SPECIAL REPORT No 13/98 concerning the audit of the use of risk analysis techniques in customs control and the clearance of goods accompanied by the replies of the Commission(pursuant to Article 188c(4), second subparagraph, of the EC Treaty) (98/C 375/02)

1. INTRODUCTION

1.1. The Community Budget is part financed by customs duties and agricultural levies, known as 'Traditional own resources`. For 1997 the budget estimate for traditional own resources was ECU 14,2 billion, approximately 17,7 % of the total budget. Responsibility for the collection of these resources is devolved to the Member States (1). The front line collection is carried out by the customs services of the 15 Member States. In addition to the collection of own resources these customs services are responsible, among other things, for the implementation of economic policy measures, the protection of the Community from imports of prohibited or dangerous goods and illegal drugs, monitoring trade in endangered species under the CITES (2) convention and the control of exports which give rise to charges on the Community budget under the common agriculture policy.

1.2. The completion of the single market on 1 January 1993 was accompanied by the coming into force of the Community Customs Code, which was established by Council Regulation (EEC) No 2913/92 of 12 October 1992 (3). The Code, together with its implementing provisions (4), was intended to consolidate existing Community customs regulations, one of whose objectives is to guarantee a uniform application of customs legislation throughout the Community. This, and subsequent amending legislation, lay down in specific terms the nature of controls to be exercised by the customs authorities in the protection of the Community's financial interests and is designed to ensure the correct application of customs legislation.

1.3. To put the role of the European Union's customs services into perspective, it is necessary to understand the nature of the environment in which they are required to operate. Their work is dictated by Community regulations underpinned in each of the Member States by national legislation and instructions. They are required to process each year some 30 million import declarations, a similar volume of export declarations and some 20 million Community transit operations. They apply the Combined Nomenclature TARIC (5) containing in excess of 30 000 product positions many of which comprise a number of different duty rates dependent on the country of origin of the product. Customs legislation, in so far as it applies to specific goods, is constantly being updated by the amendment of duty and levy rates and the imposition of anti-dumping duties and tariff quotas. All these factors, together with the objective of traders to maximise their profits, combine to ensure that the timely collection of the correct duties and levies are 'at risk`.

2. THE NEED FOR RISK ANALYSIS

2.1. Given these operating circumstances and the will in all Member States to reduce the cost of public services, the need to introduce efficiency measures in customs control while ensuring that these services remained effective in protecting the Community's interests became obvious.

2.2. To this end and because of the imminent commencement of the single market on 1 January 1993, the European Institute of Public Administration of Maastricht carried out, at the request of the Commission, a study on 1992 and the organisation of Member States' and Community customs administrations. The Customs Questions Committee (6) agreed that the recommendation of the study to improve the working methods of the customs services deserved more consideration. An ad hoc group of experts set up to pursue this question, concentrated its efforts on the use of risk analysis in customs control.

2.3. The Ecofin Council approved, on 23 November 1992, the use of risk analysis to target financial controls. It asked the Commission to adopt, before the end of 1993, measures to introduce the principle of targeted-risk audits to all sectors subject to controls. However, the Community Customs Code and its implementing provisions only provided for the use of risk analysis techniques in the control of goods imported under simplified customs procedures by air or by sea. The regulations governing the monitoring of agriculture products receiving refunds at export (7) also require the use of risk analysis techniques if examination rates fall below 5 %.

2.4. In addition, there is one legislative imperative on customs authorities to use risk analysis techniques in determining how and where to exercise their control mandates: Decision No 210/97/EC of the European Parliament and of the Council of 19 December 1996 adopting an action programme for customs in the Community (Customs 2000) (8) which arose out of the pilot action Customs 2000 initiated in December 1993 when the heads of the customs administrations in the Member States stated their intention to improve cooperation between the customs services to implement controls and procedures of equivalent effectiveness at every point of the external border of the Community customs territory; the decision stated:

'The action programme shall be called the "Customs 2000 Programme" and shall be implemented for the period 1 January 1996 to 31 December 2000`. (Article 1(2)).

'The common framework of objectives provided for in Article 1(3) within which the Commission and the Member State shall establish plans and priorities in order to define and carry out coordinated action to ensure that customs action matches the needs of the Community's internal market, shall aim to:

1. guarantee that Community law is applied in such a way as to achieve equivalent results at every point of Community customs territory in order to:

- . . .

- protect the Community's interests, particularly its financial interests; . . . ` (Article 4).

'In accordance with the objectives defined in point 1 of Article 4, the Member States and the Commission shall:

. . .

4. in order to achieve these objectives, adjust working methods, equipment and materials so as to reinforce the effectiveness of controls; foster in particular the development of targeting, risk analysis and a posteriori auditing techniques and where necessary, determine the nature of controls to achieve equivalent results; . . .` (Article 5).

