National Audit Authorities and their impact on the assurance level of the EU Cohesion Policy spending

The European Union’s Cohesion Policy plays a key role to reduce development disparities between regions in the EU in order to increase economic, social and territorial cohesion. Cohesion Policy in the EU is planned and implemented within a multilevel governance structure embracing a wide range of stakeholders. For the 2007-2013 programming period, a formal three-level control structure was introduced at national level for the first time, with the audit authority as the first “independent” audit layer. Preliminary studies indicate that the introduction of the AAs has been a milestone in making the management and control system of the EU’s budget implementation more robust, thus contributing to protect the EU budget from ineligible spending. This research paper analyses whether the work of the national audit authorities in the framework of the EU’s Cohesion Policy has an effect towards the assurance level of the EU spending in this area. The analysis links the development of the irregular spending with the development of the tasks entrusted to the audit authorities. The main objective of the article is to conclude on whether the evolution of the responsibilities of the audit authorities contributed towards a more robust management and control system for the delivery of Cohesion Policy.


Introduction
The European Union's Cohesion Policy is defined by the Treaty on the Function-ing of the European Union (TFEU) (European Union, 2012). Title XVIII attributes powers to the EU in the policy field of "Economic, Social and Territorial Cohesion" (Bode, 2020, pp. 10-13).

ISSN 2534-9228
(2021) VUZF review, 6(1) one third of the total EU spending is dedicated to this policy field. It is a catalyst for investments that has become increasingly important for cofinancing public investments, while compensating for shrinking national and regional investments in the aftermath of the economic and financial crisis. However, the effectiveness of Cohesion Policy funds on the living conditions (here expressed as growth rate of GDP) differs widely between the Member States (Bode, 2020, pp. 26-28, 91-94, 251).

Material and methods
Cohesion Policy is implemented using a multilevel governance approach and it involves a wide range of stakeholders at various levels.
Essentially, Cohesion Policy is implemented under shared management be-tween the Commission and the Member States (Article 32 GR; Articles 26-29 CPR). Under this management mode, responsibility for implementing policy fields and the related funds, including implementation monitoring, is shared by the Commission and the Member States. Although the Commission remains ultimately responsible for implementing the EU budget, responsibility for the actual management and control of EU funds and programmes lies with Member State authorities (Bode, 2019a, p. 25; Article 59 of Regulation (EU, Euratom) No 966/2012 of the European Parliament and of the Council of 25 October 2012 on the financial rules applicable to the general budget of the Union and repeal-ing Council Regulation (EC, Euratom) No 1605/2002(Financial Regulation 2012, FR (2012), European Parliament;Council, 2012, European Court of Auditors, 2017a. Discharge is granted to the Commission alone (Article 319 TFEU). The Treaty also stipulates in Article 317 that the financial rules shall lay down the control and audit obligations of the Member States in the implementation of the budget and the resulting responsibilities (European Commission, 2014, p. 224). In conclusion, the shared management mode of budget implementation is the result of the practical application of a multilevel governance architecture.
As the bulk of EU policies, including Cohesion Policy (Article 14(1) GR for the 2007-2013 period; Article 4(7) CPR for 2014-2020 programme period; Recital 179 FR (2018); European Commission, 2014, p. 223), are implemented under the principle of shared management, this type of management dominates the implementation of the EU budget and is primarily responsible for delivering political results. Responsibility for implementing Cohesion Policy and the related funds, including control activities, is therefore shared by the Commission and the Member States.
Since the 2007-2013 programming period, three separate authorities have been established at Member State level, each with specific tasks: a managing authority (MA), a certifying authority (CA), and an audit authority (AA). Each of these bodies fulfils a specific task at OP level. The tasks of the different authorities are stipulated in the sectoral regulations (Articles 60, 61 and 62 GR for the 2007-2013 programme period and Articles 125, 126 and 127 CPR for the 2014-2020 programme period). Together, they must ensure the legality and regularity of the co-financed operations, under the Commission's supervision and final responsibility (Bode, 2020, pp. 2-3, 51-52).
With this set-up, for the 2007-2013 programming period, a formal three-level control structure was introduced at national level for the first time, with the AA as the first "independent" audit layer. In principle, the audit authorities are a further development of the socalled control bodies, which, until the 2000-2006 programming period, were merely in charge of expost checks Latopoulou, C.;Guevara-López, J., 2014, p. 41; See Articles 10 and 15 of Com-mission Regulation (EC) No 438/2001, European Commission, 2001. Compared to these bodies, however, audit authorities are more independent and have a much broader and more robust mandate. Together with other structural reforms, the formalisation of the internal control structure with a clear separation of functions and greater independence for management and control bodies is regarded as a material contribution towards a more robust governance system in relation to the audit of Cohesion expenditure (Bode, 2019a, pp. 2-3, 51-52;Latopoulou, C.;Guevara-López, J., 2014, p. 41;Karakatsanis G., Weber M., 2016, pp. 172-174).
The need for an audit body independent from the programme implementing authorities is very well reflected in the fact, that the audit authorities still report a high volume of errors for expenditures that were already subject to manage-ment verifications by the implementing entities. For example, for the 236 trans-actions, the European Copurt of Auditors sampled for its Statement of Assurance audits for the accounting 2019, the audit authorities had reported 64 quantifiable errors in the assurance/closure packages. The number and impact of the errors detected demonstrate that there is a high inherent risk of error in Cohesion Policy area and that programme bodies' checks are still often ineffective for preventing or detecting irregularities in the expenditure declared by beneficiaries (European Court of Auditors, 2020, pp. 138-139).

