TABLE OF CONTENTS

1........... Introduction.................................................................................................................... 3

2........... Functioning of the Directive............................................................................................. 5

2.1........ Ad-hoc report on the correct and effective application of the Directive............................. 6

2.2........ Questionnaire on the use of data.................................................................................... 10

2.3........ Statistics provided by Member States on the application of the Directive........................ 13

2.4........ Contributions from the expert group on the Directive ('EUSD' group)............................. 22

2.5........ Determination of the start-up and recurrent costs of implementation of the Directive:....... 24

3........... Economic effects and analysis....................................................................................... 26

3.1........ International deposits – BIS international locational banking statistics.............................. 26

3.2........ Euro-area deposits – ECB MFI statistics...................................................................... 41

3.3........ Switzerland – data from the Swiss National Bank (SNB)............................................... 50

3.4........ IMF Coordinated Portfolio Investment survey – Evolution of investments in debt securities 57

3.5........ Eurostat data – household savings/investment patterns................................................... 67

3.6........ Structured retail products.............................................................................................. 73

3.7........ UCITS Funds............................................................................................................... 78

3.8........ Insurance products....................................................................................................... 80

COMMISSION STAFF WORKING DOCUMENT

presenting an evaluation for the second review of the effects of the Council Direcive 2003/48/EC

Accompanying the document

REPORT FROM THE COMMISSION TO THE COUNCIL in accordance with Article 18 of Council Directive 2003/48/EC on taxation of savings income in the form of interest payments

1. Introduction

Art. 18 of Council Directive 2003/48/EC of 3 June 2003 on taxation of savings income in the form of interest payments (the "Savings Directive" or "Directive")[1] states that "The Commission shall report to the Council every three years on the operation of this Directive. On the basis of these reports the Commission shall, where appropriate, propose to the Council any amendments to the Directive that prove necessary in order better to ensure effective taxation of savings income and to remove undesirable distortions of competition."

2011 Report and the Commission Staff Working Document

This document has been prepared by the services of the European Commission for information purposes as an accompanying document for the second (2011) Report from the Commission to the Council on the operation of the Directive. This report commits only the Commission’s services involved in its preparation.

This document relies, in particular, on data provided on exchanged information/withholding tax as well as the answers from Member States to two questionnaires on the implementation of the Directive (i.e. from the perspective of the paying agent Member State) and the use of the data exchanged (i.e. from the perspective of the beneficial owner Member State).

The data have been sourced inter alia with the Bank of International Settlements (BIS), the European Central Bank (ECB), the Swiss Central Bank (SNB), the International Monetary Fund (IMF), Eurostat, and some private data providers for specific product markets as well as a report on administrative compliance costs commissioned to Deloitte. The data sources suffer from specific limitations and pose challenges in arriving at meaningful and robust conclusions. These limitations are outlined in the relevant sections of this document where the data sources are cited. In addition, the analysis of the data has attempted to make provision, whenever possible, for the effects of the financial crisis in 2008 and from the periods thereafter although this has not been possible for every data source.

The second report and this document's aims are not to identify any loopholes in the Directive since they were already identified in the report for the first review of the Directive (the "2008 Report")[2]. Such loopholes have been addressed in an amending proposal[3] adopted by the Commission on 13 November 2008 ("the Proposal") and the latest Council compromise text elaborated on the basis of the Proposal. Rather, the aim is to provide a more in-depth analysis of the implementation and functioning of the Directive and to analyse available evidence relevant to the issues addressed in the Proposal. The main findings of this document including the widespread use of offshore jurisdictions for intermediary entities and the growth in key markets that provide products comparable to debt claims reinforce the arguments for extending the scope of the Directive and the relevant agreements. The main findings of this document are reflected in the second (2011) Report from the Commission to the Council.

2008 Report and Amending Proposal

The 2008 Report covered the transposition and implementation of the Directive. It summarised the economic evaluation and recommendations contained in the Commission's services working document[4]. The changes identified to the Directive recommended in the 2008 Report were designed to clarify certain interpretational issues and to close loopholes. To that end, the Commission adopted an amending proposal[5] on 13 November 2008 to the Directive (the "Proposal") with a view to closing existing loopholes and better preventing tax evasion. The Proposal was accompanied by an Impact Assessment[6]. In particular, the Proposal seeks to improve on two major aspects of the Directive, i.e. to (i) better ensure the taxation of interest payments which are channelled through intermediate tax-exempted structures; and to (ii) extend the scope of the Directive to income equivalent to interest obtained through investments in some innovative financial products as well as in certain life insurance products.

The European Parliament supported the Proposal by its legislative resolution[7] on 24 April 2009 and the European Economic and Social Committee did likewise by its opinion[8] adopted on 13 May 2009. The Proposal is still under discussion at Council level, building on unanimous conclusions adopted on 2 December 2008[9] and on 9 June 2009[10].

The latest compromise text[11] is considered by the Commission to have received sufficient consensus from Member States to enable the commencement of negotiations with the 5 third European countries to update the Savings Agreements with them. On 17 June 2011 the Commission adopted a recommendation to start negotiations with these third countries to bring the EU Savings Agreements in line with the amendments to the Directive. The recommendation was presented at the ECOFIN Council on 12 July 2011.

2. Functioning of the Directive

The aim of the Directive is to enable savings income in the form of interest payments made in one Member State to beneficial owners who are individuals resident in another Member State to be made subject to effective taxation in accordance with the laws of the latter Member State. To that end, the Directive's focus is to act as an incentive to compliance, that is to increase the reporting of interest income falling under the Directive in the fiscal returns of taxpayers in the EU and to provide an information source for EU tax administrations to be used in specific controls and audits on taxpayers, usually on the basis of a risk assessment.

In order to assess the functioning of the Directive, the Commission services have relied on the following material, which forms the basis of this document and the second report:

(2.1) Ad-hoc report on the correct and effective application of the Savings Taxation Directive (SEC (2011)775 final[12]);

(2.2) Results of a questionnaire sent to tax administrations on the use of the data under the Directive, complemented by a discussion at the meeting of the Commission expert group on Administrative Cooperation in Direct Taxation[13] of 28/09/2011;

(2.3) Analysis of the evolution of statistics provided by the Directive;

(2.4) Update of the contributions of the experts in the EUSD group for the second review of the Directive;

(2.5) An independent study commissioned to assess the administrative burden on paying agents as a result of the Directive.

2.1. Ad-hoc report on the correct and effective application of the Directive

In order to analyse in detail the transposition of the Directive, the Commission drafted a detailed questionnaire consisting of 38 questions about the measures taken by Member States to ensure the proper functioning of the Directive in their respective territories and about their experience of how the agreements with non-EU countries and territories were functioning.

As a detailed description of the answers to each question would not be within the scope of this Report, the replies to the questions are grouped by general topics and are summarised below.

Identification of paying agents

This topic was covered by four questions and deals mainly with the criteria for a paying agent to fall within the scope of the reporting obligations under the Directive, i.e. what determines where a paying agent is "established". To this end, 9 Member States stated that the place of effective management was partially or exclusively relevant, while 7 Member States regarded as "established" only those legal persons duly registered/incorporated in their territory and another two Member States relied on the place where the business activity is conducted as a relevant criterion. In addition, 17 Member States confirmed that persons resident abroad are also treated as paying agents if they have permanent establishments in their territory. In the reverse scenario, i.e. reporting obligations of foreign branches, most Member States' implementing texts would treat these as outside the territorial scope of the Directive. Only one Member State has specific provisions covering such situations and 7 Member States would cover these by other general provisions, the Anti-Money-Laundering obligations, anti-abuse clauses or interpretation.

With regard to the "establishment" of trusts/trustees the replies revealed that there are many interpretational issues about how trusts should be treated under the current text of the Directive and about the legal requirements to be fulfilled to act as a trustee in civil law countries. There seems to be a significant level of inconsistent treatment between Member States. Some countries reported that a trustee can be a paying agent if acting in a business or professional capacity and the beneficiary of the trust is absolutely/immediately entitled to assets/income. Regarding structures that are neither individuals nor legal entities, almost all Member States confirmed that these would be within the scope of the Directive, while only 3 Member States stated that their implementing rules do not contain specific provisions for economic operators other than legal persons and individuals.

It must be noted that the Proposal contains provisions meant to eliminate the above mentioned problems, especially for the more controversial cases of entities and legal arrangements falling under Article 4(2) of the Directive. The criteria for the place of effective management of entities or legal arrangements are detailed extensively.

Beneficial owners

Five questions on beneficial owners' issues deal with the interpretation of situations where the individual receiving the interest payment provides evidence that the payment is not received or secured for his own benefit. 4 Member States provided details on the type of evidence used if an individual claims to have received the interest payment on another’s behalf and 2 Member States cited specifically the Anti-Money Laundering Directive[14], while most Member States outlined principles rather than citing actual types of evidence.

The question on situations where an individual claims to have received the interest payment on behalf of a discretionary trust was meant for Member States whose implementing rules do not consider a trust to be a paying agent upon receipt under Art. 4(2). In 16 Member States, if the beneficiary is not known at the time the interest is paid, the trustee is treated as the beneficial owner. MT and UK did not provide explicit confirmations that non-resident individuals who are trustees of discretionary trusts are treated as beneficial owners under the Directive. Member States with extensive legislation on trusts did not provide explicit confirmation that non-resident individuals who are trustees of discretionary trusts are treated as beneficial owners under the Directive which may mean that no beneficial owner is identified in the case of discretionary trusts.

Again the above mentioned difficulties would be eliminated would the Council adopt the corresponding elements of the Proposal.

Identity and residence

The part of the questionnaire dedicated to the identification of beneficial owners and their residence considered separately contractual relationships entered into before and after 1 January 2004, as covered by Article 3 of the Directive.

For contractual relationships entered into before 1 January 2004 the paying agent would rely on the information at its disposal, in particular information to be collected in order to comply with the anti-money laundering obligations. One question was designed to investigate to what extent there is alignment with the customer due diligence obligations where an individual is identified as a beneficial owner of a customer that is a legal entity or a legal arrangement. The replies to that question were divergent with 9 Member States referring to specific obligations regarding the use of information available under the customer due diligence obligations. A second question also referred to the alignment with the customer due diligence obligations, but also sought to clarify whether a change of the permanent address for anti-money laundering purposes would also lead to a revision of residence for the purposes of the Directive. 6 Member States replied negatively to that question.

The use of information collected for anti-money laundering purposes is broadly extended and clarified within the Proposal.

The main question for contractual relationships entered into after 1 January 2004 concentrated on cases where the beneficial owner presents a tax residence certificate (or other official evidence) proving actual residence for tax purposes to be in a Member State, but other than that of the address on the passport or identity card. 4 Member States replied that they have no provisions to assign tax residence to the Member State that issued the tax residence certificate.

This issue is addressed in the Proposal prescribing that the paying agent must rely on recently issued tax residence certificates from other Member States.

Residual entities/paying agents upon receipt

While the situation of an economic operator who pays interest to a beneficial owner appears to be understood and applied correctly across the board by Member States and economic operators, this may not necessarily be the case for the provisions on the ‘paying agent upon receipt’[15]. The main questions in that regard referred to the fulfilment of the conditions for exemption from the paying agent upon receipt provisions. 6 Member States provide guidance of a varying detailed and binding nature for demonstrating the fulfilment of these conditions.

An additional question requested further information on the second residual entity test - whether the entity’s profits are taxed under the general arrangements for business taxation (‘subject-to-tax test’, Article 4(2)(b) of the Directive). Two different solutions have been retained by Member States in regard to this test. According to the first solution, the test is passed (i.e. the entity is not a residual entity) even if the entity’s profits are not taxed at entity level but are taxed at the beneficial owner level[16]. Under the second solution, the entity whose profits are not taxed through the entity itself but via its beneficial owners would not pass the subject-to-tax test and would have to be treated as a ‘paying agent upon receipt/residual entity’ by the upstream economic operator making an interest payment to it. The first solution is shared by 5 Member States.

The subject-to-tax test is also applied differently in situations where the entity is being subject to tax under the general arrangements for business taxation, but that taxation is at a zero or very low rate. The entity would still be considered subject to tax in that case according to the solutions retained by 11 Member States.

The provisions for entities and legal arrangements falling under Article 4(2) of the Directive are extensively detailed in the Proposal, specifically referring to the taxation at the level of the entity/legal arrangement and addressing the cases of zero rate taxation or the lack of effective taxation. The annex of indicative lists for such entities/legal arrangements under the Proposal and the corresponding latest compromise text is also another aspect bringing more clarity to such situations.

Definition of interest payments

The questionnaire referred to the application of the concepts of substance over form and abuse of law with regard to the definition of interest in the application of double taxation treaties[17] and asked whether the same concepts are applied when implementing Article 6 of the Directive. It also asked whether there is a list of products that could potentially be re-classified as interest-producing debt claims. Either substance over form or abuse of law is used by 16 Member States in the application of the Directive, but none of those Member States have a list of products or examples that could potentially be re-classified as interest-producing debt claims.

The Proposal includes a number of additional types of products which could be viewed as similar to debt claims from an economic perspective.

Supervision of paying agent obligations

The questionnaire dealt extensively with the issue of the proper supervision of paying agent obligations by the relevant tax authorities of the Member States of establishment of those paying agents. When the replies to these questions are summarised, the general conclusion is that all Member States have introduced significant safeguards to ensure the correct implementation of the Directive. The approaches range from inducing compliance by way of cooperation with paying agents to imposing relevant penalties and sanctions in cases of non-compliance.

The conciseness of some replies did not allow the construction of a full and consistent picture of the actual audit procedures with regard to paying agents. The Commission encourages Member States to actively engage and cooperate with each other in order to strengthen and build best practices in their audit procedures.

Agreements with non-EU countries and territories

The final set of questions concentrated on the application of the EU savings agreements with each of the five non-EU Western European countries and the bilateral agreements between the Member States and each of the 10 dependent and associated territories. In their replies, 9 Member States raised issues with one or more of these agreements. 5 of these Member States referred specifically to the format of information exchanged with some non-EU countries and territories.

Conclusion

The replies to the questionnaire suggest that some Member States interpret certain provisions of the Directive differently. Some of the risks of differing interpretations thus highlighted had already been identified in the 2008 Report. These problems would be removed through the corresponding new rules contained in the Proposal.

2.2. Questionnaire on the use of data

A questionnaire was sent to Member State experts in the Administrative Cooperation in Direct Taxation (ACDT) group13 in July 2011 regarding the use of the data provided by the Directive by Member States. A meeting was held on 28 September 2011 which examined the responses to the questionnaire. All Member States had replied to the questionnaire by the date of this meeting apart from Ireland and Romania. The meeting also called on earlier work of the Group which examined the operation of exchange of information and withholding tax mechanisms of the Directive for the fiscal years 2009, 2008 and 2007.

The Commission assists in facilitating the transmission of data between Member States under the automatic exchange of information provisions of the Directive. Member States should then decide on the processes and application tools, according to their own national requirements, in order to make best use of this data for helping to ensure that the relevant interest payments are subject to effective taxation in accordance with their laws. The replies from Member States to the questionnaire demonstrate that there is scope for Member States to learn best practice from one another both in the transmission of data and in its use.

Questionnaire on the use of data

Question 1) asked whether Member States were aware of any tax audits having their origin in the information on their tax residents received from other Member States under the Directive. The majority of Member States use the information obtained from the Directive in order to perform specific audits on taxpayers. However, the extent of these audits varies between Member States. Some Member States, including AT, BE, DE and HU have provided statistics on the number of audits while most Member States report that audits have been carried out but are not aware of the number or of the results.

The structure set up by Member States to use the data exchanged under the Directive will naturally depend on the resources available and the amount of data they receive under the Directive. Italy has reported that it has a savings directive database which is used by its Regional Directorates and has set up a dedicated working group at central level to facilitate use of the data. For the UK, the data feeds into its system for evaluating risk for audit purposes.

From the replies, it would appear that a structured process of the dissemination of data from the receiving unit to the tax collections services in the Member State, and the latter's service feedback on the use of data, could improve the efficiency of the use of data to target specific taxpayers.

Some Member States have confirmed that they import the received data directly into their national tax databases for verification purposes. The integration of a savings directive database with the national tax database can lead to a more efficient processing and monitoring system to ensure that the beneficial owners are correctly registered as taxpayers and to determine whether their savings income has been correctly declared.

In addition, the development of risk management and a more automated process of cross-checking the data should be encouraged to limit the need for costly investigations of individual taxpayers.

Question 2) asked Member States about the quality of data received under the Directive from other Member States, including the percentage of the data received which cannot be used at all for tax assessment due to its poor quality. This process would be after the Member State receiving the data had attempted to complete or rectify the data by themselves. Member States were then asked whether they had received a higher share of unusable data from certain Member States and if yes, whether they had informed the Member State concerned and what had been the reaction.

There is a wide variation reported by Member States of data that cannot be used due to its poor quality. Further investigation on this variation is needed but the possibility of a better registration of the Tax identification number (TIN) with which to identify the taxpayer has been highlighted by Member States as a major factor for improving the quality of the data received. If such a number is properly reported by the paying agent, the tax administrations of Member States can then easily identify the beneficial owner. By default or for countries where no TIN exists, the date and place of birth must be correctly reported. If these essential elements are not properly reported by paying agents, Member States face difficulties in identifying the beneficial owners. To help paying agents to better identify a TIN on an official document, the Commission is developing an online checking system for TIN that will be available on the EUROPA website together with samples of official documents with which to identify the TIN.

Most Member States did not report any specific Member State from which they receive a high percentage of files which could not be processed due to the poor quality of data and stated that this varied from year to year and from Member State to Member State. Furthermore, most Member States have not contacted the corresponding Member State as a result of the poor quality of data received. These replies would tend to indicate the need for more regular bilateral follow-up between Member States which could be undertaken to improve the quality of data received.

For the period under review, the Member States have highlighted a clear improvement in the quality of data received under the Directive that they attribute to the structured format and common rules of procedures under which the data are reported. By comparison with exchange of information under bilateral treaties, the quality of the data received under the Directive is significantly higher.

Since the first feedback of exchange of information in 2006, the Commission has repeatedly advised Member States tax administrations to organise upfront quality checks of data submitted by paying agents before sending them to the receiving Member States, including reminding paying agents of their obligations under the Directive to provide complete and correct information. Around half of the Member States conduct checks on the content of the information received from paying agents. Such checks should be made systematically, bearing in mind that many Member States indicated that they prefer to receive incomplete records rather than no records at all as this would in any case allow them to try to match these records with their own national database.

In order to facilitate the transmission of data by paying agents to their tax authorities under the Directive, the Commission could consider, if Member States so wish, using a central web-based application for paying agents to download the standard FISC 153 format using a common character encoding for exchange of information under the Directive. This application would be the one that will be developed in future for eForms for the implementation of the new Directive[18] on Administrative cooperation in the field of (direct) taxation (Council Directive 2011/16/EU of 11 March 2011) and would provide the following advantages: an on-line application with built in controls on the input of data; very low implementation costs for Member States; no deployment would be necessary; and maintenance could be simplified and centralised at Commission level.

Question 3) asked Member States whether they were aware of cases where the information received under the Directive has lead to investigations on the sources of the underlying funds even where the income tax on the interest payment concerned was not substantial or the income thereof was exempt and if so to provide details.

Most Member States replied that they had not used this data for this purpose or are not aware of the results given that it is part of the general audit process of taxpayers, as commented by them for question 1) of the questionnaire. However, even for those Member States who replied that the use of the data from the Directive was not relevant to their country as interest income is not taxed[19], this data would nevertheless be of use to tax administrations for identifying assets held abroad.

