EN

Brussels, 21 March 2000

SEC (2000) 522

Commission staff working paper

Revisiting the Dublin Convention:

developing Community legislation for determining which Member State is responsible for considering an application for asylum submitted in one of the Member States

The Dublin Convention provides a mechanism for determining which Member State is responsible for considering an application for asylum lodged in one of the Member States. On the basis of over two year's experience of implementing the Convention, there seems to be widespread agreement that it is not functioning as well as had been hoped. Following the entry into force of the Treaty of Amsterdam, the Convention needs to be replaced with a Community instrument. The Commission considers that it is sensible to use the opportunity provided by the new Treaty both to take stock of practical experience of implementing the Dublin Convention to date and also to reflect again on the principles on which the Convention is based, in the light of the objectives established by the Treaty of Amsterdam in the field of asylum. This working document analyses the extent to which the Dublin Convention system delivers a series of objectives and presents some policy options. In addition, the Commission is ready to work in conjunction with the Member States to complement this working document with a detailed practical evaluation, which could make use of the results of evaluative work on the Dublin Convention which is being carried out in the framework of the Odysseus programme. On the basis of the discussion which this document is intended to facilitate, and taking account of the results of further practical evaluation, the Commission will draw up a proposal for a Community legal instrument to replace the Dublin Convention.

The Dublin Convention

On 15 June 1990, the Member States concluded the Convention determining the State responsible for examining applications for asylum lodged in one of the Member States of the European Communities, which is generally known as the "Dublin Convention". The Convention entered into force on 1 September 1997 for the twelve original signatories, on 1 October 1997 for Austria and Sweden, and on 1 January 1998 for Finland. It effectively replaced similar arrangements for determining responsibility for considering asylum applications which had been in force since 26 March 1995 amongst the states applying the Schengen Agreement.

Article 18 of the Dublin Convention establishes a committee which is responsible for examining questions of application and interpretation. This Committee has adopted three decisions:

Decision 1/97 concerning provisions for the implementation of the Convention;

Decision 2/97 establishing the Committee's Rules of Procedure;

Decision 1/98 concerning provisions for the implementation of the Convention;

Decision 1/97 laid down some important substantive provisions relating to the implementation of the Convention, which were subsequently supplemented by decision 1/98. These decisions are an integral part of the overall Dublin Convention system.

The Dublin system has two distinct components with different purposes: the criteria for determining which Member State is responsible for considering an asylum application (Articles 4 — 8); and the readmission rules which apply when a person who has previously lodged an asylum claim in one Member State and is subsequently present in a second Member State (Articles 3(7) and 10). These two sets of rules are described in more detail below.

The Dublin Convention establishes a hierarchy of responsibility criteria. The first criterion is based on the principle of family reunification. The rest of the main criteria are based on the principle that the Member State which is responsible for a person's presence on the territory of the Member States should also be responsible for considering any subsequent asylum application. The criteria are as follows:

If the applicant for asylum has a family member who has been recognised as having refugee status within the meaning of the Geneva Convention in a Member State, and is legally resident there, that state will be responsible, provided the person concerned so desires (Article 4);

If the applicant is in possession of a valid residence permit or valid visa, the Member State which issued it will be responsible for examining the application for asylum. Provision is made in relation to cases where the applicant is in possession of more than one valid residence permit or visa issued by different Member States and in relation to cases where the visa(s) and/or residence permit(s) have expired (Article 5);

If it can be proved that the applicant for asylum irregularly crossed the border into a Member State by land, sea or air, having come from a non-member state, the Member State thus entered will be responsible, unless it can be proved that the applicant has been living in the Member State where the application for asylum was presented for at least six months before making the application (Article 6);

The Member State responsible for controlling the entry of the applicant into the territory of the Member States will be responsible for examining the application for asylum unless the applicant first entered a Member State where the visa obligation is waived, before presenting an application for asylum in another Member State where the visa obligation is also waived (Article 7);

If none of these criteria applies, the first Member State in which the application for asylum is lodged is responsible for examining it (Article 8);

The readmission criteria contained in the Dublin Convention are essentially concerned with ensuring that an asylum applicant can not pursue an asylum claim in a Member State other than the one which is responsible for considering his or her application. The Convention therefore provides for the applicant to be readmitted by the state responsible in the following circumstances:

If the applicant is irregularly in a second Member State whilst his/her claim is under examination in the State responsible (Article 10(1)(c));

If the applicant withdraws his/her application while it is under examination in the State responsible and lodges another asylum application in a second Member State (Article 10(1)(d));

If the applicant's claim has been rejected in the State responsible and he/she is illegally in another Member State (Article 10(1)(e));

Additionally, the Convention contains a provision which is designed to prevent asylum applicants from circumventing the mechanism contained in the Convention for determining responsibility. Article 3(7) deals with the situation where a person lodges an asylum application in one Member State, which begins to apply the Dublin procedure for determining the Member State responsible, but he/she then withdraws that asylum claim, travels to another Member State and claims asylum there. It provides that in this case he/she would be taken back by the first Member State with a view to completing the Dublin procedure for determining the Member State responsible.

The Convention also contains two important discretionary provisions. Under Article 3(4), the so-called "sovereignty" clause, each Member State retains the right to examine any asylum application submitted to it, even if it is not the Member State responsible under the criteria set out in the Convention, provided that the applicant agrees. Under Article 9, the so-called "humanitarian" clause, any Member State, even if it is not the Member State responsible under the criteria set out in the Convention, may for humanitarian reasons based particularly on family and cultural grounds, examine an application for asylum at the request of another Member State.

The Treaty of Amsterdam, the Vienna Action Plan and the Tampere conclusions

The Treaty of Amsterdam introduced important amendments to the Treaties which effectively require the replacement of the Dublin Convention with a Community legal instrument within five years of the entry into force of that Treaty (i.e. by 1 May 2004). Article 63 of the amended Treaty establishing the European Community sets out a clear programme of work in the field of asylum and protection. Article 63(1)(a) requires the Council to adopt "criteria and mechanisms for determining which Member State is responsible for considering an application for asylum submitted by a national of a third country in one of the Member States". These measures must be in accordance with the Geneva Convention of 28 July 1951 and the Protocol of 31 January 1967 relating to the status of refugee and other relevant treaties.