Results and discussion
The Audit Authorities' role and tasks In this multi-level governance system, and based on the single-audit con-cept (Bode, 2020, pp. 57-68;Bode, 2019b), AAs act as cornerstones of the over-all assurance process by providing the Commission with assurance as to the effective functioning of the management systems and internal controls for operational programmes (OPs) -and thus the legality and regularity -of the expenditure certified by the Member State to the Commission (Europe-an Court of Auditors, 2016, p. 184; European Commission-DG Regio, 2009, p. 2; European Commission, 2018, p. 8; For details on the AAs destinct role within the multi-level governance system see Bode, 2020, pp. 53-56).
AAs are in charge of independent ex-post controls and act as an extended arm of the Commission in the Member States. In this way, the AAs help to protect the financial interests of the EU (Bode, 2020, p. 252).
In the context of that responsibility, the AAs are entrusted with a multitude of tasks, which, due to their specificity and diversity, call for significant requirements in terms of organisation and expertise. The range of tasks is very extensive, and the subsidiary legal framework can differ from one Member State to another and even from one region to another within one Member State. Moreover, the combination of ongoing, yearly, and one-off audit and reporting tasks is challenging for their internal organisation, in particular for the allocation of human resources. The AAs' spectrum of responsibilities is very broad and het-erogeneous, ranging from current verifications and reporting requirements to annual reports and one-off statements. Furthermore, the AAs need to cope simultaneously with different programming periods, which have different functional requirements. This is a highly challenging professional and organisational combination (Bode, 2020, pp. 104-106, 252).
This main AA's responsibilities are: -compliance assessment of the management and control system; -audits of operations; -system audits; -preparation and submission of the annual control report; -preparation and submission of the annual summary resp. the annual closure package; -preparation and submission of the final closure declaration.
The multiple roles and the changing legal environment necessitate a high level of professionalization within the AAs. This includes specialised training courses and developing theoretical knowledge in different fields. A high degree of turn-over within the AAs and frequent changes of AA should therefore be avoided (Bode, 2020, p. 252).
The AAs' responsibilities have increased . From a beneficiary's point of view, however, it must be highlighted that it is not the European regulatory frame-work, but rather the Member States' organisational structures and administrative inefficiencies, which constitute a significant proportion of the administrative burden (Bode, 2020, pp. 252-253; European Court of Auditors, 2018b, p. 21; European Court of Auditors, 2018a, pp. 34-37).

The European Court of Auditors and the EU budget
Article 287(1) TFEU states that: "The Court of Auditors shall provide the European Parliament and the Council with a statement of assurance as to the reliability of the accounts and the legality and regularity of the underlying transactions [...]. This statement may be supplemented by specific assessments for each major area of Union activity." The same Article also states: "The Court of Auditors shall draw up an annual report after the close of each financial year." Current practice is that this Statement of Assurance is included in the European Court of Auditors' Annual Report on the implementation of the general budget, which also contains specific assessments of the main Multiannual Financial Framework (MFF) headings (European Court of Auditors, 2017b, p. 10).
In its audits on "legality and regularity of transactions underlying the accounts", the European Court of Auditors' (ECA) overall audit objective is to establish whether the transactions comply, in all material respects, with the applicable laws and regulations (i.e. the TFEU, the Financial Regulation, Implementing Rules, specific regulations, financing decisions and contractual provisions) (European Court of Auditors, 2017b, p. 11).This is the field of audit where the ECA issues an overall error rate for the EU budget in general and for the different MFF-headings in particular. Although the ECA's audit approach have changed slightly over time (the current version of the ECA's audit methodology can be found on the webpage of the ECA (https://www.eca.europa.eu/en/ Pages/AuditMethodology.aspx; The current version of the relevant "Financial and Compliance audit Manual" dates on September 2017, European Court of Auditors, 2017b)), the way the results can be interpreted did not. Whenever, the ECA identifies errors at Member State level, it essentially means that the management and control system is not working effectively. Since the AA's form the last line of defence on national level within the governance structure, errors spotted by the ECA are ultimately due to deficiencies in the AA's work, even if the errors are caused by beneficiaries or by managing authorities or other involved bodies. The assumption here is, that if the AAs conducted a properly risk assessment and carried out well and soundly planned audits, errors rooted at other levels (e.g. at the level of the MA or the beneficiary), should have been identified before (Bode, 2020, pp. 140-142).

