52003SA0006

Special Report No 6/2003

concerning twinning as the main instrument to support institution-building in candidate countries together with the Commission's replies

(pursuant to Article 248(4), second subparagraph, of the EC Treaty)

(2003/C 167/02)

TABLE OF CONTENTS

>TABLE>

SUMMARY

I. A strong administrative capacity (institution-building) is vital for the candidate countries of central and eastern Europe in order to be able to adopt and implement Community law (acquis communautaire), one of the criteria for accession. The Commission launched twinning in 1998 as the main instrument to assist candidate countries in strengthening their administrative capacity. Twinning involves the provision of national experts by Member States to candidate countries. Up to February 2002, 503 twinning projects in total had been approved, with an overall budget of 471 million euro.

II. The Member State administrations and public institutions have unique knowledge and specific experience concerning the implementation and enforcement of Community law. The introduction of the twinning instrument in 1998 provided for transferring knowledge and experience from the public sector of the Member States to the candidate countries. Twinning is therefore a positive initiative by the Commission to assist candidate countries in acquiring the capacity to adopt, implement and enforce the acquis communautaire(1).

III. The twinning projects acted as a catalyst in setting the candidate countries' reform in motion, bringing together specialists from Member States and candidate countries' administrations and promoting the adoption of the acquis communautaire through legislation. However, progress was rather less in implementing and enforcing the acquis. The objectives stated in the twinning covenants (the so-called "guaranteed" results(2) were often unrealistic, and could often be achieved only partially within the project period. In practice it proved overly optimistic to expect that a fully functioning, efficient and sustainable candidate country organisation would exist in a given field after one twinning project, the average duration of which is 18 months (see paragraphs 16 to 22).

IV. Twinning is a complex activity involving a variety of functions and participants (the national experts or Pre-Accession Advisers (PAAs), different Member State and candidate country administrations and the Commission (Brussels and Delegations)). The desired results can only be achieved if all parties perform as required. The audit showed that the limited achievements are in general due to a combination of factors involving all parties (see paragraphs 23 to 24):

- setting of unrealistic objectives,

- poor candidate country commitment and ownership,

- management shortcomings at the level of Member State administrations and the Commission.

V. Twinning is a relatively new approach in delivering institution-building assistance to the candidate countries. A positive feature of it is the fact that the Commission assumed its responsibility for organising a learning process and as a result introduced many changes to improve the instrument (see paragraphs 25 to 32). However, a negative side effect was increased complexity as different sets of rules from those applying to older ones had to be applied to new projects (see paragraph 43).

VI. The interaction of the numerous public administrations involved in twinning created administrative complexity, diminishing efficiency and effectiveness. Too much time is spent on purely administrative issues, to the detriment of the main task, namely advising candidate countries on institution-building. The still lengthy periods between needs assessment and project realisation, as well as the highly complicated payment systems, are two of the more significant of the current difficulties (see paragraphs 33 to 44). There was also a tendency to over-emphasise twinning at the expense of other mechanisms that are eligible for support. This sometimes resulted in a departure from the instrument's original aim and in an insufficiently selective use of twinning (see paragraphs 45 to 51).

VII. While the instrument certainly contributed towards strengthening the candidate country's institutional and administrative capacity with regard to the requirements of EU accession, there is considerable room for further improvement. It is recommended that (see paragraphs 60 to 64):

- the rate of delivery should be increased by a real results-orientated approach. The basic components needed for effective implementing capacity should be better identified and the objectives for individual projects should be defined more realistically and precisely,

- the different stages, from needs assessment up to project realisation, should be completed more quickly and be less bureaucratic,

- payment procedures should be simplified and accelerated,

- the use of fixed-price or lump-sum contracts requiring specified deliverables should be considered,

- the use of twinning should be the result of a conscious choice between different instruments: the Commission should increase its efforts to ensure the most appropriate mix of the different instruments available for institution-building,

- the Commission should create a PAA network to safeguard the store of specific knowledge, to spread good practice and to reduce the risk of errors recurring.

INTRODUCTION

1. The ability of the candidate countries of central and eastern Europe to adopt and implement Community law (acquis communautaire) was established at the European Council in Copenhagen in 1993 as one of the criteria for accession. Twinning, introduced in 1998, is one of the means by which the Union has sought to assist the Candidate countries in strengthening their administrative capacity.

2. In November 1999, the Commission adopted new guidelines for the implementation of the PHARE programme for the period 2000 to 2006, in which PHARE support is targeted on two main priorities:

(a) institution-building (around 30 % of PHARE resources);

(b) investment support (around 70 % of PHARE resources).

3. The first priority, institution-building, is defined as the process of helping the candidate countries to develop the structures, strategies, human resources and management skills needed to strengthen their economic, social, regulatory and administrative capacities. PHARE contributes to the financing of institution-building in all sectors.

4. The main instrument for institution-building is twinning. Twinning involves the provision by public administrations of Member States of the EU of seconded experts to the candidate countries on a full-time and long-term basis(3). The overall purpose is to help candidate countries to put in place or strengthen the institutions required to incorporate the acquis communautaire(4), through a transfer of knowledge and experience from Member State administrations and public institutions. While the long-term national expert or pre-accession adviser (PAA)(5) forms the core of a twinning project, the twinning package may also include additional short or medium-term experts and the supply of training and in duly justified cases small items of office equipment (unit cost below 5000 euro). The candidate country chooses the twinning partner.

5. Key elements of the twinning instrument are the twinning covenants signed by candidate countries and Member States and endorsed by the Commission to make them eligible for funding by PHARE. These covenants contain the specific objectives of the relevant twinning projects which, in the fields selected, cover a wide range of issues (such as organisation, training, legislation, information, technology, etc.). They constitute the legal documents which commit both parties to achieving the results desired through the secondment of Member State experts to candidate country administrations.

6. The management of twinning is complex. The Commission is responsible for the overall management and financing of the twinning instrument, and its Delegations also have a role in preparing and monitoring individual covenants (see Diagram 1). However, the main responsibility for implementing the individual covenants has been delegated to the Member States and the candidate countries.

Diagram

Main steps and responsibilities in twinning

>TABLE>

Abbreviations

CC: Candidate country

MS: Member State

CFCU: Central financing and contracting unit of CC

HQ: Headquarters

Source:

European Commission.

7. The Commission developed a Twinning Manual(6), which sets out in detail the legal, financial and procedural framework for twinning projects.

PROJECT OVERVIEW AND BUDGETARY SIGNIFICANCE

8. There is no specific article devoted to twinning in the general budget of the European Union. The instrument is financed from chapter B7-03 (PHARE pre-accession instrument) together with infrastructure/investment projects.

9. The Commission launched the twinning concept in 1998 and the first projects started in the summer of 1999. Up to February 2002, a total of 503 twinning projects were approved by the Commission with a total budget of 471 million euro (see Table 1), giving an average budget per project of nearly 1 million euro.

Table 1

Overall budget for twinning projects 1998-2001

>TABLE>

Source:

European Commission.

10. The Commission's central management unit in Brussels has all information on the amounts contracted for each twinning project. As to the actual payments made, the Commission's central management unit had no up to date information at the beginning of the Court's audit because it intended to wait for the closure of the first round of twinning projects. During the Court's audit the central management unit obtained such information from the Delegations. The responsibility for making payments for twinning projects lies with central financing and contracting units (CFCU), based in the Ministry of Finance, and other implementing agencies in each candidate country.

11. Table 2 shows the twinning projects by candidate country and by sector (agriculture, environment, etc.). Poland, Romania and the Czech Republic have the highest number of twinning projects. Each project may have more than one Member State participating, either as project leader or partner (see Table 3).

Table 2

Twinning projects per sector in the CCs 1998-2001

>TABLE>

Source:

European Commission.

Table 3

Number of twinning projects in which the Member States are involved (as leaders or partners)

>TABLE>

Source:

European Commission.

THE COURT'S AUDIT

12. The main objective of the audit was to assess whether the implementation of the twinning instrument by the Commission, the Member States and the candidate countries had been efficient and effective. The audit did not include an assessment of the legality and regularity of the commitments and payments. To support the Court's assessment of effectiveness of management, a sample of 45 projects out of all those completed up to February 2002 which numbered 98, covering a variety of sectors of the acquis communautaire, was examined to establish how far they had met their objectives as defined in the covenants. Most of these completed projects were initiated in 1998 which was the start-up year for the twinning instrument. Special attention was paid to the area of "Structural Fund preparation", as this will have particular impact on the future spending of EU funds. A separate audit on environmental measures funded through the pre-accession instruments(7) included an examination of twinning projects in this area, and the main findings are incorporated into this report.

13. To assess the efficiency of the implementation of the instrument the procedures and structures were analysed, especially for projects which started in the second half of 2001 and at the beginning of 2002. The assessment included the Commission's readiness to learn lessons and take corrective action.