Question 4) asked whether the certificate procedure under Article 13(1)(b) provided any useful information for tax assessment and collection. Most Member States replied that they had not used this for tax assessment. This was reinforced by the comments in the ACDT meeting of 28 September 2011 which indicated that Member States preferred other sources of information for tax assessment or that few certificates had been issued by their tax administrations. The Proposal addresses this issue and no longer provides an option for the Member State of the paying agent to allow only the certificate procedure to be used for exemption from the withholding tax under the transitional regime of the Directive.

Question 5) addressed a similar issue by asking whether the certificate under Article 13(2) was issued by the central office or by the local office which is responsible for dealing with the assessment of personal tax due by the taxpayer requesting it. If issued by a central office, it was asked to the Member State to explain how would its local office which is responsible for that taxpayer be informed about the issuance and the detailed content of the certificate. The replies indicated that the certificate was issued in most Member States by the local office. When it was issued by the Central office, Member States had a system in place to inform local offices about the issuance of the certificate.

Question 6) asked whether Member States had any evidence that the introduction of the Directive has led to better compliance by their taxpayers in the recording of interest payments coming within the scope of the Directive (i.e. interest income received from abroad actually reported in fiscal declarations) and, if so, to provide details.

Although most Member States have not yet undertaken a quantitative assessment, those that have carried out an assessment have indicated positive compliance results. It appears nevertheless that the collection/control systems of Member States may not be able to provide them with information that is specific enough for income amounts falling under the Directive to gauge increased compliance.

Establishing criteria (for example the reporting of interest income earned abroad in the fiscal declaration) for measuring how far the Directive gets as an incentive to compliance, is crucial for measuring the effectiveness of the Directive, but also for the effectiveness of the validation/control procedures of the Member States. DK has provided an example of best practice by sampling the compliance of taxpayers for various source incomes, including that of the Directive, to assess its effectiveness. Given that Member States have data available to identify the taxpayer, that fiscal declarations ask taxpayers to record amounts falling under the Directive and that they know the population affected by the Directive, then Member States should be able to statistically assess the effectiveness of the Directive.

It would also be important for Member States to consider the use of indices or identifying benchmarks for gauging assets held abroad that give rise to the income being reported under the Directive through the use of national and EU statistics (refer to work done on the simulation exercise on ECB data performed in section 3.2 of this document).

Finally for Question 7) Member States were asked whether they had performed an analysis on the administrative costs borne by their tax administration for its obligations under the Directive including the data collection and exchange and also the costs of control for its correct application. If so, they were asked to provide the main results of any such studies.

Most Member States did not report any analysis on administrative costs due to the Directive. The Commission encourages Member States to make such an analysis in order to perform a cost/benefit analysis and to exchange best practice with other Member States.

Conclusion:

Member States have expressed satisfaction with the overall system of automatic exchange of information provided for by the Directive to enable them to ensure that interest payments are effectively taxed. For the period under review, Member States have indicated a clear increase in the quality of data received that they attribute to the structured format and common rules of procedure under which the data is reported.

However, there is still some progress to be made in reporting by paying agents to their tax administrations notably by the correct recording of the TIN and/or the date and place of birth. There is also scope for Member States to work towards more effective use of the data through best practices. In particular, regarding the quality of data, Member States should consider the use of controls on data submitted by their paying agents to ensure that it is correct according to their obligations under the Directive.

2.3. Statistics provided by Member States on the application of the Directive

In the first review of the Directive, the analysis of statistics was limited to the second half of 2005 and 2006. At that stage the Commission had received 16 Member States' statistics for 2006 and 8 Member States' statistics for 2005. The report to Council suggested that this lack of data and the fact that definitions on data had not yet been agreed had restricted the analysis that the Commission could perform. The second review of the Directive includes the years 2007, 2008 and 2009 as well as updated data for 2005 and 2006.

In its Conclusions[20] of the meeting of 12/05/2008, the Council lists for the first time a set of statistics which could be used by the Commission services to assess the efficiency and effectiveness of the Directive. The Commission may thus quote data contained in those statistics in its report under Article 18 of that Directive, if it deems necessary for such an assessment. The statistics in question are the following:

· Amount of interest payments/sales proceeds between MS[21]: For the Member States exchanging information or availing of the voluntary disclosure provision (of Article 13(1) of the Directive), the amount of interest payments and sales proceeds within their territory subject to exchange of information (under Article 9 of the Directive), split by Member State. The amount of interest by type according to the categories of Article 8(2) of the Directive split by Member State.

· Withholding tax: Luxemburg and Austria (and Belgium until 31.12.2009) communicate the total amount of tax revenue shared from the withholding tax country, split by Member State of residence of the beneficial owners.

· Number of beneficial owners resident in other Member States, split by Member State, if this information is available to the authorities.

· Paying agents per sending Member State involved in the exchange of information or withholding tax.

In addition, there is also information about elements like reporting on Article 4(2) and certificates for exemption from the savings withholding tax.

Table (1) below shows the evolution of interest payments and sales proceeds reported by Member States using information exchanged and sent to other Member States. Data submitted for this review has increased significantly from the first review of the Directive. Furthermore, in the ACDT meeting of 28 September, Member States reported a high submission rate of data from other Member States as stipulated in the working arrangements of the ACDT group. All Member States have submitted information to the Commission up to and including the fiscal year 2009[22], apart from Ireland that had technical problems in compiling the data for 2009, and Sweden which has not provided the Commission with the exchange of information data element since the inception of the Directive.

As with the first review, the largest economies provide the highest amount of information exchanged although in large financial centres like Luxemburg this figure has also been significant. Peak year for the exchange of information was 2007 when information amounting to EUR 38.9 billion was exchanged. The majority of this came from the United Kingdom (EUR 24.4 billion), which mostly related to information on sales proceeds reported under Article 8.2 (b) (EUR 23.4 billion) exchanged with Germany. In 2009, the total amount of information exchanged went down to EUR 9.9 billion with significant decreases in Luxemburg (down to EUR 1.6 billion from EUR 5.7 billion in 2008) and the United Kingdom (down to EUR 808 million from EUR 3.4 billion in 2008).

Table 1: Interest payments and sales proceeds reported by countries using information exchange* (million Euros)

EU Member States || || || 2005 (2nd half) || 2006 || 2007 || 2008 || 2009

Austria || || || n.a. || n.a. || n.a. || n.a. || 202,56[23]

Belgium || || || n.a. || n.a. || n.a. || n.a. || n.a.

Bulgaria** || || || n.a. || n.a. || 1,54 || 2,03 || 3,64

Cyprus || || || 5,26 || 15,05 || 25,41 || 32,39 || 35,62

Czech Republic || || || 2,92 || 17,81 || 26,75 || 24,49 || 14,50

Germany || || || 660,73 || 1392,06 || 942,09 || 761,75 || 1125,00

Denmark || || || -. || 415,31 || 693,10 || 168,94 || 164,91

Estonia || || || -. || 4,40 || -. || 1,34 || 0,89

Spain || || || 488,11 || 423,42 || 421,21 || 383,02 || 295,61

Finland || || || 26,02 || 60,93 || 7,80 || 18,13 || 3,49

France || || || 568,14 || 2020,04 || 2485,41 || 2495,89 || 1576,99

Greece || || || 6,85 || 23,11 || 19,46 || 25,21 || 13,98

Hungary || || || - || 5,22 || 6,60 || 11,13 || 20,56

Ireland || || || 258,88 || 771,00 || 1901,24 || 1403,27 || n.a.

Italy || || || 280,53 || 1615,92 || 1232,97 || 1913,71 || 3407,78

Lithuania || || || - || 0,09 || 0,18 || 0,37 || 0,47

Latvia || || || 0,18 || 0,65 || 1,38 || 2,12 || 1,93

Luxemburg || || || 1119,79 || 4188,68 || 6344,34 || 5650,77 || 1641,40

Malta || || || 1,02 || 2,10 || 1,73 || 4,75 || 2,68

Netherlands || || || 107,83 || 800,14 || 389,14 || 408,06 || 481,18

Poland || || || 0,00 || 15,40 || 10,05 || 5,57 || 7,38

Portugal || || || - || 0,56 || 5,18 || 17,68 || 21,12

Romania** || || || n.a || n.a. || 7,34 || 10,75 || 22,39

Sweden || || || - || -. || -. || -. || -.

Slovenia || || || 0,59 || 1,35 || 1,48 || 3,07 || 37,93

Slovakia || || || 1,87 || 4,76 || - || 8,25 || 3,25

United Kingdom || || || 9132,49 || 485,38 || 24425,01 || 3356,86 || 808,29

Total || || || 12661,21 || 12263,40 || 38949,40 || 16709,55 || 9893,52

Third Countries || || || || || || ||

Switzerland || || || 70,50 || 549,12 || 500,89 || 516,80 || 444,30

Dependent and Associated Territories || ||

Anguilla || || || -. || - || - || - || -

Aruba || || || 0,01 || 0,09 || 0,18 || 0,24 || 0,12

Guernsey || || || - || - || - || 775,95 || 148,90

Jersey || || || - || - || - || 410,61 || -

Cayman Islands || || || 8,81 || 18,02 || 25,55 || 17,48 || 8,78

Netherland Antilles || || || - || 0,05 || - || 0,09 || 0,63

Montserrat || || || - || - || - || - || -

|| || || || || || ||

*Amount of interest payments and sales proceeds under Art. 9 subject to exchange of information/ voluntary disclosure, (Beneficial owners and residual entities) ||

**Bulgaria and Romania joined the EU in 2007 therefore no data is available for 2005 and 2006. ||

In analysing the data, one cannot just compare the overall amounts of information to be exchanged, for example the drop in information exchanged from 2008 to 2009, to verify the effectiveness of the Directive. Rather, account should be taken of a number of factors discussed below which include interest rates on deposits, maturities, the cross-border deposit base, and the relative weight of sales proceeds compared to interest income reported under the Directive.

A noticeable feature of the data is the high variability of amounts reported by Member States. The principal reason for the variation is that table (1) includes both interest income and sales proceeds. The amounts of sales proceeds may be reported as the gross proceeds from the sale in one year and as net gains in another year leading to a different basis of comparison. This effect also applies if gross income received from UCITS distributions are used in one year and only the interest element of those distributions is used in the following year. In order to mitigate the possible effects outlined above, table (2) below has been refined to detail only the interest income elements reported by Member States (Article 6 (1) (a) and (c)). This is derived from the national statistics supplied by Member States (reporting table 2.1 and, where possible, table 3) which relate to amounts of interest payments subject to exchange of information and voluntary disclosure under Article 9 of the Directive split by residence of the beneficial owner.

Table (2): Information exchanged on interest payments sent by Member States to other EU Member States, fiscal years (million Euros) – Bulgaria and Romania joined the EU in 2007 therefore no data is available for 2005 and 2006. n.a. – not applicable.

Member State || 2005 (2nd half) || 2006 || 2007 || 2008 || 2009

AT || n.a. || n.a. || n.a. || n.a. || 47,6823

BE || n.a. || n.a. || n.a. || n.a. || n.a.

BG || n.a. || n.a. || 1,42 || 2,03 || 3,64

CY || 5,23 || 15,01 || 25,44 || 32,39 || 35,61

CZ || 2,92 || 7,82 || 9,68 || 11,52 || 8,33

DE || 109,08 || 181,82 || 158,14 || 196,48 || 287,70

DK || - || 119,90 || 124,90 || 51,26 || 58,37

EE || - || 4,40 || - || 1,34 || 0,89

EL || 6,69 || 22,83 || 19,46 || 25,20 || 13,98

ES || 224,25 || 198,86 || 273,30 || 256,21 || 220,91

FI || 26,02 || 60,93 || 7,80 || 2,86 || 1,85

FR || 41,1 || 113,55 || 123,63 || 149,88 || 123,78

HU || - || 3,72 || 3,06 || 7,07 || 17,10

IE || 131,14 || 426,96 || 795,36 || 290,5 || -

IT || 11,3 || 124,91 || 141,84 || 84,12 || 803,16

LT || - || 0,09 || 0,18 || 0,37 || 0,48

LU || 76,79 || 270,46 || 368,01 || 432,54 || 102,83

LV || 0,17 || 0,64 || 1,32 || 2,12 || 1,93

MT || 1,01 || 2,10 || 1,73 || 4,62 || 2,65

NL || 88,08 || 789,42 || 200,30 || 232,98 || 236,73

PL || - || 7,28 || 10,05 || 5,57 || 7,37

PT || 2 k || ,56 || 3,74 || 16,93 || 20,22

RO || n.a. || n.a. || 7,31 || 10,75 || 22,14

SE || - || - || - || - || -

SI || 0,6 || 1,35 || 1,48 || 3,07 || 37,54

SK || 0,95 || 4,52 || - || 7,76 || 2,53

UK || 293,63 || 293,90 || 1.282,93 || 562,69 || 259,39

Total || 1.018,96 || 2.651,03 || 3.561,08 || 2.390,26 || 2.316,81

While 2009 with EUR 2.3 billion in total interest payments reported has a very similar amount to the total reported in 2008, the differences in individual Member States from one year to the next are significant. The decrease in interest payments reported could be due to a change in the interest rate that households receive on deposits. Table 3 below provides an overview of the interest rates over the period 2005-2009 for euro-area Member States and the UK:

Table 3 Evolution of interest rates for household deposits (source: ECB)

|| 2005 || 2006 || 2007 || 2008 || 2009

Euro (interest rate over all maturities || 2,42% || 2,54% || 3,04% || 3,58% || 2,84%

UK (interest rates maturities up to 2 years) || 4,59% || 4,52% || 5,29% || 5,59% || 3,87%

Interest rates have remained relatively stable until the end of 2008 when the financial crisis emerged after which interest rates on deposits fell significantly. Interest payments received by households also depend on the maturities of interest deposits within the respective years. Data from the ECB on deposits held by households in the euro-area has shown an increase in the share of overnight deposits[24] from 24,51% of total deposits in 2003 to 46,38% in 2011. This has largely been at the expense of the share of deposits with a maturity up to one year which have fallen from 47,37% in 2003 of total deposits to 25,55% in 2011. This would normally result in lower interest payments received due to the lower interest rate on overnight deposits compared to deposits with a maturity.

The evolution of interest rates should also be put in context with the evolution of cross-border deposits. Data from the ECB has shown that euro-area household cross-border deposits have been relatively stable in recent years peaking at EUR 287,4 billion in September 2008, before dropping slightly to EUR 280,7 billion in May 2011. Data for cross-border deposits of households outside the euro-area were not available.

The decrease in interest rates, if not the level of cross-border deposits, would tend to justify a lower amount of interest payments in 2009. However, for many of the Member States there is no correlation between this interest rate decrease and a decrease in the amounts of interest income reported. This could be explained by investors shifting to domestic deposits or investing in products outside the scope of the Directive but again, the level of stability of cross-border deposits at least for Euro households would tend to argue against this. In addition, Eurostat data (refer to section 3.5) does not provide evidence of a decrease in deposits held by households in the EU.

There has also been a high fluctuation from one year to the next in the number of paying agents reporting information to the tax administrations of their Member State, although for most Member States this has stabilised in recent years. Similarly, there is noticeable fluctuation from one year to the next of the number of beneficial owners reported by many Member States. Given the stability of cross-border deposits, one would have expected to observe more continuity in the number of beneficial owners reported by Member States under the Directive.

Separate work using the ECB figures (refer to section 3.2) on the completeness of interest payments being reported under the Directive according to cross-border deposits of households in the euro-area highlights the importance of Member States to make checks on the quantity of data submitted to them by their paying agents. Within the framework of the preparation of the ad-hoc report (section 2.1 of this document), Member States were requested to provide details on specific audits on paying agents regarding their reporting responsibilities and on the capacity of tax administrations to verify whether all income reported by paying agents under Article 6 (1) is complete. The replies by Member States indicated that most Member States rely primarily on the correct application of their guidelines by paying agents for the correct identification of relevant income to be reported, although they have provisions in place to perform audits and apply sanctions for non-compliance with the Directive.

It would be therefore beneficial for Member States to consider upfront controls on the data exchanged by paying agents. The following is a non-exhaustive list of possible controls that could be implemented to ensure the completeness of information exchanged under the Directive:

(i) Central register to be established by each Member States which lists paying agents established in their jurisdiction. This would ensure correct monitoring of the data supplied by paying agents is complete;

(ii) Examination of the data received by paying agents for completeness: variations of number of beneficial owners and amounts exchanged under the Directive should be compared from one year to the next and fluctuations investigated; comparison of interest income reported to the size of the bank (capitalisation) and the client base;

(iii) Cooperation between Member States in order to strengthen the audit procedures relating to paying agents including paying agents' systems and internal control guidelines;

(iv) the development of benchmarks and comparisons to other sources of data to ensure completeness of data – for example national statistics on the reporting of cross-border deposits.

Member States should continue their work in the ACDT group on how to improve the reporting of the income elements exchanged under the Directive. This work and the collaboration between Member States is important not just for this Directive but for the technical framework to be designed for automatic exchange of information to be provided by the new Directive[25] on Administrative cooperation.

In addition, it also appears that no data for residual entities (defined under Article 4(2)) has been submitted by a number of Member States. The 2008 Review has identified problems with the interpretation of this Article therefore verification should be undertaken by the Member States concerned to identify the reasons for non/under reporting of this data element.

Withholding tax:

Until 31.12.2009, 3 Member States levied a withholding tax on the interest income reported under the Directive: AT, BE and LU. Since 01.01.2010 BE has changed over to the exchange of information mechanism. In LU, beneficial owners have the option of (i) authorising voluntary exchange of information; or (ii) making use of the exemption certificate provided by their country of residence; or (iii) being subject to the levying of the withholding tax. Table (6) below gives details of the amounts of tax revenue shared on the basis of the 75% of the withholding tax transferred to the Member States of residence of beneficial owners :

Table 4: Tax revenue shared by countries with withholding tax regime (million Euros)

EU Member States || || 2005 (2nd half) || 2006 || 2007 || 2008 || 2009

Austria || || || 9,48 || 44,32 || 59,53 || 67,01 || 49,48

Belgium || || || 7,51 || 19,61 || 25,92 || -. || -a.

Luxemburg || || || 35,90 || 124,59 || 153,00 || 174,72 || 122,95

Total || || || 52,89 || 188,52 || 238,45 || 241,72 || 172,43

Third Countries || || || || || || ||

Andorra || || || 3,50 || 12,77 || 16,34 || 18,84 || 13,94

Liechtenstein || || || 1,94 || 7,08 || 9,06 || 11,30 || 7,17

Monaco || || || 3,75 || 11,70 || -. || -. || -.

San Marino || || || 1,13 || 7,47 || 10,75 || 15,40 || 14,02

Switzerland || || || 77,23 || 255,92 || 298,23 || 348,90 || 265,64

Total || || || 87,54 || 294,93 || 334,39 || 394,43 || 300,78

Dependent and Associated Territories || || || || ||

British Virgin Islands || || 0,00 || - || - || -. || -

Turks and Caicos || || || 0,01 || 0,02 || - || - || -

Guernsey || || || 4,93 || 16,83 || - || 14,24 || 5,51

Jersey || || || 13,26 || 32,15 || 38,34 || 33,55 || 9,94

Isle of Man || || || 13,26 || 20,35 || 23,39 || 16,89 || 7,12

Netherlands Antilles || || - || 0,05 || 0,15 || 0,08 || 0,11

Total || || || 31,47 || 69,39 || 61,88 || 64,76 || 22,67

Overall total || || || 171,90 || 552,84 || 634,73 || 700,92 || 495,87

Tax withheld under the Directive and the agreements providing for it has decreased from EUR 700,9 million in 2008 to EUR 495,9 million in 2009, although there has been a steady increase in the amounts withheld from the inception of Directive in 2005 until 2008. There is no analysis from the countries involved on why there has been a decrease, but it would seem likely that part of the reason would be the fall in interest rates in 2009. Switzerland contributed the most tax (53,57% of total) with Luxemburg being the next largest (24,8%). The benefit of aforementioned controls on completeness of information exchanged under the Directive could also be applied and adapted to the specifics of the withholding tax regime.

