Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia

ABSTRACT We analyse European Union (EU) funds management systems performance using the cases of Hungary and Slovakia – two Central European countries with a common history and background, yet different concepts of EU funds management. We explore the role of administrative capacity and institutional stability as key factors affecting performance in both countries. We also provide evidence for the relationship between administrative capacity and non-compliance, measured by financial corrections, with a model robust to different programming periods and model specifications. Using the resulting baseline model applicable to all EU countries, we assess EU funds management system performance in Hungary and Slovakia.


INTRODUCTION
Several studies and evaluations by the European Commission 1 have confirmed that Cohesion Policy (CP) has positively contributed to economic growth (e.g., Pellegrini et al., 2013).However, this effect is conditional on national and regional institutions' capacity to design robust strategies, effectively allocate resources and administer European Union (EU) funds (Bachtler et al., 2014;Ederveen et al., 2006).It has also been suggested that it is essential for the efficiency of implementation systems to clearly define powers and responsibilities and set up well-functioning coordination mechanisms that are adequately documented and properly implemented (Musiałkowska et al., 2020).Significant variations in management and implementation structures are seen across the EU member states, attributable to different domestic systems of regional policy governance, managerial experience, the scale of EU funding, and the scope and orientation of the programmes (Ferry et al., 2007).
Accession to the EU opened eligibility for a relatively wide range of EU funding opportunities for Central and Eastern European member states (CEECs).The capacity to absorb the funds allocated to the countries depends on their economic situation, availability of resources to cofinance the operational programmes (OPs), and the applicants' ability to prepare good-quality projects.Member states also have to create management and control systems that ensure effective use of the funds.As noted in Seventh Report on Economic, Social and Territorial Cohesion, 'the quality of governance and institutions is a fundamental precondition for sustained increases in prosperity, wellbeing and territorial cohesion in the EU' (European Commission, 2017, pp. 136-137).The Eighth Report on Economic, Social and Territorial Cohesion stressed too that 'to work well, institutions need high levels of administrative capacity that in turn enhance the effectiveness and transparency of public spending, including of EU funds' (European Commission, 2022, p. 214).
This research provides a comparative analysis of EU funds management systems in Hungary and Slovakiatwo Central European countries with common history and background, yet different concepts of EU funds management.To evaluate the performance of the CP implementation regime, we have focused on both qualitative and quantitative aspects, such as government changes, the strength of coordination and management authorities, administrative capacity, irregularities, and financial corrections.
The remainder of the paper is structured as follows.The next two sections provide a literature review on the concept of administrative capacity and its relation to CP managements system performance, which is followed by a section on the research sample and methodology used for analysis of EU funds management systems.The discussion and results section provides the case studies' comparisons and analyses institutional establishment, administrative capacity, irregularities and financial corrections.In this section we also establish an administrative capacity and compliance performance evaluation model.The last two sections highlight the theoretical and managerial implications and provide the concluding discussion of our main results.

ADMINISTRATIVE CAPACITY
Although the concept of administrative capacity may seem intuitive, a formal definition is somewhat elusive.As Addison (2009, p. 1) states in a review paper on the subject: whilst the concept has appeared intermittently over time, it is not established at the centre of any research field.The concept serves different purposes; cross-references among different sub-disciplines are rare; and specific definitions, labels and means of observing it vary.
For this paper we find the concept by Boeckhout et al. (2002, p. 2) particularly useful, defining the term as: the ability and skill of central and local authorities to prepare suitable plans, programmes and projects in due time, to decide on programmes and projects, to arrange the co-ordination among principal partners, to cope with the administrative and reporting requirements, and to finance and supervise implementation properly, avoiding irregularities.
To operationalize the concept, we use a proxy for administrative capacity in the form of the 'government effectiveness' index from the World Bank's Worldwide Governance Indicators (WGI) dataset.The indicator has been used in studies on EU funding (e.g., Mendez & Bachtler, 2017;Van Wolleghem, 2022) and summarizes the quality of the administration for policy formulation and implementation, and quality of public services.