14. The audit work included a review of central Commission management files, project files from selection rounds held in 1998,1999 and 2000 and external monitoring reports by the OMAS Consortium(8). The auditors attended a PAA meeting in Brussels in February 2002. They also interviewed the representatives of all stakeholder groups (Delegation, PAAs, project leaders, national contact points (NCP)(9) and CFCUs) in Estonia, Hungary and Poland. The NCP in Finland (particularly involved in projects in the Baltic States) and twinning partners in the United Kingdom (playing a leading role in the area of external audit) were also visited. The environmental measures audit included visits to all 10 candidate countries benefiting from PHARE and ISPA assistance.

15. The Court evaluated the replies to questionnaires received from the NCPs of eight Member States and six candidate countries and 54 separate replies to questionnaires received from PAAs and project leaders during the Court's visits to Estonia, Hungary and Poland. Further information was obtained from SIGMA(10). The Court also took note of the status reports issued by the Commission in February 2002.

ACHIEVEMENT OF RESULTS

16. The majority of the twinning projects from the first selection round 1998 started in the second half of 1999. About a quarter of them started in the beginning of 2000. As the average project duration is around 18 months, most of the first round of twinning projects were completed in the second half of 2001 or the beginning of 2002. Thus it was necessary for the Commission to extend the disbursement deadline for some PHARE 1998 twinning projects until 31 March 2002. The Court's assessment of the final results of projects was concentrated on those which the Commission indicated were completed by February 2002 (see paragraph 12).

Concept of "guaranteed" results

17. The fundamental basis of twinning projects is that they must result in specific and "guaranteed" progress in candidate country's institutional and administrative capacities to adopt, to implement and to enforce the acquis communautaire. Candidate countries must establish a modern, efficient administration and institutions that are capable of applying the acquis to the same standards as present Member States. According to the Twinning Manual, candidate country organisations must be fully functioning, efficient and sustainable after twinning, although no criteria were set as to what constitutes a strong administrative capacity.

18. In order to emphasise this objective of a real, operational and sustainable outcome in a particular field, the Commission established the term "guaranteed" results. The design of a twinning project should be such that results are ensured or "guaranteed". However, in reality the Commission has no specific means to ensure that the "guaranteed" results are actually delivered. There is only the possibility of withholding 10 % of the budget until the final project report is approved by the Delegation. Even this provision has not been very effective, as projects' budgets were calculated rather generously, meaning that budget execution rates were under 90 % in many cases. Hence there are no significant financial consequences in the case of only partial success or failure to reach the "guaranteed" results. In practice, the approval of the final report is not dependent on success or failure of the outcome.

Limited achievement of ambitious objectives

19. In general, the twinning projects acted as catalyst in setting the candidate countries' reform in motion, bringing together specialists from Member States and candidate countries' administrations. Most progress was made in the adoption of the acquis communautaire through legislation. However, progress was rather less regarding its implementation and enforcement. The Court's analysis of 45 completed projects showed that it was often possible to achieve the objectives stated in the covenants only partially within the project period, which made extension or follow-up projects necessary. It was generally too optimistic to expect that a fully functioning, efficient and sustainable Candidate Country organisation would exist in a given field after one twinning project. Nevertheless, such unrealistic objectives were often stated in the twinning covenants. The findings in some major areas were that:

- in the area of "Preparation for Structural Funds" important parts of the stated objectives concerning the sound management of the Structural Funds and the Cohesion Fund in time for accession to the European Union could not be achieved (see details in the Annex),

- in the Environment sector performance has often been below expectations. Sectoral implementation plans could not be established as planned and "guaranteed" in the twinning covenants,

- other areas, such as Agriculture and Fisheries or Justice and Home Affairs, provide examples of projects which were over-ambitious in conception, and/or did not achieve the stated objectives.

Concern for sustainability

20. Many projects included large scale training of staff in candidate country administrations. However, PAAs and project leaders pointed out that many of the younger staff in particular would not stay in public service after training but move to the private sector. Despite recent improvements, employment in the civil service in the candidate countries remains relatively unattractive because of uncompetitive salaries and poor management(11). Even if the twinning projects did deliver the "guaranteed" results in such cases (i.e. the training of a defined number of staff), the long-term impact on the strengthening of the administrative capacity may not be achieved because of this high rate of turnover of staff.

21. The impact of twinning is reduced if adequate attention is not paid to improving the general framework for public administration in the candidate countries. The capacity of the civil service in most candidate countries is not yet developed to a level where twinning can be fully effective. The Commission itself pointed out in its PHARE 2000 Review that the instruments used in the PHARE programme risk being undermined by systemic failings in candidate countries national administrations. The OMAS Consortium concluded in October 2001 that the impact and sustainability of the results of twinning were being reduced because of the continuing inadequacy of candidate countries' public administration culture, systems and funding(12).

22. Sustainability could also be at risk if Member States with differences in culture and approach are involved in the same twinning project. For example, the risk was apparent in the start of the Polish "Enhancement of the internal financial control of public spending" project. It was hampered by different views on how to establish adequate internal control systems in the public sector.

Reasons for limited achievements

23. The "guaranteed" results can only be achieved if all parties perform as required. The audit showed that limited achievements are in general due to a combination of factors involving all main parties:

(a) The Member State administrations were not always equipped to cope with all the extensive requirements of the twinning rules. Resources required for coordination were underestimated, particularly for consortium projects (two or more Member States involved). In some cases, the performance of the PAA did not reach the level required, due to problems of language and integration. Also delays in recruiting PAA assistants(13) sometimes had a negative effect on the progress of twinning projects. The PAA was also not in all cases sufficiently supported by the Member State project leader and staff changes among the Member State short-term experts also hindered effective project implementation.

(b) Commitment and ownership by the Candidate Country was not in all cases sufficiently developed. The candidate countries sometimes did not sufficiently support the required reforms and changes. PAAs were not always well equipped with adequate office facilities and equipment from the beginning of the project. High-level candidate country officials were in some cases unable to devote sufficient time to project activities because of their regular work.

(c) The Commission, which has a final responsibility for twinning, sets the framework but is not directly involved in the management of twinning projects. The Commission continues to be in a process of restructuring and trying to optimise its system for sharing management between its central offices and its Delegations. Joint efforts were undertaken to advice on individual projects, however, these efforts were sometimes hindered by inefficient exchanges of views and demarcation disputes. The Commission did not always respect the preferences of the candidate country, one of the fundamental requirements of the twinning rules (partnership principle): It heavily promoted twinning even in situations when the candidate country was convinced that twinning could not offer the best solution. A further criticism by the candidate countries was that the Commission did not always make sufficient effort to counteract political pressure by Member States, with the result that the choice of the twinning partner was not entirely left to the candidate country. There was also insufficient monitoring of the quality and punctuality of project reports and unclear information flows to line DGs. This meant that the Commission was less able to propose practical and effective recommendations aimed at improving performance. In addition PAAs criticised the fact that the Commission did not set up a PAA network.

24. One overall effect of the situation described above is that efficiency and effectiveness are reduced (see paragraphs 33 to 44). However, twinning is a relatively new initiative, and all parties involved are in the process of learning from experience.

EXTERNAL EVALUATION IN 2000

25. One positive feature is that the Commission assumed its responsibility for organising this learning process by ordering an external evaluation(14) of the twinning instrument at the beginning of 2000, even if at that time none of the projects had been completed and most had been operational for less than 12 months.

26. A team of four independent experts from different Member State administrations established a "report on an assessment of the twinning instrument under PHARE" in July 2000. On the evidence from an examination of 18 twinning projects and the consultation of all the main stakeholder groups, the experts concluded that improvements could be made in three main aspects:

(a) measures to make twinning more user-friendly: delays in payments and in decision-making were criticised as well as the lengthy and complicated covenanting process, which was found to be bureaucratic, inflexible, cumbersome and too detailed;

(b) measures to make it easier for Member State administrations to participate in twinning activities;

(c) measures to help increase the commitment of the candidate country administrations to twinning.

CORRECTIVE ACTION BY THE COMMISSION AS A RESULT OF THE EXTERNAL EVALUATION

27. The Court recognises that the Commission has introduced many improvements as a result of the external evaluation. The corrective actions undertaken are reflected in several significant revisions of the Twinning Manual, the last two of which became effective in June 2001 and in February 2002 (see paragraph 43).

28. In order to improve the drafting of covenants (see paragraph 26(a)), the external evaluation recommended that the costs of preparing the covenant should be funded by the PHARE programme. The revision of June 2001 permitted this possibility and also introduced greater flexibility, notably concerning the budgetary rules as well as the introduction of "twinning light" procedures(15).

29. The latest revision of February 2002 foresees further user-friendly measures:

(a) new PAAs are invited by the Commission to attend a two-day training seminar at Commission Headquarters in Brussels, the costs for which can be charged to the twinning budget;

(b) the Member State body responsible for the implementation of the project may request a higher advance payment than was previously allowed.