The "Action Plan of the Council and Commission on how best to implement the provisions of the Treaty of Amsterdam establishing an area of freedom, security and justice" of December 1998 (hereafter referred to as the "Action Plan") specifies that the measures to be taken within two years of the entry into force of the Treaty of Amsterdam include "continued examination of the criteria and conditions for improving the implementation of the Convention and of the possible transformation of the legal basis to the system of Amsterdam".

The Action Plan also identifies two specific issues which should be addressed in this context. These are to what extent the mechanism should be supplemented inter alia:

by provisions enabling responsibility for dealing with the members of the same family to be conferred on one Member State where the application of the responsibility criteria would involve a number of States; and

by provisions whereby the question of protection when a refugee changes his country of residence can be resolved satisfactorily.

The Action Plan also identifies the need to take measures in the field of asylum to limit "secondary movements" by asylum seekers between Member States. This is a relevant consideration in relation to the content of an instrument to replace the Dublin Convention.

The European Council held a special meeting in Tampere on 15 — 16 October 1999 on the creation of an area of freedom, security and justice. In the Tampere conclusions, the European Council called for the establishment of a Common European Asylum System. It specified that in the short term this System should include a clear and workable determination of the State responsible for the examination of an asylum application.

Other developments since the conclusion of the Dublin Convention

In reviewing the operation of the Dublin Convention with a view to replacing it with a Community legal instrument, it is appropriate to note three relevant developments which either have already occurred in the decade since the Dublin Convention was signed or else can be expected to take place in the foreseeable future:

The gradual realisation in practice of an area without internal frontiers in which the free movement of persons is ensured. Considerable progress towards an area without internal frontiers has been achieved in the context of Schengen co-operation. The Schengen acquis has been integrated into the framework of the European Union by virtue of the relevant protocol in the Treaty of Amsterdam. The area without internal frontiers is in the process of being extended to include a tenth Member State, and in the near future is expected to cover thirteen Member States. As internal frontiers are progressively abolished, the potential for secondary movements of asylum applicants is increased, and flanking measures are required. The Dublin Convention itself was linked conceptually to the objective of creating an area without internal frontiers, as its preamble makes clear. The Treaty of Amsterdam specifically envisages that measures aimed at ensuring the free movement of persons will be accompanied by directly related flanking measures, covering inter alia criteria and mechanisms for determining which Member State is responsible for considering an application for asylum submitted in one of the Member States.

The enlargement of the European Union. There were just twelve Member States when the Dublin Convention was signed. Austria, Sweden and Finland have acceded to the Union since then and significant further expansion will take place in the future when the candidate countries accede to the Union. Enlargement will both redefine the Union's external land borders and create a larger area without internal frontiers.

The development of the Eurodac system. Even before the Dublin Convention entered into force, it was apparent to the Member States that it would be difficult to implement the Convention solely on the basis of the evidence provided by travel and identity documents, since these can easily be disposed of or destroyed. In December 1991, Ministers responsible for immigration meeting in The Hague therefore agreed that a feasibility study for a Community wide fingerprint system for asylum applicants should be undertaken. Work has been ongoing since then to develop a system for the computerised comparison of fingerprints to assist the implementation of the Dublin Convention.

Between March 1996 and March 1999, the Council drew up the Eurodac Convention and Protocol, in accordance with the provisions of Title VI of the Treaty on European Union. The Convention provided for the collection and comparison of fingerprint data on applicants for asylum in order to facilitate the identification of cases where an applicant for asylum had previously claimed asylum in another Member State. The Protocol was intended to further facilitate the application of the Dublin Convention in two ways. Firstly, it provided for the collection of fingerprint data relating to persons apprehended in connection with the irregular crossing of an external border to provide evidence if the persons concerned later claimed asylum. Secondly, it provided a facility to make checks with Eurodac in certain circumstances to determine whether a person found illegally present within a Member State had previously claimed asylum in another Member State. The Council agreed on and "froze" these two texts in December 1998 and March 1999. Following the entry into force of the Treaty of Amsterdam, the Commission drew up a proposal for a Eurodac Regulation. Negotiations on this proposal are at an advanced stage, and arrangements for the practical implementation of the Eurodac system are being made.

PART II - EVALUATION OF THE DUBLIN CONVENTION IN THE LIGHT OF THE OBJECTIVES OF A SYSTEM FOR ALLOCATING RESPONSIBILITY FOR ASYLUM APPLICATIONS

Now that the Dublin Convention has been in force for over two years, it should be possible to arrive at a considered verdict on whether or not it has operated successfully in practice. On the basis of the experience which they have reported so far, few if any Member States appear to regard the Dublin Convention as an unqualified success, and many have expressed concerns about the efficiency of the Dublin system. Indeed, the Article 18 Committee drew up and in June 1998 approved a "Dublin Convention: Programme of Action" in response to concerns about the operation of the Convention.