Concerning which Member States received the tax, Germany (41, 5%) and Italy (28, 96%) continue to be the Member States receiving the most of this revenue, which is explained by their size and proximity to major financial centres that operate the withholding tax mechanism.

Conclusion:

Information exchanged under the Directive peaked in 2007 with total of EUR 38.9 billion and has decreased to EUR 9.9 billion in 2009. However, further refinement of the data is needed, for example separating sales proceeds from interest payments and assessing the evolution of interest rates/maturities on cross-border deposits, in order to arrive at meaningful conclusions on the data exchanged. The variability of the data exchanged and the results of the ECB coverage analysis (section 3.2 of this document) indicate that Member States could benefit from better monitoring of the completeness of data submitted by paying agents.

2.4. Contributions from the expert group on the Directive ('EUSD' group)[26].

A group of business experts (Expert Group on Taxation of Savings "EUSD Group" - EUSD = "European Union Savings Directive") had been set up to assist the Commission's Services in their first review of the functioning of the Directive as provided under Article 18. Details of the contributions from the expert group for the first review can be found on the TAXUD website26.

On 17 February, 2011 a meeting of the Group took place concerning the second review of the Directive. In the meeting, the Commission outlined the elements to be contained in the second review, including the methodology to be used and data sources. Experts were asked to provide their feedback on the elements contained in the second review, to provide data input to facilitate the second review and to update, where appropriate, the comments they provided for the first review of the Directive. A summary of the responses is given below.

The contribution of EBF[27] related to its comments on the text of the Presidency compromise elaborated on the basis of the Proposal (version FISC 170 text[28]) and released on 25 November 2009. . The EBF has graded its comments according to their level of importance including the following which they deem as being very important:

Paying agents obligations (Art. 3 (2) (b)): the implementation date of the Proposal should be fixed to around 1 year after the Proposal enters into force including for the new products to be included in the scope of the Directive (life insurance/securities) under the Art. 6 definition of an interest payment. In addition, for contractual relations entered into, or transactions carried out in the absence of contractual relations, on or after 1 January 2004 and before the date referred to above, EBF recommends that the current text should be amended so that no TIN or equivalent is considered to be available when it was not on the passport or the official identity card or any other documentary proof of identity presented by the beneficial owner.

Tax certificates and residence (Art. 3 (3)): EBF is concerned about the burden put on paying agents to obtain a tax certificate for determining the tax residence of the beneficial owner.

Place of effective management and residual entities (Art. 4 (2)): EBF considers that economic operators, even those with a direct relationship with the payee, may not be able to determine the place of effective management, particularly where this is by reference to where the key management decisions are taken relating to the assets producing interest payments. Consequently, paying agents should be able to decide between the place of establishment or the place of effective management in Articles 2 (3) and 4 (2) of the Proposal.

Approval of data providers for implementing measures: Under Art. 18a (1), EBF are concerned about whether the Commission should have a role in "approving" data providers as stipulated in sub-paragraph (a). EBF proposes that the Commission and the relevant industry representative bodies might consider working together to produce an indicative list of approved data providers.

In its contribution, CEA[29] was concerned about the possible risk of the duplication of reporting requirements of the Proposal and the new Directive on administrative cooperation, Council Directive 2011/16/EU, in addition to the consequences for the competitiveness of the European insurance industry in comparison with non-EU financial centres. CEA considers that the Commission should wait until the new Directive on administrative cooperation25 has been implemented, in particular on the list of information to be exchanged under Article 8 of the Directive, before considering the new reporting requirements contained in the Proposal. CEA also considers that the new reporting requirements should have been included in the scope of the study to measure the start up and recurrent costs of implementation of the Directive (see following section 2.5). CEA also provided details on cross-border insurance premiums, which are given in Section 3.8 of this document.

EFAMA[30] stated that their comments submitted in July 2009 remain valid. They stressed the importance of achieving a level playing field between competing investment products like investment funds and certain insurance products which are marketed to retail investors and fulfil the same investment needs. In particular EFAMA considered that the Proposal did not achieve this aim by containing a grandfathering provision for insurance contracts issued before 1 July 2010. Secondly, EFAMA considered that the Commission should reflect on providing generally applicable guidelines for the new rules regarding insurance products that will enter the scope of the Directive. Finally, EFAMA provided data on the UCITS/non-UCITS ratio and the investment pattern of those (between debt and equity) details of which can be found in Section 3.6 of this document.

EFSA[31] confirmed its comments for the first review made to the Commission of the letters dated 24 April 2008 and 21 May 2007. EFSA also supported the comments, including the need for clarity and practicality in the Directive, made by EBF (see above). In order to maintain a level playing field with market operators in non EU jurisdictions, EFSA stressed that any tightening of the provisions of the existing Directive must at the very least be matched by corresponding tightening for the "equivalent measures" which are in place (Savings taxation agreements) in Switzerland, Liechtenstein, and the other "third countries" and dependent territories. According to EFSA the lack of a level playing field is already present with the Directive and related Savings tax agreements and EFSA is concerned that the Proposal might accentuate these effects. Secondly, any changes that are to be made to the Directive should ensure that adequate time is provided for Member States to draw up and to implement the necessary local fiscal rules in order to give paying agents sufficient time to implement the necessary system changes.

In its contribution, FECIF[32] provided a cost study on administrative burden among its members which indicated that 20% of their costs are taken up with legal, compliance and regulatory obligations including those linked to the Directive. Therefore, FECIF stressed the importance of ensuring that a disproportionate administrative burden is not imposed on paying agents by the Directive.

The aforementioned points and earlier comments of the Expert Group have been taken into account in the Council discussions on the Proposal.

EuroInvestors[33] proposed an expert who was nominated to the group in 2011. In order to safeguard the interests of investors, EuroInvestors suggests that the Directive should contain an obligation or at least a recommendation from the Member State of residence of the beneficial owner to grant full relief from double taxation by taking into account the taxes already paid in the country of the source of income. Secondly, there should be an obligation for paying agents to disclose, free of charge and in a timely manner, a spreadsheet detailing for each 'saving income' the withholding tax levied pursuant to the Directive. The Commission considers that both proposals do merit further consideration, but would not fit within the scope of the current Proposal.

2.5. Determination of the start-up and recurrent costs of implementation of the Directive:

An evaluation was produced by Deloitte and had the following objectives:

· Estimate the start-up and recurrent costs incurred on the implementation of the Directive according to the EU Standard Cost Model;

· Identify the most burdensome provisions in the current text of the Directive; and

· Summarise the feedback obtained from the economic operators on more cost efficient solutions that would retain and enhance the accuracy and integrity of the information exchanged, without undermining the objectives of the Directive.

The study focuses on the administrative burden incurred under the Directive by paying agents rather than the tax authorities. The latter were asked to provide their feedback to this for question 7) of the questionnaire on the use of the data (refer to section 2.2).

Essentially the study broke down the requirements of the Directive into 2 key IOs (information obligations):

IO1: based on article 8 (1) which requires the Paying Agent to report details on the Beneficial Owner and the interest payments to the competent authority; article 3(2) requires the Paying Agent to establish the identity of the Beneficial Owner: name, address, tax identification number or date/place of birth (by means of the ID card/passport); and article 3(3) which requires the Paying Agent to establish the residence of the Beneficial Owner.

IO2: based on Article 4(2) which requires paying agents to report the name, address and interest paid to 'residual entities' established in another Member State.

Results of the study:

In terms of the start-up and recurrent costs, the Directive is not considered to be particularly burdensome. For all EU Member States (apart from Austria, Belgium and Luxemburg), IO1 has been estimated at 114 million Euro per annum which represents 1,2% of the total interest payments/sales proceeds[34]. For Luxemburg, using a different method of extrapolation, the administrative burden was estimated at EUR 35 million. No extrapolation was possible for Belgium and Austria due to data limitations. It should be noted that these figures should be interpreted with caution due to the low participation rate of paying agents in the study. Nevertheless, the majority of respondents noted that the information needed to fulfil this obligation is already available to them under Anti-money laundering legislation14, domestic legislation or internal business practices.

For IO2, it was not possible to obtain data as most paying agents surveyed indicated that, due to their business model, all their clients were individuals. However, in terms of administrative burden, none of the respondents for which IO2 applies indicated that they would collect such information if not required to do so by the Directive.

In terms of the provisions for IO1, respondents reported that the most burdensome activities involve understanding the requirements of the Directive, training, and getting acquainted with the IO. In terms of products, classifying the proceeds from sales of UCITS income to see whether they fall within the provisions of the Directive are difficult to monitor as they are not business as usual costs and therefore bring a high administrative burden[35].

The report recommends that the provisions on the classification for UCITS and paying agent on receipt (IO2) should be clarified in order to reduce the administrative burden for paying agents. It should be noted that these points have already been taken up by the first review and are incorporated in the Proposal for the Directive by providing, for example, Annexes which list potential intermediary structures to be classified as look-through entities or paying agents upon receipt.

The report indicated that the cost for paying agents with a low number of Beneficial Owners were higher than for paying agents with a high number of Beneficial Owners primarily due to more automated processes for the latter and the lower recurrent costs per beneficial owner. It would therefore seem appropriate for paying agents to consider whether they can centralise the reporting of information under the Directive in order to take advantage of economies of scale.

Some respondents also noted it may be useful for Member States to consider streamlining different types of reporting for the same information (towards national authorities and for the purposes of the Directive) which could potentially reduce the administrative burden.

3. Economic effects and analysis 3.1. International deposits – BIS international locational banking statistics

Description

The International Locational Banking Statistics of the Bank for International Settlements (BIS) includes quarterly data on assets and liabilities of domestic banks and branches of foreign banks situated in the 43 reporting countries broken down on a bilateral basis according to the vis-à-vis[36] country of their foreign counterparts. The positions are reported on a gross[37] and unconsolidated[38] basis and in USD million.

The reporting countries include most OECD countries[39], several developing countries and some offshore financial centres[40]. From the financial centres reporting to the BIS that are part of the Savings agreements network, only Switzerland and Guernsey agreed to disclose their bilateral data through the BIS for the purpose of the Directive's review. None of the offshore financial centres[41] which are not part of the Savings agreements network agreed to disclose their bilateral data.

A breakdown of the data by instrument, currency, sector, country of residence of counterparty, and country of the reporting banks is available. The breakdown by instrument provides details for principal items of international assets (loans) and liabilities (deposits) that are not represented by negotiable securities. These loans and deposits reflect both interbank positions and inter-office/branch balances. The sectoral breakdown includes the international positions with non-banks as an "of which" item, which allows a separate analysis of non-bank balances. Non-bank balances would include any non-bank borrower/depositor, i.e. non-bank financial institutions, non-financial companies, individuals and other entities[42].

In addition, fiduciary/trustee businesses, i.e. funds received by banks on a fiduciary basis and re-deposited on the banks' name but on the account of the depositor, are included in the balances of liabilities and assets respectively.

Studies that use the data

In Johannesen (2010)[43], based on BIS data, the author estimates the effect of the entry into force of the EU-Swiss Savings Tax Agreement[44] on Swiss bank deposits held by EU residents. The Swiss bank deposits held by EU residents are used as a treatment group, while the Swiss bank deposits held by non-EU residents is used as a control group. The study finds that around the time of the entry into force of the Agreement the deposits of EU residents decreased by more than 40% and claims that the estimated treatment effect is large and very robust, with implied elasticity in the range 2.5-3. The entry into force of the Agreement is seen as a turning point in the evolution of deposits of EU residents and non-EU residents. The author finds that the average annual growth rates for the quarters after the entry into force of the Agreement (-9.1% for EU vs. 19.2% for non-EU) as proof of the impact on deposits of EU residents. Johannesen finds similar reactions, although not as strong and definite, with regard to Luxemburg and Jersey.

While acknowledging that the sectoral breakdown for the "non-bank" balances does not allow a separate data series for balances with individuals[45], the author argues that, at least for Switzerland as a reporting country, the "non-bank" group would consist predominantly of households, and even if there would be also companies in that group it would tend to bring down the estimated elasticity.

Previous Savings Directive review

The 2008 Report and the accompanying working document on economic evaluation used the BIS data to measure the trends in the international deposits across three groups of countries: (i) EU Member States that exchange information; (ii) EU Member States and third countries or jurisdictions applying a withholding tax; and (iii) other jurisdictions outside the network of Savings Agreements. The comparison of the quarterly growth rate for the three categories revealed that until 2004 growth rates varied much less in the countries in groups (i) and (ii) than in the jurisdictions outside the network of Savings Agreements (group (iii)). After that period the development for all groups appeared similar.

Secondly, the share of each of the groups in the total external non-bank deposits from the Member States and the third countries and territories in the network of Savings Agreements was observed for the period 2002-2007. It was found that the share of the Member States and third countries or jurisdictions applying a withholding tax (group (ii)) had decreased from 35% to 29.3% between mid-2003 and mid-2005, with a corresponding increase in the share of other jurisdictions outside the network of Savings Agreements (group (iii)). These shares have been stable since this period.

Thirdly, a comparison was made between the three groups with regard to the share of non-bank deposits[46] in the total of all deposits (i.e. from banks and non-banks). This share was found to be around 35% for groups (ii) and (iii) and around 20% for the Member States that exchange information (group (i)). This, coupled with the fact that the findings from the previous paragraph were not matched by similar findings with regard to deposits from banks, had suggested that non-bank depositors have a preference for countries maintaining some form of bank secrecy.

Limitations of the data

Several limitations exist which must be taken into consideration when analysing the results stemming from an analysis of BIS data. These are summarised below.

With regard to the sectoral breakdown, it is not possible to differentiate between deposits from individuals, companies or other structures within the total amount of the non-bank balances. Therefore, any changes to the deposits should be interpreted with caution. As mentioned above, Johannesen assumes in his paper that most deposits are held by individuals. While it is impossible to know the exact share of individuals in the depositor base, the country split could give an indication as to whether Johannesen's assumption is correct. Indeed, the top five non-EU vis-à-vis countries/territories are offshore centres with another offshore centre at ninth place. While many entities established in those countries/territories may be held directly or indirectly by individual beneficial owners, deposits may not necessarily be held directly by individuals but rather by intermediary structures. Therefore, it appears that Johannesen's assumption appears rather strong from that perspective. Some indication on the share of banks' business with private individuals may be derived from the publication "Banks in Switzerland 2010"[47] where the share of private customers varied between 30% and 50%. Another indication may be derived from comparing the BIS data (for all non-financial balances) with the ECB data (for households) where the share of private customers for UK and LU varies between 25% and 40%.

While many Member States report to the BIS, there are several which do not report. In addition, many third countries/territories, both in the network of Savings Agreements and some important financial centres in South-East Asia that report to the BIS, did not agree to the disclosure of their bilateral claims/liabilities. In fact, only Switzerland and Guernsey from the Savings Agreements network have agreed to disclosure through the BIS. This would enable only a partial analysis from the point of view of the jurisdictions within the scope of the Savings Agreements as locations of potential paying agents. Nevertheless, the vis-à-vis country data for the countries that agreed to the disclosure of their bilateral positions is a good basis for evaluating the deposit base attributable to depositors from most offshore centres. It can reasonably be assumed that the non-bank sector in those jurisdictions does not consist primarily of industrial companies or individuals, but of intermediary entities.

Despite the limitations highlighted above, the data could be used to give an approximation of global trends on the evolution of the residence of the holders of debt claim liabilities in key jurisdictions inside and outside the scope of the Directive. In addition, the data can be used to measure the interconnectedness of different groups of countries in financial terms. This is especially important with regard to intermediary entities from offshore centres.

Evolution of financial centres – aggregate BIS data

As mentioned above, the previous (2008) review of the Directive grouped (i) EU Member States that exchange information; (ii) EU Member States and third countries or jurisdictions applying a withholding tax; and (iii) other jurisdictions outside the network of Savings Agreements and performed comparisons of the evolution of international deposits between these categories. Unfortunately, from the second group, only the relevant Member States, Switzerland and Guernsey have agreed to disclose their bilateral data for the purpose of the review. From the third group some important offshore financial centres like Singapore, Hong Kong, Macao, Malaysia and Taiwan have not agreed to disclose their bilateral data for the purpose of the review. Therefore, the results reproduced for this review may not be comparable to those produced for the 2008 review.

Publicly available data[48] allows for a full comparison along the lines of the 2008 review, i.e. grouping counties into different groups and observing the evolution of their shares in non-bank positions. There is an additional limitation in that the amounts compared are for positions with all counterparties and not only the EU. As mentioned previously, Johannesen contrasted the evolution of deposits by EU non-bank sector against the non-EU non-bank sector for Switzerland and to a lesser extent for Jersey and Luxemburg. Such differing trends would not arise in this comparison, since it is based on the aggregated data for all vis-à-vis countries. For the purposes of the review, countries have been divided into the following categories:

Table (5): Categories used for the purposes of the review:

Category || Countries included

MS info exchange || EU Member States that exchange information under the Directive

MS WHT || EU Member States that apply a withholding tax under the Directive

Agr WHT/info exchange || Countries or jurisdictions with a Savings agreements

RoW onshore || Third countries considered to be onshore financial centres

RoW offshore || Third countries considered to be offshore financial centres

Figure (1): Share of BIS reporting countries in non-bank deposits, by category

Figure (1) demonstrates that the shares remain relatively stable at around 43% (2.8% average quarterly growth rate) for the Member States that exchange information, and 7% (2% average quarterly growth rate) for the Member States that levy a withholding tax[49]. The share of the offshore third countries group[50] also remained relatively stable at around 10% (1.95% average quarterly growth rate). However, the share of the onshore third countries group[51] increased from 8.85% in 2000 to almost 20% in 2010 (5.02% average quarterly growth rate) mostly at the expense of the third countries group that are part of the network of savings agreements whose share fell from 27.78% in 2000 to 20.69% in 2011.

The publicly available BIS data for those countries that did not disclose their bilateral data could in addition give an indication of their relevant importance in terms of international deposits[52] (figure (2)). Figure (2): Trends for the major offshore jurisdictions within the network of Savings Agreements, based on all foreign non-bank deposits

The balances for Cayman Islands start at values comparable to those for Switzerland, but with an average quarterly growth rate of 3.72% until the third quarter of 2007 exceed these values consistently after 2003 and, unlike Switzerland (average quarterly growth rate of 2,07%) and Jersey (average quarterly growth rate of 4,76%), do not seem to be affected during the last financial crisis. The balance for Cayman Islands is actually the second largest of all BIS reporting countries and comparable to that for the United States, as seen from the data for the first quarter of 2011 below in figure (3).

Figure (3): 5 Largest BIS reporting countries/jurisdictions based on foreign non-bank deposits

The share of EU positions in total worldwide positions with Switzerland and Guernsey, which ranges between 25% and 60%, may serve as a general point of departure is assessing the share of EU customer related business in addition to some of the above-mentioned centres.