2.5. However, several Community documents and pronouncements either encourage or exhort the use of risk analysis or express the desire that the Community customs services 'act as if one`. The most significant of these are the following:

(a) the communication from the Commission to the European Parliament and the Council (9) outlining an action plan for transit in Europe;

(b) Council resolution of 23 November 1995 on the computerisation of customs transit systems (10);

(c) Council Regulation (EC, Euratom) 2988/95 (11) on the protection of the Community financial interests;

(d) Council resolution of 25 October 1996 on the simplification and rationalisation of the Community's customs regulations and procedures (12);

(e) the Commission replies to observations made by the Court in its 1993 and 1996 Annual Reports (13);

(f) the European Parliament resolution of 25 November 1995 (14);

(g) 'second report on the operation of the inspection arrangements for traditional own resources` 11 December 1997 (15);

(h) Council resolution of 18 December 1997 (16) setting out priorities for cooperation in the justice and home affairs field for the period 1 January 1998 to the date of entry into force of the Treaty of Amsterdam.

2.6. All the documents listed in paragraph 2.5 stress the need for using state-of-the-art customs techniques and more effective working methods such as information technology, risk analysis and targeted control using risk profiles, common and coordinated risk management policy, for example harmonised qualitative and quantitative control by degree of risk, and advanced audit systems. They insist on the requirement that national customs services act as if they were one. Member States are invited to cooperate closely and to coordinate their efforts with a view to attaining these common objectives. The Commission notes that there is a lack of homogeneity in the inspection techniques from one Member State to another and expresses its belief that the development of risk analysis methods may remedy this state of affairs. Under the Customs 2000 programme it hopes to draw up Community-wide 'risk profiles`. For some customs regimes it will ask Member States to make systematic use of risk analysis methods in their control measures.

3. WHAT IS RISK ANALYSIS?

3.1. Risk analysis may be defined as a systematic process of making decisions based on structured collection and analysis of information, structured risk identification and analysis to target intervention at the most serious risks and to evaluate how effective the targeting is in minimising or eliminating the risks. Alternatively, risk analysis is a system for the identification and use of appropriate management information, in a consistent and transparent manner, to comprehensively identify, document and assess risk. The system should lead to the minimisation of risks by coordinating control efforts to which priorities have been assigned. It should include performance information to monitor progress and allow effective review of the entire system.

3.2. Any system of risk analysis must adhere to certain fundamental principles if it is to meet the objectives set out in the aspirations of the European Parliament, the Council, the Commission and the heads of customs administrations (see paragraph 2.5). Such principles are:

(a) the context within which risk analysis operates must be clearly defined as the customs territory of the European Union as a single market. This implies that the fragmented nature of Community customs administrations must be acknowledged and steps must be taken to counterbalance this with appropriate coordinating actions;

(b) because of the freedom of movement given to goods within the single market, risks faced by the Community as a whole as well as purely local risks must be identified. Risk identification depends very significantly on the ease with which information and other worthwhile intelligence is shared between all the affected services;

(c) access to the information of risks identified must be ensured Community wide so that adequate and coordinated risk analysis can be carried out. Any results of analysis work carried out in individual Member States should be shared with all other Member States and analysis work must be coordinated to avoid unnecessary duplication and waste of valuable resources;

(d) the relative priority of the risks identified needs to be determined so that uniform treatment of these risks can be effected at all points of the territory, with sufficient resources being dedicated to eliminating those risks of greater impact to Community own resources;

(e) finally, no structured system would be complete without an effective monitoring and review process to ensure that all pertinent information, underlying the assumptions and decisions taken, is accurate, up-to-date and relevant. Such monitoring must be carried out across the customs services of the Community, in a spirit of cooperation. The review must use information feedback from the process and lead to updating the analysis and prioritisation of control work.

4. COORDINATION OF RISK ANALYSIS ACTIVITIES TO DATE

4.1. The ad hoc group of experts, referred to in paragraph 2.2, commenced work on the question of risk analysis in March 1991. With the help of the Commission services and with contributions from most Member States, this group produced guidelines on risk analysis to be used across the Community.

4.2. Draft guidelines were presented to the Commission by the group in February 1993. Before completion of the guidelines, a number of drafts were made available to Member States, which were encouraged to implement them as they stood. Non-binding guidelines were issued to all Member States in October 1997. The guidelines are to be the focal point of a pilot project to be conducted in six Member States until 31 December 2000.

5. THE COURT'S INQUIRY

5.1. The Court's inquiry focused on the use of risk analysis in customs controls exercised in the physical inspection of goods and documentary checking of import declarations which affect the Community's own resources.

5.2. Despite the absence of an all-encompassing statutory basis for using risk analysis, the Commission has advocated the use of risk analysis in customs controls, because 'it is materially impossible for national customs administrations to carry out physical inspections and documentary checks on all consignments of goods` (17). Further, as it is traders and not their goods which perpetrate fraud, risk analysis focused on traders is an important consideration. The Court considered that some six years after starting the risk analysis programme and four years after the opening of the single market the state of implementation of risk analysis techniques in the Member States needed to be audited.