The Dublin Convention does not lay down any monitoring and evaluation criteria according to which its success or failure can be judged. It is therefore necessary to go back to first principles and to consider what the objectives of the Dublin Convention are or should be. The preamble to the Convention identifies the purpose of the Convention in quite narrow terms, related purely to providing certain guarantees for asylum applicants and addressing the problem of "refugees in orbit". It must be recognised however that many interested parties consider that the Dublin Convention either has or should have certain other objectives. This working document therefore examines the Dublin Convention in the light of the following possible objectives of a mechanism for determining responsibility for asylum applications:

to avoid any situations where applicants for asylum are left in doubt for too long as regards the likely outcome of their applications;

to provide all applicants for asylum with a guarantee that one of the Member States will promptly assume responsibility for considering their asylum application and to ensure that applicants for asylum are not referred successively from one Member State to another without any of these States acknowledging itself to be competent to examine the application for asylum;

to prevent asylum applicants from being able to pursue multiple asylum applications, either concurrently or consecutively, in different Member States;

to create a direct link between the allocation of responsibility for asylum applicants and the success or failure with which a Member State discharges its responsibilities (in an area without internal frontiers) to carry out effective pre-entry and entry controls on persons seeking to enter the territory of the Member States;

to deter misuse of asylum procedures by preventing asylum applicants being able to choose in which Member State they seek asylum;

to maintain the unity of families and to reunite separated families;

to ensure an equitable distribution of asylum applicants between the Member States, in proportion to each Member State's capacity to receive asylum applicants;

It should be stressed that these are only possible objectives, that some of them are controversial, and that not all of them are compatible with each other. The Commission's approach in Section II of this Working Document has therefore been to consider which of the objectives mentioned above are appropriate and realistic, to examine to what extent the Dublin Convention achieves and is designed to achieve each of these objectives, and to consider what improvements or changes to the system could be made in order to better attain the objectives which are appropriate and realistic.

In addition, it is clear that the system for allocating responsibility for asylum applicants should be clear, workable and effective. The extent to which the Dublin Convention meets this test is analysed in Section III of this Working Document.

(1) To avoid any situations where applicants for asylum are left in doubt for too long as regards the likely outcome of their applications. This objective is specifically mentioned in the preamble to the Dublin Convention. A speedy asylum system is an appropriate aim of European asylum policy as a whole. It can not be achieved solely by an instrument concerned with criteria and mechanisms for allocating responsibility for considering asylum applications. Apportioning responsibility for asylum applications is only a preliminary step, and a speedy system depends mainly on efficient substantive asylum procedures. Nevertheless, an important requirement for the allocation of responsibility is that it must be done quickly.

The Dublin Convention is frequently criticised for operating too slowly in practice. The quarterly statistics on the Dublin Convention which are exchanged between Member States under the terms of decision 1/97 are not sufficiently detailed to provide a complete picture of well justified this criticism is, since they relate to just one stage of the Dublin process. They do however indicate that the average time for responding to a transfer request frequently exceeds the target of 1 month set down in Article 4(1) of decision 1/97, and in relations between some Member States the average delay is as much as 90 days, which is the maximum allowed under Article 11(4) of the Convention before the state to which the request was made is automatically deemed to have accepted responsibility. No statistics are compiled on the following: the average time which elapses between the lodging of an asylum claim and the making of a request for another Member State to take charge of an applicant; the average time it takes a Member State to respond to a request for information made under Article 15 of the Dublin Convention prior to a formal request for another Member State to take charge of an applicant; the average time which elapses between a Member State agreeing that is responsible for an asylum applicant and the actual transfer of the applicant to that Member State.

Under the provisions on time limits in the Dublin Convention, it is possible for a period of up to 9 months to elapse between the lodging of an asylum application and a decision on responsibility. Article 11(1) of the Convention provides that a Member State has a maximum of six months from the date when an asylum claim is made to call on another Member State to take charge of the applicant, and Article 11(4) provides that the second Member State then has a maximum of three months to reply. Under Article 11(5), there is then a further month for the transfer to take place, unless there is a suspensive appeal against the decision on responsibility, in which case the period of one month for the transfer to take place does not start running until the conclusion of the appeal proceedings. In recognition of the need for quick decisions, Article 4(1) of decision 1/97 instructed Member States to make every effort to respond to a request to take charge of an applicant within one month rather than three. Article 4(4) envisaged a review one year after decision 1/97 had been in effect with a view to considering whether one month could constitute a maximum time limit, but this has never taken place. Article 5 of decision 1/97 established an urgency procedure, but the Member State to which a request is made under this procedure is not bound to respond within the deadline laid down by the requesting Member State. Little has been done to encourage Member States to make any request to a second Member State to take charge of an applicant under Article 11(1) of the Convention at the earliest possible moment, or to facilitate this. In particular, the Convention itself does not lay down a time limit within which a Member State must reply to a request for information made under Article 15 of the Convention, and decision 1/97 simply instructs the Member State concerned to "make every effort" to reply to the request if possible immediately and in any case within one month.

In conclusion, it is essential that the system for determining which Member State is responsible for considering an asylum applicant operates quickly in order to provide certainty both for applicants and for Member States. There would appear to be scope for improving the Dublin Convention system in this respect.

(2) To provide all applicants for asylum with a guarantee that one of the Member States will assume responsibility for considering their asylum application and to ensure that applicants for asylum are not referred successively from one Member State to another without any of these States acknowledging itself to be competent to examine the application for asylum. This objective is specifically referred to in the preamble to the Dublin Convention. It can be considered the fundamental purpose of the Dublin Convention (despite the fact that the Convention is regarded by many as an instrument concerned primarily with control rather than with protection). The hierarchy of responsibility criteria in the Dublin Convention theoretically guarantees that one of the Member States accepts responsibility for an asylum applicant. In addition, the implementing rules contain specific arrangements for the applicant to report to or be taken in charge by the authorities of the second Member State. This should ensure that the phenomenon of "refugees in orbit" between the Member States can no longer occur. This represents a considerable step forward from the situation which prevailed previously, when an asylum applicant could be transferred between Member States without an agreement on which Member State was responsible for considering the claim, and indeed without prior contact between the authorities of the Member States concerned.

Whilst on paper the Dublin system appears to provide a comprehensive solution to the problem of refugees in orbit between the Member States, it relies on rapid practical implementation to achieve this result. The slow operation of the system can mean that an applicant is informed by a Member State that it considers that is not responsible for considering his or her claim, but there is a considerable delay before the transfer is carried out. Decision 1/97 of the Article 18 Committee says that if the Dublin procedure ended up taking longer than the substantive asylum procedure "the Convention would fail totally to have the desired effect and would even contradict one of its objectives since the delays would create a new category of "refugees in orbit", asylum-seekers whose applications would not be examined until the procedure laid down under the Convention had been completed".