COHESION POLICY MANAGEMENT SYSTEM PERFORMANCE: THE ROLE OF ADMINISTRATIVE CAPACITY
Cohesion Policy implementation systems are complex, involving several levels and numerous implementing actors operating in an intricate framework of conditions.The relationship between inputs, outputs and outcomes may not be sufficiently clear and measurable, and data availability may not be guaranteed.In general, the three main requirements for implementing CP include compliance with EU law (regularity), timely delivery (absorption), and capacity to define strategies, select and deliver sound investments using EU funds to achieve policy objectives (effectiveness).These factors constitute the basis for evaluating the EU funds management systems performance.
3.1.Non-compliance with rules and capacity to achieve policy objectives EU compliance research has expanded over the past decade.However, studies have devoted scant attention to the financial dimensions of compliance regarding infringements of EU rules on the regularity and legality of EU expenditures (Mendez & Bachtler, 2017).
A judgment of non-compliance with EU CP could be the result of an error, 2 irregularity 3 or fraud 4 (Hajdu et al.,  2017).A financial correction has the consequence of a total or partial withdrawal of EU funds from either a programme or a project.In both cases, the consequences are the same: the European Commission ceases the release of EU funds for the amount affected by the financial correction.Thereafter, either the member state or the project owner would be responsible to finance that part of the programme or project.
The European Commission's annual activity reports, the European Court of Auditors (ECA) audit reports and analysis by Mendez and Bachtler (2017) all conclude that administrative capacity impacts financial compliance.When implementing EU funds under shared management, it is primarily the member states' function to establish management and control systems (Ferry & Polverari, 2018), define strategies and select and deliver sound investments.In addition, member states must detect, investigate and sanction irregularities (Tátrai & Vörösmarty, 2020).The decisions on the principal elements and nature of the management structure (centralized versus decentralized, the number of authorities, coordination function, etc.) fall under the member states' responsibility.
While the management and control system's stability is one of the critical factors for successful implementation, the reality is that these structures are not set in stone, nor are the relationships between implementing bodies.The administrative architecture described in programming documents often evolves due to political factors (including government changes following elections) or performance factors (usually due to reviews that suggest the current setup is inefficient or insufficiently effective).Effective coordination among managing authorities and different government actors (Ambrusz et al., 2018) participating in the implementation process is essential to optimize public investment outcomes (Organisation for Economic Co-operation and Development (OECD), 2020).

Absorption performance
From a theoretical point of view, the cohesion and structural funds were created to address a primary objectiveto achieve economic, social and territorial cohesion by reducing regional disparities in the EU.As European integration progressed, the funds grew in importance and volume to achieve the primary policy goal.Although the concept was sound in theory, early experiences pointed out several problems in the implementation of the EU Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia 705

REGIONAL STUDIES
CP, such as a struggle to attain high absorption levels and efficiently use the funds, and to address emerging fraud.
As EU member states operate in a harmonized legal environment with binding legislation translated into national legal frameworks, the explanation for the variation had to be sought in other factorsnamely, the national implementation framework itself.
Several studies confirm that administrative capacity can explain the variation of performance in implementing structural funds (e.g., Bachtler et al., 2014;Milio, 2007).More generally, weaknesses in administrative capacity and the quality of public administration introduce barriers to CEECs in implementing EU policy (Tiganasu et al., 2018).However, some studies question whether macroeconomic capacity or political decentralization represents the driver responsible for absorption performance (e.g., Tosun, 2014).Surubaru (2017) argued that it is necessary to go beyond administrative capacity and strengthen the relationship between political and administrative governance.Surubaru provides evidence that Bulgaria's higher absorption performance over Romania could be interpreted as evidence of how political factors influence administrative capacity.Incaltarau et al. (2020) have shown that government effectiveness and combating corruption boost the absorption of structural and cohesion funds, especially in CEECs that have had a shorter experience complying with EU policies.Other determinants of absorption performance include fiscal decentralization and territorial economic preconditions (Kersan-Škabić & Tijanić, 2017), including the ability to co-finance the programmes and projects.

Effectiveness
Extensive literature has covered CP's effectiveness, usually employing formal econometric modelling focusing on the evolution and lowering of regional disparities and contributions to regional economic growth.Whether CP funds induce regional convergence has been studied since the 1980s and 1990s (Dall'erba & Le Gallo, 2008; for other periods, see Becker et al., 2013).As Rodríguez-Pose and Novak (2013) revealed, CP policy is changing.Greater internal monitoring and external scrutiny have been followed by an increase in policy effectiveness over time.The empirical evidence for positive effects has been provided, for instance, by Maynou et al. (2016), Pellegrini et al. (2013), Fiaschi et al. (2018) and Crescenzi and Giua (2020).Several macroeconomic models have confirmed that the impact is greatest in the main beneficiary countries and spillover benefits mainly affect member states with strong trade links to cohesion countries (European Commission, 2017).
Recent research has acknowledged that CP's impact depends on 'conditioning factors', which include macroeconomic and financial absorption capacity, fiscal and political decentralization (Kersan-Škabić & Tijanić, 2017), institutional and structural regional factors, interaction between CP and other policies with territorial impact, local political factors (Crescenzi & Giua, 2017), sufficient human capital and goods institutions (Becker et al., 2013).
Italian and Spanish case studies show that CP effectiveness depends on the institutional characteristics of domestic authorities involved in the management of funds and policymakers' decisions (Casula, 2020).Having sound institutions may prevent distortion caused by certain interest groups, avoid clientelism and corruption, hinder rent-seeking practices and beneficially affect CP effectiveness (Breidenbach et al., 2019).
The long-term effectiveness of EU CP was studied by Bachtler et al. (2017), who have demonstrated positive effects of European Regional Development Fund (ERDF) funding but highlight the need for a greater concentration of resources and coherent strategies, noting that many of the problems have been related to administrative capacity.Rodríguez-Pose and Garcilazo (2015) have estimated a panel econometric model on EU regions, suggesting that after some level of funding has been reached, government quality becomes more important for CP efficiency than additional spending.Gorzelak (2015) emphasized a crucial distinction between administrative capacity as a precondition for achieving high absorption rates and potential longer term effects on socioeconomic efficiency.The authors specifically note that in the CEECs, 'substantial amounts of EU funding are being spent against a background of traditionally weak policy management and implementation systems and constrained domestic public expenditure' (Gorzelak, 2015, p. 49).