5.3. Using as a model the components of a risk analysis system set out in paragraph 3.2, the inquiry focused on the methodologies adopted in eight Member States to identify and protect the Community's own resources from risks. The inquiry also examined the controlling and coordinating role of the different services within the Commission. It also sought to focus on the management and effectiveness of the techniques employed, the use to which sources of intelligence are put and the extent of feedback mechanisms in place. Where formal risk analysis techniques were not in use, the inquiry attempted to assess the effectiveness of alternative methods in protecting the Community's financial interests.

5.4. In the eight Member States visited (Belgium, Denmark, Germany, Spain, France, Italy, the Netherlands and the United Kingdom) the Court sought information on:

(a) national control policies and instructions and assessed the extent to which risk analysis is used;

(b) the role played by national customs computer systems in the selection of customs declarations for special examination. This included the systems' ability to isolate transactions by use of profile testing (filtering);

(c) the information sources and methodology used in the collection of risk data;

(d) the methodology used to analyse the risk data provided and assign resources to counterbalance the highest risks; and

(e) the systems in use to distribute risk details to outfield customs staff and to obtain feedback on the use of this data.

5.5. The Court tested the application of the methodology in a number of customs offices in each Member State visited.

5.6. The Court selected a sample of information notices issued by the Commission to Member States under Council Regulation (EC) No 515/97 of 13 March 1997 on mutual assistance (18), to assess the use made of these notices by Member States and the extent to which they are made available to operational customs staff.

6. FINDINGS

6.1. The Court acknowledges that a number of Member States are wholly committed to and supportive of attempts being made to develop a Community-wide risk analysis methodology for use in customs control. The findings presented in this report should assist in these efforts.

General findings

6.2. In most Member States the need for using risk analysis techniques is recognised. Operational staff have reasonable access to information on risks. However, in most instances this information is not analysed and used in an adequate way to ensure that resources are allocated to the areas of highest risk.

6.3. Apart from the communication of information on detected fraud cases through the mutual assistance provisions, there are no structured lines of communication to ensure that information or intelligence on common risks are shared between the Member States and the Commission.

6.4. Mutual assistance information is used by the Member States essentially for a posteriori controls of customs transactions. It is only used exceptionally by the customs services as a source of information of potential risks in current or future imports.

6.5. While, the Commission guidelines formally acknowledge that 'people defraud, goods don't` risk analysis at present focuses nearly exclusively on risky goods. However, there is no significant effort throughout the Member States to coordinate the identification of risky goods. Even though traders and customs regimes are inherently more risk prone, there is no endeavour to concentrate risk analysis on traders or customs regimes.

6.6. In most Member States the feedback process consisted of financial details of the amounts collected as a result of a posteriori controls. In most cases it was not possible to say if the controls giving rise to the collection of additional duties were carried out as a result of the risk analysis exercise, nor did data transmitted by the collection centres give details of the circumstances and specific risks which gave rise to the collection of additional duties.

6.7. The extent of convergence of the level of physical and documentary controls carried out in the Member States is low. In three Member States national instructions provide that direct customs declarations are subjected to a 100 % documentation check, in one Member State visited no statistical data on documentary checks were available. In the other Member States visited, the level of documentary checks carried out varied from 17 % to 70 %. In a number of Member States it was not possible to determine precisely the level of physical inspections carried out on directly imported goods. Estimates made by the Court at the customs offices visited indicate that the level of physical checks carried out varied from 1,4 % to 10,3 %.

Specific findings

6.8. All the Member States visited would contend that they use risk analysis techniques in their customs control activities and support this contention with national control plans or instructions. However, the emphasis placed by them on the distinct risk areas of goods, traders and customs regimes, is variable. This variability in emphasis is further heightened by the different levels of central involvement in the risk management process.

6.9. Responsibility for the performance of risk analysis is wholly devolved to local offices in four of the Member States visited, while the others had elements of local and central or regional input to the methodology.

6.10. In most of the Member States the distinction between risk identification (see paragraph 3.2(b)) and risk analysis (see paragraph 3.2(c)) is not made. This lack of clarity, evident at central administration level, is carried down to operational level by the absence of clear national guidance or training.

6.11. Only in two Member States did the Court find evidence of on-going analysis of risks related to the importation of goods into the Community. In another Member State, although the system is said to operate on the basis of risk analysis, no evidence could be produced to support this. In a further three Member States there is ample distribution of information on risky goods, primarily under mutual assistance arrangements, but there is no real provision for analysis of the risks involved to ensure the targeting of those goods of major risk.

6.12. In three of the Member States visited all customs declarations except those presented under simplified procedures must receive a documentary check before final clearance of import goods. In one of these, risk analysis is used to select declarations for physical and documentary checks at the time of clearance. Other declarations are subjected to ex post documentary checks.

6.13. The very fact that 100 % checks are carried out in Member States purporting to use risk analysis in their control procedures calls into question the respective customs authorities' confidence in the quality of the risk analysis carried out. According to the Commission 'this constitutes an unwarranted obstruction to the flow of legitimate trade and places a disproportionate burden on the staff and budgets of customs administrations` (19).