Another aspect of the system which has given rise to difficulty in some cases is the provision in Article 3(5) of the Convention whereby Member States retain the right to send an asylum applicant to a third state. Problems can arise in cases where the Member State to which a transfer request is made would apply the safe third county concept in a case where the requesting state would not do so because it does not consider that the third country can be regarded as safe for the applicant. This has given rise to concerns in some Member States about so called "chain refoulement", and is only one of a number of examples which illustrate that if a mechanism for allocating responsibility for asylum applicants is to operate effectively, it must be accompanied by common standards in procedural and substantive areas of asylum law.

(3) To prevent asylum applicants from being able to pursue multiple asylum applications, either concurrently or consecutively, in different Member States. The Dublin Convention is generally regarded as an instrument designed to tackle the problem of multiple asylum applications in different Member States. In paragraphs 5 and 6 of this document, a distinction was drawn between the responsibility criteria and the readmission criteria in the Dublin Convention. The purpose of the readmission criteria is clearly to prevent a person from pursuing asylum applications concurrently or consecutively in different Member States. This is based on the sensible principle that an asylum applicant should have his or her case dealt with by one Member State only.

At present it is not possible to assess what proportion of the total number of multiple applications which are lodged in the EU are being successfully identified and dealt with under the Dublin Convention. But this is an area where the Eurodac system can make a decisive contribution. Eurodac will in due course provide proof of a previous asylum application in another Member State in every case where such an application has been made (although the system will start empty and will only become fully effective after several years of operation). It is not possible to evaluate the extent to which the Dublin Convention has deterred multiple asylum applications in the past, but it is reasonable to assume that the introduction of Eurodac will increase the deterrent effect. Irrespective of what the rules are on determining initial responsibility for considering an asylum application, it is difficult to see how any system could function without rules on readmission ("taking back"). In conclusion, the readmission rules are necessary and Eurodac will make it possible to identify all cases in which they could be applied. Whilst it may be appropriate to look again at the provisions in Article 10 of the Convention which specify the cut off point when the obligation to take back an asylum applicant ceases to apply in order to ensure that they are consistent and justifiable, in general the current readmission rules do not need a major overhaul.

The application of the rules on taking back asylum applicants can be hindered by differences in national legislation and policy. National courts have in some cases prevented transfers where they have had concerns about the circumstances in which asylum was refused by the Member State which was responsible for dealing with the claim. This is a further illustration of the need for a common approach in areas such as the treatment of repeat claims or further representations from applicants whose claims have been refused and the substantive interpretation of the refugee definition.

(4) To create a direct link between the allocation of responsibility for asylum applicants and the success or failure with which a Member State discharges its responsibilities (in an area without internal frontiers) to carry out effective pre-entry and entry controls on persons seeking to enter the territory of the Member States. The Dublin Convention establishes a link between the performance of controls on entry to the territory of the Member States and responsibility for subsequent applications for asylum. Its preamble refers specifically to the objective of an area without internal frontiers in which the free movement of persons shall be ensured and states that the measures in the Convention are taken in pursuit of this objective. The criteria set out in Articles 5 — 7 of the Dublin Convention are based on the premise that the Member State which is responsible for controlling a person's entry onto the territory of the Member States should also be responsible for considering any subsequent asylum application.

The questions which arise are first whether this is an appropriate basis for allocating responsibility and second whether it can be achieved effectively. In response to the first question, the answer is that this is a political choice for the European Community. In signing the Dublin Convention in 1990, the Member States decided that it was appropriate to link responsibility for considering asylum applications to the responsibility for controlling the external borders and to seek to limit the impact of removing internal frontier controls on each Member State's asylum system. The Dublin system was therefore designed to counterbalance the increased scope for secondary movements of asylum seekers in an area without internal frontiers. In the decade since the Convention was signed, the conditions in which such secondary movements can take place more easily have steadily come about. But experience of the operation of the Dublin Convention suggests that it can be difficult in practice to obtain sufficient proof to demonstrate that a particular Member State is responsible for a person's presence on the territory of the European Union and should therefore assume responsibility for considering their asylum application. (Problems relating to evidence and proof are discussed in detail below.)

The Treaty of Amsterdam identifies a link between achieving free movement of persons in an area without internal frontiers and adopting directly related flanking measures on criteria and mechanisms for assigning responsibility for asylum applicants. It does not, however, state whether there should be a link between the duty to control the external border and the allocation of responsibility for an asylum applicant. The Community will need to decide whether the current approach is workable, once the difficulties relating to proof have been fully evaluated.

(v) To deter misuse of asylum procedures by preventing asylum applicants being able to choose in which Member State they seek asylum. The structure of the Dublin Convention system should theoretically deny asylum applicants the freedom to travel within the European Union before deciding where to lodge an application for asylum. Articles 6 and 7 are intended to apply in all cases where an applicant has moved from the Member State which he or she first entered to a second Member State before claiming asylum. The Convention would be an effective mechanism for redressing secondary movements if there was evidence to apply it in all cases. In many cases where these is evidence that an asylum applicant was previously in another Member State, however, there is insufficient evidence to demonstrate which Member State was responsible for the applicant's entry into the European Union. (See the analysis below on problems relating to evidence).

The principle that a person seeking international protection should claim asylum in the first safe country which he or she reaches is frequently invoked by administrations which are anxious to combat "forum shopping". Member States are concerned that if people are free to choose the Member State in which they lodge an asylum application, there is increased scope for them to go to a particular state for economic reasons and to lodge an asylum application to ensure that they cannot be removed. There is also a concern that, if asylum applicants were able to choose the Member State in which they claimed asylum, there might be a tendency for them to treat some Member States as transit countries rather than as countries of destination. As mentioned above, the Action Plan identifies the need to take measures in the field of asylum to limit "secondary movements" by asylum seekers between Member States.