RESEARCH SAMPLE
Our analysis is performed on Hungary and Slovakia, two neighbouring countries who simultaneously joined the EU in 2004.They share a similar development level, and macroeconomic figures mainly illustrate similar trends (see Table A1 in Appendix A in the supplemental data online).The two countries are significant net beneficiaries of European Structural and Investment Funds (ESIF), with a total allocation in 2014-20 of €2995 per capita in Slovakia and €2552 in Hungary, with €2289 per capita in Slovakia and €2217 in Hungary in the 2021-27 period.CP funding is similarly important source of development funding in both countries, where it represents approximately 55% of public investment (see Figure A1 in Appendix A in the supplemental data online).The European quality of government index 5 is low in both countries (for EU countries, Hungary is ranked 23rd and Slovakia 20th) with a noticeable decline in Hungary since 2017 (Charron et al., 2021).According to European Commission (2017), government effectiveness and citizens' confidence in institutions diminished in Hungary and increased in Slovakia between 1996 and 2015.Moreover, the perceived level of corruption is high in both countries (Transparency International, 2021).So, comparative analysis of these two countries with similar weak level of government effectiveness may identify reasons for the differences in their CP performance.
Our interest in the case of Hungary and Slovakia is based also on the fact that while they both share common history and cooperate in terms of policy under the Visegrád group of countries, they exhibit marked differences in territorial and administrative structure at the regional level as well as independence at the local municipality level.In Slovakia, fiscal decentralization has strengthened the role of self-government in regional development.Despite decentralization in national policy, EU funds management remained centralized until the period 2021-27.
As part of the harmonization efforts for the EU regional development system, decentralization initiatives in Hungary from 2000 to 2010 saw the establishment of regional development councils and agencies to oversee the regional development system.However, the regions, their bodies and responsibilities disappeared totally in January 2013.EU funds management systems in both countries have been influenced by different historical path, where Hungary moved from decentralized to centralized system, Slovakia has always had a centralized system.

METHODOLOGY
This paper uses quantitative and qualitative data to present a descriptive and comparative analysis of institutional systems and their performance when implementing CP programmes in Slovakia and Hungary.Given the operational and structural complexity of CP procedures, we focus on the most relevant, quantifiable and measurable aspects of CP implementation to evaluate the systems' overall performance.In the absence of more detailed information on other indicators for this analysis (e.g., lack of recent, regular and comparable data on the cost of the implementation), our examination is aimed towards the following: . Political and institutional changes and amendments to the coordination arrangements. .

Administrative staff as an input indicator.
. Absorption rate as an output indicator.
. Irregularity and financial correction rates as a performance indicator.
Our analysis is based on regulations and publicly available data presented by the European Commissionevaluations, guidelines, annual management and performance (AMPR) reports, the European Commission's annual reports on the protection of EU financial interests (PIF) 2014-18, reports from the European Commission to the European Parliament, the Council and the Court of Auditors as well as working papers and studies. 6Additional data sources include European Court of Auditors (ECA) (2014-18) reports, publicly available and specifically required documents and data from the Hungarian and Slovak Government institutions and European Commission (2020).
As the review in the preceding sections points to the importance of the human resources part of administrative capacity in general, and given the evolution of EU fund management systems in Hungary and Slovakia in particular, we formulate our first research question in the following way: Question 1: Did organizational changes have had an adverse effect on administrative staff?
As the question is qualitative in nature, we address it by combining the information on the relevant EU and national documents with in-depth interviews with 17 officials directly involved in the implementation system (managing authorities, intermediate bodies, and control authorities) and nine local stakeholders (mayors and project managers) in 2019-20 to gather information on the impact of organizational changes on implementation, including information on the problems and risks affecting implementation capacity.The in-depth interviews have provided an additional insight for our analysis.
Question 2: Has administrative capacity had a significant relationship with non-compliance with EU CP funding rules?
To evaluate the second research question as a formally testable hypothesis, we have performed an econometric analysis of non-compliance, measured as a rate of financial corrections to contributions by estimating fractional logistic regression models, as well as beta regressions of Ferrari and Cribari-Neto ( 2004), using data from the AMPR reports for EU contributions and financial corrections, WGIs, country scores from the regional authority index (RAI) of Hooghe et al. (2016) and government expenditure data from Eurostat (2021).We use the quantified econometric models to evaluate the performance of EU fund management systems, where we compare the predicted level of non-compliance based on the practice in other EU member states with the actual financial corrections incurred in Hungary and Slovakia in 2000-06 and 2007-13 programming periods.

