowers, with the result that lenders charge a higher average interest rate, reflecting the average risk they are taking on when lending to a borrower. Good borrowers then (on average) pay a price that is too high and bad borrowers one that is too low as good borrowers cross-subsidise good ones that should receive lower prices.

[37] JAPPELLI, T. and PAGANO, M. (2005): Role and Effects of Credit Information Sharing , Dipartimento di Scienze Economiche – Università degli Studi di Salerno, Working Paper No 136.

[38] WORLD BANK (2004). Doing Business in 2004 – Understanding Regulation (Oxford Univeristy Press, Oxford).

[39] Case C-238/05, Asnef Equifax, Servicios de Información sobre Solvencia y Crédito, SL v Asociación de Usuarios de Servicios Bancarios (Ausbanc), 23 November 2006.

[40] The register in Cyprus collects only information on cheque defaults.

[41] E.g. Experian, one of Europe's largest credit information providers, is owned by Great Universal Stores plc.

[42] Only four of the private credit registers have foreign ownership and this share is usually below 20%.

[43] ECB (2002): Memorandum of Understanding on the Exchange of Information among National Central Credit Registers for the Purpose of Passing it on to Reporting Institutions . Available at:

http://www.ecb.int/pub/pdf/other/moucreditregistersen.pdf

[44] A detailed analysis of laws has been conducted by JENTZSCH, N. (2006): The Economics and Regulation of Financial Privacy (Springer-Physica, Heidelberg/New York). The author rates countries in terms of rights of the supervisory authority, the data subject, obligations of credit bureaus, trans-border data flows and obligations of information furnishers and sanctions.

[45] The relevant text of the Proposal stipulates: “In the case of cross-border credit, each Member State shall ensure access for creditors from other Member States to databases in that Member State under non-discriminatory conditions.”

[46] Firms that offer deferred payment schemes such as hire purchase may also be considered as credit providers.

[47] Though this does not exclude that such foreign banks were subject to discrimination through other means, such as fee structures for the credit register.

[48] Moreover, widespread access may also present the risk for some consumers that a payment default to one type of provider (e.g. a telecomms company) could affect their credit rating with other providers such as banks or credit card companies.

[49] For analysis of the discriminatory effects of joining fees in payment card networks, see Chapter B.8 of Interim Report I; and in relation to retail payment systems, see Chapter B.10.

[50] The latter two registers are reportedly run on a non-profit basis, although it should be noted that their joining fees are considerably higher than those applied by most for-profit credit registers.

[51] Since negative data relates to borrowers who are in default, it covers only a small share of all retail borrowers; typically less than five per cent. The coverage of positive and full data reports will approach 100 per cent of the credit-active population.

[52] For example, because the technical definitions of payment defaults and delinquencies differ, consumers in different Member States will have their data stored in credit registers under different circumstances.

[53] Thus the ratios shown in Figure 10 may be modest underestimates of the actual level of coverage.

[54] The six public credit registers cover Austria, France, Italy, Latvia, Lithuania and Slovenia.

[55] The three public credit registers cover Belgium, Portugal and Spain.

[56] Belgium and France do not have private credit bureaus.

[57] For example some banks may not disclose credit balances only the credit limit. A rival lender’s credit scoring model would then take the highest-balance level as the credit limit. “Substituting the highest-balance level for the credit limit generally results in a higher estimate of credit utilization because the highest-balance amount is typically lower than the credit limit; the higher estimate leads, in turn, to a higher perceived level of credit risk for affected consumers.” (Avery, Bostic, Calem and Canner (2004: 306).

[58] Avery, B.R., Bostic, R.W, Calem, P.S., Canner, G.B. (2004). Credit Report Accuracy and Access to Credit, Federal Reserve Bulletin (Summer 2004): 297–322.

[59] The register’s management claimed this was a temporary situation.

[60] There is clearly no strong commercial incentive if the register is run on a non-profit basis; and if run on a for-profit basis, such profits would be generated from fees levied on its members (i.e. the banks owning the register).

[61] A specific basis for abuse of dominance might be ‘refusal to supply’, since the access conditions set by the credit register could implicitly exclude some types of credit provider from accessing the register.