Use of filtering systems

6.14. The alternative to 100 % checking of declarations is the use of filtering systems. These systems rely on the creation of filtering profiles or parameters and can only effectively be implemented in a computerised clearance environment. Customs clearance systems in all the Member States visited, have some filtering capability to determine whether selected declarations are subjected to physical or documentary control. In one Member State, however, only those customs offices connected to the computerised customs clearance system can make use of filtering information and even then control of the selected declarations is not mandatory.

6.15. The degree of effectiveness, sophistication and reliability of the filtering systems in use in the Member States varies from one Member State to another. Filtering of customs declarations is effected by establishing profiles of those transactions most prone to risk and programming the clearance system to isolate those matching the profiles. Profile setting is the responsibility of the central customs administration in two Member States, while in a further five, local customs office have the facility to add profiles which deal with locally determined risks.

6.16. In one Member State, while the facility to set profiles centrally exists, it is not currently used for selection of imports for physical checks. A customs value limit criterion of greater than ECU 12 500 is set for documentary control. This criterion selects so many declarations as to render virtually ineffectual local risk-based criteria, where they exist. Local criteria are not used in some offices; in others they are managed applying the principles set out in paragraph 3.2. The pre-clearance of goods using commercial documents (20) before entry into the computerised system, may result in customs being unable to perform the checks required by the system. As a consequence, the control process may be rendered uneven.

6.17. The system in another Member State only allows profiles to be set using three out of eight valid parameter types. This restriction results in broad category selection with the consequence that very few national or central profiles are used. The system, comprising a registration and clearance module allows users of simplified procedures in the Member State to easily bypass a profile selection or any user to avoid being caught by a profile by clearing goods at a customs centre other than their home office. There is no coordination of locally set profiles with the result that homogenous control is not possible in the customs offices.

6.18. A third Member State's nationally set profiles are mainly designed to isolate declarations subject to customs control mandatory under Community regulations. Only local profiles are expected to be based on risk analysis. There is little coordination of locally set criteria, with the result of unequal treatment of declarations of similar risk in different customs offices. Risk analysis is not used in the regular revision and updating of locally established risk profiles.

6.19. The risk profiling system of another Member State, which is carried out by a central service, is concentrated mainly on risks associated with smaller irregular importers. A 'customer` based system is used for audit control of major traders. National profiles, which rarely give rise to mandatory physical inspections, allow local customs offices autonomy in setting the level of physical inspections and documentary control using local risk factors.

6.20. The system in yet another Member State, which is centrally administered, uses predetermined, statutory risk criteria together with a mathematical algorithm (random factor) to select declarations for physical control. On-going risk analysis is not used to update the profile criteria in use, as these are set in legislation and are of a general character.

Random controls

6.21. The Commission guidelines recommend the integration of random controls in the overall risk analysis methodology as a means of detecting new or previously undetected risks. Only in three Member States was there evidence of elements of random selection of declarations for control. One Member State's system by its nature incorporates an element of randomness in its selections. Local customs staff in another Member State are not encouraged to use random tests.

6.22. Even the most rigorous risk-analysis system runs the risk that new important risk factors may be overlooked or that knowledge of the criteria in use may be discovered by potentially dishonest traders or agents. It is critical therefore that all risk analysis systems contain an element of randomness in their selection procedures.

Effectiveness of filters

6.23. The effectiveness of filters is very much determined by the quality of the analysis work which has gone into setting the profiles. The setting of profiles which are insufficiently focused results in the selection of declarations which are not inherently risky and as a consequence subvert the original intentions of the profile-setting exercise. Examples of this were cited above in relation to two Member States (see paragraphs 6.15 and 6.16).

6.24. To properly assess the effectiveness of the filtering process it is necessary that adequate and relevant data is available on which to base such an assessment. This means that information on the types of declarations selected by the filters, the results of the ensuing controls (value and circumstances) and an interpretation of this information must be fed back to the analysis service for inclusion in future profile exercises.

6.25. In none of the Member States visited was there evidence of adequate feedback of the results of the filtering process. In two Member States local customs staff have the power to override nationally set profiles without the need to report the fact. As a consequence, national data and statistics on the effectiveness of filtering are unreliable.

6.26. Where local customs staff carry out controls in response to local risk factors, there is generally no formalised mechanism for recording the factors for future local or national use. The system in two Member States has, outside the 'risk analysis` system, provision for the recording of incidence of locally selected controls. It also provides for feedback of the results of serious fraud or irregularity cases.

6.27. In one customs office in a Member State, ECU 22,5 million in additional duties was recovered in respect of underdeclarations which had initially been cleared, under the risk-analysis system, without check. The balance of ECU 15 million recovered arose from declarations selected by the system as being risky. The Court is concerned that such a significant sum could be collected from declarations initially deemed to be 'low risk.`

6.28. In all cases, the absence of adequate feedback to central administrations means that these administrations are unable to confirm that a truly homogenous coverage of customs control exists within their own customs areas, let alone ensure such homogeneity at Community level.