Political and institutional changes
Both Slovakia and Hungary in the period 2014-20 operate centralized systems for managing EU funds; however, the historical path taken by the two countries has been rather different.
In the Slovak Republic (SR), the Ministry of Construction and Regional Development coordinated the regional policy of SR and EU funds from 2004 to 2010.The involvement of the subnational level in CP management was limited.The first programming period was characterized by gaining experience in drawing down the funds.After 2010, the coordination between the regional policy of SR and the CP became significantly more difficult.The regional policy competencies were divided between two ministries and the Government Office.Several amendments were made to EU funds coordination arrangements due to political influences.Transferring Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia 707 REGIONAL STUDIES the agenda between public administration bodies resulted in increased staff turnover and delays in the implementation of EU funds (Government Office of the Slovak Republic, 2015).Coordination of the funds remained a weak feature in Slovakia in 2007-13 (Position of the Commission Services, 2012) together with other related problems, such as the large number of implementing bodies, high level of political influence and delays in public procurement (Applica et al., 2016).In 2016, coordination of EU funds was transferred to the Office of the Deputy Prime Minister for Investments and Informatisation, followed by a transfer of coordination of regional development policies and support for less-developed districts in 2019.That consolidation created an opportunity for better coordination.The latest election in 2020 brought another change via deepening centralization to secure higher efficiency and effectiveness.
In Hungary, regional policy decentralization was a basic element of political democratization and pluralism until 2010.Several partly decentralized domestic funds were available targeting various cohesion objectives.To manage the EU funds, Hungary set up a parallel and centralized system.Parallel institutions existed for national and EU funds management and the management of EU funds operated outside the traditional Hungarian ministerial system even though funds followed sectorial logic.From EU accession in 2004 until 2010, coordination was first delegated to the Prime Minister's Office (PMO) then to the newly established National Development Agency (NDA).After 2010, a strong centralization process began (Pálné Kovács et al., 2017).The political government dimension has been markedly strengthened, and all the relevant decisions have been made at the governmental level.The role of territorial actors has been steadily declining.After 2010, and particularly after 2012, the approach to the use of EU funds changed in Hungary. 7In 2012, Hungarian policymakers started to prioritize developments that the Hungarian government considered necessary, and for which it was possible to obtain EU financing.

OPs and power of the managing authorities
Almost all OPs within Slovakia are centrally managed by several ministries (see Table A2 in Appendix A in the supplemental data online).During the first programming period 2004-06, Slovakia did not prepare regional programmes.Three sectoral OPs and OP Basic infrastructure followed the national sectoral approach.As only the Bratislava region was eligible for Objectives 2 and 3, 8 separate programmes followed the EU regulations.In the 2007-13 programming period, besides the sectoral OPs, there was a regional OP and an OP for the capital region.The higher number of OPs and increased number of intermediate bodies led to fragmentation.There were many changes in managing authorities in 2007-13, influenced by the elections and political priorities with consequences of delays in implementation of EU funds and fluctuations in employees (Government Office of the Slovak Republic, 2015).
In 2014-20, Slovakia had a lower number of OPs and more continuity in managing authorities, although deficiencies and considerable bureaucracy remained.Slovakia has only one 'Integrated regional development programme (IROP)' for all four NUTS-2 regions.The regions and eight regional cities serve as intermediate bodies for several priority axes in the IROP.Only 3.7% (108 out of 2942) specific jobs in coordination, management, implementation, control and audit of ESIF were created at the self-government level in 2019 (Government Office of the Slovak Republic, 2020).
The most problematic OP with low absorption and high automatic de-commitments was 'OP Research and Innovation 2014-2020'.To eliminate the risk of another round of de-commitments, the programme was merged with the most effective programme in terms of absorption, 'OP Integrated infrastructure', in December 2019.Changes in intermediate bodies within programming periods were influenced mainly by political changes and a stronger focus on the urban development in the period 2014-20 (see Table A2 in Appendix A in the supplemental data online).
During the period 2004-06 in Hungary, one regional OP was implemented in addition to the sectoral OPs.The managing authorities 9 all worked at the central level in the line ministries under the coordination of the PMO.However, the involvement of multiple agencies 10 with overlapping authorities created a complex system with difficulties in maintaining a nationally consistent approach to the interpretation and employment of regulatory requirements.
In the 2007-13 programming period, a complete overhaul of the system occurred.The new National Development Agency established at the central level hosted all managing authorities, and the new national regulations laid down unified standards for programme delivery.Strong centralization and standardization were behind the reorganization.However, more than 25 intermediate bodies and overlapping regulations led to inconsistencies and fragmentation.
A mid-term revision, together with the political change in 2010, included reprogramming and institutional redesign.Coordination and control functions of the managing authorities had to be strengthened, and provisions increasing efficiency had to be introduced.As a final change, towards the end of the programming period, the supervisory authority over the NDA was transferred from the Ministry for National Development to the PMO.
The start of the 2014-20 programming period again brought substantial changes in the CP institutional system.The NDA was abolished in January 2014, and its functions were distributed between the specialized ministries and the PMO.Managing authorities were again transferred back to the line ministries.The PMO was also entrusted with central coordination tasks. 11Intermediate bodies were dismantled, and their duties were integrated into the ministries responsible for managerial authority functions.An exception was the Hungarian State Treasury, which acts as an intermediate body for the Territorial and Settlement Development OP (see Table A3 in Appendix A in the supplemental data online).
When comparing the programming periods, the number of OPs in both countries (and thus managing authorities) was the highest in 2007-13 (11 in Slovakia and 15 for Hungary for objective 1 of CP).Therefore, the need for coordination during this period was the greatest.The number of intermediate bodies has been high in all programming periods in Slovakia (with a maximum of 27 in the period 2014-20) and has permanently declined in Hungary (only one in the 2014-20 period).