[62] See estimates in the evaluation of the joint venture NatIxis created by the Groupe Banque Populaire and the Groupe Caisse d’Epargne ; in Bulletin Officiel de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) No. 7 bis du 15 septembre 2006 and Commission's own estimates.

[63] Creation of the joint venture NatIxis; see Bulletin Officiel de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) No. 7 bis du 15 septembre 2006.

[64] Rabobanks own estimates; see: http://www.rabobank.com/content/investors/longterm.jsp .

[65] OP Bank Group’s own estimates; see: https://www.op.fi/eng?cid=160115224 .

[66] The banks’ own assessments in terms of assets and/or liabilities; see http://www.creditocooperativo.it/template/default.asp?i_menuID=2396

http://www.icbpi.it/main.asp?tipo=1&ID=71 .

[67] See : http://gb2005.rzb.at/ereport.asp?fCompanyID=12&fAction=SHOWREPORT&freportid=94&fpageid=2542&fLangID=2 ; http://www.oenb.at/de/stat_melders/statistische_publika/Finanzinstitutionen/fi_oesterreichische_banken_entwickeln_sich_weiterhin_dynamisch_20060123.jsp .

[68] See : http://www.uni-marburg.de/fb02/ifg/aktuelles/news/plogmann and Commission’s own estimates.

[69] See also Article 1 (3) of Council Regulation (EC) No 1435/2003 of 22 July 2003 on the Statute for a European Cooperative Society (SCE); OJ L 207, 18.8.2003, p. 1–24.

[70] http://www.rzb.at/eBusiness/rzb_template1/1026359884948-1026359885014_1026067924320-1033775677229-NA-NA-EN.html .

[71] See: http://www.credit-agricole.fr/about-us-172/organisation-173/the-first-banking-group-in-france-174/leading-european-bank-with-operations-worldwide-567.html .

[72] See also Article 3 of Directive 2006/48/EC in OJ L 177/8 of 30.6.2006.

[73] See, for instance, evaluation of the joint venture NatIxis in: Bulletin Officiel de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) No. 7 bis du 15 septembre 2006.

[74] See, for instance, the internet presentation of the groupe Banque Populaire regarding its regional banks: http://www.banquepopulaire.fr/scripts/groupe/publigen/content/templates/show.asp?P=371&L=FR&SYNC=Y .

[75] This, for instance, applies to the Italian 'Banche Popolari'.

[76] Section 2 (2) 3 Kartellgesetz.

[77] Plus the Caisse d’Epargne which has been re-formed into a co-operative bank under state influence.

[78] See, for instance, Art. L 512-5 and 512-31 ( http://www.admi.net/jo/codemonetaire.html ).

[79] See press release IP/06/746 IP/06/746 of 7 June 2006.

[80] Own estimates of the national associations for different product markets (see: http://www.dsgv.de/download/aktuelles/Geschxftszahlenflyer_2005.pdf and http://www.ceca.es/CECA-CORPORATIVO/en/caja_b.html ).

[81] See press release of 11 April 2005 in http://www.bcee.lu/fr/decouvrir_la_bcee/pressroom .

[82] See: http://www.oenb.at/de/stat_melders/statistische_publika/Finanzinstitutionen/fi_oesterreichische_banken_entwickeln_sich_weiterhin_dynamisch_20060123.jsp.

[83] See: http://www.swedbank.com/sst/www/inf/out/fil/0,,348319,00.pdf .

[84] See evaluation of the joint venture NatIxis in Bulletin Officiel de la Concurrence, de la Consommation et de la Répression des Fraudes (DGCCRF) No. 7 bis du 15 septembre 2006.

[85] For instance, the Spanish Cajas are required to allocate at least half of their profits to reserves and the remainder to projects that fall under their social mandate (Obra Social). In Germany savings banks also have 'public missions' ruled in the individual savings banks laws of the States (Länder); for instance, savings banks' profits that are not allocated to reserves are supposed to serve general welfare purposes, mainly in as far as the surpluses distributed to the owners (Träger), e.g. municipalities or cooperations and associations of municipalities, the latter use them to finance general welfare tasks.