30. With regard to the Member State administrations' interest in participating in twinning (see paragraph 26(b)), the Commission also allowed more and more private bodies, which act in lieu of public administrations, to be mandated to implement twinning projects. This wider definition takes into account the general trend in some Member States towards the privatisation of public activities (but see paragraphs 46 to 48) and has increased the interest of these Member States in participating in twinning.

31. In respect of the commitment of candidate country administrations (see paragraph 26(c)), the Commission took measures to ensure that Candidate Country staff input is presented in detail in the twinning covenant. The revision of the Twinning Manual in June 2001 introduced the obligation to ask for prior approval from both the Commission and the CFCU if a change in candidate country project leader or PAA counterpart is intended.

32. In summary, many start-up or teething problems have been identified and addressed by the Commission. However, major problems of a more structural nature still persist, which are described in the following paragraphs.

CONTINUING STRUCTURAL PROBLEMS

Administrative complexity

33. Many new PAAs who started work in 2001 were still expressing frustration with the extent of the bureaucracy in all stages, from project programming to project implementation. In their opinion too much of their time was spent on purely administrative issues to the detriment of the main task, namely advising candidate countries on institution-building. Indeed, the interaction of the numerous public administrations involved in twinning (the Commission in Brussels, Commission Delegations, various public bodies in Member States and candidate countries) each with its own administrative culture creates administrative complexity that diminishes efficiency and effectiveness. The continuing overly long periods between needs assessment and project realisation (see paragraphs 34 to 36), as well as the highly complicated payment systems (see paragraphs 37 to 44), are two examples of such difficulties.

Inefficient procedures from needs assessment up to project realisation

34. Before the PAA can start his work in a candidate country, a project needs to pass through many different stages (see Diagram 1). This results in long periods between the initial needs assessment and the subsequent practical action to meet those needs. 30 % of all projects in the 2000 selection round had not yet been started in February 2002. It took two years in these cases to complete the various phases (needs identification, successive drafting of the project outline, call for proposals, selection of twinning partners, drafting of the twinning covenant, approval of financing by the Commission's Steering Committee, signature of the twinning covenant, endorsement by the Commission, notification of the twinning covenant and, finally, starting the PAA's work in the candidate country). As a result, some elements in the project design are often already outdated by the time the project is able to start so that changes to the covenant become necessary from the very beginning. The real risk of the withdrawal of the designated PAA and other key personnel before the start of the project is a further serious consequence of the lengthy procedures.

35. Many PAAs criticised the excessive precision required in the planning process. They considered a waste of resources and time the need to specify exactly, many months in advance, such matters as the location of a training event, the number of persons attending, room hire, travel costs, per diem allowances or the number of pages to be translated for handouts. In the event of any deviation from the plan, time-consuming procedures have to be followed to amend the covenant. If the change concerns a mandatory clause (e.g. an article on "guaranteed" results) it takes the form of an addendum to the covenant(16), which requires formal approval from headquarters. Side letters have to be drawn up for changes to indicative clauses of the covenant. In the past many addenda or side letters were often made during the lifetime of a project.

36. In Poland the precise planning of all cost elements for twinning projects were made redundant in July 2001 by a 60 % increase in the per diem rates for Member State experts (from 229 euro to 367 euro). This high increase of per diem rates within the lifetime of twinning projects led to serious disruptions in their budgetary management. The Polish National Fund was of the opinion that the increase was not justified and not in line with the economic reality.

Complicated payment systems

37. Originally, payments for twinning to reimburse Member States' costs were made directly by the Commission to Member States. Responsibility for payments was then transferred to the candidate countries within the Commissions' general policy of decentralisation (increased responsibility for candidate countries). Thus the main money flow is now from Brussels to the candidate countries and then from the candidate countries to the Member States. The PAA's allowances and expenses are paid directly to them by the CFCU in the Candidate Country, which is the paying agent for the EU contribution to twinning projects.

38. Member State administrations still complain that the reimbursement of their costs often takes months due to failures in the payment systems. PAAs also experienced considerable delays in receiving payments. In some cases advances were not received in time so they had to use their own funds at the start of projects to pay for flights, accommodation costs or for their assistant.

39. The overall system has not functioned properly because of the overly complex interrelated procedures. The coexistence of the twinning financial rules with other financial or budgetary rules at the different levels of the Commission, Member States and candidate countries causes major dysfunctions.

40. PHARE funds in Hungary are administered through the central budget. The payment procedure at the Hungarian treasury foresees 27 steps, which resulted in considerable delays in PHARE payments being made, including twinning payments. The Delegation in this country set up a task force in 2001 to tackle these payment problems.

41. The CFCUs have to check whether cost statements submitted by the Member State twinning partners are accurate and conform to the twinning rules. The Polish CFCU in particular was reproached by Member State twinning partners for carrying out this control in an excessively strict way, thereby causing payment delays. The Polish CFCU retorted that their checks showed that many cost statements were inaccurate. Different interpretations of the twinning rules were one cause of the problems. The Court found evidence of shortcomings on both sides.

42. Preparatory work performed in the Member States is not compensated on the basis of real costs. Instead a flat rate compensation is given, calculated on the basis of the work performed in the candidate country. Irrespective of the nature of the project or of the amount of preparation necessary, this flat rate is always 150 % of the fees paid for the work of short-term experts in the candidate country. Although it is proper that Member States should be compensated for essential preparatory work carried out by experts in the Member State, the Commission has not reviewed the justification of the 150 % rate since its introduction in 2000. The external evaluation points out that the flat rate compensation system creates perverse incentives. Since it is mainly up to the project management in the Member State to assess how much time is to be spent in the candidate country, it encourages the experts to do preparatory work in the candidate country instead of in the Member State.

43. The application of the twinning financial rules presents a challenge for the administrative capacities in both Member States and candidate countries. The Commission changed the rules several times in order to improve the instrument. However, this also increased complexity, as different sets of rules had to be applied for old and new projects. Further confusion was created because some changes (clarifications) also applied retroactively(17), whereas others did not. Coherent advice was not always given by the Commission, with the result that the twinning financial rules were not always applied in a consistent way(18).

44. In general, complicated administrative procedures for twinning take up a disproportionate amount of resources of the Delegations, the CFCUs and the PAA's.

Over-emphasis on twinning

45. Institution-building is not equivalent to twinning. However, some representatives of the candidate countries criticised the Commission's tendency to over-emphasise twinning at the expense of other mechanisms directed towards institution-building that are eligible for support (general horizontal support to public service(19) (see paragraph 21), private sector service contracts, participation in Community programmes). The strong emphasis put on twinning by the Commission sometimes resulted in a departure from its original aim and in an insufficiently selective use of it.

Extended use of twinning for mandated bodies

46. Twinning is meant to represent cooperation between public administrations. The PAA should have a specific knowledge of implementing the acquis communautaire in the public sector administration of a Member State which is not available in the private sector. According to the Commission, twinning should offer something that no other tool can, namely help in those areas that are the exclusive preserve of Government. Twinning should be used in cases where no alternative solution by using private consultants exists. That is the reason why a derogation from standard PHARE tendering procedures is made for twinning, allowing the candidate country to make its choice of Member State twinning partners irrespective of cost.

47. However, more and more Member State twinning partners are so-called mandated bodies which are defined as bodies entitled to act in lieu of public administrations. The list of mandated bodies accepted by the Commission has grown considerably to more than 300. In this way the Commission has recognised the Member States' tendency to entrust public services to private bodies. There was no formal decision by the Commission to accept or refuse an organisation as mandated body.

48. These mandated bodies are often privately owned and they may also have a commercial interest, leading them to offering their services via commercial calls for tenders on other occasions (including to the Commission in the framework of its technical assistance). The status of a mandated body allows it to charge over 100 % more than the standard fees for experts from public administrations. However, for twinning they are not required to submit to competitive tendering procedures. In these cases the original idea of twinning is lost and it is questionable whether the derogation from the standard PHARE procedures is appropriate.

Automatic use of twinning

49. The Court found the use of twinning to often be a default choice rather than the result of a detailed evaluation between different instruments. Enough consideration was not always given to whether the conditions for a successful implementation of twinning had been fulfilled. Experience shows that twinning activities are not well suited to projects where the candidate country has not yet decided which ministries should have responsibilities for the activities or where the candidate country has not yet adopted the necessary legislation. Sometimes twinning was chosen even if there was no obvious need for an uninterrupted PAA residence of at least 12 consecutive months (a basic requirement of the twinning instrument).

50. There were also cases where, after a lot of preparatory work, the selected twinning projects failed to start and were cancelled. The Commission then resorted to the use of private contractors. This shows that twinning and private consultancy were in practice sometimes interchangeable. Not enough thought had been given in each case, to whether the expertise required could best be provided by twinning or by private operators.