Administrative staff fluctuation and absorption rate
When focusing on the human resources aspect of administrative capacity in relation to our Question 1, a significant problem in both countries was identified as the lack of administrative employees with relevant experience in the labour market, as well as the absence of a network for continuous training for the existing administrators and recruits in the first years.Insufficient staff and high turnover resulted in non-compliance with deadlines in Slovakia (Ministry of Construction and Regional Development, 2006).Despite a significant amount of technical assistance in 2007-13, problems with staff remained along with a significant administrative burden, corruption and issues with procurement procedures and project selection (Position of the Commission Services, 2012).Despite the seemingly increasing staff up until 2018 (Table 1), the period was characterized by high fluctuation, reaching its peak in Slovakia in 2016 (20.3%).
Based on the interviews and opinions of former employees in Hungary and reports of the Government Office of the Slovak Republic, the reasons for the high turnover rate in both countries are generally rooted in the lack of administrative staff, which led to the creation of cumulative functions and increases in the workload of employees who were not adequately remunerated in relation to their performance.Inadequate incentive factors, insufficient support for human resources development through relevant education, and frequent organizational changes constitute other causes for high employee turnover (see Figure A2 in Appendix A in the supplemental data online).
In Slovakia, a KPMG (2014) study also identified corruption, conflicts of interest and unsatisfactory performance in protecting the EU financial interests in 2007-13.Moreover, a relatively high number of position holders were found to have been recruited for temporary government service, and problems concerning the selection process were revealed as well.Slovakia demonstrated its highest turnover rates in 2010 and 2016, which correlates with parliamentary elections.In the case of Hungary, the turnover was slightly higher than usual in 2016, when a substantial bonus was paid based on successful programme closure in the preceding year to compensate for the heavy workload of 2015.Thus, many employees intending to leave had waited until the payment of this bonus, then instantly resigned from their jobs.
Finding qualified personnel remains a problem in Slovakia, although the reason differs from the circumstances prevalent at the early stage of EU funds implementation.A possible explanation is offered by relatively lower financial motivation than in the private sector.Most of the positions are situated in the Bratislava region, where average wages are higher than the rest of Slovakia.
In Slovakia, the institutional system for implementing the ESIF in 2014-20 has been based on the experience gained in managing the previous period's OPs.To ensure smoother implementation with lower irregularities, an 'integrated network of information and advisory centres Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia 709 for ESIF' was established in 2016.In Hungary, there was a complete reorganization and restructuring, attempting to save accelerate implementation.Hungary introduced strong salary incentives for employees and significant pre-financing for project owners.
The results in Table 1 show that the Hungarian administration manages higher amounts of sources, €25 billion (Hungary) versus €15 billion (Slovakia).In addition, Hungary had higher amounts of payments, €7.85 billion (Hungary) versus €3.69 billion (Slovakia), and almost the same number of employees 2962 (Hungary) versus 2902 (Slovakia). 12Slovakia belongs to countries with the lowest share of ESIF spent (as of July 2021, only 45%), while Hungary is average, with more than 59% spending.In the previous periods performance of both countries has been satisfactory as an absorption rate of 99% was achieved in 2004-06 and 96.08% in 2007-13 in Slovakia.The high number of OPs and bodies involved in the implementation and other implementation problems such as the late adoption of programmes, bureaucratic procedures and difficulties in public procurement influenced slow absorption in 2007-13 (Applica et al., 2016), accelerating only in 2015 following political pressure.In 2004-06, Hungary achieved a 92% total absorption and 99.64% in the period 2007-13.
For a proper evaluation of EU funds management system performance, administrative costs are also relevant.Based on Polverari et al. (2020, p. 12) Hungary technical assistance (TA) allocations of ERDF, CF and ESF, which are closely linked to the funds' implementation are the lowest within the EU (0.9% of all funds, €193 millions) and Slovakia is one of the countries with the highest allocations (4% of all funds, €545 millions).Both countries have allocated the most resources to management interventions. 13In addition to separate axes of technical assistance within each OP, Slovakia also had a separate 'OP Technical Assistance 2014-2020' whose goal was to create highly professional and effective support for managing and implementing ESIF.Comparing this structure of expenditure with the Hungarian 'Public Administration and Civil Service development OP' using Polverari et al. results (2020, p. 85), Slovakia has allocated the most TA resources on personnel (55%), whereas Hungary prioritized organizational structures and resources (46%).