[86] If the term ‘publicly owned’ is used in this document, this means de facto owned by state bodies such as municipalities. Even though these savings banks, for instance, in Austria or Germany, often describe themselves as ‘ownerless’, their capital is provided by public bodies which also exercise decisive influence on the management and supervisory boards of an individual savings bank. Consequently, they are regarded as publicly owned. In Spain the presence of persons appointed by public entities in the General Assemblies is legally limited and cannot exceed 50%. In Italy savings banks have been reformed, transformed into foundations and/or privatised; some of them have merged with private banks, in other there is still public influence.

[87] There are very few exceptions to the rule; these are Sparkassen (‘Freie Sparkassen’) that are controlled by a foundation based on private law.

[88] The maximum representation of public entities such as local or regional governments in the governing bodies is legally limited to 50%.

[89] Though, as already mentioned for co-operative banks, the French Caisses d'Épargne act as a group and are regarded as a group.

[90] Section 2 (2) 4 Kartellgesetz in combination with section 30 (2a) of the Banking Act.

[91] Haftungsverbund Erste Bank / Sparkassen, Oberlandesgericht Wien als Kartellgericht, Interim decision 27 Kt 83/04-71 of 13 June 2006 (appealed); see http://www.bwb.gv.at/BWB/Aktuell/haftungsverb_070706.htm .

[92] See press release IP/06/1692 of 06/12/2006.

[93] Speech on the occasion of the 2006 World Savings Day, see http://www.bis.org/review/r061115b.pdf .

[94] See press release IP/06/746 IP/06/746 of 7 June 2006.

[95] 2006/736-738/EC: Commission Decisions of 20 October 2004 on aid granted by Germany to 7 Landesbanken, OJ L 307, 7.11.2006, p. 1–193; 2005/345/EC: Commission Decision of 18 February 2004 on restructuring aid implemented by Germany for Bankgesellschaft Berlin AG, OJ L 116, 04/05/2005 P. 0001 – 0054.

[96] See press release IP/03/49 of 15 January 2003 (Germany); IP/03/476 of 2 April 2003 (Austria).

[97] Commission decision to initiate the procedure concerning Dornbirner Sparkasse; OJ C 92 of 20 April 2006.

[98] Particularly where the banks in question are controlled or significantly influenced by public bodies.

[99] Section 2 (2) 4 Kartellgesetz in combination with section 30 (2a) of the Banking Act.

[100] Haftungsverbund Erste Bank / Sparkassen, Oberlandesgericht Wien als Kartellgericht, Interim decision 27 Kt 83/04-71 of 13 June 2006 (appealed); see http://www.bwb.gv.at/BWB/Aktuell/haftungsverb_070706.htm .

[101] Handelsblatt of 9 October 2006.

[102] http://www.elpais.com/articulo/economia/Competencia/expedienta/cajas/vascas/navarra/repartirse/mercado/elpepieco/20061028elpepieco_6/Tes.

[103] Page 96 of interim Report II provides a theoretical discussion of the main issues.

[104] For example, a mono-line provider of deposit accounts or current accounts would find it more difficult to capture market share where such products were typically tied to a mortgage or personal loan.

[105] For example, price competition in retail banking is generally held to be most intense in the mortgage market, with considerable pressure on interest rates. However, customers are typically also required to take out current accounts, for which prices are much less transparent. Thus it is difficult for customers to be sure which is the cheapest offer overall, based on the costs of the mortgage, current account and perhaps other tied products.

[106] Readers should note that these figures differ from those presented on pages 109 and 110 of Interim Report II. Those figures were based on simple averages; i.e. if one bank of three in a Member State reported tying, the simple tying ratio would be 33% irrespective of relative market shares. The figures reported below are weighted based on the sum of the sample market shares of banks reporting tying.

[107] Cross-selling ratios for the lead products can be used to check which is likely to be the dominant effect, with a cross-selling ratio close to 1.0 suggesting a high proportion of tying.

[108] See OECD paper (2000): Mergers in Financial Services DAFFE/CLP (2000)17, p. 22; See also DoJ Banking Merging Policy, US Department of Justice (1996): Consolidation in the Banking Industry: an Antitrust Overview , available at: http://www.usdoj.gov/atr/public/speeches/0657.pdf

[109] Commission decision of 11 March 1997 in Case IV/M.873 – Bank Austria/Creditanstalt, OJ C 160, 27.5.1997.