Rate of controls effected under filtering systems

6.29. It was not possible to obtain a breakdown of the rate of controls effected in all of the Member States visited, either generally or by traders or by risky goods.

Simplified procedures under a filtering system

6.30. Under simplified clearance procedures the use of filtering systems is problematic for the exercise by customs authorities of the right to physically inspect goods. In the first instance, the simplified procedure of direct trader input (DTI) using electronic data interchange, only allows the customs authorities a short time in which to indicate a desire to carry out physical inspection of the goods before they are removed from customs control. The procedure permits customs authorities to block the removal of goods. In practice, the pressures on operational customs officials to carry out controls, often in geographically remote locations, present logistical problems. On the other hand, physical controls on imports by traders with approval to use period entry systems, i.e. one global monthly declaration of all goods imported in a given month, are at best difficult. Post import control of these traders is the only possible option.

6.31. In one Member State national instructions, in use for a number of years, require that risk analysis be used in the control of importations by large traders using simplified procedures. However, the Court found that the selection of importations for on-the-spot control is carried out on a daily ad hoc basis without any obvious use of risk analysis techniques.

Strengths and weaknesses of filtering systems

6.32. In principle, filtering systems provide a basis for adequate controls on the Community's own resources. However, this is only the case when they are carried out with appropriate analysis of identified risks and where there is reasonable autonomy in local offices to hone the outputs of such systems.

6.33. The effectiveness and efficiency of the filter systems in the Member States is lessened by a number of factors, in particular, the lack of coordinated transnational profiles, the absence of real analysis of risks in many areas, the absence of feedback on the results of controls, the inadequacies of certain computer systems and in some instances the lack of understanding of the purposes and objectives of a targeting system. In one case the Court noted the use of a risk profile declaration where the maximum amount of duty had been paid, rather than those where preferential rates were claimed.

6.34. These shortcomings can result in:

(a) risky goods or traders not being identified because of inadequate feedback;

(b) real risks not being properly attended to because of a swamping of filter hits by use of blanket criteria, for example value limit (see paragraph 6.15);

(c) traders bypassing, either accidentally or deliberately, filter controls arising from inflexible computerised systems and procedures (see paragraph 6.16) and

(d) inconsistencies in the treatment of like risks not only across the customs territory but within individual Member States. This can arise from the use, as local profiles, of risk factors which should have been regarded as of national or even Community importance - as in the two Member States where there is no coordination of the risk criteria used between customs offices.

Post import controls and risk analysis

6.35. At a national level appropriate risk ratings of traders is of primary importance in the effective control of Community own resources.

6.36. It is clear from the DAS audit work carried out by the Court in all Member States in recent years that a significant level of own resources is paid by a small number of large traders in all the Member States. Many of these traders make their customs declarations on a deferred monthly basis, using the period entry system. The Commission guidelines advocate the use of risk analysis techniques in the selection of traders for control.

6.37. All Member States visited use post clearance control visits to supplement controls carried out at the time of importation or, for the control of traders using simplified procedures.

6.38. In one Member State, the annual control plan for the valuation and accounting control service is not based on formal risk analysis. The national standards set for the level of inspections are not being met.

6.39. The national control plan of one Member State's customs service provides for the use of risk analysis in post import control work. This work is carried out by a national and regional inspection service in accordance with preselected financial limits. Although traders for control under the annual programmes are to be selected by using risk-analysis techniques, the criteria used in specific cases under review could not be provided to the Court. The Court noted that where post import controls carried out under these plans detect underpayment of duties the procedures in operation do not permit the control to extend to a current date, thus requiring a subsequent control under the next programme and may result in additional duties not being established at the earliest opportunity.

6.40. In one Member State, the authorisation of traders to use simplified procedures is deemed to be determined under risk analysis conditions, using all recorded fiscal knowledge of the trader. National instructions require an annual control visit (an audit) to be carried out in every instance and include provision for extraordinary visits if deemed necessary. Risk analysis is not formally used to select the traders for extraordinary visits.

6.41. Control plans in one Member State provide for the selection of traders for post import controls on the basis of potential revenue loss. A system based on quantified revenue at risk is designed to give assurance in terms of 'confidence that any gap between what is intended to happen and what does happen is acceptable`. The system, which targets the traders and customs regimes of greatest risk, has been examined in a recent National Audit Institution review and as a consequence is currently being revised. The criticisms levelled at the system are that: risk is adjusted to give a desired level of audit work; it is inflexible; systems evaluations are incomplete; tests are insufficient; the training and guidance given under the system needs to instill a better understanding of auditing techniques and principles; resourcing is static and does not reflect risks, and other competing objectives of the customs administration may result in a reduction in the achievement of the assurance objective.

6.42. Customs control plans for period entry traders in another Member State are required to be formulated using formal risk analysis procedures. In most cases this is done, although some inconsistency was detected in a number of offices visited.

6.43. In a further Member State, although the national control plan requires the use of risk analysis in selecting traders for control, there was little evidence of a formal risk-analysis approach. In the customs offices visited, local knowledge of traders was applied in the selection process, but in an informal way.