Evolution of international deposits of EU residents

As mentioned above, an entire replication and an update of the 2008 review based on a split among different groups of Member States is not possible as some of the dependent territories no longer allow the disclosure of their bilateral data for the purpose of the review. Nevertheless, the split mentioned below in figure (4) yields comparable results.[53]

Figure (4): Share of investment location of deposits of EU non-bank depositors in BIS reporting countries.

The share of Member States that exchange information in the deposits by EU non-bank sector remains relatively stable at around 56% (reaching 60% at the end of 2010), while the share of third countries increases from 4% in 2000 to above 20% in the years 2006-2008, primarily at the expense of Member States that levy a withholding tax and jurisdictions within the network of Savings Agreements. However, it should be noted that the absence of data on Cayman Islands and Jersey from the group of jurisdictions within the network of Savings Agreements is a serious limitation to the comparison. As demonstrated in figure (3), based on publicly available BIS data, the non-bank deposit balances in the Cayman Islands is larger than that of Switzerland.

Evolution of international deposits by the non-bank sector from jurisdictions in Annex I and Annex II

The Proposal that resulted from the 2008 Report and is currently under discussion in the Council includes an "Indicative list of categories of entities and legal arrangements which are considered to be not subject to effective taxation, for the purposes of Article 2(3)". Annex I consists of a list of jurisdictions where the entities established would be subject to look-through provisions[54]. Annex II contains a list of "Entities and legal arrangements whose place of establishment or place of effective management is in a country or jurisdiction listed in Article 17(2), to which Article 2(3) applies pending the adoption by the country or jurisdiction concerned of provisions equivalent to those of Article 4(2)"[55].

In order to put the deposits from those jurisdictions into perspective, figure (5) compares the evolution of deposits in EU Member States and Savings Agreement countries ("EUSD->") by EU non-bank depositors ("->EU") with those BIS reporting countries/jurisdictions included in Annex I ("->Annex I") and Annex II ("->Annex II") of the Proposal.

Figure (5): Evolution of deposits of EU non-bank depositors vs. depositors from Annex I and Annex II.

The results show that the deposits of the non-bank sector entities from those jurisdictions is rather significant (an average of 34.44% for the two annexes combined) and the implementation of an approach for intermediary entities from those jurisdictions in the Directive would therefore appear justified and necessary.

It is also important to look further into the breakdown in figures (6) and (7) below, i.e. deposits of the non-bank sector from the jurisdictions in Annex I ("->Annex I") into the three different groups (i) EU Member States that exchange information ("info-EU->"); (ii) EU Member States levying a withholding tax ("WHT-EU->") and (iii) jurisdictions from the Savings Agreements network ("Agr->").

           

Figure (6) Share of non-bank deposits          Figure (7) Amount (million USD) of non-bank

                of BIS reporting countries category per         deposits of BIS reporting countries category per

                amending Proposal annex                I                               amending Proposal annex I

The comparatively large share (44% on average) of the third group (iii) jurisdictions from the Savings Agreements network in the deposits of the non-bank sector from the jurisdictions in Annex I (figure 6) and the fact that the balances of the three groups with the jurisdictions in Annex I have also (figure 7) nearly doubled from 2002 until 2007[56] is an indication that the implementation of the look-through provisions for entities in Annex I is well justified and necessary not only in the context of the amendments to the Directive but also for updating the Savings Agreements.

A third approach which also focuses on the updating of the Savings Agreements is to look at the interconnectedness of each of the jurisdictions within the network of Savings Agreements with the other jurisdictions from the same group. Currently, there are 270 bilateral agreements between each of the 27 Member States and each of the 10 dependent and associated territories and 5 separate EU agreements with Andorra, Liechtenstein, Monaco, San Marino and Switzerland. The bilateral agreements feature provisions of paying agent upon receipt in the limited version of the current text of the Directive. Nevertheless, unlike the Directive, there are certain interpretational issues with regard to the application of the paying agent upon receipt in triangular situations involving an upstream economic operator from one Member State/territory, a paying agent upon receipt from another Member State/territory and a beneficial owner from a third Member State/territory. These are primarily due to the bilateral nature of the agreements with the dependent and associated territories. The 5 EU agreements do not currently include paying agent upon receipt provisions at all.

Figure (8) below shows the shares of non-bank deposits from the three groups of countries: (i) EU Member States ("->EU"), (ii) Annex I countries ("->Annex I") and (iii) Annex II countries ("->Annex II"), in the jurisdictions within the network of Savings Agreements ("Agr->") which is a means to measure the latter's interconnectedness with the three groups.

Figure (8): Deposits by EU, Annex I and Annex II countries in Savings Agreements countries

The share of deposits from the EU Member States shows a steady decrease from 50% in 2004 to 35% in 2007, mainly to the advantage of the share of deposits from Annex I jurisdictions, which increased from 24% in 2004 to 34% in 2007, but also to the share of deposits from other Annex II jurisdictions, which increased from 26% in 2004 to 31% in 2007. These results, in addition to the aforementioned necessity to include look-through provisions for Annex I entities in the agreements, also supports the need for a solution to triangular situations for Annex II jurisdictions involving upstream economic operators, paying agents upon receipt and beneficial owners from three different Member States, third countries or territories.

In order to identify individual jurisdictions from Annex II as the jurisdiction of the non-bank depositor, the analysis must rely solely on publicly available BIS data[57] even though the information is aggregate. To put their vis-à-vis balances in perspective, it is useful to have a look at the first 10 countries ranked by the amount of non-bank deposits by investors from those countries.

Figure (9): Share of top ten vis-à-vis countries of non-bank depositors in all BIS reporting countries

As seen from figure 9, the shares of Cayman Islands and West Indies (collectively)[58] average 12% and 4% respectively of all non-bank deposits in all BIS reporting countries.

Evolution of international EU deposits vs non-EU deposits

For this part, the vis-à-vis countries are divided into two groups (figure 10): (i) EU and (ii) non-EU. The evolution of deposits from the non-bank sector for both groups is provided for the United Kingdom, Switzerland, Germany, Ireland, Netherlands, Belgium, Luxemburg and France, being the eight largest BIS reporting counties that have agreed to the disclosure of their bilateral data. This approach is used by Johannesen and the underlying assumption is that EU and non-EU depositors differ in that the Directive applies only to EU depositors. Johannesen claims that the growth of EU deposits following the entry into force of the Directive slowed down in comparison to the control group of non-EU depositors. The deficiencies of relying on data that aggregate all types of non-bank depositors have already been outlined above. The result for each country is listed below. The countries are divided into three groups: (i) countries that reflect the trend described by Johannesen (the United Kingdom, France and Switzerland); (ii) counties that do not reflect the trend described by Johannesen (Germany, the Netherlands and Luxemburg); (iii) counties that reflect a trend opposite to the one described by Johannesen (Austria, Ireland and Belgium). However, if the data series are aggregated, the overall trend reflects the trend described by Johannesen.

Figure (10) Aggregated time series for foreign EU vs. non-EU deposits for the 8 Member States and Switzerland

Group (i) countries that reflect the trend described by Johannesen (UK, FR and CH) (figures 11-13)

Figure (11)………………………………………………… Figure (12)

The United Kingdom || France

Figure (13)     

Switzerland[59] ||

Group (ii) counties that do not reflect the trend described by Johannesen (figures 14-16)

Figure (14)………………………………………………………………Figure (15)

Germany || Netherlands

Figure (16)

Luxemburg

Group (iii) counties that reflect a trend opposite to the one described by Johannesen (figures 17-19)

Figure (17)………………………………………......................Figure (18)

Austria[60] || Ireland

Figure (19)

Belgium

For details on break-in-series please refer to the document at:

http://www.bis.org/statistics/breakstables17.pdf

Conclusion

As demonstrated, 3 out of the 9 countries show (to a varying degree) a decreased growth of EU deposits at the time of the entry into force of the Directive compared to the control group of non-EU depositors. For three others the comparison is inconclusive. For three Member States the opposite effect is observed, although the growth rate difference is not so pronounced after the end of 2005. The per-country analyses show that the effect described by Johannesen can indeed be observed in at least three of the 9 countries with significant deposits from the foreign non-bank sector, while the analysis shows that the effect described by Johannesen can also be observed if all the 9 data series for the countries are aggregated.

The vis-à-vis results indicated that the amount of foreign non-bank deposits for many offshore centres is quite significant on a worldwide basis, especially for the Cayman Islands, Switzerland and Jersey. The vis-à-vis results revealed that a significant share of the non-bank deposits in Member States, and in within the network of the Savings agreements, are held by customers located in offshore jurisdictions. These results indicate that the implementation of look-through and paying agent upon receipt provisions for some relevant legal structures located in offshore jurisdictions is justified and necessary for both the Directive and the Savings agreements.

3.2. Euro-area deposits – ECB MFI[61] statistics

Description

The data provided by the ECB is based on MFI balance sheet statistics. It comprises monthly balances of deposits into monetary financial institutions in each of the Member States placed by non-residents from the euro-area, i.e. the balances are aggregated at euro-area depositor level without a detailed split among individual euro-area Member States. The data are split by the sector of the depositor, including a detailed split for households. The data are also split into seven different maturities, i.e. (1) Overnight deposits24, (2) Deposits with an agreed maturity up to 1 year, (3) Deposits with an agreed maturity over 1 and up to 2 years, (4) Deposits with an agreed maturity over 2 years, (5) Deposits redeemable at notice up to 3 months, (6) Deposits redeemable at notice over 3 months, (7) Repurchase agreements.

The major advantage of the data is that it is very similar to the Directive's product scope, personal scope (individuals), cross-border scope and the level of reporting:

(i) Product scope: According to Annex I, Part I, Subpart I, point 7 of Regulation ECB/2001/13:

bank deposits shall mean cash deposits made with credit institutions, repayable on demand or upon prior notice of up to three months, or at agreed maturities of up to two years, inclusive of sums paid to credit institutions in respect of a transfer of securities under repurchase operations or securities loans

From the definition of bank deposits it transpires that those would normally be classified as debt claims[62] and the income thereof would fall within the scope of Art. 6(1) (a) together with income from other debt claims from debt securities.

(ii) Personal scope: According to Annex I, Part I, Subpart III, point 9 of Regulation ECB/2001/13

In order to calculate a monthly sector disaggregation of the monetary aggregates and credit counterparty, other resident sectors are further broken down by the following subsectors: other financial intermediaries + financial auxiliaries (S.123 + S.124), insurance corporations and pension funds (S.125), non-financial corporations (S.11) and households + non-profit institutions serving households (S.14 + S.15).

The sectoral split provides a separate time series for households and non-profit institutions serving households. This is a sufficiently detailed split that allows the approximation of deposits held by individual beneficial owners as defined in the Directive.

(iii) Cross-border scope: According to Annex I, Part I, Subpart III, point 8 of Regulation ECB/2001/13:

Counterparties located in the domestic territory and in the other participating Member States are identified separately and treated in exactly the same way in all statistical breakdowns.

The split between resident and non-resident depositors from other euro-area Member States allows the filtering of only those deposits held by non-resident beneficial owners, which is another element that very closely matches the Directive.

(iv) Level of reporting: Since the product covered is bank deposits, it is reasonable to expect that the bank/debtor is the paying agent and the household depositor/creditor is the beneficial owner under the Directive, i.e. an intermediary would normally not be involved in that setup unlike other products under the Directive (e.g. securities). In addition, the paying agent reporting level (i.e. paying agents report in their Member State of establishment and not that of their head office) is also matched by the MFI data since there is an explicit prohibition of cross-border consolidation in Annex I, Part III, and p.8 of Regulation ECB/2001/13:

No consolidation for statistical purposes is permitted across national boundaries.

Therefore no deposits in foreign branches/subsidiaries are included in the reported data.

Limitations of the data

There are very few limitations inherent to the data. The data covers only Member States at the level of the MFI/debtor/paying agent and only aggregated euro-area at the level of the households/creditor/beneficial owner. Therefore, any comparison with non-EU MFIs or analysis of non euro-area households is not possible. Secondly, the data refers only to MFI deposits and not to the other types of savings products which come within the scope of the Directive, e.g. debt securities and units/shares in UCITS and other collective investment undertakings. Nevertheless, as long as the data are considered as the minimum amount of savings income to be explored, that last limitation would be mitigated. Thirdly, the sectoral breakdown aggregates households with non-profit institutions serving households. However, it is assumed that this allows for a sufficient approximation of deposits held by individual beneficial owners. Finally, there have been indications from individual central banks that the sectoral split is based on an estimation made by the statistical bureau of the relevant central bank. It is expected that the lowered expectations for the coverage would accommodate any possible variations within the confidence interval for those estimations.

Estimation of the coverage of derived interest income from the tax base

The Commission obtained ECB data on cross-border bank deposits of households in the euro-area. The data includes each Member State as the country of the deposit and the euro-area (as one group) as the place of the household depositor. This data was selected as the best possible of all available alternatives in order to provide a limited evaluation of whether the data exchanged (or tax withheld) under the Directive reflect a satisfactory coverage of the potential tax base involved.

While not covering all types of interest income under the Directive, all the balances for deposits should normally be within the scope of art. 6(1) (a) thereof, which would mean that the comparison would give the absolute minimum of the underlying tax base.

As a second step, the monthly data on the deposits were matched with the corresponding interest rates, also available on a monthly basis from the ECB[63] and broken down according to the aforementioned maturities. Where those were not available, reasonable estimations were made, with the preference for an underestimation of the particular interest rate where necessary in order to provide the best-case scenario in terms of the reflection of the underlying tax base. Since the MFI data reflects stocks and not flows, the interest rates on the outstanding amounts are preferred and sometimes substituted for similar maturities[64]. In some cases a sensitivity analysis was performed with the most favourable reference rates (even 0%). It must be noted that the rates used are those that apply to both residents and non-residents in the euro-area Member States, but it is reasonable to assume the same level of deposit interest rates apply to both residents and non-residents in the euro-area Member States. From the matching of the monthly balances per maturity with the relevant interest rates an annual average of the underlying tax base was derived.

As a third step, the estimated annual average of the underlying tax base was compared to the information exchanged with euro-area Member States for interest income under Art. 8(2) (a)[65], or when the Member State has opted for the last sub-paragraph of Art. 8(2)[66] the underlying tax base was compared to the part on "total amount of interest or income" for each of the fiscal years 2006-2009. From the comparison a % of coverage of the potential tax base by the exchanged data (or tax withheld) under the Directive is derived.

Simplified illustrative example:

Member State X has a MFI deposit amount of 100 EUR for each of the 7 maturity breakdowns for each month in the year 2009 (therefore total deposits for all maturities are 700). The interest rates for all of the maturities and months in 2009 equals 2%. For each of the months in 2009 the derived interest rate received/credited would be 700*2%=14 EUR. This result is then compared to the actual information exchanged on interest payments by Member State X for that year to euro-area Member States which in this example is 16. The resulting coverage is then (16/14) 114,3%. As previously outlined, it is expected that coverage results above 100% would be common.

This method gives only the absolute minimum of coverage of the underlying tax base, since the scope of the ECB data would always be smaller than the scope of classical interest income under Art.8(2)(a), the latter also including income from other types of debt claims than bank deposits. This effect would be even stronger for Member States that report only the "total amount of interest or income" under the last sub-paragraph of Art. 8(2), giving a certain bias for a higher estimated coverage for those Member States due to the fact that "total amount of interest or income" would also include distributions from UCITS under Art. 6(1)(c) and also potentially the income under Art. 6(1)(b) and Art. 6(1)(d) which is reported on a net basis (and not gross proceeds). Therefore, any coverage above 100%, even substantially higher, should be common. On the contrary, actual coverage, which is substantially lower than 70%[67] consistently across all of the four years surveyed would merit further investigation and follow-up with the relevant Member States.

In addition, it must be noted that this method gives some bias towards better % coverage for the withholding Member States, because the withholding tax figures reflect not only the classical interest income under Art. 8(2)(a) or "total amount of interest or income“ under the last sub-paragraph of Art. 8(2), but also sales proceeds from both debt instruments and units/shares in UCITS. The bias would be strongest for those Member States whose paying agents deal extensively with cross-border UCITS[68].

The results of the simulations are summarised in the tables (6) to (8) below:

Table (6) Coverage of interest income reported under Art.8(2)(a) as a % of notional income per ECB cross-border deposit:

|| BG || CY || DK || ES || HU || LT || LV

2006 || 0,00% || 26,04% || 129,06% || 108,21% || 99,73% || 42,83% || 110,75%

2007 || 85,96% || 68,51% || 200,87% || 72,51% || 68,56% || 72,78% || 75,45%

2008 || 58,08% || 44,47% || 40,31% || 82,00% || 143,08% || 49,92% || 93,57%

2009 || 50,69% || 49,72% || 137,37% || 103,68% || 309,14% || 58,14% || 60,57%

Average || 64,91% || 47,18% || 126,90% || 91,60% || 155,13% || 55,92% || 85,08%

|| MT || NL || PL || SI || UK[69]

2006 || 59,13% || 580,43% || 148,23% || 116,79% || 7,82%

2007 || 56,47% || 143,08% || 165,91% || 69,90% || 25,21%

2008 || 56,21% || 134,45% || 67,68% || 90,65% || 9,23%

2009 || 120,66% || 124,44% || 72,89% || 1478,97% || 9,07%

Average || 73,12% || 245,60% || 113,68% || 439,07% || 12,83%

Table (7) Coverage of interest income reported under the last sub-paragraph of Art.8(2) as a % of notional income per ECB cross-border deposit:

|| CZ || DE || EE[70] || EL || FI[71] || FR || IE

2006 || 1133,01% || 112,82% || 0,00% || 285,68% || 9624,82% || 61,70% || 266,64%

2007 || 825,46% || 72,19% || 0,00% || 166,63% || 167,10% || 58,39% || 358,28%

2008 || 578,29% || 73,19% || 62,35% || 129,48% || 56,54% || 53,90% || 574,72%

2009 || 489,47% || 154,90% || 103,01% || 100,39% || 88,02% || 64,28% || 71,68%

Average || 756,56% || 103,27% || 82,68% || 170,55% || 2484,12% || 59,57% || 317,83%

|| IT || LU[72] || PT || RO || SE[73] || SK

2006 || 1515,05% || 49,58% || 1,59% || 0,00% || 713,80% || 472,95%

2007 || 1083,70% || 44,56% || 6,74% || 37,21% || 711,71% || 72,81%

2008 || 381,13% || 36,66% || 26,51% || 30,48% || 1933,60% || 241,97%

2009 || 856,51% || 25,51% || 74,56% || 40,83% || 15573,38% || 139,51%

Average || 959,10% || 39,08% || 27,35% || 36,17% || 4733,12% || 231,81%

Table (8) Coverage of withholding tax levied as a % of notional withholding tax per ECB cross-border deposit:

|| AT || BE || LU

2006 || 245,04% || 158,89% || 253,01%

2007 || 204,66% || 133,33% || 213%

2008 || 151,93% || 112,66% || 173,33%

2009 || || || 298,36%

Average || 200,54% || 134,96% || 234,40%

The method for this analysis was discussed in the ACDT meeting on 28 September 2011. The estimations are based on the latest updates received from Member States after that meeting. It is hoped that Member States will be able to use this data source in future in order to verify the completeness of the records they exchange under the Directive. From the estimations it transpires that three Member States (UK, PT and RO) fall significantly and consistently below the 70% benchmark. In addition, there are 4 Member States that show results that are below the 70% benchmark.