[110] That is to say, products that customers would otherwise buy separately.

[111] Potential cost savings arising from tying would most likely result from a lower average level of marketing, distribution and administration costs across the two products, compared to when they are sold separately.

[112] It should be remembered that the consumer accepting a tied mortgage or loan would also have to incur the switching costs associated with closing their existing current account.

[113] However the same bank is domiciled in an EU15 country. Its parent bank did not report any technical difficulties in unbundling its products and did not report tying any of the products surveyed by the Commission.

[114] As mentioned on page 9 in relation to the UK’s prohibition on tying of SME loans, the CC has found that servicing an SME loan from a current account at another UK bank is both feasible and efficient. It is not clear why this practice would not be viable in any other Member State.

[115] There is a distinction between current accounts required for ‘technical’ reasons (i.e. holding accounts for payments to the loan account); and current accounts to and from which all the customer’s payments are made. However, both types of account generate income for banks; the former at least through the ‘float’ value of holding a large balance for several days. Moreover, banks will be aware that a significant proportion of customers holding ‘technical’ accounts may eventually make all their payments through that bank.

[116] This proposed remedy was partly based on evidence suggesting that 84% of UK SMEs had a current account with only one bank. This suggested that the requirement of banks to use a current account in conjunction with a loan was likely to restrict competition.

[117] Where UK banks stated that an SME current account (or deposit account) was initially a technical requirement alongside the SME loan, the Competition Commission recommended that: “ the Bank shall use all reasonable endeavours to overcome or remedy such technical constraint within 12 months of the publication of the Report and notify the Director as soon as it has overcome or remedied it”. In other words the Competition Commission was not convinced of any fundamental technical reason why SME current accounts or deposit accounts had to be held with the same bank offering an SME loan. They stipulated that where such an account was necessary for technical reasons, it should be provided free of charge.

[118] For more information see: http://www.ifsra.ie/frame_main.asp?pg=%2Fnews%2Fnw%5Frecs%2Easp&nv=%2Fnews%2Fnw_nav.asp

[119] This section refers to banking 'prices' in general, though it should be recognised that the overall price or user cost of banking products is a composite of several elements, including interest rates and fees for payments and other services.

[120] The A recent study by McKinsey (McKinsey Quartely (2006): How Europe’s banks should prepare for payment reform -February 2006) compares different approaches to banks’ pricing policy. Three broad approaches are illustrated: institutions which are “balance earners” (revenues are largely earned from the interest on credit card balances or from interest rate margins on current accounts); “fee-oriented” banks (that charge their customers for everything from transactions to account maintenance) and “efficiency focused” banks that charge lower fees and earn a lower income from account balances but also keep processing costs down.

[121] See for example Klemperer, P. (1995). “Competition when consumers have switching costs” Review of Economic Studies 62: 515-539.

[122] These types of incentive are offered particularly by UK banks.

[123] Countries with less than three observations are not included in the present analysis.

[124] Chapter 6 of Interim Report II presents the inquiry's data on the use of various current account facilities in the Member States.

[125] The average number of current accounts for 2004 has been used.

[126] In these countries, annual fee for account management generally include a packet of free of charge services.

[127] http://www.agcm.it/agcm_eng/COSTAMPA/E_PRESS.NSF/0af75e5319fead23c12564ce00458021/06cf64f45f96ad23c12571ef003aaf92?OpenDocument

[128] It is possible that these figures underestimate the use of excess borrowing fees and default charges by banks in Europe. This may be the case where, in their responses to the inquiry, banks have reported income from 'excess borrowing fees' merely as the charges levied when consumers first exceed their borrowing limit, and not also as the charges levied on subsequent payments made while in excess of agreed borrowing limits.

[129] For further details on the OFT see: http://www.oft.gov.uk/News/Press+releases/2006/130-06.htm

[130] The amount has been chosen on the basis of the average value of ATM withdrawals provided by the RBR in “ATMs and Cash Dispensers Western Europe 2005”.

[131] The term 'off-us' denotes that the withdrawal is made at an ATM belonging to another payment network.