6.44. A post import control system is widely used by the customs authorities in one Member State to counter the risks faced. The system is characterised by a computer-assisted risk selection process using predefined risk factors. An annual plan of trader audits is produced. The Court noted however, that the data on which the risk analysis is based were not kept up to date and that different levels of initial risk assessment are made for new traders depending on which customs region makes the assessment. For the regions visited on audit, the Court noted that the level of audits planned was not being achieved and that the results of previous audits were generally not being used to update the risk factors in the planning process.

6.45. It is not clear how the procedures in use in the Member States visited to ensure the selection of risky traders in the annual audit plan, are capable of identifying traders who, although they have significant customs transactions, clear their declarations at different customs points both within the country and in neighbouring Member States.

National control plans

6.46. A precondition for the creation of a homogenous approach to customs control throughout the Community is the existence of consistent national control plans.

6.47. Although all the Member States have national control plans, strategies or instructions, only in three Member States was there evidence of formal national control planning. Despite this, it was evident from the Court's audit, that the diversity of control practices found arises from the absence of adequate coordination at a central level of the local or regional inputs to the planning system.

Post import control problems

6.48. The Court has already reported on post import controls in the Member States in its report on the 1993 financial year. This report notes that a number of Member States are using risk analysis in determining their control plans.

6.49. The Court finds that five years later there is little progress in the use of risk analysis to target those risky traders. In fact, even in those Member States where there is formal expression given to the use of risk analysis, the findings above indicate that there is urgent need for a more formalised targeting in this area.

6.50. In many cases the risk criteria are adjusted to match a predetermined resource allocation. Rather, under a proper risk-analysis system risks should, as stated in the preamble, be identified, assessed and prioritised before the allocation of resources.

Test sample of customs declarations matching notices issued pursuant to the mutual assistance arrangements

6.51. In all cases, the Member States carry out a retrospective examination of transactions matching the notices and put in place post clearance recovery procedures where appropriate. However, in relation to the use of this information as a basis of on-going risk analysis the Court's findings are more varied.

6.52. There is widespread distribution of mutual assistance information notices to customs offices in three Member States. However, in each of these three Member States, customs officials at these offices must rely on their memories to target declarations matching the criteria involved. In one system, the declarations are subject to documentary check and under the random selection mechanism integrated into the system each declaration made has a chance of being selected for physical examination.

6.53. In three other Member States some analysis is carried out on the detail of the information notices, to determine whether national profiles should be set in the declarations' filtering system.

6.54. In another Member State mutual assistance information is not used to create national profiles. Selective distribution of the information is carried out, but not on the basis of risk analysis.

6.55. One of the Member States visited only issues mutual assistance notices to customs offices where there has been a past history of affected transactions. Thus this information is not available for future determination of risks in all the regions.

7. CONCLUSION

7.1. The aspirations of the European Parliament, the Council, the Commission and the heads of customs administrations that the customs services of the Community act as one by taking coordinated action to achieve equivalent results in the Community customs territory, by use of effective, efficient and homogenous application of Community rules are far from being met (see paragraphs 6.2 and 6.6).

7.2. While there is a widespread willingness to further the use of risk analysis in the customs services, the lack of coordination in particular on the determination of risky goods and circumstances renders ineffectual a good deal of the control work being undertaken at local level (see paragraphs 6.4, 6.14 to 6.19).

7.3. There was sufficient evidence to confirm that much information on risky subjects is being passed out to the frontline of the customs services. However, in the absence of really effective analysis of the relevance and potential of these risks for the individual frontier posts, the availability of this large mass of information does nothing to improve the direction of the control effort available (see paragraphs 6.9, 6.32 and 6.50).

7.4. There are as many manifestations of risk-analysis methodologies as Member States visited on this enquiry. While there may be common agreement on the need for a risk methodology and while the Commission's guidelines give a good road-map for its implementation, present trends indicate an uncoordinated and uneven application of Community legislation (see paragraphs 6.6, 6.33(d), 6.46 and 6.49).

7.5. Five years after the coming into being of the single market and after six years of operation of the Matthaeus Programme (21), a greater amount of cohesion would be expected from the training and exchange programmes initiated. Alternatively, some integration of customs administrations by creating common specialised units or services at Community level for example planning and control activities, may improve coordination among Member States (see paragraph 6.2).