Evolution of non-resident euro-area household deposits

The MFI statistics allow an analysis of the dynamics of non-resident euro-area household deposits for all Member States. Deposits from all maturities, including overnight deposits typically producing low interest, start at over EUR 164 billion in 2003 and peaked at over EUR 247 billion in October 2008 (a 50% increase for the period). The figures below outline the evolution by groups of Member States. Deposits with agreed maturity are first compared with deposits with no agreed maturity (Fig. 20), assuming that the deposits with agreed maturity produce relatively higher interest income and are therefore more sensitive to the tax treatment thereof. The deposits with agreed maturity have decreased from EUR 72 billion in January 2003 to EUR 60 billion in November 2005 – 15% for the period and a monthly decrease of -0.52%. This is not replicated by the deposits with any agreed maturity.

Figure (20) Trends of cross-border deposits with agreed maturity and deposits with no agreed maturity aggregated for all MS

A finer distinction between the trends for deposits with agreed maturity in euro-area Member States and non-euro area Member States reveal that the downward trend until November 2005 came primarily from euro-area Member States (-38.84% for the period and a monthly decrease of -1.51%). During the following three years from November 2005 to November 2008 the two categories show an almost perfectly matched increase in cross-border deposits from EUR 60 billion to EUR 81 billion (an increase of 25% for the euro-area and 30% for the non-euro area respectively).

After the end of 2008, deposits with agreed maturity in euro area Member States appear to have suffered the strongest decrease (from EUR 41 billion in November 2008 to EUR 21 billion in March 2010), primarily due to the financial crisis.

Figure (21) Trends of cross-border deposits with agreed maturity in euro area and non-euro area Member States

Looking into individual Member States, for several Member States (LU, ES, FR, DE, BE, PT, NL and IT) there was a marked decrease of household deposits from 2003 until mid-2006 – the total amounts for those Member States were almost halved during that period. The estimated average monthly decrease is -1.7% per month.

For three Member States (UK, IE and DK) there was an average 61.54% increase of household deposits for the same period (average monthly increase of 1.3% per month).

Figures 22-23 below show the evolution of euro-area cross-border household depositors in the Member State where the deposit is held.

Figure (22)………………………………………………Figure (23)

From the Member States with the highest amount of cross-border depositing by euro-area households, there is a decrease in Luxemburg and an increase in the United Kingdom. The other group of Member States show a general decrease of cross-border depositing by euro-area households[74].

Evolution of household vs. company deposits

If the MFI statistics[75] for the non-financial companies sector are compared to those for the household sector, this could highlight any shifts of investor preferences towards investing directly or through an intermediary company (figures (25)-(26)). The shares of each of the two sectors would give a practical estimation of the described effect. Out of all Member States, a decrease of the share of households for the benefit of non-financial companies is observed for France (average monthly decrease 0.82%), Spain (average monthly decrease 0.78%), and for Luxemburg (average monthly decrease 0.71%).

Figures 24-26 show the evolution of household sector compared to the non-financial company sector for non-bank deposits in specific Member States.

…Figure (24)………………………………………… Figure (25)

…Figure (26)

Conclusion

Overall, in several Member States there has been a decrease in cross-border savings in deposits with agreed maturities by euro-area households around the time of the entry into force of the Directive both as an absolute amount and as a share of households in the total of non-financial depositors. Nevertheless, that trend appears to have started already in 2003 and is not matched by the trends for overnight deposits and deposits redeemable at notice.

3.3. Switzerland – data from the Swiss National Bank (SNB)

Description

The Swiss National Bank (SNB) issues an annual publication "Banks in Switzerland"[76] with very detailed statistics based on the financial reports of Swiss banks. The 2010 publication covers a total of 320 Swiss banks and local branches of foreign banks from various banking sectors. The banks and their balance sheet totals are listed in the publication. The figures are provided in CHF millions (the figures for custody accounts are in CHF billions) and the 2010 publication contains data series for the period 2002-2010.

The level of consolidation of the reporting entity is mostly "parent company" which includes all domestic offices as well as foreign branches of the same legal entity. Nevertheless, the statistics on securities holdings in custody accounts are reported only at the level of domestic offices.

Most importantly, the publication provides some very detailed geographical and/or client breakdowns for the following:

· Table 32 "Geographical breakdown of assets and liabilities shown in the balance sheet"[77] provides a geographical breakdown of a set of assets and liabilities and most importantly "Liabilities towards customers in the form of savings and deposits"[78]

· Table 38 "Fiduciary business, by country"[79] provides the 2002-2010 time series for fiduciary assets[80] and liabilities[81]

· Table 38c "Holdings of securities in bank custody accounts, by domicile of custody account holder, category of security and business sector"[82] provides the 2002-2010 time series for custody accounts distinguishing between resident and non-resident, private, commercial or institutional custody account holder for bonds[83], shares[84], units in collective investment schemes[85] and other securities[86].

Studies that use the data

The SNB data was used in the study "The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors?"[87] by Gabriel Zucman to supplement IMF data on portfolio investment and argue that if tax havens were taken into account then the euro-area would turn into a net creditor.

Limitations of the data

There are certain misalignments between the data and the Directive's scope. Firstly, only in the case of Table 38c is there a breakdown between private customers and other sectors. For the other statistics the data is available for the aggregate "customers" group, which nevertheless allows the explicit (for table 32) or implicit (for the "liabilities" part of Table 38) exclusion of other banks as customers. Secondly, the reporting in Banks in Switzerland is consolidated at the level of the parent company and therefore also includes balances at the level of foreign branches of a Swiss bank. Those foreign branches, if situated in the EU, would normally be considered paying agents in the Member State of their establishment and not a Swiss paying agent. It must be noted that the statistics in Table 38c are reported only at the level of the domestic offices and do not suffer from that limitation.

Liabilities towards customers in the form of savings and deposits

This data enables a comparison based on the split by country of the customer. It is assumed that those savings and deposits are considered as Swiss source income for the purposes of levying the anticipatory tax on Swiss source interest payments. Therefore, it is unlikely that the income produced by these liabilities would fall within the scope of the retention defined in article 1(1) of the current EU-Swiss Savings Agreement[88].

The largest savings and deposits are held by customers from Germany, France, Italy, the United Kingdom, Austria and Spain from the EU and the United States. The second group of countries includes Greece, the Netherlands, Belgium and Sweden from the EU, Canada, Israel and Russia from the third countries and Panama and Singapore as offshore jurisdictions. Panama and Singapore showed a remarkable increase from 2007 to 2010 with an average annual increase of 167% for Panama and 45% for Singapore[89]. Another important offshore centre, the West Indies UK (collectively)[90], which is not shown in the figures because of lower absolute amount of the balances, showed an average annual increase of 252% over the 2007-2010 period. For comparison, the deposits from German clients have increased by 22% over that period.

Figures (27) and (28) show the evolution of savings and deposits per country of depositor in Swiss banks (amounts CHF millions)

Figure (27)…………………………………………………………Figure (28)

Foreign savings and deposits appear to have recovered in the last two years from the slowdown in 2007 and 2008.

Fiduciary business

The definition of fiduciary business for the purposes of the "Banks in Switzerland" report can only be found in the 1997 edition of the publication, which is not available in English.

"Les opérations fiduciaires englobent les placements, crédits et participations que la banque effectue ou accorde en son propre nom, mais pour le compte et aux risques exclusifs du client, sur la base d'un ordre écrit. Le mandant supporte le risque de change, de transfert, de cours et de recouvrement et il lui revient la totalité du rendement de l'opération; la banque ne perçoit qu'une commission. Les fonds que les banques reçoivent à titre fiduciaire proviennent en majeure partie de l'étranger et sont placés presque exclusivement à l'étranger. Il s'agit essentiellement de placements à court terme à l'étranger, dans des banques tierces ou dans les propres comptoirs juridiquement dépendants des banques suisses. Dans ce dernier cas, les opérations doivent figurer dans le bilan, étant donné qu'il en résulte un engagement du comptoir envers le siège en Suisse."[91]

From the definition it is clear that fiduciary business consists primarily of deposits from non-residents that are afterwards re-deposited abroad in the name of the bank, but for the account of the depositor. The re-depositing may be done into a third-party bank, or with branches of the Swiss bank abroad. Since the Swiss bank is acting in its own name, the bank where the re-depositing takes place would not be aware that the re-deposited funds belong to the original depositor. Due to the fact that the reported level of consolidation encompasses foreign branches, the re-depositing into the Swiss bank's foreign branches would not appear in Table 38. Rather, the liability to the original depositor will appear on-balance sheet, for example in Table 32, while there will be no reported asset/claim to the foreign branch. That would mean that the positions in Table 38 would in fact reflect only the fiduciary business with third-party banks.

Since Swiss-source income is currently not subject to the "retention" under the EU-Swiss Agreement, fiduciary deposits are quite important in the context of the EU-Swiss Agreement since they frequently do not represent Swiss-source income. To that extent, they do not suffer the Swiss anticipatory tax (and are in fact marketed with that advantage), but would rather be subject instead to the "retention" referred to in Article 1(1) of the EU-Swiss Agreement[92].

Fiduciary arrangements are usually offered to high net worth customers who have substantial portfolios of assets, including cash. There are usually commercial and tax reasons for the arrangements given that the rates received are closer to those normally paid only to banks or large institutions, while benefiting from the spreading of risk and easier administration.

From an accounting perspective, the received funds under a fiduciary arrangement by the Swiss bank are reported as off-balance sheet liabilities, while the re-deposited funds into the third-party bank are reported as off-balance sheet assets. The country breakdown for both is available from the SNB publication.

Fiduciary liabilities

The geographical breakdown and ranking for fiduciary liabilities (assumed to consist primarily of fiduciary deposits) appears substantially different from the breakdown under SNB Table 32 (figure (29)). There are 8 countries that may be considered as offshore centres among the first 15 countries and they represent an average of 67% of the fiduciary liabilities attributable to that group (the share of West Indies UK alone is almost 26% on average). It must be noted that although fiduciary deposits are within the scope of the EU-Swiss Agreement, there is currently no obligation for the Swiss paying agent to report payments to countries outside the EU due to the absence of look-through or paying agent upon receipt provisions in the EU-Swiss Agreement.

Figure 29: Evolution of fiduciary customers' liabilities in Swiss banks (million Euros) per country of creditor/depositor:

It is also important to put into perspective the fiduciary funds managed by Swiss banks by comparing the positions in Table 38 "Fiduciary business, by country" with the positions in Table 32 "Geographical breakdown of assets and liabilities shown in the balance sheet" and in particular "Liabilities towards customers in the form of savings and deposits". On average, the share of fiduciary business in the sum of the two series is 93% (figure(30)).

Figure (30): Evolution of share of liabilities towards Swiss-source income depositors compared to liabilities towards fiduciary business creditors/depositors in Swiss banks.

Fiduciary assets

The fiduciary assets geographical breakdown represents the place where the fiduciary funds are being re-deposited, i.e. the country of the third-party bank.

Figure (31): Evolution of fiduciary assets invested by Swiss banks (million Euros) in country of debtor:

The fiduciary funds are placed primarily in third-party banks in Luxemburg, the Netherlands, The United Kingdom and France. As previously outlined, the fact that a Swiss bank acts in its own name when re-depositing the funds received under a fiduciary arrangement would likely prevent the bank where the funds are re-deposited from tracing any individual depositor. Jersey, Guernsey and Isle of Man appear as the primary offshore centres into which fiduciary funds are re-deposited into third-party banks.

As mentioned in the section on the limitation of the SNB statistics, there is a disadvantage of the data in that only re-depositing in third-party banks is covered. Fiduciary positions received by foreign branches of the Swiss bank are consolidated and do not figure as an off-balance sheet item. The Commission has received from the BIS under a confidentiality agreement, and with the assistance of the SNB, the entire set of fiduciary positions, including inter-office positions. The data generally confirms the country breakdown officially available in the SNB publication with the addition of one country from South-East Asia. The data also confirms that the prevailing share of fiduciary funds are received from non-banks and re-deposited in banks.

Custody accounts

The custody accounts are another aspect of the SNB statistics that is relevant for the EU-Swiss agreement. A Swiss bank that offers custody account services to its customers would normally deal with the beneficial owner directly and would therefore in many instances act as a paying agent. The SNB publication distinguishes only between resident and non-resident custody account holders, i.e. no detailed country breakdown is available. It is therefore impossible to isolate EU-resident custody account holders from the aggregated values attributable to non-residents. Nevertheless, the data series provides a detailed sectoral split for the account holder distinguishing between private, commercial and institutional clients. Another advantage is that there is a separation of securities by bonds, shares, units in collective investment schemes and other securities.

It is first useful to concentrate the analysis on the series for bonds (figure (32)). There is a clear downward trend for the private customers sector, which starts already in 2002 and is matched by a corresponding increase of holdings by institutional investors. The total decrease from 2001 to 2010 is -52%, i.e. the holdings are halved over that period. The series for resident private customers may serve as a control for that trend. The decrease there was only -26%.

Another way to explore the evolution of the custody account business with non-resident private customers is to look at the development of their aggregate portfolio, i.e. in which products they would invest (figure (33)). The bonds part of the portfolio shows a steady decrease from 2002, which is matched by an increase in investments in shares and especially in UCITS.[93] Those have however decreased sharply after the beginning of the financial crisis.

Figure (32) evolution of bonds for non-residents per   Figure (33) evolution of private customers category of customer in Swiss banks (billions CHF)                                                                       investments per product in Swiss banks (billions CHF)

Conclusion

The SNB data as a whole contributed to the conclusions already drawn on the basis of the BIS data outlined earlier, i.e. the need of an update of the relevant agreements with third countries to address the cases of intermediary entities established in offshore jurisdictions.

3.4. IMF Coordinated Portfolio Investment survey – Evolution of investments in debt securities

Description

The IMF in its Coordinated Portfolio Investment survey (CPIS) reports the cross-border investment positions (holdings at market prices prevailing at the end of the year) of investors resident in 74 participating countries (country of investor) with a split by the corresponding country of investment (country of issuer). There are separate data sets for (i) Total investments, i.e. debt and equity, (ii) Total equity securities, (iii) Total debt securities, (iv) Long-term debt securities and (v) Short-term debt securities. Data are generally available on an annual basis from 2001 to 2009.

The CPIS data also provide a split based on categories of "sector of holder" for some countries, which enables an analysis of investments made by individuals (sector "households"). Nevertheless, it must be stressed that, of the EU Member States, only Cyprus, Denmark, Spain, Finland, France, Hungary, Italy, the Netherlands, Romania and Sweden provided data for the sector "households" for the period 2001-2009. Of those, only Denmark, Spain, Italy, the Netherlands and Sweden post sizable investments by their resident households (Finland does not disclose investments by households in Austria and Luxemburg due to confidentiality reasons). Austria, Germany and the United Kingdom provide the data for the sector "households" only for 2009.

Studies that use the data

IMF data was used in the study "The Missing Wealth of Nations: Are Europe and the U.S. net Debtors or net Creditors?" by Gabriel Zucman to measure, amongst others, the shift of undeclared assets towards off-shore jurisdictions. The study came to the conclusion that recorded liabilities were significantly higher than recorded assets in the data due to investors not declaring assets located in tax havens.

In Rixen-Schwarz (2011)[94], based on IMF CPIS data, the authors examined any changes to the investment patterns of individual investors, thereby attempting to evaluate the effectiveness of the Directive using the CPIS data for Sweden, France, Italy and Spain. Their conclusions were that:

· Individual investors adapted to the institutional changes to be implemented before the Directive became effective in 2005.

· The effect of product substitutability (debt products into equity products) is more pronounced than the geographical relocation effect (from the Directive network countries into third countries).

· The withholding Member States and countries did not experience an outflow of portfolio capital, whereas Member States exchanging information lost capital relative to third countries outside the scope of the Directive.

They found strong behavioural responses by French investors, and no such evidence in regard to Italian investors, which they explained by "the absence of a national automatic reporting system on capital income in Italy".

Limitations of the data

As already mentioned, many Member States, including the United Kingdom and Germany, which report this data only for 2009, do not report a breakdown of debt claims by sector for the period 2001-2009. Therefore, the analysis would be restricted to only certain countries that report this information.

The data is in principle compiled according to where the debtor (issuer) is located and the creditor (investor) is located which does not necessarily replicate the paying agent-beneficial owner stage in the payment of savings income.

Illustrative example:

It is possible that the security is issued by Jurisdiction A and owned by an investor in Jurisdiction B, but held in a custodian account with an intermediary (on behalf of the investor) based in a Jurisdiction C. In that case the CPIS data would show an amount of investment in A by investor in B, while the Directive's information exchanged/tax withheld would be concerned with the payment from the intermediary in Jurisdiction C to the investor in Jurisdiction B.

That means that the paying agent Member State would not necessarily coincide with the debtor country/Member State. Any change of the country of the intermediary from a jurisdiction within the geographical scope of the Directive to a non-cooperating jurisdiction would be therefore omitted. Issues regarding how custodians should report the relevant CPIS data, in particular those custodians who hold securities on behalf of other custodians or nominees, are indeed mentioned in the CPIS Guide[95]. However, the CPIS Guide suggests that if the trust and the beneficiaries are located in two different countries, then it is unlikely that any data will be collected by either of them. It also appears that the custodian reporting may only apply in practice to situations where the custodian and the investor are in the same reporting country.

Although product substitutability has been explored by Rixen-Schwarz (2011), it appears more plausible that capital-protected fixed-income securities would rather be substituted with products of a similar risk profile. Unfortunately, CPIS data does not include financial derivatives, but would arguably cover some of the structured products and other wrappers for debt securities, while omitting other products (e.g. any kind of insurance product)[96]. Therefore, some of the most important substitutable products are either not present in the CPIS database, or would be classified as the product they were meant to substitute, i.e. as debt securities. There is another misalignment between the Directive's scope and the CPIS classification being that shares/units in mutual funds are classified as equity, whereas income from debt claims distributed (or realised at the sale of the units thereof) by UCITS or other undertakings for collective investment that invest in debt claims is within the scope of the Directive.

With regard to the product scope, it must be noted that income from debt securities is only a part of all savings income covered by the Directive. To that end, any impact on income from other types of debt claims or other savings income would be omitted from the analysis.

The CPIS represents the year-end holdings according to market prices. Therefore it would be impossible to distinguish between increases driven by increased investment in the particular asset class from an increase driven by increased asset prices.

In addition, the tables for some of the Member States explored exhibit what appears to be an inconsistent use of the values denoting lack of data ("…."), zero or close to zero values ("--") and confidential data ((c)) for the same country of investment in different years, leading to some important structural breaks.

EU Member States

As outlined above, a limited number of countries have been identified as having detailed information by sector (household etc) and posting sizable investments by their resident individuals: Denmark, Spain, Finland, France, Italy, the Netherlands and Sweden over the relevant time period since the introduction of the Directive[97].

The evolution from 2001 to 2009 of portfolio investments for the Member States mentioned above is described below.

From the surveys for Member States one general pattern emerges being the increase in the share of debt securities issued by Member States that exchange information. The average annual growth of the share of debt securities issued by Member States that exchange information ranges from 3% in the Netherlands to 10% in Spain and Finland.

France

The banks, insurance companies and mutual funds are the principle sectors of investors representing a combined share that averages around 90% of all investments in debt securities. The household sector represents a very small part compared to other sectors. Nevertheless, it displays a marked relative decrease in 2003, the year of the adoption of the Directive, to one fifth of the 2002 values. This decrease affects investments only by households and only in debt securities.