[132] The reason is due to some technical difficulties met in the interpretation and assessment of data received by banks. Whereas for banks operating in a Euro country the level of fees indicated for off-us transactions are valid for both national and cross border payments as consequence of Regulation 2560/2001, for the non-euro countries many banks did not explicitly indicated whether the level of fees listed are applicable uniquely to off-us transactions conducted within the country or also abroad.

[133] When a bank has indicated more than one relevant price, the fee charged for transactions in compliance with Regulation 2560/2001 has been considered.

[134] Greece is also among the EU Member States where interchange fees for credit transfers are highest.

[135] This helps explain why consumers that switch banks tend to respond to ‘push’ factors such as poor service or refusal of a loan rather than the ‘pull’ of better interest rates or product range.

[136] Financial Services Action Plan: Progress and prospects, Expert group on banking, 2004, p 6.

[137] KPMG: Banking beyond borders: will European customers buy it? 2004 http://www.us.kpmg.com/microsite/FSLibraryDotCom/docs/209-246%20FS%20Banking%20Beyond%20Borders%20survey%20v3.pdf

[138] Nederlandse Vereniging van Banken: NVB Bulletin June 2005, page 9

Other surveys include: Federation Bancaire Francaise: http://www.fbf.fr/web/internet/content_fbf.nsf/(WebPageList)/Observatoire+de+l+opinion/$File/IREQ_20emevague_6nov06.pdf

EPSI - European Performance Satisfaction Index (2006): Pan European Customer Satisfaction 2005 , Compiled by Rating Editorial Board, Göteborg, Sweden

[139] KIM, M., KLIGER, D., and VALE, B. (2003): Estimating switching costs: the case of banking , Journal of Financial Intermediation, Volume 12, Issue 1, pp. 25-56.

[140] E.g. in the responses from consumer associations in Finland, France, Germany, Greece, Slovenia and Sweden.

[141] DG Competition’s discussion paper on the application of Article 82 of the Treaty to exclusionary abuses, Rn 40. Available at: http://europa.eu.int/comm/competition/antitrust/others/discpaper2005.pdf

[142] KLEMPERER, P. (1995): Competition when consumers have switching costs , Review of Economic Studies 62 (4): 515-539.

[143] For example, it may be more difficult for foreign banks to enter and gain scale in a mature market (e.g. the personal loans market in Sweden) than in a maturing market (e.g. the credit cards market in Hungary). The overall degree of customer mobility on a particular product market may be higher than on other product markets, which could favour the entry of niche players.

[144] Such price comparisons can be made using common formats such as the European Single Information Sheet, and applying standard price variables such as the Annual Percentage Rate of Charge (APRC).

[145] Law of 4 August 2006 n. 248 – Decree Citizen-Consumer (Bersani Decree)

[146] http://www.pattichiari.it/inside.asp?id=387

[147] A number of responses on the consultation referred to observations of higher switching rates for products other than current accounts.

[148] We may draw a parallel to the relationship between average age and expected lifetime in a given population in population demography.

[149] See for example the low figures reported in Latvia, Slovakia and even Greece, where a large proportion of newly opened accounts produces a low average longevity.

[150] Eurostat Yearbook 2005, 2003 data http://epp.eurostat.ec.europa.eu/portal/page?_pageid=1334,49092079,1334_49092794&_dad=portal&_schema=PORTAL

[151] ECB Blue Book Addendum incorporating 2003 figures, August 2005, page 10.

[152] High account management fees in Italy have a pronounced effect on the consumption patterns of customers. According to ECB data, Italian customers hold only 0.63 current accounts per head; the lowest ratio in the EU15. Joint accounts are common in Italy; perhaps partly to spread the cost of high account management fees.

[153] The annual weighted average growth rate of the personal current account market for the EU25 between 2002 and 2005 was 2.41%.

[154] This suggests that differences across countries are probably more associated to different competitive strategies across countries than to the indicators commonly used as measures of market power.

[155] Reliable estimates of switching costs are not available across the EU Member States. As explained above, surveys of the level of customer satisfaction do not appear robust or comparable at country level (i.e. across Member States), nevermind across individual banks. In relation to financial stability, it is not clear how a specific indicator could be constructed across Member States to measure the stability of the retail banking sector.