Irregularities, financial corrections and their causes
Looking at the irregularities and financial corrections in the 2000-06 and 2007-13 programming periods, it could be concluded that in the first period after the accession, Hungary was quite successful with the implementation of the EU funds.In the 2000-06 period, Hungary performed in accordance with the expectations and rules of the EU with a low financial corrections rate, in contrast to Slovakia (see Figure A3 in Appendix A in the supplemental data online).
However, in the 2007-13 programming period, both countries worked with definitely higher financial correction rates (see Figure A4 in Appendix A in the supplemental data online).In the beginning, Hungary improved its compliance level in terms of EU regulations and objectives.After 2010, however, and especially after 2012, the trend changed.Hungary was ranked third behind Slovakia with the highest financial corrections rate at the end of the 2007-13 period, which has consequently resulted in a higher national co-financing requirement.
Further examining the irregularities and financial corrections in the current period, we can observe additional new elements.According to the figures on reported irregularities, we can see that the financial implication of irregularities linked to the reported data is significantly higher in Slovakia than in Hungary.Although the number of irregularities has declined in Slovakia, the amount affected has been high.Furthermore, Slovakia stood for 78% of all fraudulent irregularities in the EU in 2018, whereas in 2014, it was only 0.4%.The high fraud detection rate in 2017 and 2018 is due to a few cases (four and two) with exceptionally high amounts involved (PIF report 2017 and2018).
Slovakia has taken several measures to deal with corruption, conflict of interest and the high volume of irregularities, such as enhanced transparency via the online available register of beneficiaries; central register of contracts; new legislation of protecting whistleblowers; strengthened controls of the Office for Public Procurement and the adoption of an action plan to enhance transparency and simplify the implementation of the ESIF.Slovakia has also introduced an 'Anticorruption Policy of the Slovak Republic for 2019-2023'.According to the PIF Report 2017, a national antifraud strategy exists only in ten of the 28 member states, including Hungary.
Table 2 shows the reported irregularities.Lower Hungarian figures could be explained by both the lower rate of cases and reduced reporting levels.The methodical issue would be evident if we compare the financial corrections data from the past three years linked to the two countries' implementation performance, as presented in Figure A5 in Appendix A in the supplemental data online.
There are apparent differences between the PIF report and AMPR reports data, between the financial corrections imposed by Commission Decisions, and what financial corrections the member states themselves applied in the accounts following their operations audits.Hungary, compared with Slovakia, seems to be reporting weakly, but according to the Commission's level of financial corrections, implementation is progressing alongside many irregularities.
In 2019, Hungary accepted a 10% flat-rate financial correction on all contracts in exchange for the continued financing of eight OPs, which had been in reservation since 2018 (so payment was stopped by the European Commission).The reason was severe deficiencies in the management and control system's functioning concerning the control of public procurement procedures identified in a horizontal public procurement audit.performance evaluation model We established a formal model for financial corrections on the EU-28 level based on the 2000-06 and 2007-13 programming periods (Question 2).The estimation of fractional logistic models explaining the rate of financial corrections allows for inferences on the impact of administrative capacity and an evaluation of Hungary and Slovakia's compliance performance given the performance of other European countries (Question 3).
In our approach, we extend on the model used by Mendez and Bachtler (2017), with the dependent variable in our baseline model represented by the ratio of financial corrections confirmed to the contribution amount: ERDF + ESF contribution for the period 2000-06, and ERDF + CF + ESF for the 2007-13 programming period.
We consider the set of WGI as explanatory variables.The WGI consists of indicators on voice and accountability, political stability and absence of violence or terrorism, government effectiveness, regulatory quality, rule of law and control of corruption.A country score from the RAI of Hooghe et al. (2016) is also used as a control variable.All dependent variables are averaged over the respective programming periods.Mendez and Bachtler (2017) introduce two additional variables in their analysis -'regional programming' and 'goodness-of-fit', which might control for additional factors affecting compliance.As for the programming variable, the authors note that the decision taken at a subnational level might also be relevant.While allocating some of the funds for programmes run at the regional level might prove beneficial as opposed to a more central form of governance, particularly in the ability to tailor the programmes to local specific conditions and needs, such programme structuring might also lead to additional implementation and management problems.As for goodness-of-fit, the authors utilize the idea explored by Haverland (2000) of a potential misfit between domestic status quo and policy vis-à-vis EU policies and objectives.In countries where the national and EU objectives are not completely harmonized, there is a stronger potential for creating conflicts and frictions resulting in non-compliance.Additionally, the divergence in policy may be related to the size of the share CP funding hold in public investments, with larger share making closer alignment more likely (Davies & Polverari, 2011).
In our case, we have opted to not include the variable on regional programming and only include goodness-of-fit (defined as the ratio between EU contribution to government gross fixed capital formation, obtained from Eurostat) as a robustness check, favouring a more parsimonious model specification given the limited sample size and prior research results.Most importantly, both variables were shown to be statistically insignificant when explaining the non-compliance in EU CP on the full sample covering the 2000-06 period (Mendez & Bachtler, 2017).
Methodologically, the constructed models have to consider the nature of the dependent variable.As the financial corrections are expressed as a ratio to the total contribution, the values are restricted to the unit interval, making the usual methods, such as ordinary least squares (OLS) regression unsuitable.Instead, we employ a fractional logistic model and beta regression of Ferrari and Cribari-Neto (2004).Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia 711