51. Twinning cannot stand on its own. In some cases there was a lack of complementarity and synchronisation with private sector input. Twinning could not progress because linked tendering, procurement and contracting procedures for equipment were uncoordinated or delayed. Two examples illustrate the uncoordinated deployment of different means for institution-building:

(a) in the Polish "Internal financial control at regional level" project (project duration: October 2001 to October 2003) the corresponding tender for a database was delayed. The database was to cover EU structural interventions, with particular emphasis on the development of a system for monitoring community support at regional and local level;

(b) the twinning component of the Hungary project in the area of Structural Fund preparation (see the Annex, paragraph 4) was not coordinated with the private consultancy component thus increasing the risk of overlaps and the unnecessary duplication of efforts.

52. In general it may be argued that twinning is not necessarily more cost-effective than technical assistance contracts following tendering procedures. The following factors hamper the achievement of optimum cost-effectiveness of twinning:

(a) although Member States and candidate countries are the contracting parties, the funds are provided by the Commission. Therefore Member States or candidate countries have no financial incentive to end the contract if one of the parties does not fulfil its obligations. In some cases poorly performing contracts have been continued with in order to avoid political problems and final payments have been made despite results not being achieved;

(b) due to the number of functions and participants involved in a project, accountability for shortcomings is diffuse (see paragraphs 23 and 24), which increases the risk of failure or underperformance;

(c) twinning projects involve many administrations at different levels, which means that resources are devoted to bureaucratic matters to the detriment of the main task, namely advising candidate countries on institution-building (see paragraphs 33 to 44);

(d) an uninterrupted PAA residence of 12 months is not necessary or desirable for every project (see paragraph 49). Quick, flexible and tailor-made solutions to specific problems are very difficult to fit into the twinning framework. However the "twinning light" procedures have introduced more flexibility (see paragraph 28).

TWINNING AND NEW ACTION PLANS 2002

53. The Commission announced in its 2001 Strategy Paper on Enlargement that it would reinforce the institution-building actions in 2002 by an Action Plan for each of the 10 candidate countries of central and eastern Europe. In June 2002 the Commission reported(20) that these Action Plans were complete and implementation was underway. A 250 million euro supplementary institution-building facility under the PHARE programme is provided.

54. The Action Plans are now considered by the Commission to be a major instrument and a key tool in helping the candidate countries to reinforce their administrative and judicial capacities. Exactly what role twinning should play within this supplementary assistance is still to be decided.

55. The 2002 regular reports from the Commission on the candidate countries' progress towards accession identified the needs still remaining in terms of institution-building before accession. Under the current mechanism new twinning projects could not contribute to addressing these needs before 2005. The twinning mechanism is therefore certainly not an instrument for responding immediately to institution-building needs.

CONCLUSIONS AND RECOMMENDATIONS

A positive initiative

56. The strengthening of administrative capacity (institution-building) became a priority of PHARE support in November 1998. A strong administrative capacity is vital for the candidate countries of central and eastern Europe, not only because of the specific needs of EU accession but also for the general transition towards increased competitiveness in the global economy. Twinning is the main PHARE instrument for strengthening administrative capacity in the context of accession.

57. Member State administrations and public institutions have unique knowledge and specific experience concerning the implementation and enforcement of the acquis communautaire. Before 1998 the Commission mobilised advice mainly from the private sector. However, best advice from the private and public sectors is needed for the institution-building process, so twinning is a laudable initiative by the Commission to assist candidate countries in acquiring the capacity to adopt, implement and enforce the acquis.

58. The Commission has a final responsibility for the twinning programme but it is not directly involved in the twinning projects and thus has limited possibilities to influence them. The concept of twinning, although positive overall, implies that the shared responsibility between the Commission, candidate countries and Member States hampers a clear and straightforward accountability structure (see paragraphs 23 and 24).

Limited achievements of ambitious objectives

59. In general, the twinning projects acted as catalyst in setting the candidate countries' reform in motion, bringing together specialists from Member States and candidate countries' administrations. Most progress was made in the adoption of the acquis communautaire through legislation. However, progress was rather less regarding its implementation and enforcement, The Court's analysis of completed projects indicates that the achievements have in general fallen short of the stated objectives. The so-called "guaranteed results" could often be achieved only partially within the project period. It was too optimistic to expect that a fully functioning, efficient and sustainable candidate country organisation would exist in a given field after one twinning project, the average duration of which is 18 months. Success was often jeopardised by the multitude of functions and participants in the process and a lack of accountability (see paragraphs 19 to 24). The effect of administrative complexity also hampered the full achievement of the desired results. Too much time was spent on purely administrative issues, to the detriment of the main task, namely advising candidate countries on institution-building (see paragraphs 33 to 44).

Improving the delivery

60. To increase the value for money and to improve the achievement of results twinning should be more focused on the timely delivery of results. The capacity to deliver could be strengthened in the following ways:

(a) greater attention should be given to the formulation of a limited number of realistic and achievable objectives, which clearly address identified weaknesses in administrative capacity;

(b) the procedures for assessing the achievement of results should always be fixed in each covenant, possibly through the use of performance indicators or benchmarks;

(c) the covenants should also stipulate that the sustainability of the results must be evaluated after the completion of a project. The term "guaranteed" is misleading as long as payments are not tied to the achievement of these targets and it is recommended that the use of the word "guaranteed" should be discontinued.

(d) the Delegations should monitor the achievement of project benchmarks more promptly, in order to allow timely action to be taken, such as reorientation of the project or even its termination.

Making twinning quicker and less complex

61. Twinning involved lengthy procedures and was too much occupied with the micro-management of inputs. 30 % of all projects in the 2000 selection round had not yet been started in February 2002 (see paragraph 34). The Commission has already demonstrated its readiness to take corrective action (see paragraphs 27 to 32). However, further improvements are still necessary to simplify the lengthy procedures at the various stages from needs assessment up to project finalisation:

(a) all stages of project preparation (involving the candidate countries and Member States as well as the Commission) should be rationalised without compromising the respect of sound administrative practice;

(b) the Commission should speed up its own internal consultation procedures between Headquarters, Delegations and line DGs by respecting pre-set deadlines;

(c) the Commission should simplify and accelerate the payment procedures without disregarding the risks involved;

(d) in general, the Commission should consider the use of fixed-price or lump sum contracts requiring specified deliverables.

Choosing twinning more selectively

62. Institution-building is not only twinning. The use of twinning is too often the default choice when considering institution-building (see paragraphs 45 to 51). The Commission should increase its efforts to develop a coordinated and balanced deployment of different instruments for institution-building (general horizontal support to the public service (see paragraph 21), private sector input, participation in Community programmes).

63. For cases where mandated bodies are acting as possible twinning partners, the Commission should establish a procedure enabling it to pay more attention to the costs involved (see paragraphs 46 to 48).

Building on knowledge and experience

64. Four years of twinning have created a store of specific knowledge and relationships. PAAs from the first projects are leaving and new PAAs are arriving. The Commission should create a PAA network (perhaps via the Internet) to preserve the store of specific knowledge, to spread good practice and to reduce the risk of errors recurring.

Outlook

65. Twinning will remain relevant far beyond the envisaged accession. Projects starting just before accession will last for up to two more years. The normal PHARE support will have to be continued for candidate countries which are not part of the first wave of accession. In addition the Commission foresees the funding of a transition facility for institution-building (available to new Member States after accession) covering actions not eligible for the Structural Funds. Improvements to the instrument will therefore be worthwhile because twinning is likely to continue for a considerable time.

This report was adopted by the Court of Auditors in Luxembourg at its meeting of 10 April 2003.

For the Court of Auditors

Juan Manuel Fabra Vallés

President

(1) The entire body of EU law as expressed in the treaties, in the secondary legislation and policies of the Union, as well as in the jurisprudence of the European Court of Justice.

(2) The Commission established the term "guaranteed" results in order to underline the objective of a concrete and fully operational outcome in a particular field. However, there are no financial consequences in the case of only partial success or failure in reaching the "guaranteed" results.

(3) The Commission has established the rule that Member State experts should be statutory civil servants or acting civil servants (Twinning Manual, part A, Article 5.3.2). However, certain mandated non-administrative bodies, usually those to which a Member State has delegated its powers, may also provide twinners.

(4) See footnote 1.

(5) PAAs assist the administration or other government bodies in the Candidate Country in the context of a pre-determined work programme.

(6) The latest version is dated February 2002.

(7) Special Report No 5/2003 on PHARE and ISPA environmental programmes (see page 1 of this Official Journal).

(8) The OMAS Consortium was charged by the Commission to assess the work in progress of the EU PHARE programmes.

(9) Each candidate country and Member State has a NCP for institution-building activities.

(10) SIGMA (Support for Improvement in Governance and Management in Central and Eastern European Countries) is a joint initiative of the OECD and the EU and is principally financed by the EU's PHARE programme.