[156] The interpretation of these estimates is as follows: the estimate is the variation of one variable when the other variable changes in one unit. For the purposes of regression coefficient estimates we have measured the relevant variables in the following units:

- churn in percentages

- profits in percentages

- market share in percentages

- longevity, in number of years

- cross-selling in ratios

[157] As reported in Chapter A.2 on banks' financial performance, average EU pre-tax profitability as a share of gross income was 28.8% in 2004.

[158] Indeed, the two measures in the two customer groups are highly positively correlated; 0,66 for churn, and 0,92 for longevity.

[159] One would expect that overall customer churn would have a stronger negative effect on profitability than SME churn. This is because around 80 per cent of retail banks' gross income (using the inquiry's definitions) is derived from consumers, with the remainder from SMEs.

[160] These results are significant at reasonable degrees, less than 1% and 5% respectively.

[161] For example, it is notable that banks in relatively concentrated markets such as Sweden and Finland report high profits and fairly low levels of customer churn.

[162] Source: ECB Blue Book (2006): Payment and Securities Settlement Systems in the European Union and in the Acceding Countries – Addendum Incorporating 2004 data. Available at: http://www.ecb.int/pub/pdf/other/bluebook2006addenden.pdf

[163] Same source as previous footnote.

[164] http://europa.eu.int/comm/internal_market/payments/docs/framework/com_2005_603_en.pdf

[165] In POS systems, interchange fees are typically paid by acquirers to issuers, but in principle they could go either way and there are systems where no such fees are charged.

[166] Interchange fees tend to flow from acquirer to issuer but in principle could flow in either direction (or be set at zero).

[167] For a full description of the data, methodology and sampling approach used, please refer to Interim Report I, page 13 to 16.

[168] The response rate was close to 95% for the questionnaire addressed to acquirers and issuers and virtually 100% for that addressed to payment card systems.

[169] It includes, for example, the ABR report on payment cards. DG Competition also consulted seminal economic literature on the payment card industry.

[170] Store cards (cards issued by non-banking institutions for use for payment in specified stores) were explicitly excluded from the scope of the questionnaire.

[171] These numbers rely solely on data provided by the network. The network reported the aggregate acquiring

turnover under the name of a licensing company. This aggregate turnover may in fact represent the cumulative acquiring turnover of at least 3 network licensees. The individual turnover data of these licensees were not supplied by the network and therefore are not considered in the analysis. Cross-border acquirers are not accounted for either. For these reasons, the level of concentration shown in the graph 50 may in fact significantly overestimate the true one. For the sake of consistency, however, it was decided to present the network’s reported numbers ‘as is’ (with no adjustments).

[172] This graph does not include acquiring volumes on co-branded cards if another facility (not that of the network considered) was used for transactions. Thus, no volumes are reported for some countries.

[173] For example, the collection of a signature on a receipt for the acceptance of a POS transaction.

[174] As well as the risk of incurring chargeback losses if a merchant goes bankrupt after a transaction is contested by a cardholder.

[175] Two countries were excluded due to data unavailability.

[176] A typical “business” cardholder pays 32 euros for a standard/classic card while a typical “consumer” cardholder pays 19 euros in the MasterCard network. Similarly, a typical “business” cardholder pays 34 euros for a standard/classic card while a typical “consumer” cardholder pays 18 euros in the Visa network.

[177] It should be noted that the issuance fee is not weighted by the validity period of the card.

[178] Only MasterCard and Visa are analysed, due to data unavailability.

[179] It should be noted that for some countries the sample is only for one network.

[180] The econometric analysis is provided in Annex 5 of Interim Report I.

[181] A card transaction is considered to be executed when the transaction amount is debited from the consumer’s account and, after deduction of the MSC, is credited to the merchant’s account.

[182] For some countries, no cross-reference to Visa levels was possible, as the set of countries with available MSC levels for the two networks differed somewhat in each case. Generally speaking, the maxima and minima found in this analysis should be treated as “local” rather than “global”, given that no entire set of data was available. Nonetheless, the variation noted will, if anything, be increased by adding countries to the analysis.