7.6. There is still a considerable amount of work to be done to ensure there is an effective, efficient and homogenous risk-analysis system in operation in all the Member States. The Commission should:

(a) ensure its guidelines are incorporated in binding regulations and framed to include the best possible protection to the Community. Community legislation governing all aspects of customs activity and control should be amended without delay to provide for mandatory use of risk-based methodologies and targeted controls;

(b) ensure its services have adequate resources to provide for the effective coordination of meaningful analysis of all data for use in all Member States and in particular in the areas of goods and regimes at risk. In particular, the Commission should undertake the coordination of risk-analysis and consequential control plans with regard to intra-Community transactions or traders operating in more than one Member State;

(c) take steps to ensure that in proposing any new draft customs legislation, it does not increase risks by introducing inappropriate (burdensome or unmanageable) legislative requirements. The Commission must also ensure effective on-going oversight of national legislation introduced in the Member States, which in some instances gives effect to Community legislation, to ensure that this national legislation does not create additional risk;

(d) ensure that the activities of its inspection and monitoring services are planned on a risk basis so as to contribute to the achievement of equivalent controls in all the Member States and to an acceptable level;

(e) take the actions necessary to incorporate into its guidelines and legislation the many effective innovative initiatives taken in the Member States, for example centralised analysis of products risks, analysis of regime risks, the use of freight intelligence units, and random selection using risk algorithms. It is essential that the best of these innovations are brought together to form a comprehensive protective barrier to the Community's own resources;

(f) require the harmonisation of the level of documentary checks carried out in the Member States. The Court noted that documentary checks varied from 17 % in two Member States to 100 % in three Member States. Clearly, using risk analysis in a 100 % checking environment is unnecessarily wasteful of resources and also imposes an additional distraction to the carrying out of the checks involved. Often post import controls suffer in these circumstances. The interests of harmonising controls, to place an even burden of control on all economic operators in the Union, must be stressed;

(g) assist Member States to design more structured and scientific risk analysis systems, by the provision of expertise, training, research and coordination facilities.

This report was adopted by the Court of Auditors in Luxembourg at its meeting of 16 and 17 September 1998.

For the Court of Auditors

Bernhard FRIEDMANN

President

(1) Council Decision 94/728/EC, Euratom of 31 October 1994 on the system of the European Communities' own resources (OJ L 293, 12.11.1994, p. 9).

(2) Convention on International Trade in Endangered Species (CITES).

(3) OJ L 302, 19.10.1992.

(4) Commission Regulation (EEC) No 2454/93 of 2 July 1993 laying down provisions for the implementation of Council Regulation (EEC) No 2913/92 establishing the Community Customs Code (OJ L 253, 11.10.1993, p. 1).

(5) Council Regulation (EEC) No 2658/87 on the tariff and statistical nomenclature on the Common Customs Tariff (OJ L 256, 7.9.1987, p. 1), latterly amended by Commission Regulation (EC) No 2086/97 of 14 November 1997 amending Annex I to Council Regulation (EEC) No 2658/87 (OJ L 312, 14.11.1997, p. 1).

(6) The Customs Questions Committee is composed of the heads of the customs administrations in the Member States and the Director-General of DG XXI.

(7) Council Regulation (EEC) No 386/90 of 12 February 1990 on monitoring carried out at the time of export of agricultural products receiving refunds or other amounts (OJ L 42, 16.2.1990, p. 6).

(8) OJ L 33, 4.2.1997, p. 24.

(9) OJ C 176, 10.6.1997, p. 3.

(10) OJ C 327, 7.12.1995, p. 2.

(11) OJ L 312, 23.12.1995, p. 1.

(12) OJ C 332, 7.11.1996, p. 1.

(13) OJ C 327, 24.11.1994, p. 295 and OJ C 348, 18.11.1997, p. 39.

(14) OJ C 308, 20.11.1995, p. 57.

(15) COM(97) 673 final, 11.12.1997.

(16) OJ C 11, 15.1.1998, p. 1.

(17) Written question No 951/96 by Viviane Reding to the Commission. Problems with customs checks (OJ C 322, 28.10.1996, p. 16).

(18) OJ L 82, 23.3.1997, p.1.

(19) Written question No 951/96 by Viviane Reding to the Commission. Problems with customs checks (OJ C 322, 28.10.1996, p. 16).

(20) Community Customs Code, article 76.

(21) Council Decision 91/341/EEC of 20 June 1991 on the adoption of a programme of Community action on the subject of the vocational training of customs officials (Matthaeus programme) (OJ L 187, 13.7.1991, p. 41).

REPLIES OF THE COMMISSION

1. INTRODUCTION

Although the Court's observations provide a clearer idea of customs controls, the conclusions it draws are based on how the customs union will be administered by the final phase of the Customs 2000 programme or even beyond.

1.2. Member States do have a general obligation to undertake controls but they clearly also have an opportunity to assess the way in which such controls are carried out both in general and in specific cases.

The Council has introduced rules governing control methods for agricultural exports, although they cover customs export procedures and are tailored to the needs of the EAGGF, for which they were designed. For only one product, bananas, are there rules specifying checks for the purposes of collection of import duties.

Commission Regulation (EC) No 89/97 of 20 January 1997 introduced an Article 290a and Annex 38b to the implementing provisions laying down the methods of checking bananas entered for free circulation. The circumstances surrounding this Regulation are highly specific and its adoption cannot be interpreted as proof of a willingness on the part of the Member States to accept that this area be administered horizontally by Community legislation.

1.3. The Customs 2000 decision defines the institutional environment in which national customs authorities are required to operate. Its objective is to establish a balance between the role of Community institutions and that of the Member States and to maintain this balance in a spirit of unity and partnership. The Commission will shortly present a report on the implementation and impact of the Customs 2000 programme to Parliament and the Council.