(11) See SIGMA publications on fulfilment of civil service baseline in candidate countries.

(12) Ad hoc Report on the Twinning Instrument, OMAS Consortium, 24 October 2001.

(13) To facilitate the work of the PAA, a full time assistant (candidate country national) may be provided, to deal with translation/interpretation problem and other project management tasks.

(14) Report on an assessment of the twinning instrument under PHARE, July 2000, European Commission.

(15) "Twinning light" was introduced in June 2001 for projects of limited scope. The financial ceiling for "twinning light" projects has been set at 150000 euro and their duration limited to six months; in exceptional cases this can be extended to eight months.

(16) The Delegation is entitled to agree or refuse endorsement of addenda on behalf of the Commission. It must request formal approval from headquarters before agreeing addenda concerning a change e.g. of the "guaranteed" results. If the Commission agrees, the Delegation also has to request the written approval of the CFCU.

(17) With each change the Commission indicated that "clarifications of implementation rules may apply to earlier covenants, where appropriate, or be subject to prior agreement with the CFCU".

(18) E.g. the revision of the Manual in June 2001 introduced maximum rates for Member State Project Leaders and short/medium term experts. According to a Member State administration these maximum rates would sometimes be respected and sometimes not in projects from the 2000 selection round, depending on the advice given by different Delegations.

(19) Horizontal support to public service is given by SIGMA and TAIEX (Technical Assistance Information Exchange Office). TAIEX is a Commission office and part of DG ELARG.

(20) Communication from the Commission on the Action Plans for administrative and judicial capacity, and the monitoring of commitments made by the negotiating countries in the accession negotiations, COM(2002) 256 final of 5.6.2002.

ANNEX

LIMITED ACHIEVEMENT OF AMBITIOUS OBJECTIVES

1. Twinning brought together specialists from Member States' and candidate countries' administrations, creating new links of technical expertise throughout Europe and initiating first steps to strengthen the administrative capacity in the candidate countries. Most progress was achieved concerning the alignment of legislation with the acquis communautaire, of which about 50 % concerns the agriculture sector. In this sector, twinning pushed forward the transposition and adoption of legislation in subjects such as "Integrated administration and control systems" (IACS), "Paying agencies" and "Veterinary administration". In other areas too, twinning contributed to the strengthening of the legal and institutional framework, notably in the fields of "Border control and management", "External audit" and "Migration and asylum".

2. However, the following examples show that many objectives stated in the twinning covenants were unrealistic and could only be partially achieved.

Example: Preparation for Structural Funds

3. The "Special preparatory programmes for Structural Funds" (SPP) in the candidate countries usually include twinning arrangements. For example, the 1998 PHARE National Programme for Estonia allocated an amount of 3 million euro to the SPP, of which 21 % was for a twinning project and the remaining 79 % for non-twinning measures. Both the non-twinning and twinning measures were intended to contribute to the general objective of the SPP, namely the development and implementation of the necessary structures within the Estonian administration for the sound management of the Structural Funds and the Cohesion Fund in time for accession to the European Union.

4. The main PAA concluded that the objectives set for the twinning project were unrealistic and that the benchmarks were insufficiently precise. Indeed the covenant describes in 30 pages a confused list of around 70 often imprecise objectives, goals, contents, tasks, expected results, specific purposes, benchmarks and activities. This makes the subsequent evaluation of results extremely difficult. As a consequence, the final report of the twinning project did not systematically compare the results with the set of objectives and outputs described in the covenant and nor did it distinguish between those results due to twinning and those due to technical assistance.

5. The Hungarian SPP twinning project (2,6 million euro) was also very ambitious in content and time allocation. Out of the 40 objectives or results in the four finalised components, only 15 were described as fully completed by the Delegation. Achievement was especially weak in the "European Social Fund" and "Sapard"(1) components. The purpose of the Sapard component was to enable the Ministry of Agriculture and Rural Development (MARD) to administer the Sapard pre-accession measure. At the end of the project (January 2001 for the Sapard component), the Sapard Agency had not yet been provisionally accredited and a large amount of work was still required before this could be achieved. This was due primarily to the decision to move the Agency from the Agricultural Intervention Centre to the MARD mid-way through the project. As such only half of the stated objectives were achieved in the Sapard component.

6. The SPP project in Poland (6,75 million euro for phases one and two) was not only very ambitious but also extremely complex (eight PAAs, six Member States, 16 actions covered by three components). Some of the project objectives, mainly in the second component comprising the establishment of a legal and administrative basis for Structural Funds/Cohesion Fund including ISPA and Sapard, were not achieved. This was in part because of this complexity. In particular the procedures, administrative capacities and the role of organisations at regional level were not clarified.

7. Irregularities committed in the Agency for Restructuring and Modernisation of Agriculture (ARMA) had a strongly adverse effect on the achievement of the objectives in all two agricultural-related twinning projects(2). An external audit (October 2001) as well as investigations by OLAF and by the Polish Supreme Chamber of Audit (February 2002) reported major incidents of mismanagement in ARMA: EU-funded equipment and staff trained under the twinning project were transferred from the public administration to a private company. The Commission issued a recovery order of 2,8 million euro in February 2002. Because the PAAs only operated at a relatively low level within ARMA, they were not aware of this transfer to a private company.

8. In the SPP project in Lithuania some benchmarks were deleted, namely those related to the testing of the established administrative systems and procedures. It became clear during the implementation period that all the necessary structures to implement the pre-accession funds could not be in place before the end of the project.

9. The "guaranteed" results could to a large extent not be achieved for the SPP/1 projects in Bulgaria (according to the Commission, due to a lack of commitment on the part of the beneficiary, an inexperienced PAA and inadequate back-up from home administration) and in Latvia (very poor working relations between partners and weak project management).

(1) Special action programme for agriculture and rural development, the pre-accession instrument for this sector.

(2) PL98/IB/AG-01 (Introduction of integrated administration and control system) and PL98/IB/AG-03 (Animal identification and registration system).

The commission's replies

SUMMARY

I-II. The Commission takes note of the Court's positive and constructive appraisal of twinning as a positive initiative for institution-building. Twinning was created as a new instrument for targeted international/European administrative cooperation. Prior to its launch, assistance in ublic administrative reform in the candidate countries was delivered through consultancy contracts, which neither drew on practical "hands-on" experience nor resulted in effective transfer of practical expertise. Very often, these contracts merely yielded theoretical reports, devoid of practical relevance. Twinning thus grew as from 1998 as an experiment. All this underscores the need to analyse twinning (and more especially its key aspects: targeted programming and implementation through public sector experts) in relation to the alternative of traditional technical assistance.

III. The concept of "guaranteed results" was the key feature of twinning. Both project partners commit themselves to work towards a commonly agreed result in a joint project implementation process. The Commission indicated from the outset that twinning projects should focus on limited and well-defined institutional targets. This realistic approach, however, conflicted at times with the high level of ambition of the beneficiary countries. It is not, therefore, disputed that the earlier projects in particular were over-ambitious, and at times failed to achieve their targets. However, even these ambitious projects acted as a catalyst in setting the reform process in motion, and compared to the results achieved by the technical assistance-based approach from 1990-1996, twinning has a good track record.

Finally, the positive feedback from administrations involvedin twinning was essential in building Member State's confidence and secured closure of the negotiations in Copenhagen. This widely acclaimed success was the ultimate guaranteed result delivered by twinning.

IV. Two of the most important, non-measurable but visible results of twinning are network building and change of attitudes and behaviour. During those four years of implementation, twinning brought together specialists from Member States' and candidate countries' administrations who were seldom international experts, creating a network of technical expertise throughout Europe. This is a fundamental process for building an enlarged Europe and for bringing Europe closer to its citizens. Most of the Member State experts and the candidate country civil servants were involved for the first time in an EU project and had for the first time the responsibility of managing EU funds. In many of the cases, the links formed continued after the closing of the respective project, both between Member States and candidate countries as well as between different Member State administrations.

V. In 1998 there was insufficient knowledge about the specific rules and principles in twinning with any of the actors involved (be it EC Delegations, Member States or Candidate Countries). Subsequently twinning project design and implementation evolved through a process of "learning by doing". The Twinning Manual embodies the evolution of a practice, which is now well settled. The aim of all amendments to the Twinning Manual has indeed been to maintain twinning as a pragmatic instrument and to simplify procedures as much as possible.

VI. Bringing experts from different countries and administrative cultures, with no previous experience in international cooperation, to work together efficiently has been and is a complex and difficult task. Many of the rules governing twinning projects have been introduced to ensure proper spending of PHARE funds allotted to twinning. The Commission has also repeatedly insisted with the twinning task managers in Delegations and headquarters to maintain a flexible yet sound approach during the implementation of the projects. Great efforts have been made to speed up payments. More selective use of twinning can be addressed at the programming stage. The Commission refers in this respect to Annex 3 of the annual Phare Programming Guidelines entirely devoted to adequate and targeted programming of Twinning and Twinning light.