Long term development of error rates reported by the European Court of Auditors
In general, the current and the previous programming periods saw accountability arrangements strengthened for Cohesion Policy. The reinforced internal control arrangements have contributed to a significant drop in the level of irregularities identified by the EU's external auditor, the European Court of Auditors. A drop that translates into billions of euros spent without irregularities (Bode, 2020, pp. 253-254; European Court of Auditors, 2018a, pp. 24-25). As regards the AA, Especially the following factors materially contributed a higher assurance level on EU spending in the Cohesion area: -Compliance assessment including a report and an opinion for each OP by an independent audit entity; -The elaboration of an audit strategy by the AA; -The submission of annual control reports and annual opinions, where the AA discharges on its responsibility for providing reasonable assurance on the effective functioning of the management and control systems of the programme, and as a consequence on the regularity of the expenditure certified to the Commission (Bode, 2020, pp. 140-141, 253; European Court of Auditors, 2010, p. 98; Karakatsanis G., Weber M., 2016, pp. 173-175).
The error rates reported by the European Court of Auditors for the reference years can serve as an indicator to measure the effectiveness of the enforced accountability arrangements and the AAs' work and whether the improvements in the management and control system actually materialises (Bode, 2020, p. 141). Figure 1 shows the long-term development of the ECA' error rates. A long-term analysis of the error is not easy to conduct, as not only the presentation of the figures changed over the years, but also the level of detail and aggregation, which limit the exactness of the analysis and ultimately the conclusions drawn out of it. However, despite these constraints, it is still possible to draw the following general conclusions and to identify a long-term trend (Bode, 2020, pp. 144-147): -Overall, the last decade demonstrates a marked improvement in the legality of EU spending: between 2006 and 2010, the Commission succeeded in improving financial management, bringing down the error rate from 7 % in 2006 to 3.7 % in 2010 (See also European Parliament, Fact Sheets on the European Union, 2019). While the error rate rose to 3.9 % in 2011 and to 4.8 % in 2012, it has been improving every year since then (see In 2018, this trend was broken as the error rate rose slightly to 2.6 %, and to 2.7 % in 2019. -The overall error rate for the EU budget has been below the one for Cohesion. This is comprehensible, as the Cohesion Policy is one of the most complex, thus an error prone field. This derives mainly from the special set-up of this area. In Cohesion, the Member States receive payments from the EU budget based on reimbursement of actual expenditures in a multi-level governance system. In the agricultural policy, in comparison, the payments are mainly based on entitlements, which is far less error prone. However, the gap between these two error rates diminished from 4 % in 2006 to meanwhile 0.6 % for the year 2017. In 2018, the gap stagnated to 2.4 % (see Figure 2) and was again reduced to 1.7 % in 2019. Furthermore, until 2017 there was a clear and robust trend towards an alignment of the two rates. With an outweighed increase in the Cohesion Policy's error rate (compared to the overall error rate increase) for the year 2018, the trend still shows towards an alignment, but to a much flat slope.
-There has been a significant drop of error rates in the Cohesion Policy field since the year 2000, from roughly 11 % in 2000 to 3 % in 2017 and recently to 4.4 % in 2019; -For the entire programming period 2000-2006, the Annual Reports estimate an error rate of more than 5 %. Together with ancillary sources, a more precise estimate of the error rate is 11 %.
-In the programming period 2007-2013, when the AA were formally established, the error rate plunged from 11 % at the beginning to 6.8 % in 2012, which is a decline of more than 38 %. This is a clear indicator that the new governance framework, with the AAs at the core of it, proved to be effective.
The biggest drop was in from 2009 to 2010 (-3.3 %), i.d. three years after the outset of the programming period. This was exactly the time, when the AAs became actually operational. This has two reasons: o First, the first years in the programming period the Member States still deal with winding-up the "old" programming period and do not yet focus on projects to be supported under the current period. Accordingly, there are no material payment applications from the Member Stated to the Commission. These leads, in turn, to little or no material audit work of the AAs; o Since the AAs were first time officially set up in 2007-2013 period, the Member States purely needed time to create the administrative prerequisites to establish a new authority and to equip it with the necessary resources. Additionally, the AA needed to build the administrative capacity and to implement sound processes. All this needs some time.
-In the programming period 2014-2020, the error rate further decreased from 5.3 % at the beginning to 3 % in 2017, which is a decline of more than 43 %. The latest figures for 2019 show an error rate of 4.4 %, which represents a decline of nearly 17 %.
Taken the development until the year 2017, the figures indicate that the governance framework with the AAs at the core of it has further strengthened. Moreover, a clear learning curve of the AAs could be assumed. The AAs internal processes have become established and there has been an intense interaction with the Commission. Audits of the ECA further contributed to the lessons learnt and to adapt the procedures and methodologies.
However, the quite severe increase in the error rates from 3 % in 2017 to 5 % in 2018 put this assumption under question. For the 2018 budget year, the ECA identified in the 220 transactions refereeing to 15 AAs 36 quantifiable errors, which had not been detected by the MAs, and ultimately neither by the AAs (European Court of Auditors, 2019, pp. 140, 144 o Incorrect treatment of errors with an impact on the reported error rates. Since the number and impact of the additional errors the ECA revealed for transactions in the 2014-2020 period were higher than for transactions relating to closure of the 2007-2013 period (European Court of Auditors, 2019, p. 140), the AAs might still have problems with properly conducting their duties under the legal framework for the 2014-2020 programming period. This indicates a need for further capacity building measures.
In addition, the Commission concluded in its draft reports of its 26 carried out compliance audits in 2018 that in 81 % of the cases (21 audits) that the error rates reported in the AAs' annual control report were underestimated (European Court of Auditors, 2019, p. 140). In 2019 the situation deteriorated since the Commission concluded in all its draft reports for the 26 compliance audits it carried out in 2019 in 11 Member States, that the residual error rates reported in the audit authorities' annual control reports for the 2017/2018 accounting year were underestimated. In 15 cases, this led to revised residual rates above the 2 % materiality threshold (European Court of Auditors, 2020, p. 150) -Although, the long-term trend in the error rates unambiguous show a further decline, if and when the final goal to achieve an error rate below 2 % will be achieved is still unsure given the recent stagnation in the error rates.