REGIONAL STUDIES
The beta regression is a powerful tool with great flexibility in modelling due to the beta distribution properties, which allows for a wide range of shapes for the overall distribution.However, the method is only appropriate for data strictly inside the unit interval.Since several countries (such as Finland and Malta for the 2000-06 period, and Finland, Denmark and the Netherlands for the 2007-13 period) had zero confirmed financial corrections as of 2018, these observations would have to be dropped.Instead of disregarding these observations, a supplementary fractional logistic regression model is used, which accommodates the boundary value cases and serves as an additional robustness check.
Figure 1 shows that the WGI indicators are mutually highly correlated.They are also negatively correlated with the financial corrections ratio, providing evidence that better governance leads to fewer financial corrections.The high correlations also mean that it would be prudent not to include all WGI explanatory variables in the final regression model.Lower model dimensionality is also preferred due to the smaller available sample size.To choose a parsimonious model representation, our baseline model only includes the WGI variable on government effectiveness.Mendez and Bachtler (2017) also justify its use in their analysis as a proxy for administrative capacity, as the 'government effectiveness' index is used on the 'capacity of governments to formulate and implement sound policies effectively'.
Table 3 shows the results of the logistic models.First, we may see that the results of both modelling approaches (beta regression and fractional logistic regression) come to the same conclusions, higher administrative capacity is accompanied by a lower rate of corrections confirmed; the relationship is also always statistically significant (hypothesis corresponding to Question 2).While the control variable of regional authority is insignificant in the beta regressions, the converse is valid for the logistic models.Here, the results may be caused by a slightly larger sample in the latter case, where also best-practice countries with zero confirmed corrections are included.The RAI variable sign is positive in all cases, suggesting that a higher rate of financial corrections accompanies a higher regional authority.
To assess parameter stability and the influence the countries under analysis have on the overall results, we reestimated both models without Hungary and Slovakia.The results are robust to their exclusion.Table 4 provides a comparison of actual financial corrections confirmed for the two countries and two programming periods, as well as in-sample (models with all countries) and out-of-sample forecasts (models with Slovakia and Hungary excluded).The forecasts may be interpreted as expected financial corrections levels obtained if the relationship was the same as in the other EU countries.The results suggest that financial corrections were lower than expected only for Hungary during the 2000-06 period (Question 3).
The stability of the financial corrections' relationship with the explanatory variables is explored in Figure 2, where separate bivariate models have been fitted for each variable and programming period.The downward slope in the case of administrative capacity (left) is notable in both periods.
The results support the proposition that building better administrative capacity advances better compliance, the latter measured as a lower share of financial corrections   4), ( 7) and ( 8) include the 'goodness-of-fit' explanatory variable.The sample in models (2), ( 4), ( 6) and ( 8 Administrative capacity and EU funds management systems performance: the cases of Hungary and Slovakia 713 REGIONAL STUDIES on contributions.This outcome is in line with the results of other studies highlighting the role of administrative capacity.Although a regression analysis does not warrant an interpretation of direct causality, both the inferential statistics in Table 3 and visual inspection in Figure 2 seem to support investment into administrative capacity as a measure of enforcing compliance. Several estimations have been performed as robustness checks to verify our results.The fractional logistic model serves as an alternative model allowing for the inclusion of countries with no financial corrections incurred, which the beta regression is unable to accommodate.In the fractional logistic models, we use robust standard errors estimators to account for possible heteroscedasticity.While relying on asymptotic properties of such estimators in small and moderate size datasets might not be warranted, this robustness check does indicate validity of the main results, particularly when all WGI regression coefficients remain significant and negative.To account for the panel nature of the dataset, estimations of the baseline model for individual programming periods separately are included in Table A4 in Appendix A in the supplemental data online.In all but one case the results support our main finding.In one particular case, the WGI coefficient is not  Sources: AMPR reports (European Commission, various dates), worldwide governance indicators (WGI, 2020) (government effectiveness) and regional authority index (RAI) (Hooghe et al., 2016).
714 Eva Výrostová and Györgyi Nyikos significant, but even then, the estimated coefficient is negative.
Despite the statistical evidence, there were several observations (Figure 2) where the quality of government measures has been relatively high, accompanied by high levels of irregularities, seemingly contradicting the previous interpretation.The effect is more clearly seen in the 2000-06 programming period, rather than the 2007-13 period, which might suggest that the relationship is dynamically evolving in time.Furthermore, the outliers in the relationship consist mainly of two groups of countrieslarger or more developed countries (e.g., Spain, Italy and Greece) and CEECs (e.g., Hungary and Slovakia).
According to the models, the results are as follows: . Administrative capacity (WGI government effectiveness) has a significant negative relationship with financial corrections. .Hungary did better than expected (given the practice in other EU countries) in 2000-06, but worse in 2007-13.Slovakia has always performed worse than expected. .The relationship of the regional authority and government effectiveness with non-compliance (financial corrections) is robust across models and programming periods.