The 2003 decrease apples to debt securities from all jurisdictions. As previously described, unlike the CPIS the latter may be explained in that the Directive operates at the paying agent rather than at the creditor level. However, it must be noted that in 2005 the share of the Member States that apply the withholding tax doubled from 10% to 22%. The increase in investments in equity securities by households of 16% is below the increase for all sectors (70%), but also below that for the non-financial sector (more than 22%), which encompasses the household sector, suggesting that an equity substitution scenario cannot be demonstrated either.

Fig (34) Evolution of investments in debt securities by FR households per category of the issuer (million USD)

Netherlands

Whereas other Member States provide a detailed split by sector of investor from 2001, the Netherlands data only provides such a split as from 2003. Therefore any effects prior to that date could not be analysed. The cumulative growth from 2003 to 2007[98] of investments by the household sector into debt securities as compared to equity securities is remarkably similar at 148% and 147% respectively. Therefore, for the observed period a substitution of debt securities with equity securities is present.

If the jurisdictions of investment, i.e. where the creditor is established, are compared, the share of jurisdictions within the Savings Agreements network demonstrates a very strong increase for the years 2004 and 2005. This is mainly due to the investments in Cayman Islands, which at almost 25% of all household investments in debt securities for 2004 and 2005 is the number one investment location for debt securities for the household sector, surpassing the US, UK, DE and FR, followed by Jersey with 6% for 2004 and almost 5% for 2005.

Figure (35) Evolution of investments in debt securities by NL households per category of the issuer

Another observed effect is on the side of equity investments by households and it is related to the rapid increase of equity investment into the group of Member States that apply the withholding tax, largely driven by equity investments in Luxemburg. It must be noted that equity investments also cover investments in shares/units in mutual funds and although the CPIS data lacks the level of detail, it is assumed that this increase is largely due to the increase of investments in Luxemburg-based mutual funds – figure 36.

Fig (36) Evolution of investments in equity products by NL households per category of the issuer (million USD)

Italy

Compared to other Member States for the observed time period, third-country debt securities (mainly US and International Organisations) held by Italian households represent a larger share of all investment locations, peaking at 51% in 2004 and 52% in 2007.

Fig (37) Evolution of investments in debt securities by IT households per category of the issuer (million USD)

Similarly to equity investment by other Member States' households, equity investments in Luxemburg (presumably in mutual funds) were the main driver for the observed nearly tripling of the equity investments from 2001 to 2007. The share of equity investments in Luxemburg grew steadily from 69% of all household equity investments in 2001 to 84,49% in 2007.

Fig (38) Evolution of investments in equity products by IT household per category of the issuer (million USD)

Spain

The direct holdings of debt securities of ES households were invested mainly in Cayman Islands, which is the most favoured investment location for the years 2001-2007, although its share of total debt securities investment by households gradually dropped both in relative and absolute terms from almost 70% in 2002 to 30% in 2007. This is reflected below in the dynamics of the share of debt securities investments in jurisdictions within the Savings Agreements network.

Fig (39) Evolution of investments in debt securities by ES households per category of the issuer (million USD)

The growth of the Luxemburg fund industry already outlined for the other Member States is also reflected in the ES data for households' portfolio equity investments.

Fig (40) Evolution of investments in equity products by ES households per category of the issuer (million USD)

Denmark

Like in France, a decrease in household investments in debt securities is observed for the year 2003, although it is not as pronounced and as long-lasting. It also affects all investment locations across the board. The observed rebound in 2004 and 2005 is driven partially by investments in Bermuda – 6% of all investment locations for both years and in Jersey and Guernsey with a combined 8,5% share in 2005. For the case of Bermuda, it must be noted that investments in debt securities by the household sector decreased rapidly in 2006 and 2007, while households' investments in equity securities increased at the rate of almost 44% and 60% in 2006 and 2007, at the time when the investments by other sectors remained largely stable.

Fig (41) Evolution of investments in debt securities by DK households per category of the issuer (million USD)

Sweden

The household sector in Sweden investing in debt securities directly represents an almost negligible share of all investor sectors (the insurance and banking sectors being strongest) – 0.19% in 2002, but grows to around 3% in later years. The household investors are largely biased towards investing in equity – an average of almost 95% of household investments in securities is in equity securities. Luxemburg equity products (presumably mutual funds) represent a very large share of all equity investments, peaking at almost 82% in 2003.

Finland

Like the Netherlands, Finish data covers only years after 2003. In addition, the data for some important locations (e.g. LU) is listed as confidential for 2008 and 2009. For the remaining years Luxemburg is the leading location for household investment in debt claims, which increases the share of the Member States that apply the withholding tax.

Fig (42) Evolution of investments in debt securities by FI households per category of the issuer (million USD)

Jurisdictions within the Savings Directive agreements network

Unfortunately, the level of detail of the CPIS data would not enable an EU-wide analysis (from the perspective of the investor) of the major jurisdictions within the Savings Agreements network. Nevertheless, it is possible to identify the major jurisdictions within the Savings Directive agreements network for all investor sectors from the EU Member States participating in the CPIS (essentially all Member States except Lithuania). In addition, the Member States of residence of the investor from which the major investments originate do provide the split by sector for 2009. Of those, only Luxemburg and Ireland do not and it could be expected that the major part of the investors from those are from the financial industry.

The following table summarizes the findings for the Member States of residence of the investor from which the major investments originate, based on 2009 data:

Table 9: Share of country of issuer of financial asset categories per EU investors

Inv. from Inv. in || EU total || of which from

UK** || FR** || LU || IE || DE** || NL** || IT**

All || % * || Debt || % * || All || Debt || All || Debt || All || Debt || All || Debt || All || Debt || All || Debt || All || Debt

Cayman Islands || 392.090 || 2,09 || 257.981 || 2,01 || 114.526 (19.907) || 96.801 (20.158) || 61.430 (296) || 41.446 (95) || 57.186 || 14.557 || 35.123 || 15.534 || 28.958 (640) || 26.419 (158) || 21.609 (959) || 11.690 (745) || 5.436 (289) || 2.306 (230)

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from Japan, the US, Hong-Kong, Bermuda and from the jurisdictions within the Savings Directive network of agreements – from Switzerland, Guernsey and Jersey.

|| UK** || LU || DE** || FR** || NL** || IE || SE**

Switzerland || 236.739 || 1,26 || 35.378 || 0,28 || 60.487 || 13.433 (2.621) || 41.890 || 3.420 || 30.647 (9.737) || 2.210 (323) || 23.940 (3.473) || 1.944 (41) || 19.725 (1.096) || 3.561 (1) || 18.777 || 6.807 || 14.891 (653) || 697 (16)

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from the US, Japan, Norway, Iceland, Bermuda and from the jurisdictions within the Savings Directive network of agreements – from Jersey and Guernsey.

|| UK** || DE** || FR** || LU || IE || NL** || IT**

Jersey || 147.091 || 0.78 || 111.509 || 0.87 || 35.756 (28.260) || 25.708 || 22.947 (3.217) || 21.700 (3.058) || 18.880 (324) || 10.129 (249) || 18.505 || 10.699 || 10.274 || 8.848 || 7.971 (550) || 6.258 (527) || 5.749 (439) || 4.836 (435)

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from the US, Japan, South Africa, Canada, Hong Kong and from the jurisdictions within the Savings Directive network of agreements – from Switzerland, Guernsey and Isle of Man.

|| FR** || LU || DE** || NL** || BE || AT** || UK**

NL Antilles || 62.084 || 0,33 || 53.222 || 0,42 || 46.494 (1.737) || 43.444 (58) || 3.967 || 2.486 || 2.773 (925) || 2.321 (820) || 2.069 (562) || 526 (101) || 1.683 || 1.352 || 899 (199) || 824 (167) || 630 (263) || 452 (248)

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from the US, Japan, Canada, Norway, Hong Kong and from the jurisdictions within the Savings Directive network of agreements – from Switzerland, Guernsey and Jersey.

|| UK** || FR** || DE** || NL** || LU || IT** || FI**

Guernsey || 35.093 || 0,19 || 18.141 || 0,14 || 9.785 (927) || 3.179 (548) || 5.951 (468) || 3.992 (355) || 5.129 (468) || 4.581 (375) || 3.273 (192) || 827 (63) || 2.958 || 1.331 || 1.664 (110) || 1.571 (109) || 1.531 (27) || 315 (26)

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from the US, Japan, South Africa, Canada, Hong Kong and from the jurisdictions within the Savings Directive network of agreements – from Switzerland, Isle of Man, Jersey and NL Antilles.

|| LU || FR** || UK || PT || IE || NL** || BE

BVI || 18.604 || 0,10 || 5.054 || 0,04 || 5.724 || 1.026 || 3.150 (195) || 900 || 2.499 || 1.101 || 1.790 || 1.372 || 1.348 || 102 || 863 (84) || 366 || 759 || 13

The large portfolio investments from outside the Savings Directive network of agreements come, by order of magnitude, from the US, Hong Kong, Russia, Canada, Bermuda, Japan, and from the jurisdictions within the Savings Directive network of agreements – from Guernsey, NL Antilles, Isle of Man and Jersey.

* % of all investment destinations by EU investors

** figures in brackets show the values for the household sector, if available

All figures in USD million.

Conclusion

The two major trends that can be highlighted from the CPIS data on the Member States that provide a detailed sectoral split are that:

(i) The introduction of the Directive did not drive individual investors away from investing in securities issued in EU member States and particularly those exchanging information. In fact the trend was opposite for most of the Member States surveyed;

(ii) The Luxemburg fund industry has experienced a remarkable increase as a share in the equity investments of households in the EU.

3.5. Eurostat data – household savings/investment patterns

Description

The sectoral accounts data which is used for income elements in this section of the review combines institutional units with broadly similar characteristics and behaviour: households and non-profit institutions serving households (NPISHs), non-financial corporations, financial corporations, and the government.

The households sector comprises all households and includes household firms which cover sole proprietorships and most partnerships that do not have an independent legal status. In the European accounts, non-profit institutions serving households (NPISHs), such as charities and trade unions, are grouped with households. Their economic weight is relatively limited.

The data outlined is based on the national statistics submitted to Eurostat by the Member States, therefore cross-border transactions and financial claims among euro-area/EU Member States have not been removed as is normally the case to arrive at the EU wide statistics.

Regarding the review of the assets held by households, the European System of Accounts [99](ESA 95)2 include balance sheets per sector to describe the stocks of assets and liabilities at the beginning and at the end of the accounting period. The ESA provides not only a classification system and accounts structure, but also various rules (valuation, time or recording, etc.) which should be followed when compiling national accounts.

In the first review of the Directive, an analysis was made of the evolution of interest payments vis-à-vis other types of savings income and the evolution of household recipients vis-à-vis financial and non-financial corporate recipients ((i) and (ii) below). The report concluded that, on the basis of the data, the introduction of the Directive did not appear to have led to major changes in the composition of savings income.

Limitations of the data

The analysis given below can only be regarded as providing a broad evolution of the investment patterns of households in the EU due to the following limitations:

(a) The sectoral income elements include both cross-border income amounts which would fall under the Directive and the domestic income amounts which would not. There is a lack of data regarding the proportion of assets held cross-border by households. Using data obtained from the ECB, the Commission has calculated that for households in the euro-area 3,48% of deposits held by households were cross-border with other EU Member States in 2009 (equivalent figure for 2008 is 3,16%). Therefore cross-border deposits represent a relatively small proportion of total deposits of households in the euro-area. Unfortunately, due to data constraints, no estimate could be determined for all EU Member States. Given the small proportion of deposits held cross-border with other EU Member States, it should be underlined that only a broad interpretation can be deduced from the figures obtained from Eurostat as to overall investment patterns;

(b) The income reported to Eurostat stems from the national authorities which compile this data from a variety of sources including tax authorities and surveys; therefore it is unlikely that interest payments that evade tax will be part of the data since these payments are probably not reported by households;

(c) The level of detail on both the income and asset statistics is too broad to allow us to determine a very refined analysis of the evolution of products which have been deemed as substitutable to debt claims and which are included in the Proposal (structured products, insurance wrappers, etc). Nevertheless, data on the evolution of assets has demonstrated the importance of financial derivatives as a component for household assets.

Analysis

Data from the sectoral accounts of Eurostat was used to monitor the following elements:

(i) The evolution of the interest income received by households that would fall under the Directive compared to interest income received by all sectors of the economy;

(ii) The evolution of interest income that would fall under the Directive, received by households as compared to other forms of income (dividends etc) received by households which would not fall under the Directive;

(iii) The evolution of the type of assets held by households, the income of which would fall under the Directive (debt claims etc) as compared to assets which would not fall under the Directive (equity).

The objective is to measure whether there has been a shift to assets, and their derived income streams, which do not fall under the Directive.

Analysis of income received by total economy compared to households over the period (2000-2009) (tables 10-13):

Data source: Eurostat sectoral data obtained from the national accounts of Member States. Figures are based on the sum of the national accounts of Member States with the exception of the following: data incomplete for Bulgaria and Malta; data for Ireland not available in 2000 and 2001 and not available for Luxemburg for 2000-2005.

Total economy (tables 10 to 13):

|| Table 10 || Table 11 || Table 12 || Table 13

Year || Gross savings to GDP || Property Income to GDP || Interest Income in total property income || Share of Dividend income in total property income

2000 || 20,40% || 36,11% || 58,01% || 31,16%

2001 || 20,16% || 36,20% || 57,13% || 35,09%

2002 || 19,84% || 31,86% || 55,45% || 34,77%

2003 || 19,65% || 30,08% || 52,82% || 37,00%

2004 || 20,23% || 30,59% || 51,10% || 37,19%

2005 || 19,90% || 32,84% || 50,52% || 37,76%

2006 || 20,61% || 37,62% || 53,51% || 35,63%

2007 || 21,37% || 42,45% || 56,85% || 32,50%

2008 || 20,39% || 42,19% || 59,42% || 32,41%

2009 || 17,57% || 31,81% || 51,45% || 39,04%

The figures for gross savings[100] (table (10)) as a % of GDP generally remain constant since 2000 with a dip in 2009 due to the impact of the financial crisis, Table (11) shows that the % of property income received by the total economy has followed a U-curve since 2000 before falling back considerably in 2009 due to lower returns on financial assets. Table (12) shows that the proportion of interest income in property income[101] has followed a similar U-curve between 2000 and 2006 before peaking in 2008 then falling back significantly in 2009 as a result of lower yields on deposits and bonds[102]. Table (13) demonstrates that the fall in interest income has been matched by a similar increase in dividend income over the same period.

Households (tables 14 to 17):

|| Table 14 || Table 15 || Table 16 || Table 17

Year || Share of interest income of households in total interest income || Share of dividend income of Households in total dividend income || Share of interest income in Households' total property income || Share of dividend income in Households' total property income

2000 || 18,86% || 50,95% || 31,89% || 46,27%

2001 || 18,41% || 49,28% || 30,75% || 48,48%

2002 || 18,02% || 50,16% || 28,50% || 49,76%

2003 || 18,29% || 49,47% || 27,00% || 51,15%

2004 || 18,08% || 48,85% || 26,52% || 52,16%

2005 || 16,36% || 46,90% || 24,84% || 53,22%

2006 || 14,56% || 43,66% || 26,21% || 52,33%

2007 || 13,87% || 42,49% || 28,65% || 50,17%

2008 || 13,95% || 43,05% || 29,66% || 49,93%

2009 || 13,82% || 44,16% || 22,54% || 54,64%

Table 14 shows that interest income received by households as a % of total interest income received by all sectors of the economy has been on a downward trend since 2000 but this is also matched by a similar decrease in dividend income received by households as a % of total dividend income received by all sectors of the economy (Table 15). Table (16) shows a U-curve between 2000 and 2008 for interest income as % of total property income of households with a significant dip in 2009. Table (17) shows an increase between 2000 and 2009 for dividend income as a proportion of total property income of households. The decrease of interest income received reflects the financial crisis and the significantly lower yields on debt claim assets the income of which falls under the Directive rather than a decrease in debt claim assets held by households as disclosed in tables (18) and (19) below of assets held by households.

Analysis of assets held by households:

Table (18): financial assets held by households in millions of Euro.

|| 1999 || 2000 || 2001 || 2002 || 2003 || 2004 || 2005 || 2006 || 2007 || 2008 || 2009

Currency and deposits || 5.111.033 || 5.250.815 || 5.615.770 || 5.835.193 || 6.055.882 || 6.432.663 || 6.863.744 || 7.364.008 || 7.744.310 || 7.891.101 || 8.213.711

Loans || 70.465 || 70.941 || 72.158 || 70.178 || 75.539 || 74.974 || 77.071 || 83.092 || 86.115 || 103.643 || 99.152

Securities other than shares || 1.079.001 || 1.162.407 || 1.266.706 || 1.345.627 || 1.316.202 || 1.389.431 || 1.392.621 || 1.442.313 || 1.470.748 || 1.510.045 || 1.501.396

Other accounts receivable/payable || 442.924 || 486.348 || 511.769 || 539.494 || 520.767 || 537.062 || 580.491 || 648.469 || 644.414 || 631.878 || 680.417

Financial derivatives || 228.321 || 260.297 || 328.113 || 328.839 || 493.292 || 581.548 || 703.566 || 3.432.587 || 4.943.715 || 11.604.382 || 6.926.641

Mutual funds shares || 3.048.777 || 3.217.797 || 3.418.226 || 3.204.138 || 3.662.827 || 4.014.914 || 4.719.310 || 7.181.492 || 7.492.072 || 5.720.192 || 6.871.170

Shares and other equity || 5.957.315 || 5.917.406 || 5.409.356 || 4.574.892 || 5.007.312 || 5.402.883 || 6.195.332 || 6.815.047 || 6.853.032 || 4.874.815 || 5.559.698

Insurance technical reserves || 5.853.554 || 6.064.254 || 6.276.737 || 6.024.420 || 6.367.949 || 6.915.769 || 7.904.450 || 8.641.655 || 8.760.537 || 7.592.160 || 8.567.133

Total || 21.791.390 || 22.430.265 || 22.898.835 || 21.922.780 || 23.499.770 || 25.349.244 || 28.436.585 || 35.608.662 || 37.994.941 || 39.928.217 || 38.419.317

Table (19): category of financial asset as a % of total household financial assets

|| 1999 || 2000 || 2001 || 2002 || 2003 || 2004 || 2005 || 2006 || 2007 || 2008 || 2009

Currency and deposits || 23,45% || 23,41% || 24,52% || 26,62% || 25,77% || 25,38% || 24,14% || 20,68% || 20,38% || 19,76% || 21,38%

Loans || 0,32% || 0,32% || 0,32% || 0,32% || 0,32% || 0,30% || 0,27% || 0,23% || 0,23% || 0,26% || 0,26%

Securities other than shares || 4,95% || 5,18% || 5,53% || 6,14% || 5,60% || 5,48% || 4,90% || 4,05% || 3,87% || 3,78% || 3,91%

Other accounts receivable/payable || 2,03% || 2,17% || 2,23% || 2,46% || 2,22% || 2,12% || 2,04% || 1,82% || 1,70% || 1,58% || 1,77%

Financial derivatives || 1,05% || 1,16% || 1,43% || 1,50% || 2,10% || 2,29% || 2,47% || 9,64% || 13,01% || 29,06% || 18,03%

Mutual funds shares || 13,99% || 14,35% || 14,93% || 14,62% || 15,59% || 15,84% || 16,60% || 20,17% || 19,72% || 14,33% || 17,88%

Shares and other equity || 27,34% || 26,38% || 23,62% || 20,87% || 21,31% || 21,31% || 21,79% || 19,14% || 18,04% || 12,21% || 14,47%

Insurance technical reserves || 26,86% || 27,04% || 27,41% || 27,48% || 27,10% || 27,28% || 27,80% || 24,27% || 23,06% || 19,01% || 22,30%

Total || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00% || 100,00%

The income from the following categories of assets would come within the scope of the Directive if it is cross-border income: deposits, loans, securities other than shares, and other accounts receivable/payable. In the category currency and deposits currencies are estimated at 2,86% in 2009 of the total amount (3,13% in 2008). For mutual funds, this would depend on whether the assets of the fund would fall within the scope of the Directive in terms of their investment in debt claims[103]. Given the current scope of the Directive, it would be unlikely that financial derivatives would fall within this scope.