The proposition that allowing asylum applicants the freedom to choose where to lodge their asylum claim gives rise to misuse of asylum procedures is controversial, and frequently contested by the non-governmental sector in particular. It is difficult to establish conclusively what factors cause asylum applicants to lodge their claims in particular Member States. Research in this area has suggested that a range of factors other than economic considerations also play a part. On the other hand, Member States report concerted attempts to circumvent their immigration controls through exploitation of their asylum systems, including cases where traffickers in human beings are targeting particular destinations and advising their clients to claim asylum both to prevent removal and to secure economic benefits. In such a complex and controversial area, the question of the extent to which an instrument for apportioning responsibility for asylum claims has a role in preventing misuse of asylum procedures will remain a matter of opinion.

Other components of the Common European Asylum System will have a significant role to play in reducing differences between Member States which may influence the distribution pattern of asylum applications within the European Union. Substantive asylum law and asylum procedures have not yet been approximated and the recognition rates for certain nationalities can vary significantly from one Member State to another, so it is understandable that people in need of international protection may find one Member State a more attractive destination than another. The conditions in which asylum applicants are received also vary considerably between the Member States.

(vi) To maintain the unity of families and to reunite separated families. There is general agreement that a system for determining responsibility for considering asylum applications should respect the principle of maintaining the unity of families and reuniting separated families, even if there are divergent views on the exact circumstances in which the system should do this in practice.

The provisions on family unity and reunification in the Dublin Convention are limited, but they do override all other criteria on responsibility contained in the Convention. Article 4 is designed to ensure family reunification or maintain family unity in cases where the asylum applicant has an immediate family member who has been recognised as a refugee and is legally resident in one of the Member States. For the purpose of Article 4, a family member is a spouse or minor child, or a parent in cases where the applicant is a minor. Article 3(4) (the "sovereignty clause") and Article 9 (the "humanitarian clause") may be used at the discretion of a Member State to maintain family groups and to reunite families in a much wider range of cases, but the Convention does not automatically ensure that this is achieved. The Article 18 Committee has considered but been unable to agree on a draft decision concerning the transfer of responsibilities for family members in accordance with Article 3(4) and Article 9 of the Dublin Convention.

The Convention does not contain objective criteria to preserve family unity or reunite the family in cases where an asylum applicant has a family member who is legally resident (but not recognised as a refugee) in one of the Member States. The Community is currently developing general rules on family unity on the basis of the Commission's recent proposal for a directive on the right to family reunification. It will be important to ensure that a Community act on responsibility for applicants for asylum is consistent with future Community legislation on family unity.

In cases where several members of the same family seek asylum in the European Union, the application of the Dublin Convention criteria can result in two or more Member States being responsible for taking charge of the applicants. The Action Plan specifies that further work on the Dublin Convention should investigate the scope for additional provisions enabling responsibility for dealing with the members of the same family to be conferred on one Member State where the application of the responsibility criteria would involve a number of States. The aim should be to develop binding rules which ensure family unity in such cases, rather than guidelines which make the application of this principle discretionary.

(vii) To ensure an equitable distribution of asylum applicants between the Member States, in proportion to each Member State's capacity to receive asylum applicants. Despite the claims which are sometimes made about the Dublin Convention as a burden sharing instrument, it is not designed to distribute asylum applicants between the Member States in accordance with an objective indicator of each Member State's capacity to receive applicants. The Dublin system is incompatible with an approach under which each Member State would take responsibility for a fixed proportion of the total number of asylum applicants in the European Union. The Convention has been criticised in some quarters on the grounds that it puts too great a burden on Member States which have external borders which are particularly exposed to migratory pressures. The statistics on implementation of the Dublin Convention confirm that the Member States with external land and sea borders to the south and east are net recipients of asylum applicants under the Dublin system. At the same time, the number of transfers carried out is relatively modest, and at present the system can not be said to be putting an excessive burden on any Member State.

The Treaty of Amsterdam specifically introduced the objective of promoting a balance of effort between Member States in receiving and bearing the consequences of receiving refugees and displaced persons. As a first step towards achieving this objective, the Commission has made a proposal for a Council decision creating a European Refugee Fund, under which the Community would provide financial assistance to the Member States in proportion to the number of asylum applications which they receive and the number of refugees they recognise. This issue will also need to be addressed in the context of future discussions on temporary protection. As mentioned above, the Common European Asylum System will also have a role to play in removing certain differences between the Member States which appear to contribute towards the unequal distribution of asylum applications between the Member States. The most pragmatic approach to the question of burden sharing would appear to be to address it further in the contexts mentioned above, and not to seek to replace the Dublin Convention with a mechanism for distributing asylum applicants between the Member States in proportion to each Member State's capacity to receive them, particularly since discussions on physical burden sharing according to factors such as each state's population, population density or GDP have not produced any concrete results.

III. THE CLARITY, EFFICIENCY AND EFFECTIVENESS OF THE DUBLIN CONVENTION AND ITS IMPLEMENTATION

The Tampere conclusions call for a clear and workable system for determining which Member State is responsible for considering an asylum application. It is fair to assume that the need for the system to be clear, workable and effective is uncontroversial. Experience of the implementation of the Dublin Convention has given rise to a range of concerns about its clarity, efficiency and effectiveness. These are examined in detail in this section.

Transfers: the statistical evidence (1). Under Article 28(1) of decision 1/97, the Member States conduct quarterly exchanges of statistical information on the practical implementation of the Convention in a format prescribed under that decision. Unfortunately, the statistics which have been submitted under this decision are incomplete, and this makes it very difficult to draw up an overall analysis. Certain conclusions can, however, be drawn. In 1998, the incomplete figures record approximately 3,000 transfers out of Member States and around 4,500 transfers into Member States under the Dublin Convention (if the figures were complete, these figures should of course correspond). On this basis, it would appear that less than 2% of asylum applicants are transferred between the Member States under the Convention. On the basis of the figures available for 1998 and 1999, Germany and Austria are the main net recipients of asylum applicants under the Convention. For both of these Member States, applicants transferred in under the Convention account for about 4% of all asylum applicants. The main net "exporter" of asylum applicants under the Convention is Denmark, which succeeds in transferring about 18% of all asylum applicants. The next best performer in numerical terms, the United Kingdom, transfers only about 1% of asylum applicants. The performance of Denmark apparently reflects the successful operation of a bilateral agreement with Germany, and it will be instructive to examine and build upon the experience of bilateral agreements between the Member States to improve the application of the Dublin Convention.