Conclusions
The work of the national audit authorities constitutes an essential foundation for appraising legality and regularity, and the proper use of Cohesion Policy re-sources. In this way, the AAs help to protect the financial interests of the EU. In the context of that responsibility, the AAs are entrusted with a multitude of tasks, which, due to their specificity and diversity, call for significant requirements in terms of organisation and expertise (Bode, 2020, pp. 251-252).
Strengthened internal controls and accountability arrangements have contrib-uted reduce the level of irregular spending (European Court of Auditors, 2018a, pp. 24-25). The error rates reported by the ECA have significantly plunged, with recent stagnation. The long-term trend clearly shows a further decline. Howev-er, if and when an error rate below the 2 % materiality level will be achieved is hard to predict given the recent stagnancy and identified shortcomings in the work of the AAs. The weaknesses found for the year 2018 and 2019 reveals that further improvements are necessary to put more reliance on the AAs' work (Bode, 2020, pp. 253-254).
To what degree this reduction in the ineligible spending and the preventive and corrective measures of the Commission to protect the EU budget is attributable to the work of the AAs and their ongoing professionalization is unknown. How-ever, there is an indication that there is a correlation between the AAs' work and the robustness of the management and control system of the EU Cohesion Policy (Bode, 2020, pp. 253-254).
Further research needs to be conducted to approximate the impact of other factors influencing the error rates reported by the ECA (also the ECA calls the Commission for an analysis of the undetected errors, see European Court of Auditors, 2020, p. 155). Other factors might be first more endogenous, i.e. the impact of changes in the ECA's audit approach on the expressiveness of its error rate. On the other hand, exogenous factors also can have an influence on the error-proneness like: -clearer and simpler rules, -no "gold-plating" (i.e. over excessive rule setting addition to the EU legislation at national and/or/regional level), -funding mechanism (entitlements versus reimbursement of actual costs).
Moreover, other and alternative accountability measures to achieve legal, regular and performance-oriented policy implementation need to be further elaborated. For example -a clearer determination of accountability of the different actors of the Cohesion Policy down to the final beneficiary, -clearer transparency rules of the different actors, -focus on administrative capacity to assure an intended policy implementation, -incentive schemes to voluntarily achieve better result, might also be considered to be linked to an effective and efficient Cohesion Pol-icy with real impact (Bode, 2020, p. 42).