2. THE NEED FOR RISK ANALYSIS

2.3. Although the Ecofin Council underlined the importance of risk analysis to target financial controls in 1992 it was not until 1996 that a framework was introduced for customs by the Customs 2000 decision. This covers the whole range of customs legislation obviating, although not excluding, the need to incorporate references to these working methods in the Customs Code. On modern control methods the Customs 2000 decision is quite categoric about the Commission's role which is to support or encourage the measures to be worked out and implemented by the Member States (point 1 of Article 9).

2.6. In 1997, in fulfilment of its assigned role, the Commission drew up guidelines for drawing up risk profiles.

6. FINDINGS

6.1 to 6.55. The lack of structured lines of communication and shared information between the Member States and the Commission is due to the fact that the customs union is not managed by a single customs administration and does not have an appropriate structure.

The Commission's 1997 guidelines were designed to encourage the convergence of working methods which, as the Court notes in its findings, is a future objective. National working methods are still heavily contingent on other factors including departmental organisation, staff training, availability of resources and computerisation of procedures.

7. CONCLUSION

7.2. Computerisation of customs procedures will improve coordination. Member States are at different stages in this process. Development of transnational customs procedures such as the NCTS (New computerised transit system) will help achieve the desired coordination but a system of identifying goods based on a nomenclature code must be retained.

7.3. Intensive unstructured transmission of all sorts of information to outlying customs offices is a methodological problem which needs to be examined and addressed. This might be done in the future within the appropriate forums responsible for implementing the Customs 2000 decision.

7.5. Training and exchange schemes under the Matthaeus programme are designed to promote an open-minded approach and convergence of practice within customs services but will also encourage the eventual harmonisation of national working methods even though these are at present contingent on a number of other factors (see point 6).

There would be no point in setting up specialised units at Community level unless they had an operational role. Such a role would be appropriate for a single customs administration but goes far beyond the objective of the partnership-based management tool recommended by Customs 2000.

There would, however, be no reason why Member States might not consider setting up a joint operational structure of this kind between themselves.

7.6. Our remarks on the limits inherent in the Customs 2000 decision (see introduction) also apply to the Court's recommendation that the guidelines be made compulsory (point 7.6(a)) and to the innovative initiatives in a number of Member States (point 7.6(e)). On the question of documentary checks, which vary in scope from one Member State to another (point 7.6(f)), a distinction must be made between full-scale controls and preliminary selection operations. The latter will have to continue to be carried out by customs officials as long as there is no computerised procedural environment.

The Commission is very aware of the need to examine draft legislation for any risk it may entail (point 7.6(c)).

When compiling its annual control programmes identifying priority areas the Commission uses a number of objective criteria based on a risk analysis of the most sensitive customs and accounting procedures in terms of their impact on traditional own resources. As these are often followed up by further controls in the Member States they contribute to the achievement of equivalent controls in all the Member States to an acceptable level (point 7.6(d)).

Foreword to Special Report No 13/98 concerning the audit of the use of risk analysis techniques in customs control and the clearance of goods and Special Report No 20/98 on the audit of physical checks of agricultural products receiving export refunds (98/C 375/01)

1. The Court has carried out audits of customs controls over imports giving rise to own resources accruing to the European Union budget and exports of agricultural products which benefit from refunds financed by the European Union budget. Both audits have concentrated upon the use of risk analysis in targeting customs controls.

2. The need for separate reports stems from the markedly different legislative requirements for import and export controls. Legislative provisions for customs controls over imports, both current and a posteriori, are optional and there is only limited legislative provision for the use of risk analysis. The legislation for customs controls over exports of agricultural products benefiting from refunds provides for obligatory current (including physical) and a posteriori controls, whereas risk analysis is obligatory for the latter, it remains optional for the selection of consignments for physical checks.

3. This background has necessitated that the Court spell out a detailed definition of risk analysis for targeting customs controls on imports. It has also naturally led to more specific recommendations to improve the effectiveness of export refund controls. Nevertheless the audits have a number of common conclusions, particularly the need for obligatory use of risk analysis, monitoring over the execution of controls, management feedback, performance measurement and coordination in the European Union customs territory of physical and a posteriori controls.

4. In the context of the protection of the European Union's financial interests, to limit the obligatory use of risk analysis to export refunds is illogical. Risk analysis is a resource management tool to facilitate implementation of effective controls. The European Union therefore needs a broader legislative framework for use by the Member States which would ensure that national administrations do not just target transactions within customs regimes or any other Community scheme but allocate control resources to different customs regimes or Community schemes on the basis of the comparative risks to the European Union budget. At the same time there is a need to make Member States more accountable for the quality of targeting and execution of controls.

5. The joint publication of both reports is therefore intended to draw attention to the need to consider the applicability of risk analysis to all customs controls over regimes affecting the European Union budget. Under the given circumstances the Commission should try to adopt a comprehensive approach to the implementation of the Court's recommendations striving for as much consistency as possible in fields as closely related as the controls of imports and exports.