From the above analysis we can see that the assets that would definitely fall within the scope of the Directive (that is deposits, loans and (debt) securities) has fallen from 28.7% of total assets in 1999 to 25.5% of total assets in 2009. However, during the same period the proportion of equity assets has fallen by an even greater percentage from 27,34% to 14,47% which may reflect the greater risk aversion of households in the EU, although Table (17) shows a relatively higher proportion of dividend income in savings income

The category of financial derivatives[104] has increased from a very low base of 1,05% to 18,03% of total household assets. A more refined analysis of this data shows that the overwhelming majority of the balance of 6,926.6 billion Euro derives from UK households which comprise 5,878.4 billion Euro of this balance with French households the next highest at EUR 407 billion. The large amount for financial derivatives was first recorded by the UK national authorities in 2006 and according to the UK data sources this was due to more detailed survey methods[105] of allocating financial derivatives to the counterparty sectors including households. The financial derivatives acquired by the household sector in the UK have largely been in the form of interest rate swaps but since the end of 2009 there has been a shift towards credit derivatives.

It should be noted that the data provided to Eurostat comes from the national authorities therefore the accuracy of the data depends on the completeness of the data provided for this analysis. It could be that such product diversification is also a feature of other Member States but the data, or survey methodology, is not available to demonstrate this or it is already included in the other asset categories without further refinement.

Conclusion:

A comparison was made between income received by households compared to all sectors of the economy for income elements inside and outside the scope of the Directive to check whether there has been product substitutability. The data revealed that interest income received by EU households as a proportion of property income[106] was relatively stable until 2008 before falling sharply in 2009 due to lower interest rates received on debt claims following the financial crisis, a trend which also applies to the total economy. Therefore, no shift towards products outside the scope of the Directive could be observed. However, it should be noted that a major limitation of the Eurostat data is that it contains both domestic and cross-border income that falls under the Directive therefore only broad conclusions can be drawn.

An analysis of debt assets held by households revealed that these assets have been relatively stable over the period 2000-2009 compared to equity held by households which almost halved over the same period, perhaps reflecting the greater risk aversion of investors. A notable development is the large increase in financial derivatives held by households (1,05% of total households' assets in 1999 and 18,03% in 2009). This would tend to support an extension of the scope of the Directive to include structured financial products where the asset base is equivalent to debt claims.

3.6. Structured retail products

Description

The structured retail products database[107] was launched in 2001 and contains detailed data on a variety of structured products[108], claiming coverage of over 2 million structured products launched worldwide, with a dedicated coverage of 34 markets[109]. Starting from January 2005 the database is estimated to include around 90% of the structured retail products issued in Europe. The database is organised according to the market of the investor, allowing extracts per, inter alia, asset class, capital protection, distribution channels, payoffs, type of product provider, terms, underlying, wrapper group, etc. Both the number of products and volumes (in EUR million) are indicated. The breakdowns are available for product launches, while more aggregated data are available also for stocks.

Studies that use the data

The AMTE[110] report[111] of December 2007 analysed the retail savings market and structured product market until December 2006 in key EU Member States. The report focuses on the different products, their distribution to retail markets in the EU, and the main barriers to the European distribution of these products. The report notes the increasing share of structured products in the retail market, the two main reasons outlined being (i) the search for increased returns in a low interest rate environment and (ii) increase of the risk-aversion and demand for higher capital protection. In terms of distribution, it comments on the limited number of cross-border sales. Another reason for the fragmentation of the market is found in the flexibility of the structured products to provide comparable returns through different wrappers depending on their tax treatment in the country where those are marketed.

Limitations of the data

The point where the database does not reflect exactly the typical cross-border setup of the Directive and the Proposal is that the database is mainly targeted at the client market and would not distinguish between products distributed domestically and cross-border. Nevertheless, the data may reveal the evolution of investment patters of retail investors in general, regardless of the presence of a cross-border element. In addition, focusing on distributors which are primarily oriented to cross-border investments (e.g. from Luxemburg or Switzerland) could be a way to mitigate that limitation.

Evolution of structured products

The launching of new structured products has increased dramatically over the 2000-2007 period with an average annual increase of more than 30% for the structured products as a whole and with an average annual increase of 57% for the products with less than 100% capital protection. Despite the rapid growth of products with less than 100% capital protection, the products with protection of 100% and above[112] are still dominating the launches in terms of volumes with a share of 60-70% of all products (figure (43)). In that context, the capital protection condition under Art. 6(1)(aa)(i)[113] of the Proposal would arguably cover a substantial share of the structured products, if it is assumed that the capital protection breakdown and dynamics for the products that are distributed cross-border are similar to those for the aggregate volumes shown below.

Figure (43): Evolution of structured product by category of capital protection (million Euros)

Another important comparison is with regard to the asset class specification – figure (44). The asset class refers mainly to the composition of the structured product's underlying assets. The major asset classes are equity (single index, single share, share basket and index basket), interest rate, commodities and hybrid classes. The share of the interest rate asset class rose from 3.19% in 2001 to almost 30% in 2007. It must be noted that those structured products linked to an interest rate underlying would normally fall within the scope of Art. 6(1)(aa)(ii)[114] of the Proposal.

Figure (44): Evolution of structured product by category of asset class (million Euros)

In addition, the breakdown per wrapper class (figure (45)) would reveal the relative shares of launched products according to the type of instrument that is used for the product structuring. The securities wrapper class is the prevailing one with both the largest share and the strongest increase across the observed period. The other three significant types of wrappers are funds, deposits and life insurances. Of those, only the funds wrapper could potentially fall within the current scope of the Directive and only in cases where the underlying assets consist of debt claims. The expansion of the scope of the Directive would cover securities as per Art. 6(1)(aa) and life insurances as per Art. 6(1)(e), which, as mentioned, do appear as major wrapper groups.

Figure (45): Evolution of structured product by category of wrapper class (million Euros)

The "provider group" (figure (46)) and "distribution channel" (figure (47)) breakdowns are relevant to the identification of potential paying agents for structured products that would be within the scope of the Directive. From the provider group breakdown it is apparent that it is mostly banks and insurance companies that are involved in the structuring of the products. From the distribution channel breakdown it is also apparent that the large majority of products are marketed by the providers' sales force[115]. Figure (47) shows that the share of the direct[116] sales force group has increased substantially since 2005 and reached almost 30% in 2009, which in theory increases the possibility that those products are sold cross-border.

Figure (46) Evolution of structured products               Figure (47): Evolution of structured products per

Per category of provider group (million Euros)           category of distribution channel (million Euros)

Cross-border markets

Although the database does not allow a separate split for products manufactured in one country but sold in another, it is possible to focus the analysis on markets with a large cross-border potential. It has been established that within the EU Luxemburg is the main platform for manufacturing products in one country and distributing those in other countries, while outside the EU it is Switzerland.

The database shows 1707 products manufactured in Luxemburg (figure (48)). The primary markets for those are Belgium, France, Sweden, Germany, the Netherlands and Austria. In addition, 55% of all products manufactured in Luxemburg are with a protection greater than 95%. In terms of volumes, these are not public and therefore no comparison could be done. If any conclusions should be drawn from the aggregate EU data, the average volume per product for products with capital protection lower than 100% is much lower than that for products with capital protection higher than 100% (figure (49)). This would be translated in a higher share of the products with capital protection higher than 100% (or 95% as per the Proposal to the Directive) in terms of volumes.

Figure (48): number of LU structured funds marketed                Figure (49): LU manufactured products per capital

in other EU Member States               protection category

The database shows 928 products manufactured in Switzerland. The major primary market for those is Germany with 841 products distributed there. The exposure to the French, Dutch Austrian and Italian client base is much smaller. In addition, 34% of all products manufactured in Switzerland are with a protection greater than 95%. In terms of volumes, these are not public and therefore no comparison could be done. If any conclusions should be drawn from the aggregate EU data, the average volume per product for products with capital protection lower than 100% is much lower than that for products with capital protection higher than 100%. That would be translated in a higher share of the products with capital protection higher than 100% (or 95% as per the Proposal to the Directive) in terms of volumes.

Figure (50): number of Swiss structured products in EU           Figure (51): Swiss manufactured products per

Markets capital protection category

Conclusion

The rapid development of the structured products market in general and in particular the products with capital protection and interest-based underlying, as well as the development of particular European markets that are primarily serving foreign investors, means that the inclusion of those types of instruments in the amendments to the Directive and the Savings Agreements is both justified and necessary.

3.7. UCITS

For the purposes of this document, the term UCITS[117] is meant to refer to undertakings or entities authorised in accordance to Directive 2009/65/EC[118] (formerly Directive 85/611/EEC[119]). The terms non-UCITS and non-UCITS funds refers to all other collective investment funds or schemes.

Interest income distributed through UCITS is generally included in the scope of the Directive. As noted in the report for the first review of the Directive, the treatment of UCITS in the Directive differs from other collective schemes in the EU known as non-UCITS which are not fully captured by the Directive. Their treatment varies between those non-UCITS with legal personality (incorporated funds) and those non-UCITS without legal personality (unincorporated or "contractual" funds and trusts, etc). Income from the former is excluded from the scope of the Directive, whereas unincorporated non-UCITS are subject to the Directive as "paying agents on receipt", in regard to their income (Art.4 (2)). This difference in treatment under the current Directive makes it important to assess the evolution of UCITS and non-UCITS in the EU to ensure there have been no distortions.

Data from EFAMA

EFAMA30 has provided the Commission with an evolution of the value of UCITS and Non-UCITS as well as their respective sub categories for 2002 to 2010 (table 20). The % of UCITS in the total has decreased from 78% in 2002 to 75% in 2010 while the corresponding figure for Non-UCITS has increased from 22% in 2002 to 25% in 2010. Furthermore, in terms of the breakdown of sub-categories bond funds, which would typically come within the scope of the Directive, have decreased from 31% in 2002 to 23% of total funds in 2010. This decrease has resulted in an increase in the % share of other sub-categories of funds which may be outside the scope of the Directive depending on their asset composition.

Table (20) evolution of UCITS and non-UCITS funds over the period 2002 until 2010 in millions of Euro:

Notes:

(1) For all countries reporting UCITS & non-UCITS assets breakdowns

(2) For all countries

(3) Funds reserved to institutional investors

Limitations with the data supplied by EFAMA for the purposes of the analysis are the following:

(i) The data does not give any breakdown on the EU cross-border element of the funds which would come within the scope of the Directive;

(ii) These funds do not give a breakdown of the holder of the fund between retail investors who would come within the scope of the Directive and others which would not come within the scope of the Directive (e.g. institutional investors);

(iii) The data gives product category definitions and therefore does not provide sufficient refinement to determine the evolution of specific types of funds which are outside the current scope of the Directive but have a return/asset composition that is similar to funds which do come within the scope of the Directive.

EFAMA noted in its contribution that it would be difficult to ascertain any meaningful conclusions regarding investment patterns due to the existence and the operation of the Directive, including that of cross-border payments to the beneficiary.

ECB data on cross-border holdings of investment funds

With regard to (i) above the ECB has supplied data showing the evolution since the end of 2008 of cross-border funds issued in all EU Member States and held by households in Member States in the euro-area. The data is partially based on estimates. The data show that 94,3% of total cross-border amounts (EUR 466,7 billion) have been issued by Luxemburg at the end of the first quarter of 2011, which demonstrates the importance of Luxemburg as a financial centre for UCITS. However, it is not possible to say that this translates directly to the concept of cross-border holdings and income as contained in the Directive as the latter relates to paying agent while the data from the ECB relates to the issue of funds.

A recent report[120] by Oliver Wyman, commissioned by the Luxemburg Fund Association Industry (ALFI), indicated that alternative investment funds are principally domiciled in offshore locations. In the case of hedge funds, which may come under the Directive depending on their asset composition, the Cayman Islands had the largest share at 52% of global AUM (assets under management), followed by Delaware at 22%. The Proposal brings greater clarity to the current Directive's provisions by ensuring that the Directive encompasses interest and equivalent income from all third country funds, irrespective of their legal form and of how they are placed with investors.

Conclusion

From the aforementioned analysis it is not clear that there has been a significant switch to forms of funds and fund products that would be outside the scope of the Directive, particularly in light of the limitations of the data given above. However, any possible distortion between these various forms and fund products is addressed in the Proposal which has an equivalence of treatment between UCITS and non-UCITS funds.

3.8. Insurance products

The substitution of products within the scope of the Directive to those products outside the scope of the Directive was highlighted in the first review of the Directive. In the case of life insurance products, the Proposal includes a provision under Article 6(1)(e)[121] to include those life insurance products which could be comparable to interest bearing products that are currently within the scope of the Directive.

This review did not find any studies that have analysed the aforementioned substitutability. However, there is anecdotal evidence of the increased popularity of life insurance products with an investment element. The IMF report[122] Andorra: Assessment of Financial Sector Supervision and Regulation states:

"The Andorran financial sector has experienced significant changes over the past five years (Table 1). First, as in other jurisdictions, the low interest rate environment caused a shift in customer’s portfolios, away from bank deposits and into financial instruments offering higher returns and asset management, with the obvious consequence of shifting on balance sheet to off-balance sheet products. In this context, bank deposits remained mostly flat, while assets managed by banks and collective investment entities grew by 12.8, and 16.0 percent yearly average respectively from 2000 to 2005. Second, since July 2005 Andorra started the retention of EU taxes on savings income earned by EU residents. The prospect of higher taxes induced EU customers of Andorran banks to seek alternatives to savings instruments by moving into other financial instruments, including life policies offered by life insurance companies controlled by Andorran banks. As a result, life insurance premiums grew from EUR 61 million in 2004 to EUR 1.8 billion in 2005."

Similarly, there have been reports in the financial press[123] about investors switching to life insurance wrapper products in order to hide undeclared assets. By their nature, these are more difficult to accumulate statistics on.

A comprehensive overview of the life insurance market can be found in the 2010 study[124] undertaken by Europe Economics for the PRIPs (packaged retail investment products) initiative of the Commission. Although the study gives useful statistical information on the size of the EU life insurance market, including life insurance products with an investment element, it does not include an analysis of the distribution channels as far as cross-border selling is concerned. However, CEA has provided the Commission with the data in the following paragraphs for cross-border selling of insurance premiums up to and including 2007.

For our purposes, the life insurance market can broadly be divided into the following:

(i) Traditional life insurance products like term life insurance which pays a specified amount of money if the policy maker dies, or other biometric risk, during the term of the policy;

(ii) Other life insurance products which are, in substance, investments.

The report to Council for the first review of the Directive noted that life insurance products in category (ii) should come within the scope of the Directive where they are deemed as having no significant biometric risk coverage and where their performance is strictly linked to income from debt claims or equivalent income covered by Art. 6 and where their characteristics (notably liquidity) allow them to be marketed as substitute products to undertakings for collective investment.

Typically category (ii) products include unit linked, index linked and certain with-profits products. The recent initiatives on the Insurance Mediation Directive (IMD) and PRIPS by the Commission seek to obtain similar regulatory and disclosure requirements for category (ii) type life insurance products when comparable to other financial instruments with the same characteristics. Table 21 below shows the importance of the unit linked life insurance sector, particularly in important financial centres in the EU for cross-border investment in funds, such as Ireland (92,5% of life insurance premiums in 2008) and Luxemburg (77,0% of life insurance premiums in 2009).

According to the Europe Economics report, the EU life insurance market was worth €621 billion (gross premiums) in 2008. The UK was by far the largest market, with gross premiums in 2008 of €183 billion, followed by France (€141 billion in 2008) and Germany (€76 billion in 2008). The smallest market was Latvia, with gross premiums in 2008 of €27 million. Although not directly comparable, interest income received by EU households in 2009 was estimated at EUR 266 billion on total deposits and currencies of EUR 8,214 billion (source: Eurostat).

Table (21): evolution of unit-linked life insurance as a % of total life insurance premiums

The CEA has provided data (table 22) on cross-border life insurance premiums for the period 2003-2007. The CEA reports that there is no evidence of any significant new data trends in 2009.

Table (22): Evolution of cross-border premiums in selected EU Member States in millions of Euro (source: CEA)

The figures show that while the premiums of cross-border life insurance remain relatively low in 2007 (8,5% of the listed Member States total volume of premiums), important cross-border financial centres like Luxemburg (94,7%) and Ireland (53%) continue to have a significant proportion of their business in the sale of cross-border life insurance. Data from the ECB also indicated that these Member States have a high proportion of cross-border fund investments which could potentially fall under the Directive.

Conclusion:

Given the size of the unit linked life insurance market in the EU and the importance of cross-border life insurance in important EU financial centres, the data above reinforces the importance of extending the scope of the Directive to include these life insurance products which are equivalent to financial products already within the scope of the Directive.

[1]               http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2003:157:0038:0048:en:PDF

[2]               Report from the Commission to the Council in accordance with Article 18 of Council Directive 2003/48/EC on taxation of savings income in the form of interest payments, Brussels, 15.9.2008, COM(2008) 552 final, http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/com%282008%29552_en.pdf

[3]               Proposal for a Council Directive amending Directive 2003/48/EC on taxation of savings income in the form of interest payments, Brussels, 13.11.2008, COM(2008) 727 final, http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/com%282008%29727_en.pdf

[4]               Detailed in the Commission Staff Working Document presenting an economic evaluation of the effects of Council Directive 2003/48/EC on the basis of the available data, Brussels, 15.9.2008, SEC(2008) 2420 http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/sec%282008%292420.pdf

[5]               Proposal for a Council Directive amending Directive 2003/48/EC on taxation of savings income in the form of interest payments, Brussels, 13.11.2008, COM(2008) 727 final, http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/com%282008%29727_en.pdf

[6]               Commission Staff Working Document Impact Assessment, Brussels, 13.11.2008, SEC(2008) 2767, http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/sec%282008%292767_en.pdf and Commission Staff Working Document Impact Assessment Summary Brussels, 13.11.2008, SEC(2008) 2768, http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/savings_directive_review/sec%282008%292768_en.pdf

[7]               European Parliament legislative resolution of 24 April 2009 on the proposal for a Council directive amending Directive 2003/48/EC on taxation of savings income in the form of interest payments (COM(2008)0727 – C6-0464/2008 – 2008/0215(CNS)) http://www.europarl.europa.eu/sides/getDoc.do?type=TA&reference=P6-TA-2009-0325&language=en

[8]               Opinion of the European Economic and Social Committee on the Proposal for a Council Directive amending Directive 2003/48/EC on taxation of savings income in the form of interest payments COM(2008) 727 final - 2008/0215 (CNS) http://eescopinions.eesc.europa.eu/EESCopinionDocument.aspx?identifier=ces\eco\eco242\ces884-2009_ac.doc&language=EN

[9]               http://www.consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/ecofin/104530.pdf

[10]             http://www.consilium.europa.eu/uedocs/cms_data/docs/pressdata/en/ecofin/108392.pdf

[11]                http://register.consilium.europa.eu/servlet/driver?page=Result&lang=EN&typ=Advanced&cmsid=639&ff_COTE_DOCUMENT=6946%2F11&ff_COTE_DOSSIER_INST=&ff_TITRE=&ff_FT_TEXT=&ff_SOUS_COTE_MATIERE=&dd_DATE_DOCUMENT=&dd_DATE_REUNION=&fc=REGAISEN&srm=25&md=100&ssf=DATE_DOCUMENT+DESC

[12]                http://ec.europa.eu/taxation_customs/resources/documents/taxation/personal_tax/savings_tax/implementation/sec(2011)775_en.pdf

[13]             ACDT: Commission expert group of national experts on Administrative Cooperation in the field of Direct Taxation.