VII. Through twinning the Commission pioneered an unprecedented transfer of know-how on complex and technical issues from the diverse administrative traditions in Member States to the equally diverse candidate countries. It was also essential in mobilising political will and resources in the candidate countries.

- Rate of delivery: while the recommendation of the Court is supported, twinning projects are not assimilated to supply or investment agreements. Meeting the twinning objectives is often dependent upon external, political factors.

- Time frame of stages in the project: the Twinning Manual now provides for a general deadline of six months between selection ands start of the project. All other procedural deadlines, including those set for twinning steering committee proceedings do not exceed 10 working days.

- The Commission has already taken steps to ensure simplification of payment procedures by increasing the amount of advance payments

- The recommendation to use fixed-price or lump sum contracts requiring specified deliverables would not be practical in all situations since the financing of Twinning projects is based on recovery of expenses incurred by administrations or mandated bodies.

- The identification and design of institution-building twinning projects is the result of a well balanced process and dialogue between the Commission and the candidate country under the final scrutiny and approval of the PHARE Management Committee. Proper programming is of the utmost importance and this has been duly stressed in the PHARE Programming guidelines at least since 2000.

- PAA network: the Commission welcomes this suggestion and is prepared to refine its data base where all the professional data and contacts of PAAs are already available so that a network could be set up.

INTRODUCTION

4. The pre-accession advisers (PAAs) are the "core" of the twinning projects, but one should also highlight that both project leaders (one from the Member State and one from the candidate country) are also key actors for the success of a twinning project.

6. The Commission designs the twinning instrument, provides financing, advice and ensures monitoring and evaluation but the project implementation lies exclusively with the project partners. This highlights two essential features of twinning (i) its interlinkage of cooperation between Member State and candidate country administrations and the (ii) the traditional role of the Commission as facilitator and coordinator.

7. The Twinning Manual has been widely recognised as a very useful record of good twinning practices which provides steady yet flexible guidelines for effective project management.

ACHIEVEMENT OF RESULTS

16. The Commission realises that, especially in the start up period considered in the sample of projects examined by the Court, some twinning projects require considerable preparatory work. Nevertheless, in Poland, for example, eight out of nine twinning covenants were endorsed before the end of October 1999. The one remaining covenant was endorsed in February 2000.

The latest revision of the Twinning Manual in February 2002 led to a swifter covenanting process, which has considerably shortened the effective start of twinning projects.

17-22. More than one twinning project can indeed be required to reach proper implementation of the Community acquis in a given field or sector. The "guaranteed result" can therefore easily be an intermediate benchmark, which constitutes a specific criterion in relation to administrative capacity, as long as there is a jointly agreed target. This target must be measurable and precise.

18. The implementation of a twinning project is based on the recovery of costs incurred.

The amount of funds which can potentially be withheld following a refusal of the final report has been increased to 20 % (February 2002 revision of the Twinning Manual). Non-approval of the final report would in fact only be a penalty for the Member State involved in the twinning project whereas the implementation of the project is a joint responsibility, together with the authorities of the candidate country. The consequences of failure for the Candidate Country are more indirect. If a twinning project failed to set up institutions crucial for the implementation of the Community acquis or failed to incorporate parts of the acquis into national legislation, this was usually reflected in the Regular Reports on progress made by the candidate country and therefore had consequences in the accession negotiations.

Furthermore, the Twinning Manual contains provisions to reorient the project (especially after the first six months) or to halt a project before finalisation in case of ineffectiveness or other failure. This last alternative is something both candidate countries and Member States endeavour to prevent at all cost because this could reflect badly on their perceived commitment to the enlargement process. Indeed, the Commission decided to halt two projects before their end date (BG99/IB-CO-02 National Health Insurance Fund and BG99/IB-CO-01 Coordination of social security schemes)

19. In particular during the start-up period, the beneficiaries undertook important commitments in the negotiations, the size and importance of which were reflected in some very ambitious and sometimes over-ambitious twinning covenants. Hence it is not disputed that especially the earlier projects were over-ambitious in design. Nevertheless, even these ambitious projects acted as a catalyst in setting the reform process in motion.

- The programming of Structural Funds projects is in general a more complex exercise than for other "acquis-related" projects, as it covers "practices" rather than specific Directives or regulations. Furthermore the design of a Structural Funds project fiche involves very detailed horizontal consultations. Once the implementation of a project has been awarded to a Member State, it is up to the beneficiary and the selected Member State partner to set-up the proper activities/strategy, which will enable the achievement of guaranteed results. Reference is made to the Commission's detailed response to the Annex.

- Furthermore, institution-building is a particular challenge in the environment sector because Ministries of Environment are relatively new, while regional and local authorities as well as environmental agencies also require assistance to meet their responsibilities in relation to the acquis. Coordination between the Ministry of Environment and these other institutions and coordination between the Ministry and other ministries is often very difficult and has to be developed.

20 and 21. The problem of sustainability is not specific to twinning; it derives from the general environment of public administration in the candidate countries with high turn over of personnel, low salaries and sometimes poor management. The acute human resource constraints in public administration can only be fully overcome through growth and development, underpinned by the economic security and solidarity provided through membership. Nonetheless, the Commission does not overlook the general environment of public administration, which is targeted through SIGMA Programme (Support for Governance and Management in PHARE Candidate Countries).

22. There is also a very positive side to the involvement of more than one Member State in a given twinning project as such involvement presents the Candidate Country with different alternatives rather than with just one national model. In fact the different approaches often enrich the project.

In the case of the specific twinning project referred to by the Court, the Polish beneficiaries were in the end very satisfied that the Member State partners applied a mixture of Swedish and French approaches. In order to safeguard an efficient project management, the Commission has nevertheless introduced limits on both the numbers of Members States (maximum two Member States) involved in the same project and on the size of the projects.

23 (a) During the start-up phase, the Member State' administrations had to learn about this new instrument and to adjust their own structures so as to provide efficient assistance. With time, the necessary adjustments have been made and most Member State administrations have indeed become very efficient in setting up and managing twinning projects.

According to the latest revision of the Twinning Manual (February 2002) the assistant can be selected before the notification of the covenant whereas the service contract shall be signed with the selected assistant after notification. This innovation is one of the attempts to shorten the duration of administrative procedures.

(b) The commitment and ownership by the candidate countries improved over time. The required joint signature of all quarterly reports is one of the requirements to increase this commitment.

The EC Delegation delayed the implementation of some projects for some situations, where no visible signs of adequate commitment on the beneficiary side were manifest.

(c) Re: Role of the Commission

The identification and design of institution-building twinning projects is the result of a well balanced process and dialogue between the Commission and the candidate country under the final scrutiny and approval of the PHARE Management Committee.

The Commission designs the twinning instrument, provides financing, advice and ensures monitoring and evaluation but the project implementation lies exclusively with the project partners. (See reply to paragraph 6).

Re: Political pressure in the selection process

Delegations convey the same message: political pressure in the selection process is experienced from the side of some Member State administrations. The Commission has actively discouraged such pressure.

Re: Quality of reports

Delegations frequently write to twinning project leaders about the quality of the quarterly reports and the timing of their delivery with recommendations for improvement. The Commission supports this by asking the Delegations to add their own short appraisal on each report.

Re: Transparency of selection

In order to avoid lack of transparency during the selection procedure, the Commission insists that the beneficiaries fill in a standard table showing the advantages and shortcomings of each of the Member State proposals. This table is then sent to the unsuccessful Member States explaining why they have not been selected. As the beneficiary institutions know in advance that their "evaluation table" will be disclosed to the competing Member States, their evaluation/justification is therefore meant to be objective and detailed.

24. See the replies to paragraphs 33 and 34.

EXTERNAL EVALUATION IN 2000

26. The Commission has duly acted upon the recommendations of this report

(a) In the 2001 and 2002 revisions of the Twinning Manual, the Commission introduced more substantial advance payments (Point 7.2 of the Manual), an overall limit on the time needed for drafting of covenant (Point 3.6 of the Manual), introduction of audit certificate to limit paperwork (Point 7.3.2 of the Manual).

(b) Possibility to recover preparatory costs from the twinning budget (Point 5.2 of the Manual) and clearer financial arrangements for PAA training (Point 5.2.2 of the Manual).

(c) Information sessions on twinning have been organised in most candidate countries. The Court fully recognises that corrective measures have been taken as mentioned in the paragraphs below.

CORRECTIVE ACTION BY THE COMMISSION AS A RESULT OF THE EXTERNAL EVALUATION

29 (b) A major step taken to make twinning more user-friendly was indeed the change of the payment procedures in February 2002 whereby the Member State twinning partner may henceforth request an advance payment proportionate to nine months of project implementation funds from the CFCU.