Transfers: the statistical evidence (2). Another striking aspect of the statistics is the fact that the number of transfer requests accepted by the Member State to which they are directed is considerably greater than the number of transfers actually implemented. This may in part be due to successful legal challenges to a transfer decision, but it probably also indicates that the transfer arrangements themselves are not fully satisfactory. Article 20 of decision 1/97 offers a choice between an escorted transfer and a transfer under the applicant's own initiative, and it would appear that in a significant proportion of cases where the transfer is unescorted, the applicant disappears from the system altogether.

Cost-effectiveness. The statistics which are compiled on the implementation of the Convention do not indicate what human and other resources each Member State is devoting to the implementation of the Convention. It will be appropriate for each Member State to evaluate this aspect of its arrangements for implementing the Convention. Some Member States have established specialist Dublin Convention Units, and in these cases at least it should be a straightforward task to identify how many people are working on the implementation of the Convention. This will provide a basis for examining whether the Convention is costs effective: the concern which will need to be tested is that a considerable number of staff are deployed implementing the Dublin Convention yet the number of transfers carried out under the Convention is relatively small concern.

Resources: balance of efforts. It will also be appropriate to look at whether all Member States are devoting sufficient resources to the implementation of the Convention, since if the system is to operate successfully it is important that Member States are able to reply promptly to information requests under Article 15 of the Convention as well as to transfer requests. The present system relies on all Member States making an equivalent effort to implement it.

Evidence. Criticisms which have been made of the inefficient operation of the responsibility criteria contained in the Dublin Convention tend to focus in particular on problems relating to evidence. Implementing the principle that the Member State responsible for a person's presence on the territory of the Member States is responsible for any subsequent asylum claim depends on the availability of evidence relating to the individual's immigration history. But in many cases, there is no evidence at all, either because it did not exist in the first place or because it has been destroyed. In many other cases, no single item of evidence provides conclusive proof of responsibility. There are therefore very substantial obstacles in practice to applying the Dublin Convention to all the cases where it might theoretically be applicable.

Evidence: problems caused by the lack of evidence. Although the Dublin system is designed to operate on the basis that the Member State which is responsible for a person's presence on the territory of the Member States should be responsible for considering any subsequent application for asylum, the absence of documentary or other evidence relating to entry has the effect in a high proportion of cases that the Member State where an asylum claim is lodged has to assume responsibility for considering it (under Article 8 of the Convention). The problem of lack of evidence arises in particular in relation to illegal entry, which by its very nature is not generally documented. A limited but nevertheless significant step towards addressing this issue is envisaged under the proposal for a Eurodac regulation, under which persons who were apprehended in connection with the irregular crossing of a Member State's border with a third country and who were not turned back to the third country would be fingerprinted, and there data held on the Eurodac database. This measure obviously can not be applied in cases where people evade detection on entry, and will therefore only be as effective as the controls which are carried out at external borders. Short of introducing a system under which the criteria for determining responsibility for considering an asylum application are not linked to the applicant's immigration history, it is however extremely difficult to devise any modification to the Dublin Convention criteria or any alternative set of criteria which can compensate for the absence of evidence in such cases.

Evidence: incentives to destroy documents. Documentary evidence relating to entry and presence (passports/travel documents, entry stamps, visas, residence permits etc.) plays an important role in supporting the application of the Dublin Convention criteria, and there is therefore a strong incentive for people who wish to circumvent the system and claim asylum in a particular Member State to destroy their documents. The destruction of documents not only frustrates the application of the Dublin Convention criteria, but can also give rise to other difficulties elsewhere in the asylum and immigration systems. In some cases, doubts about an applicant's nationality may hamper consideration of his or her claim. The destruction of documents also makes removal of asylum applicants whose claims have been finally rejected more difficult. There is no firm statistical evidence available on the scale of the problem of document destruction in potential Dublin Convention cases. It is difficult to address the problem through modification of the Dublin Convention responsibility criteria, for the same reason that it is difficult to adjust the criteria to deal with lack of evidence of illegal entry. Responses have therefore tended to take the form of accelerated substantive asylum procedures for applicants who have destroyed their documents in bad faith, as envisaged in paragraph 9(c) of the 1992 resolution on manifestly unfounded applications for asylum.

Evidence: problems experienced in cases where indicative evidence only is available. Chapter IV of decision 1/97 of the Article 18 committee lays down rules on means of proof in the framework of the Dublin Convention. These rules draw a distinction between probative evidence, which provides conclusive proof of a Member State's responsibility for examining an application for asylum, save where rebutted by evidence to the contrary, and indicative evidence, the value of which in determining responsibility has to be weighed up on a case by case basis. The rules provide that a Member State should be prepared to assume responsibility on the basis of indicative evidence once it emerges from an overall examination of the asylum applicant's situation that, in all probability, responsibility relies with the Member State in question. The application of these rules seems to give rise to considerable difficulties, since there does not appear to be a consistent common approach to indicative evidence which is shared by all Member States. These difficulties will need to be examined in more detail during the course of the proposed evaluation exercise. Whilst there may be scope for making the rules on proof clearer and more specific, part of the difficulty in applying them seems to be related to the fact that the interpretation of the rules, or the weight attached to certain provisions, differs from one Member State to another. In this respect, the institutional changes introduced by the Treaty of Amsterdam will provide the means to ensure a consistent and uniform interpretation of the legislation.