[14]     Directive 2005/60/EC of the European Parliament and of the Council of 26 October 2005 on the prevention of the use of the financial system for the purpose of money laundering and terrorist financing (Text with EEA relevance)

[15]             This is confirmed by the scarce data received from the Member States' submissions on the information exchanged.

[16]             While full taxation would arguably be ensured if all the entity’s beneficial owners were residents in the Member State in which the entity is established (with that Member State treating the entity as transparent and taxing its beneficial owners directly), it may not always be ensured if the beneficial owners are resident elsewhere (e.g. the other Member State treats the non-resident entity as opaque or is unable to ensure taxation for other reasons).

[17]             The question refers in the first place to the application of those principles with regard to double taxation treaties because the definition of interest under Article 6(1)(a) of the Directive is identical to the definition of interest under Article 11 of the OECD Model Tax Convention (and paragraph 21.1 of the Commentary to Article 11 might therefore be invoked ).

[18]             http://ec.europa.eu/taxation_customs/taxation/tax_cooperation/mutual_assistance/direct_tax_directive/index_en.htm

[19]             EE, LT indicated in their replies that interest income is exempt in their country.

[20]             http://register.consilium.europa.eu/pdf/en/08/st09/st09467.en08.pdf

[21]             The Conclusions also refer to Dependent and Associated Territories, however this part of the review will not specifically assess the use of data between third countries and dependent/associated territories.

[22]             Deadline: 30/06/2011

[23]             Austria has reported information exchanged in their statistics in 2009 despite coming under the withholding tax regime of the Directive and not having a voluntary disclosure provision in place. At the time of the preparation of this report, the Commission has not received any explanation from AT for this statistic regarding whether it would derive from information exchanged under Art. 4(2) or could otherwise be justified.

[24]             ECB data definition: deposits which are convertible into currency and/or which are transferable on demand by cheque, banker's order, debit entry or similar means, without significant delay, restriction or penalty.

[25]             http://ec.europa.eu/taxation_customs/taxation/tax_cooperation/mutual_assistance/direct_tax_directive/index_en.htm

[26]             http://ec.europa.eu/taxation_customs/taxation/personal_tax/savings_tax/savings_directive_review/index_en.htm

[27]             EBF: the European Banking Federation

[28]             Interinstitutional File: 2008/0215 Link to Council website: http://www.consilium.europa.eu/documents?lang=en

[29]             CEA: Comité Européen des Assurances

[30]             EFAMA: European Fund and Asset Management Association Ltd.

[31]             EFSA: the European Forum of Securities Associations

[32]             FECIF: Fédération Européenne des Conseils et Intermédiaires Financiers

[33]             EuroInvestors: the European Federation of Investors/ Fédération Européenne des Epargnants – a European federation representing the interests of all individual financial services customers.

[34]             Total excludes AT/BE and LU figures and includes estimate for IE using 2008 figures. Figure in Deloitte's report suggests this is 0,5%, which is based on, where available, 2009 figures or the average of sales proceeds/interest income reported over the period 2005-2009.

[35]             It should be noted that if the paying agent has no information about the part of the distribution pertaining to interest income or the investment in debt claim by the UCITS, it would still be able to report this information using certain assumptions allowed under the Directive.

[36]             A vis-à-vis country is the country of the foreign depositor, e.g. a deposit in a French bank by a German depositor would be represented by France as a reporting country and Germany as a vis-à-vis country. Claims and liabilities vis-à-vis residents of respective reporting counties are also available in the locational banking statistics but are excluded from this analysis

[37]             Derivatives subject to a master netting agreement are reported on a net basis.

[38]             Including positions with own affiliates.

[39]             Of those, the following Member States report to the BIS – Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxemburg, Netherlands, Portugal, Spain, Sweden and the United Kingdom.

[40]             Of those, the following financial centres that are part of the Savings agreements network report to the BIS – Switzerland, Cayman Islands, Guernsey, Jersey, Isle of Man and Netherlands Antilles.

[41]             BIS definition: countries with banking sectors dealing primarily with non-resident and/or foreign currency on a scale out of proportion to the size of the host economy. Page 75 of the guide lists these countries: http://www.bis.org/statistics/locbankstatsguide.pdf

[42]             The BIS is considering adopting a breakdown among banks, other financial corporations and non financial corporations.

[43]             Niels Johannesen ' Tax Evasion and Swiss Bank Deposits', EPRU Working Paper Series, 2010-05, 2010 http://www.econ.ku.dk/eprn_epru/Workings_Papers/wp-10-05.pdf

[44]             Agreement between the European Community and the Swiss Confederation providing for measures equivalent to those laid down in Council Directive 2003/48/EC on taxation of savings income in the form of interest payments, OJ L 385 of 29.12.2004, p. 30.

[45]             It is presumed that only individuals' decisions are affected by the introduction of the Directive and not the companies' decisions.

[46]             BIS definition: Non-banks are all entities (including individuals but excluding official monetary authorities) other than those defined as 'banks'. General government and public corporations are part of the non-bank sector. Deposits should comprise all claims reflecting evidence of deposit, including non-negotiable certificates of deposit (CDs) which are not represented by negotiable securities.

[47]             http://www.snb.ch/en/iabout/stat/statpub/bchpub/stats/bankench

[48]             BIS Locational banking statistics http://www.bis.org/statistics/bankstats.htm, Table 3B External loans and deposits of banks in individual reporting countries, In all currencies vis-à-vis the non-bank sector http://www.bis.org/statistics/qcsv/anx3b.csv

[49]             Belgium is considered as a Member State that withholds a tax for the purposes of that comparison.

[50]             Consists of Singapore, Hong Kong, Taiwan (Chinese Taipei), Bahrain, Bahamas, Panama, Macao, Malaysia and Bermuda. The balances for Singapore are comparable to those for Belgium and the balances for Hong Kong are comparable to those for France.

[51]             Consists of United States, Japan, Canada, India, Australia, Norway, Turkey, Chile, South Africa, South Korea, Brazil and Mexico

[52]             BIS definition for deposits: institutions are asked to declare close substitutes for deposits in addition to deposits. It may be appropriate to also include collective investment schemes (includes hedge funds but excludes pension funds and money market funds), such as mutual funds, money market funds, in the reporting population if their cross-border activities are considered as playing an important role in a country's money creation and money transmission process.

[53]             The notation in the table represents the country of the bank on the left and the country of the depositor on the right, e.g. info-EU->EU means a situation where the bank is in EU Member State exchanging information and the depositor is from any EU Member State.

[54]             From those, the following are available as vis-à-vis jurisdictions – Bahamas, Bahrain, Barbados, Belize, Bermuda, Brunei, Comoros Islands, Costa Rica, Djibouti, Dominica, Fiji, French Polynesia, Grenada, Guatemala, Hong Kong SAR, Kiribati, Lebanon, Liberia, Macao SAR, Malaysia, Maldives, Marshall Islands, Mauritius, Micronesia, Nauru, New Caledonia, New Zealand, Palau, Panama, Philippines, Samoa, Sao Tome and Principe, Seychelles, Singapore, Solomon Islands, South Africa, Tonga, Tuvalu (formerly the Ellice Islands), United Arab Emirates, Uruguay, Vanuatu

[55]             From those, the following are available as vis-a-vis jurisdictions – Andorra, Aruba, Cayman Islands, Guernsey, Isle of Man, Jersey, Liechtenstein, Netherlands Antilles, San Marino, Switzerland, Turks and Caicos and West Indies UK (i.e. Anguilla, Antigua and Barbuda, British Virgin Islands, Montserrat and St. Christopher/St. Kitts – Nevis)

[56]             The share rebounded slightly by a 1 percentage point for the following three years.

[57]             BIS Locational banking statistics http://www.bis.org/statistics/bankstats.htm, Table 7B External loans and deposits of reporting banks vis-à-vis individual countries, Vis-à-vis the non-bank sector http://www.bis.org/statistics/qcsv/anx7b.csv

[58]             For BIS purposes that includes Anguilla, Antigua and Barbuda, British Virgin Islands, Montserrat and St. Christopher/St. Kitts – Nevis

[59]             It must be noted that Swiss-source interest income is currently outside the scope of the EU-Swiss Agreement. The Commission has obtained from BIS a separate data set only for fiduciary business, which are not considered Swiss-source income, but are subject to the EU-Swiss Agreement. The effect using the separate data set for fiduciary business produces comparable results that are even more pronounced.

[60]             There was change in reporting (more reporting banks, etc) with a break-in-series of about $112 billion for international claims and about $133 billion for international liabilities. Please see more details at http://www.bis.org/statistics/breakstables17.pdf

[61]             REGULATION (EC) No 2423/2001 OF THE EUROPEAN CENTRAL BANK of 22 November 2001 concerning the consolidated balance sheet of the monetary financial institutions sector (ECB/2001/13)

[62]             As became evident from the ad-hoc report on the application of the Savings Directive, there are different interpretations as to whether repurchase agreements should be considered as debt claims.

[63]             http://sdw.ecb.europa.eu/browse.do?node=9484269

[64]             This is especially the case for "Deposits with agreed maturity, Up to 1 year", where the "Outstanding amount" interest rate for "Deposits with agreed maturity, Up to 2 years" was used instead of the "New business" interest rate for "Deposits with agreed maturity, Up to 1 year". This was done due to the very high correlation between the "New business" interest rate for "Deposits with agreed maturity, Up to 1 year" and the "New business" interest rate for "Deposits with agreed maturity, Up to 2 years".

[65]             That was possible for BG, CY, DK, EE (for 2008 and 2009), ES, HU, LT, MT, NL, PL, SI and the UK.

[66]             Those Member States are CZ, DE, EL, FI, FR, IE, IT, LT, LU (for the voluntary disclosure), PT, RO, SE and SK

[67]             The 70% benchmark is considered sufficiently low to avoid any temporal differences and other effects outlined in the methodology section.

[68]             For a description of the developments in the UCITS markets of Member States, see Section 3.6

[69]             The sensitivity analysis for the UK if 0% is used for overnight deposits and repurchase agreements gives the following % coverage for the years 2006 – 18,81%; 2007 – 64,73%; 2008 – 22,16%; 2009 – 15,35%.

[70]             The ECB data series for EE start from 2008.

[71]             The average for FI without the 2006 figures is 103,89%

[72]             The figure for LU reflects only the % coverage of the underlying tax base by the voluntarily exchanged information.

[73]             Sweden does not provide a summary of exchanged data. The calculations are based on reports on received data by other Member States.

[74]             The significant decrease for France and Spain in mid 2005 is due to a change in classification (and not a genuine change to the deposit base) of a significant amount of deposits from "agreed maturity below 1 year" to "overnight deposits"

[75]             Again, deposits with agreed maturity are only compared, assuming that the deposits with agreed maturity produce relatively higher interest income and are therefore more sensitive to the tax treatment thereof. When the results are replicated including also deposits redeemable at notice up to 3 months, the decreasing share (average monthly decrease 2.3%) of the household sector investing in the Netherlands can also be observed.

[76]             http://www.snb.ch/en/iabout/stat/statpub/bchpub/stats/bankench

[77]             http://www.snb.ch/ext/stats/bankench/pdf/deen/Stat32.pdf

[78]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bLaender_D32_M8.xls

[79]             http://www.snb.ch/ext/stats/bankench/pdf/deen/Stat38.pdf

[80]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bTreuGesch_D38_M1.xls

[81]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bTreuGesch_D38_M2.xls

[82]             http://www.snb.ch/ext/stats/bankench/pdf/deen/Stat38c.pdf

[83]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bWertDepots_D38c_M2.xls

[84]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bWertDepots_D38c_M4.xls

[85]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bWertDepots_D38c_M5.xls

[86]             http://www.snb.ch/ext/stats/bankench/xls/en/bankench_bWertDepots_D38c_M6.xls

[87]             http://www.parisschoolofeconomics.eu/docs/zucman-gabriel/mwn27july.pdf, data appendix http://www.parisschoolofeconomics.eu/docs/zucman-gabriel/appendix26july.pdf and data sheets http://www.parisschoolofeconomics.eu/docs/zucman-gabriel/maintablesfigures.xlsx, http://www.parisschoolofeconomics.eu/docs/zucman-gabriel/appendix.xlsx

[88]             As of footnote 44 above.Art. 1(2) of the agreement stipulates: "Interest payments made on debt-claims issued by debtors who are residents of Switzerland or pertaining to permanent establishments of non-residents located in Switzerland shall be excluded from the retention."

[89]             Again, it can be safely assumed that the non-bank sector in those jurisdictions does not consist primarily of industrial companies or individuals, but of intermediary entities.

[90]             i.e. including Anguilla, Antigua and Barbuda, British Virgin Islands, Montserrat and St. Christopher/St. Kitts – Nevis

[91]             Unofficial translation: "Fiduciary transactions include investments, loans and equity interests which the bank holds or grants in its own name, but for the account and at the risk of the customer, on the basis of a written agreement. The instructing customer bears the full currency, transfer, price and collection risks and is the exclusive beneficiary of all accruals from such transactions; the bank only charges a commission. Fiduciary funds received by the banks mainly come from abroad and are almost exclusively invested abroad. They essentially consist of short term foreign investments in third banks or in branches legally dependent on Swiss banks. In the latter case, these transactions must appear in the balance-sheet, since they involve a commitment from the branch towards the head office in Switzerland."

[92]             There is an explicit reference to fiduciary deposits in the EU-Swiss agreement, Art. 7(1)(a) "interest paid, or credited to an account, relating to debt claims of every kind including interest paid on fiduciary deposits by Swiss paying agents for the benefit of beneficial owners as defined in Article 4".

[93]             On UCITS and the applicable definitions, c. section 3.7 below.

[94]             Rixen-Schwartz 'How effective is the European Union’s Savings Tax Directive? Evidence from four EU Member States', Journal of Common Market Studies, Forthcoming.

[95]             http://www.imf.org/external/pubs/ft/cpis/2002/index.htm

[96]             For more information and examples, see Appendix I to the CPIS Guide.

[97]             The EU economies that demonstrate the relatively high numbers for investment in debt securities by all types of investors are (by order of magnitude): FR, UK, DE, LU, IE, NL, IT, BE, ES, AT, DK, PT, SE and FI.

[98]             2008 and 2009 are not considered due to the large variations in valuation in the crisis years

[99]             http://epp.eurostat.ec.europa.eu/portal/page/portal/sector_accounts/concepts/institutional_sectors

[100]            Gross saving is defined as gross disposable income less final consumption expenditure.

[101]            Property income is defined as the income received less expenses accruing, during the income reference period, by the owner of a financial asset or a tangible non-produced asset (land) in return for providing funds to or putting the tangible non-produced asset at the disposal of another institutional unit. The property income is broken down into: 1. Interest, dividends, profits from capital investment in an unincorporated business (HY090G); 2. Income from rental of a property or land (HY040G).

[102]            For example the rate for outstanding amounts for all maturities in the euro-area decreased from 4,03% in December 2008 to 2,57% in December 2009

[103]            According to EFAMA statistics 23% of total funds are classified as bond funds in 2010.

[104]            The European System of Accounts (1995) defines derivatives as financial assets based on or derived from a different underlying instrument. The underlying instrument is usually another financial asset, but may also be a commodity or an index. Financial derivatives are also referred to as secondary instruments and since the hedging or off setting of risk are frequently a motivation for their creation, they can be referred to as hedging instruments. Only those secondary instruments, which have a market value, because they are tradable or can be off set on the market, are financial assets in the system of accounts and are classified as derivatives.

[105]            Bank of England survey: DQ survey of financial derivatives business and PL survey of MFIs' profit and loss

[106]            Property income: income derived from assets

[107]            http://www.structuredretailproducts.com/

[108]            Structured products are defined for the purposes of the database as "an investment product that most often uses derivatives to deliver a pre-defined return, usually on a passive or formulaic basis regardless of the wrapper tax status. Structured products typically come in two forms: growth products (which may provide an element of capital protection) and income products (that provide a fixed high income but with a risk to the capital return)."

[109]            Europe: Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Hungary, Ireland, Italy, Netherlands, Norway, Poland, Portugal, Slovakia, Spain, Sweden, Switzerland and the UK.

                North America: USA and Canada.

                Latin America: Brazil, Chile and Mexico.

                Asia-Pacific: Australia, China, Hong Kong, India, Japan, Malaysia, New Zealand, Singapore, Taiwan and South Korea.

[110]            Euro debt market association

[111]            http://www.icmagroup.org/ICMAGroup/files/1c/1cc0911c-5257-4631-aed7-bcb453627f88.pdf

[112]            i.e. providing additional fixed income to the capital protection

[113]            Article 6

Definition of interest payment

1. For the purposes of this Directive, ‘interest payment’ means

(aa) any income paid or realised, or credited to an account, relating to securities of any kind, except where the income is directly considered to be an interest payment in accordance with points (a), (b), (c) or (d), and where

                (i) the conditions of a return of capital defined at the issuing date include a commitment towards the investor that he receives, at the end of the term, at least 95% of the capital invested, or

                (ii) the conditions defined at the issuing date provide for a link of at least 95% of the income from the security to interest or income of the kinds referred to in points (a), (b), (c) or (d);

[114]            see footnote 79

[115]            Under the database definition the distribution by sales force is when "sales are made via a tied sales force owned by the Provider or Provider Group, often based in bank branches".

[116]            Under the database definition the direct distribution is when "sales are made directly to the end investor via the internet, the telephone or post."

[117]            Undertakings for Collective Investment in Transferable Securities

[118]            Directive 2009/65/EC of the European Parliament and of the Council of 13 July 2009 on the coordination of laws, regulations and administrative provisions relating to undertakings for collective investment in transferable securities (UCITS), OJ L 302, 17.11.2009, p. 32

[119]            Council Directive 85/611/EEC of 20 December 1985 on the coordination of laws, regulations and administrative provisions relating to undertakings for collective investment in transferable securities (UCITS), OJ L 375, 31.12.1985, p. 3

[120]            http://www.alfi.lu/publications-statements/press-releases/report-alternative-investment-funds-domicile

[121]            (e) benefits from a life insurance contract, if

(i) the contract contains a guarantee of income return, or

(ii) the actual performance of the contract is at more than 40% linked to

interest or income referred to in points (a), (aa), (b) (c) and (d).

(The 40% thresholds referred to in points (d) and (e) (ii) of paragraph 1 and in paragraph 3 shall from 1 January 2011 be 25%).

[122]            http://www.imf.org/external/pubs/ft/scr/2007/cr0769.pdf

[123]            Bloomberg article 06/04/2010 U:\doss\000xyz\000703 Review Savings Directive\000703 2011 Review Savings Directive\studies\substitution effect\Background\Switzerland Threatens Tax Cheats Using ‘Wrappers’ From Insurers - BusinessWeek.mht

[124]            Study on the Costs and Benefits of potential changes to distribution changes for Insurance Investment Products and other Non-MIFID packaged retail investment products. http://ec.europa.eu/internal_market/consultations/docs/2010/prips/costs_benefits_study_en.pdf