[183] These figures for the MSC in national (domestic) networks, however, do not cover other per-transaction fees that merchants in some countries may end up paying to local processors. These per-transaction processing fees may differ for small and large merchants, given their different transaction volumes and hence different “scale” factors. MasterCard and Visa MSC rates, however, seem to be “final” prices paid per transaction and therefore already include processing fees. Thus, once adjusted for the possible supplementary per-transaction processing fee, the price difference for small and large merchants in national networks may be somewhat greater.

Where the processing fee is already incorporated in the final MSC paid in national networks, the somewhat lower price difference for small and large merchants may also be explained by the fact that national processors tend to handle the entire volume of payments, including credit transfers and direct debits. This significantly raises the overall scale of the processing and therefore may, to some extent, weaken the incentive of processors (and therefore acquirers — if the fee is passed on through acquirers) to reward larger merchants for higher volumes. This therefore may, among other things, limit the price difference for larger and smaller merchants.

[184] The MSC has been replaced by a fixed annual fee paid by a merchant to an acquirer.

[185] OJ L 293-24 of 10-11-2001.

[186] Idem , paragraphs 11-12 and 54-58.

[187] See e.g., J-C. Rochet and J. Tirole, “ Cooperation among Competitors: Some Economics of Payment Card Associations ”, RAND Journal of Economics, Vol. 33, No 4, Winter 2002, pp. 549-570 (p. 562).

[188] Idem, p. 566.

[189] See e.g., Vickers, J., “ Competition Policy and the Invisible Price: How to Set the Interchange Fee? ”, 6 May 2005, p. 7-8.

[190] Commission Decision of 27 July 2002, OJ L 318/17 of 22 November 2002, pt. 9.

[191] The numbers given in this figure do not necessarily correspond to the officially announced national interchange fee level in a given country due to the existence of bilaterally agreed “on-us” fees. Thus, the levels depicted in the graph represent the weighted average levels of all fees applicable in a country (including the “on-us” fees).

[192] For some countries, the level of interchange fees is reported only for one network. However, this does not imply that the other network is not active in the respective geographical market.

[193] http://www.visaeurope.com/acceptingvisa/interchange.html.

[194] http://www.mastercardintl.com/corporate/mif_information.html.

[195] OJ Press Release of 24/07/2002, reference IP/02/1138.

[196] The Norwegian card scheme BAX also operates without an interchange fee.

[197] For instance in the Netherlands (PIN national network) or Belgium (Bancontact national network).

[198] Thus, if average transaction values with debit cards drop in the long run, Visa issuers would obtain higher revenues than MasterCard issuers due to the structure of their cross-border fees.

[199] For the a detailed presentation of the economic theory, please refer to the interim report, page 6 to 12

[200] See Rochet, J.-C. and J. Tirole, 2002, op. cit.

[201] 16 networks replied to the question as to the purpose of POS interchange fees and three indicated a specific purpose for ATM interchange fees. Where networks specifically commented on the purpose of ATM interchange fees, the reasons diverged from the reasons given for POS interchange fees.

[202] According to an industry expert consulted by DG Competition, some institutions may have difficulties in isolating the debit card business from other activities since debit cards may be an accessory product of current accounts.

[203] Some of these profit ratios are based on a limited number of observations, which means that results may not be entirely representative of profit ratios for a given country.

[204] The significant increase in the weighted profit ratio of country Q in 2004 is due to the inclusion of two important issuers in the sample only in this period.

[205] Naturally, it is straightforward to conclude that the number of profitable issuing institutions would be even greater for a reduced interchange income than for a zero income.

[206] Delaying the transfer of funds to the acquiring banks will allow the issuer to earn a return on the transfer amount and to recoup part of their costs for funding delayed payment by cardholders.

[207] Data on French card transactions are shown in the graph for domestic cards since domestic payments with a Carte Bleue (CB) card co-branded with a VISA or MasterCard logo count as pure CB transactions.

[208] Proposal for a Directive of the European Parliament and the Council on the payment services in the internal market and amending Directives 97/7/EC, 2000/12/EC and 2002/EC (COM(2005)603 final) .

[209] Based on replies from acquirers, as well as on the RBR Report on Payment Cards Western Europe, 2006. It needs to be noted, however, that some particularly large acquirers may see acquiring as their core activity and therefore might indeed have considered foreign entry for just acquiring business reasons.

[210] The share of total acquiring turnover generated cross-border rose by 2.2% in two years: .