Absence of judicial oversight. The Court of Justice of the European Communities has no jurisdiction to interpret the Dublin Convention or to rule on disputes between the Member States in relation to the Convention. Instead, the task of examining questions of application is entrusted to the Article 18 Committee, which does not have at its disposal any instruments to ensure that its decisions are applied in a consistent fashion. The Article 18 committee has not therefore been able to ensure the uniform application and interpretation of the Dublin Convention in the way that the Court of Justice ensures the proper application of Community legislation. This defect will automatically be rectified when the Dublin Convention is replaced with a Community legal instrument. The Court of Justice will have jurisdiction in accordance with Article 68 and the general provisions of the Treaty relating to the Court. This will provide a more effective mechanism for dealing with cases where rulings on questions of interpretation are required as well as for addressing possible infringements.

Scope: (1) Recognised refugees. Some Member States consider that the Convention applies to people who are recognised as refugees in one Member State and who subsequently apply for asylum in a second Member State in the same way that it applies to any other asylum applicant. Others consider that it can not be applied to recognised refugees. The Convention certainly does not address the reasons why a recognised refugee might want to apply for asylum in another Member State. The Action Plan recognises the need to focus on the causes of secondary movements by refugees, and calls for further work so that the question of protection when a refugee changes his country of residence can be resolved satisfactorily.

The Commission has already set out its position on this issue in the context of its revised proposal for a Eurodac regulation. If some people who have been recognised as refugees in one Member State are travelling to another Member State and seeking asylum there, this is likely to be because refugees do not enjoy a general right to reside in a Member State other than the one in which they were recognised and admitted as a refugee. The Commission considers that this situation should be remedied by including refugees within the scope of an instrument based on Article 63(4) of the Treaty defining the circumstances in which a third country national who is legally resident in one Member State may reside in another Member State. This approach is consistent with conclusion 15 of the Tampere European Council, which stated that in the longer term, Community rules should lead to a uniform status for those who are granted asylum valid throughout the Union.

Scope: (2) People seeking subsidiary forms of protection. The Dublin Convention applies only to applicants for asylum, who are defined as people seeking protection under the Geneva Convention by claiming refugee status within the meaning of Article 1 of the Geneva Convention. It does not apply to people who are claiming protection exclusively under the European Convention on Human Rights or under other international or national provisions. Several Member States have expressed concern that the rules on responsibility in the Dublin Convention do not deal effectively or at all with cases where people either withdraw their asylum claims and seek protection in some other capacity, or never lodge an asylum claim at all and from the outset lodge a claim for a subsidiary form of protection only.

At present, arrangements for dealing with claims for protection which are not based on the refugee definition in the Geneva Convention vary significantly from one Member State to another, and in some Member States there are no formal arrangements. The Tampere conclusions call for the Common European Asylum System to include measures on subsidiary forms of protection offering an appropriate status to any person in need of such protection. The Commission has previously stated that it sees a good case for establishing a single procedure for examining all the protection issues raised by an individual cases, but noted that a considerable amount of work still needs to be done on defining which types of cases should be covered by subsidiary protection arrangements. Progress in this area is a prerequisite for any extension of the scope ratione personae of the system for determining responsibility for asylum applicants.

Problems arising as a result of differences in policy and practice between the Member States. The difficulties examined above relate to factors which make it difficult for the Member States to apply the Dublin Convention's criteria for determining responsibility to a significant number of cases and in a consistent fashion. But even in cases where the Member States concerned are able to agree that an applicant should be transferred from the state where he has lodged an asylum application to the one which is responsible under the criteria set out in the Convention, there may nevertheless be obstacles which prevent the transfer taking place. National courts have been prepared to intervene in cases where the asylum law or practice of the Member State to which the applicant would be transferred differs in a significant fashion from the law and practice of the first Member State. In particular, different approaches to the application of the Geneva Convention in cases where an individual's asylum claim is based on a fear of persecution by non-state agents and also to the use of the safe third country concept have prompted national courts to intervene and block transfer. This illustrates that insufficient progress was made towards approximating asylum laws in the European Union under the intergovernmental arrangements introduced by the Maastricht Treaty. For example, whilst the 1996 Joint Position on the harmonized application of the term "refugee" in some respects provided helpful guidelines on the application of the Geneva Convention, it was unable to provide the basis for a common approach to the issue of persecution by non-state agents, which remains a key policy question. The solution to this problem can not be to legislate to replace the Dublin Convention in such a way that Member States lose the discretion currently provided under Article 3(4) of the Dublin Convention to examine an application for asylum even when they are not the state responsible under the criteria laid out in the Convention. Instead, the solution lies in successfully implementing the asylum provisions of the Treaty of Amsterdam in accordance with the Action Plan and the Tampere conclusions. Paragraph 13 of the Tampere conclusions states that the European Council has agreed to work towards establishing a Common European Asylum System, based on the full and inclusive application of the Geneva Convention. Paragraph 14 makes it clear that in addition to a clear and workable system for determination of the state responsible for the examination of an asylum application, the System must include common standards for a fair and efficient asylum procedure, common minimum conditions for the reception of asylum seekers and the approximation of rules on the recognition and content of the refugee status, and it must be complemented with measures on subsidiary forms of protection. Progress is undoubtedly required in all these areas.

Miscellaneous difficulties. A number of other difficulties with the Dublin Convention have also been encountered. In some places the Convention is unclear (for example, on its application in cases of transit through a third country after initial entry into the EU). In others, it is need of updating (for example Article 7 which anticipated the adoption of the External Frontiers Convention and is not fully consistent with Schengen rules for controlling the external border). Some measures which are included in the implementing rules would be better suited to the parent instrument (for example, the basic principles on means of proof). It is not the purpose of this Working Document to analyse each of these issues, but it should be noted that there is some scope for improvements of a technical nature.

IV: CONCLUSIONS AND POLICY OPTIONS FOR FUTURE LEGISLATION

Summary of findings on the Dublin Convention

A number of conclusions can be drawn on the basis of the analysis in the preceding paragraphs. The Dublin Convention can be considered a success or a potential success in the following respects:

In general terms, it achieves its basic objective of ensuring that asylum applicants do not find themselves "in orbit" between the Member States, without any Member State acknowledging itself to be responsible (although the slow operation of the system at present can put applicants in an unsatisfactory position);

The rules on readmission which are designed to tackle multiple asylum applications in different Member States have the potential to be extremely effective when Eurodac is introduced.