30. The reasoning behind assimilating certain non-administrative bodies to administrations is that some Member States have or are in the process of out-sourcing and privatising parts of their administration. The know-how required for twinning projects is therefore sometimes located outside the administration proper.

Since the very beginning the Commission has set four qualifying criteria for mandated bodies: 1. proven competence in a field of the acquis communautaire, 2. non-profit organisation, non-commercial business purpose 3. public ownership and 4. under the supervision of an administrative body. In some cases where criterion 2 and/or 3 are only partially fulfilled, the mandate is restricted subject to an exclusion from commercial tenders in the direct follow-up to the twinning project.

The Commission has consistently applied these criteria with the necessary transparency and objectivity.

CONTINUING STRUCTURAL PROBLEMS

33. The observations of some PAAs should be counterbalanced by the fact that some of them do not have sufficient experience in project management as such and that they sometimes lack adequate backup from their home administrations. Most of the Member State experts and the candidate country civil servants were involved for the first time in a twinning project and had for the first time the responsibility of managing EU funds.

Bringing experts from different countries and administrative cultures, with no previous experience in international cooperation, to work together efficiently is in general a complex and difficult task. Many of the rules have been introduced to guarantee proper spending of Community funds.

34. The latest revision of the Twinning Manual imposes an overall deadline of six months for the negotiation, submission and approval of the twinning covenant by the selected project partners. As a consequence, the covenanting process has been considerably speeded up during the 2002 twinning exercise.

In comparison with twinning, start-up delays are usually longer for technical assistance or investment projects. Sometimes, their cause is external (lack of political will, lack of proper conditions for starting up the project, bad planning, elections in the candidate country, change of staff, etc). The Commission took steps to avoid delays:

- Poland and Romania for instance decided to build a deadline in the Financial Memorandum (three months for the partners to come up with an advanced agreed covenant),

- the preparation costs are paid only for a limited amount of time, and

- EC Delegations got involved in supporting the partners in the drafting process, organising meetings and providing detailed comments and guidance.

There has been considerable progress since the 2001 twinning projects.

35. The requirement of a detailed work plan proved to be the only way to make partners think ahead and define their project concept and methodology. The absence of such jointly negotiated work plan would result in uncertainties and possibly disagreements during the project implementation.

The twinning manual provides for a flexible structure to amend approved covenants. As in all contractual relationships, proposed amendments have to be notified beforehand. This is a logical requirement because these amendments very often encompass an adjustment in the expenditure pattern of Community funds.

36. In the case of Poland, the increase of the per diem rates has indeed created general unease. It has therefore been decided that the per diem rates set for this candidate country would be reduced from September mid 2002.

38. Reimbursement of expenses incurred follows the time pattern of spending itself. This logically explains that reimbursements are spread out over time. The reverse side of slow payments by CFCUs is sometimes the unsatisfactory presentation of invoices submitted by the Member State administrations.

The Commission advises PAAs to wait until they receive the advance before taking up their posts.

39. The Commission wishes to stress again:

- that many of the rules have been introduced to ensure proper spending of PHARE funds,

- that the projects examined belong to the very first series of projects and that the Commission has since then taken corrective measures, as acknowledged by the Court.

Twinning evolved as a totally new instrument of targeted administrative cooperation to enhance the preparation of enlargement.

40. The Delegation in Budapest has indeed set up a task force in 2001 to tackle the payment problems. As a result of the advice of the task force the Office of the National Authorising Officer in Hungary prepared guidelines for the ministries and other parties involved in twinning payment procedures to streamline the necessary steps to be taken for smooth payment execution. Training and consultation days were organised on the same topic. If further problems occur, the task force can be convened at any time.

41. From the beginning the Twinning Manual has aimed to set clear and consistent rules for the concrete implementation of twinning projects. Efforts to streamline implementation have been upheld: as an example, the Commission regularly issues short and practical guidelines under the format of twinning news sent by e-mail to all Delegations.

42. The allocation of flat rate compensation for experts was introduced to alleviate administrative burden of checking the costs of preparatory work conducted in the Member State. The flat rate compensation is justified in comparison with expense structure of technical assistance contracts. One should also consider that the flat rate fee constitutes an element of flexibility, simplifying procedures. The amount of the flat rate (150 %) was based on a factual analysis of the proportion of expert fees that were previously charged for preparatory work in the Member State and of the average amount charged for overhead costs in relation with expert work on the project.

43. The amendments to the manual have always aimed at improving clarity. The manual specifies to which covenants amended rules should apply but in order to maintain flexibility the manual indeed states that "clarifications of implementation rules may apply to earlier covenants, where appropriate or be subject to prior agreement with CFCU". The aim of these rules is precisely to ensure a uniform yet responsive application of the Twinning rules.

44. PHARE funds allotted to twinning projects are contracted without a public tendering process and therefore without a process of price determination via a competitive process. A clear set of administrative rules to fix the prices for public sector services has thus to be defined and implemented. During the implementation of the twinning projects it has therefore to be checked whether the project partners comply with these rules. This sometimes proves cumbersome, as many twinning partners have no or very limited experience with such financial rules.

Moreover, twinning projects have been and are indeed very important in terms of preparing for accession. It therefore makes sense for the Commission to allocate appropriate resources to manage these projects.

45. This could be dealt with at the programming stage. The identification and design of institution-building twinning projects is the result of a well balanced process and dialogue between the Commission and the candidate country under the final scrutiny and approval of the PHARE Management Committee.

Regarding more precisely the "overemphasising of twinning" one should recall that in 1998 and 1999 twinning as a new instrument was indeed emphasised to get the process started.

46. The Commission underscores that the selection of twinning proposals is also based on a competitive process, which takes into account the specificities of public administrations.

Twinning should be used where it works better than commercial expertise, bearing in mind earlier, unsatisfactory experience with PHARE public administration projects.

47. The involvement of mandated bodies is consistent with the diverging administrative structures of certain Member States which tend to source out public service tasks to public or semi-public bodies on a cost recovery basis (subsidiarity principle). Moreover, in view of the growing number of twinning projects, the possibility to call on mandated bodies supplements Member States' capacity to present adequate and complete proposals. Some smaller Member States may experience capacity shortages which would prevent them from bidding on the growing number of twinning projects in the absence of mandated bodies.

48. The Commission has repeatedly insisted with the Member States to maintain public sector consistency within the implementation of twinning projects and has asked not to call on private sector experts either through CFCU or through temporary appointments in mandated bodies to a degree which could jeopardise the overall purpose of twinning, i.e. building structural links between administrations.

49. See comments under 45 regarding the programming of twinning projects.

The updated PHARE programming guidelines for 2001 and later have softened the requirement regarding the minimum duration of the stay of the PAA: Until the 2001 programming exercise, the Commission has insisted that the PAA spend at least 12 consecutive months working in the candidate country. In view of the now impending accession of the 10 Laeken countries, there may be cases where the presence of the PAA could be reduced; e.g. permanent presence of a PAA at the beginning, for a period of perhaps four to six months to kick-start the project, followed up by a monthly repeat visit of up to maybe a week by the same expert to ensure that the momentum is maintained.

The decision on the duration of the PAA's stay in the candidate country must be made at the programming stage and implementation monitored very strictly by the Commission.

50. The risks of failure of twinning projects are the same as in any technical assistance assignment or other project. If the twinning covenant is properly defined, including correct benchmarks and correct definition of the work plan, with realistic appraisal of the tasks, the risk of failure is on the contrary in many cases less than in a technical assistance project. The political factor has an important role in preventing failure: in most of the cases, the Member State administration takes pride in having a successful project. The participation, motivation of the beneficiary can be a positive or a negative factor, but would be the same in any kind of project.

51. Twinning does not stand on its own. In many cases, twinning project has an investment component attached, sometimes even a technical assistance project. These are complementary and such programming shows precisely that each instrument was dedicated to a specific purpose. A coordination effort is needed and adequate planning for the contracting to be concluded at the right time. The delay in contracting showed that in most of the cases twinning was first to start and that the risk of failure of tenders (for technical assistance or investment) was higher than for twinning. Coordination of twinning activities and investment components as well as help in the evaluation of the technical quality of offers.

(a) For Twinning project PL00/IB/FI03 this assistance was clearly spelled out in the covenant. The fact that there were delays in the tendering process was not related to twinning and therefore should not be an argument in the assessment of the twinning instrument as such. The database mentioned will cover EU structural interventions only when Poland will receive such support as a Member State, i.e. not before 1 May 2004. It can even be argued that the delay in the tendering for the IT software referred to by the Court was in this particular case beneficial because it allowed the experienced twinning experts to assist in drafting the ToR for this crucially important tender.