[211] Payment Cards, Western Europe 2006, Retail Banking Research Ltd, International Review, p.18. Western Europe covers 17 European countries: UK, France, Germany, Spain, Netherlands, Italy, Sweden, Belgium, Finland, Denmark, Portugal, Austria, Ireland, Greece, Switzerland, Turkey, Norway.

[212] This practice is discussed in detail in Chapter B.8.

[213] For a detailed analysis of network joining fees see the previous chapter.

[214] For more details see paper from BIS (2000): Clearing And Settlement Arrangements For Retail Payments In Selected Countries

[215] Among systems with targeted membership, TIPANET (for co-operative banks), IBOS; Eurogiro (originally designed for national postal organisations) and others can be listed.

[216] An exception is the system KUBAS that clears only Credit Union transactions.

[217] For instance, in Germany and in Italy, additionally to the payment systems run by the relevant National Central Banks, there are also private inter-bank payment systems, settling retail payments for specific categories of institutions.

[218] Among the surveyed infrastructures, the Czech one (CERTIS), the Lithuanian (LITAS), the Maltese (MARIS) and the Slovakian (SIPS), are both large-value and retail payment systems.

[219] For instance, in the UK, the two surveyed payment systems (C&CCC and BACS) process different types of transactions (paper-based and electronic payments, respectively).

[220] According to a study conducted by RBR (2005), 80% of cross-border bank-to-bank credit transfers are still made through correspondent arrangements or intra-bank transactions.

[221] In addition to the three systems that only handle paper-based transactions, a number of surveyed infrastructures also clear at limited extent paper-based payment instruments (e.g. cheques).

[222] It should be noted that many payment systems reported their value of “total costs” rather than “operational costs”. We do not however expect these two values to differ significantly.

[223] Czech Republic, Estonia, Germany, Italy, Latvia, Lithuania, Malta, Portugal, Slovakia, Slovenia.

[224] Some systems use different terms to indicate direct members, such as participants, settlement members; and indirect members, ancillary members, sub-participants.

[225] The inquiry has not collected evidence on membership rules specifically for large-value systems.

[226] Finnish Banking Association’s Board is currently discussing the removal of this requirement.

[227] It should be noted that charges between direct and indirect members, as a result of their own private agreements, are not included in the present analysis.

[228] For further details see page 132 of 'Interim Report II: Current Accounts and Related Services'.

[229] The Dutch system has been excluded from the simulation as fees are negotiated by Interpay and its clients on a case-by-case basis. The Swedish system has been excluded as available data are not sufficient to calculate clearing prices charged to credit institutions.

[230] In Cyprus (JCCTransfer), Greece (DIAS), Lithuania (LITAS and KUBAS), the price per transaction does not depend on the bank size.

[231] See also BIS (March 2003): Policy issues for central banks in retail payments.

[232] A separate issue is whether banks can use infrastructure arrangements to raise rivals’ costs, or indeed to exclude them entirely. A variety of mechanisms would in principle be available to support such a strategy, ranging from ownership of infrastructure through control of technical standards, intellectual property rights etc.

[233] As an example, in France only the 12 founding members participate in the main decisions concerning the system.

[234] See: http://www.ecb.int/pub/pdf/other/policystatementretailpaymentservicesen.pdf

[235] See: http://europa.eu.int/comm/internal_market/payments/docs/framework/com_2005_603_en.pdf

[236] See: http://europa.eu.int/comm/internal_market/payments/docs/framework/com_2005_603_en.pdf

[237] SIBOS issues (2005): Looking beyond the boundaries – transformation of domestic ACHs , 5 September 2005.

[238] ECB (2006): Towards a Single Euro Payments Area – Objectives and deadlines , Fourth Progress Report, February 2006

[239] The banking industry uses the following three standardised fees payment option: shared (SHA), our(OUR) and beneficiary (BEN):

“SHA” - the bank fees shall be shared and paid both by the payer and the payee,

“OUR” - all bank fees shall be paid by the payer,

“BEN” - all bank fees shall be paid by the payee.

[pic]

Proportion of Number of Debit Card

Transactions Vs Number of Credit Card

Transactions in Old Member States, 2004, %

Debit;

65%

Credit ;

35%