The operation of the Convention does, however, give rise to a number of serious concerns:

There is evidence which suggests that the system operates too slowly, which is incompatible with the objective of speeding up asylum procedures.

It is inherently difficult to assemble the necessary evidence to demonstrate which Member State is responsible for a person's presence on the territory of the Member States. Where documentary evidence exists, there is an incentive to destroy it. There is generally no evidence at all of illegal entry, and Eurodac is likely to be of limited help only in addressing this problem. Where there is indicative evidence only, there is disagreement about how to treat it. The result appears to be that in the great majority of cases, the Member State in which the asylum application was lodged is sooner or later forced to accept responsibility under Article 8 of the Convention.

In some cases where it is clear which State is responsible under the rules in the Convention, differences in national asylum policy and practice nevertheless cause problems.

It is not clear whether, at present, the Convention is producing sufficient benefits to justify the resources which are devoted to its implementation.

A number of developments have the potential to help improve the situation:

the introduction of Eurodac, as well as making a decisive contribution to the implementation of the readmission rules in the Dublin Convention, will also help to alleviate the current problems relating to evidence to support the application of the responsibility criteria. It must, however, be stressed that Eurodac will not provide a complete solution to the problem of evidence in relation to the rules on responsibility.

the communitarisation of asylum policy has ensured that Community mechanisms including the Court of Justice will be available for ensuring the uniform interpretation and consistent application of legislation;

in the medium term, the establishment of the Common European Asylum System called for in the Tampere conclusions will remove the significant differences in policy and practice between the Member States which can at present be an obstacle to transfer;

Possible alternatives to the Dublin Convention system

55. The Commission has stated previously that it is appropriate to use the opportunity provided by the transition to new treaty arrangements to consider whether a fundamentally different approach is required to the question of responsibility for considering asylum applications. A fundamentally different approach would imply a system which was not based on the principle that the Member State responsible for a person's presence on the territory of the Union should be responsible for considering any subsequent asylum application. Such a system would have to be workable and it would have to be based on coherent and justifiable principles.

56. It must be acknowledged that it is not easy to identify many models which satisfy these tests :

A system which allocated responsibility for an asylum applicant to the last known transit country within the EU. Such an approach has sometimes been suggested, because in some cases a Member State does not have evidence to demonstrate which Member State an applicant for asylum first entered, but it can prove from which Member State the applicant arrived. A system which allocated responsibility to the last known transit country would not, however, be based on a justifiable principle, since it would effectively penalise states for the removal of internal frontiers.

A system based on an aspect of the applicant's immigration history other than the Member State which was responsible for his or her entry to the European Union. Such a system would probably be arbitrary, since there is no very obvious alternative immigration criterion to employ. In addition, it would be no easier to operate than the current system, since evidence that the criterion had been fulfilled would still be a key issue.

A system based on the applicant's country or origin. Perhaps the most radical approach which has been suggested is for a system under which all asylum applicants from a particular country or origin would be the responsibility of one specified Member State. Such an approach would certainly be controversial, given its demographic consequences. Even if efforts were made to take account of cultural and historical links, there would also be an element of arbitrariness in the system. This approach would be particularly unlikely to achieve a balance of efforts between the Member States, since a crisis in a country of origin would place a very heavy burden on one Member State.

A system based on allocation of responsibility according to where the asylum application is lodged. Another alternative model which has received some support, particularly within the non-governmental sector, is to allocate responsibility to the first Member State where an asylum claim is lodged, whilst retaining readmission rules to prevent an asylum applicant pursuing applications in several different Member States. Advocates of this model argue that it removes the complex and bureaucratic rules on responsibility contained in the Dublin Convention, which have not worked well in practice, whilst still addressing the issue of multiple asylum applications. However, this approach does not address or seek to address some of the other possible objectives discussed in section II above. Self evidently, it does not establish a link between responsibility for controlling the external frontier and responsibility for dealing with any subsequent asylum application. Nor does it seek to deny an asylum applicant the freedom to choose in which Member State to pursue his or her asylum application or to address "forum shopping".

Conclusions

57. On the basis of the information which is already available, it is clear that the Dublin Convention has not operated as well in practice as its authors hoped it would. The most familiar criticisms of the implementation of the Convention were summarised in Section III of this document, and the Commission has indicated its readiness to co-ordinate a more detailed practical evaluation in order to provide a full picture of the operation of the Convention.

58. No system for allocating responsibility for asylum applicants can aspire to deliver all of the possible objectives analysed in Section II of this working document. It is therefore necessary to make certain political choices. The Commission considers that any system for allocating responsibility must provide a guarantee that an asylum application will be examined by one of the Member States, but it is legitimate to prevent asylum applicants from being able to pursue multiple asylum applications, either consecutively or concurently, in different Member States. A system for apportioning responsibility for asylum applicants clearly should not frustrate the achievement of the Treaty objective of promoting a balance of efforts between the Member States, although the achievement of this objective requires a range of measures and will be facilitated by the establishment of a Common European Asylum System. The extent to which a mechanism for allocating responsibility for asylum applicants can effectively deliver the other objectives discussed in this paper will need to be determined in the light of further evaluative work.

59. There do not appear to be many viable alternatives to the present system. The approach of allocating responsibility according to where the first asylum claim is lodged is at least clear, but it implies a choice between objectives which can be recognised as equally important. This model would certainly provide the basis for a clear and workable system in relation to the objectives of speed and certainty, avoiding refugees in orbit, tackling multiple asylum applications and ensuring family unity. But it would rely on harmonisation in other areas such as asylum procedures, reception conditions, interpretation of the refugee definition and subsidiary protection to reduce any perceived incentives for asylum applicants to choose between Member States when lodging their application.