(b) Coordination during the implementation of the project, was done through the following means:

- formal meetings of the Interministerial coordination or monitoring committee convened on a quarterly basis involved both the twinning and non-twinning partners and the delegation in Budapest,

- the Structural Funds non-twinning team organised monthly meetings to which the twinning partners were invited,

- professional meetings organised between the short term experts and the PAAs - some products were delivered in cooperation between both teams,

- any problems that occurred were not due to the lack of coordination during the implementation but mainly to the design of the programme.

52. The basic reason why twinning has been preferred over technical assistance contract to implement some institution-building projects is that twinning projects mobilise experts with more pertinent and hands-on expertise as to the actual implementation of acquis related issues.

(a) The Twinning manual contains provisions to reorient the project (especially after the first six months) or to halt a project before finalisation in case of ineffectiveness or other failure. This last alternative is something both candidate countries and Member States endeavour to prevent at all cost because this could reflect badly on their perceived commitment to the enlargement process. Indeed, the Commission decided to halt two projects before end date (BG99/IB-CO-02 National Health Insurance Fund and BG99/IB-CO-01 Coordination of social security schemes)

(b) The failure or shortcomings of twinning projects had and have almost direct consequences on progress in the accession negotiations with the candidate countries involved. Moreover, the risk failure by Member States involved in a given twinning project would jeopardise their standing and chances of being selected for subsequent twinning projects.

(c) Any project supposes management and the existence of a set of rules that may cause a degree of complexity. Twinning gets many administrations to work together even if these administrations have their own rules. This may seem difficult in the beginning, but proved feasible and manageable in the last two years, especially for those cases where adequate support was mobilised by the Member States.

(d) See Annex 3 to PHARE programming guidelines (instructions for design of institution-building projects based on input from Member State public administrations - twinning and twinning light) which indeed allow the necessary flexibility as to the length of the PAA's secondment since 2002.

TWINNING AND NEW ACTION PLANS 2002

54. The supplementary assistance provided under the action plans may be used to finance additional institution-building actions, including, where appropriate, twinning projects.

55. As already indicated, the Commission has taken numerous steps to shorten procedures with (i) deadlines for the drafting of the twinning covenant, (ii) the introduction of twinning light and the unallocated institution-building budget which can be used for needs that have suddenly arisen for short and long term measures.

CONCLUSIONS AND RECOMMENDATIONS

57. The Commission takes note of the Court's positive and constructive appraisal of twinning as a laudable initiative for institution-building. Through twinning the Commission indeed pioneered an unprecedented transfer of know-how on complex and technical issues from the diverse administrative traditions in Member States to the equally diverse candidate countries. The Commission again highlights the specific genesis of twinning as new and unique instrument for targeted international/European administrative cooperation.

58. The concept of twinning although positive overall, implies however the cooperation and commitment of many administrative partners, which may in turn explain administrative complexity, especially in the start period under review in the present report.

59. Candidate countries have indeed made quicker progress in the adoption of the acquis communautaire through legislation than through administrative implementation and enforcement. It is debatable to conclude that the overall performance of twinning is below expectation, in view especially of the conclusions of Directorate-General Enlargement's internal assessment of the 1998 and 1999 twinning projects. Twinning was never considered the sole, ultimate and general solution to all needs of the administrative capacity in the candidate countries. It will produce its (limited) effects to the extent it is correctly used and in the context of the general support to the candidate countries. The Commission underscores that the concept of "guaranteed result" was the key feature of twinning from the very beginning. In twinning, both project partners commit themselves to work towards a commonly agreed result in a joint project implementation process, possibly with the definition of intermediary benchmarks. It is precisely this joint and previously agreed commitment which distinguishes PHARE twinning from other instruments for administrative assistance and institution building. If one considers the results achieved by Twinning to those obtained in the context of the old technical assistance based approach during the years 1990 to 1996, then twinning has a good track record.

Twinning is by definition based on the cooperation between administrative partners from the Member States, the candidate countries and the Commission which entails an inevitable element of administrative complexity

The existing procedural rules have been introduced to guarantee proper spending of PHARE funds; whenever necessary, the rules have been adjusted.

60 (a) This is dealt with at the programming stage. The identification and design of Institution building Twinning projects is the result of a well balanced process and dialogue between the Commission and the candidate country under the final scrutiny and approval of the PHARE Management Committee.

(b) Twinning project partners are required to report regularly and in depth and that the implementation of twinning projects is monitored on an on-going basis.

(c) The problem of sustainability is not specific to twinning; it derives from the general environment of public administration in the candidate countries with high turn over of personnel, low salaries and sometimes-poor management. For reasons explained above, the concept of guaranteed result implies the need for the administrative partners to achieve clear and verifiable results. In this sense this commitment is central to twinning and cannot be dropped without jeopardising its essential characteristics.

(d) Delegations are already very involved in the monitoring and steering of twinning projects. The Court seems to recognise this involvement in its report.

61. The Commission welcomes however the Court's comments as to corrective actions taken on the basis of lessons learnt from 1998/1999 twinning projects.

(a) The Court has recognised in its observations 27 to 32 that the Commission has shown readiness to improve procedural requirements for twinning projects in due course.

(b) The Commission submits that the internal consultation procedures within the framework of the twinning steering committee already encompass clearly set deadlines, which are generally upheld. As such, Directorates-General are requested to provide their comments within 10 working days after which the steering committee normally finalises its decision within two or three working days. For amendments requiring the approval of the twinning steering Committee, the deadlines for consultation are even shorter - five working days. Moreover the Twinning Manual provides for a general deadline of six months between selection ands start of the project.

(c) As explained, the Commission has already taken steps to simplifying payment procedures by increasing the amount of advance payments.

62. The identification and design of institution-building twinning projects is the result of a well balanced process and dialogue between the Commission and the candidate country under the final scrutiny and approval of the PHARE Management Committee. Proper and targeted programming is of the utmost importance and this has been duly highlighted in the PHARE Programming guidelines at least since 2000.

63. The Commission has provided for a transparent and consistent procedure to remunerate the involvement of mandated bodies that is based on fixed fees and flat rate compensation.

64. The Commission welcomes this suggestion and is prepared to refine its data base where all the professional data and contacts of PAAs are already available so that a network could be set up.

65. It is indeed true that twinning, as a unique instrument for pre-accession assistance has attracted quite some attention since it is mentioned in the Commission's White Paper on European Governance (Page 25: "The Union can effectively draw on the experience acquired with the applicant countries, such as on the twinning arrangements") and even in the latest Report from the Commission on European Governance of 11 December 2002(1).

(1) Report from the Commission on European Governance of 11 December 2002; under point 3.2, sixth paragraph: In the white paper on European Governance, the Commission, drawing on experience acquired with the applicant countries, announced that it would propose twinning agreements between national authorities in order to encourage the sharing of best practices in implementing measures in particular sectors and promote awareness of Community law among national courts and lawyers. The Commission intends to propose twinning arrangements in 2003 with a view to modifying the "Twinning programme" currently applied during a transitional period.

ANNEX

LIMITED ACHIEVEMENT OF AMBITIOUS OBJECTIVES

1. Through twinning the Commission pioneered an unprecedented transfer of know-how on complex and technical issues from the diverse administrative traditions in Member States to the equally diverse candidate countries. It was also essential in mobilising political will and resources in the candidate countries.

During the period under review in the Court's report, twinning projects indeed mainly focused on four key areas (agriculture, finance, environment and justice and home affairs) whereas subsequently twinning projects covered the whole of the acquis in the wider sense.

Example: Preparation for Structural Funds

3. The programming of Structural Funds projects is in general a more complex exercise than for other "acquis-related" projects, as it covers rather "practices" than specific Directives or Regulations. Furthermore the design of a SPP project fiche involves very detailed horizontal consultations from both sides, beneficiary and Commission Institutions. Once the implementation of a project has been awarded to a Member State, it is up to the beneficiary and the selected Member State partner to set-up the proper activities/strategy, which will enable the achievement of guaranteed results.

Regarding the specific Structural Funds projects for Estonia, the Commission accepts that the initial and first Structural Funds project was not perfectly designed but the final report nevertheless assessed the results obtained and compared them with the objectives initially set.

5. The Commission refers to its general comments regarding the inherent complexity of programming the Structural Funds twinning projects, especially in the start up phase of twinning. Later Structural Funds twinning projects have been designed on a more compact basis.

6. The main reason was that it was at that time too early to establish the legal and administrative basis for SFs for two reasons: 1. candidate countries did not know the date of accession then; 2. the road map for SFs was only presented to the andidate countries in March 2000.

The Commission learnt from the difficulties of coordinating such projects and introduced limits for both the number of Member States involved in the same project and for the size of the project.

8. While acknowledging the deletion of some benchmarks as a consequence of delayed establishment of administrative structures in Lithuania, the Commission recalls that the other benchmarks have been fully met.

9. While lack of commitment, poor management and poor working relations may have affected the respective projects, these are general negative factors that can affect any kind of project. These factors are not specific for twinning as an institution-building instrument.