THEORETICAL AND MANAGERIAL IMPLICATIONS
Our results show that the Hungarian administration manages higher amounts of sources with almost the same employee numbers as Slovakia; thus, the Hungarian system seems to be more efficient from this point of view.
In further evaluation, the performance of the implementation systems through the examination of the absorption and compliance showed a more multilayered picture.
Based solely on the absorption rate, both countries' performance has been satisfactory, but the current Hungarian absorption level is higher.
Evaluating the compliance of the implementation with the funding rules shows that in 2000-06, Hungary performed in accordance with the expectations and rules of the EU with both a relatively high absorption rate and low irregularity and correction levels.Slovakia achieved a higher correction rate.However, in the 2007-13 programming period, both countries worked with definitively higher financial correction rates.Hungary's approach significantly changed after 2012.Since then, compliance has no longer constituted a priority for the Hungarian government, which has led to higher co-financing consequences.Slovakia started with a higher correction level but has taken several measures to deal with corruption, conflicts of interest and high irregularities, such as enhanced transparency.For the past few years, mainly because of public pressure and public control, the situation has slowly improved.With the new policy and decreasing administrative staff, Hungary was third-ranked behind Slovakia with the highest financial corrections rate at the end of the 2007-13 period.Thus, we can conclude that compliance has not been considered an especially important factor despite the serious consequences of financial corrections.
Organizational changes, rooted in political transitions, have negatively influenced staff turnover and led to the loss of qualified employees, a lower level of absorption and increased irregularities and financial corrections (Question 1).
At first glance, the Hungarian EU funds implementation regime's performance seems to be better.Higher absorption has been achieved with proportionally less human power and technical assistance sources.However, if occurrences of irregularities and financial corrections are evaluated as an additional layer and as indicators of regular and proper implementation, the two countries were at the same level at the end of the 2007-13 programming period.Based on the data available and presented for the 2014-20 programming period, figures of irregularities and financial corrections imposed by Commission Decisions have deteriorated significantly in both countries.Based on the data and results of our research, performance in both countries needs further improvement.
By using the available information on governance, administrative capacity and EU funds implementation, our model confirms the results of our qualitative and quantitative research and shows that government effectiveness (WGI) as a proxy of administrative capacity has a significant negative relationship with financial corrections (noncompliance).According to the model results, Hungary did better than expected (given the practice in other EU countries) in 2000-06, but worse in 2007-13.Slovakia has consistently performed worse than expected.

CONCLUSIONS
The CP implementation regime is rather complex in balancing the absorption-regularity-effectiveness during the entire programming period.The evaluation of performance based on absorption alone is thus not sufficient, and the consideration of additional aspects, such as corrections, administrative costs, and coordination of institutions is warranted.The final data on the absorption rate and financial corrections for the 2004-06 and 2007-13 programming periods show that the focus on using all allocated sources came with the cost of high financial corrections for both countries.The absorption requirement has been accompanied by irregularities in the management of the funds, which have led to financial corrections.In our research, we investigated possible causes and how organizational changes have affected administrative capacity in implementing EU funds.We point out that despite political reorganizations, the stability and professional functioning of the CP implementation system remain crucial.Our research revealed the destabilizing impact the political changes exert on administrative staff.Even in view of the political risks and opportunities carried by the outstanding ESIF allocations, current political oversight arrangements need to be reconsidered to assess the benefits of introducing institutional changes, possibly jeopardizing the regular and effective absorption of resources.Solutions that improve performance should be pursued, strengthening and stabilizing administrative capacity to unlock development potential and address current unprecedented challenges.
This paper confirms and complements to previous research in several ways.By analyzing the performance of EU funds management systems in two neighbouring countries, using Hungary and Slovakia as examples, the results of our research show that despite many similarities and common history, the countries' performance and dynamics differ and vary between programming periods.Our research explores the reasons for these differences.While our formal empirical model employed is based on EU-wide data, our analysis provides a more in-depth look into the background and causes of the overall success and failure of fund management systems, as observed in aggregate indicators such as absorption ratios and levels of financial irregularities.The model clearly illustrates the specificities of the two countries under study and the reasons for the differences.
Second, we confirm and extend the results of prior studies, such as Mendez and Bachtler (2017).The extension over two programming periods allows us to demonstrate that while the system, capacity and experience gained in one period might form a sound basis for subsequent EU fund management performance, this is not necessarily the case in practice.Our research shows that the two countries under review are not building on established and functioning institutional capacities, but are constantly restructuring their implementation systems.In terms of financial corrections, which can be considered as one of the result indicators of a funds management implementation system, it is noteworthy that Hungary has escalated from rank 15 for the period 2000-06 to three in 2007-13 period (behind Slovakia).This means that the rate of corrections has increased significantly for the 2007-13 period, despite the fact that its professional experience, capacity and effectiveness should have increased.To understand the nature of the transition and factors behind this qualitative change, we provide a detailed background on the history, organization and factors affecting EU funds management implementation, showing that institutional stability and administrative capacity seem to be particularly relevant.Our model and evaluation results show that administrative capacity (or its proxy, the WGI government efficiency indicator) is significantly negatively related to financial corrections.The results show that the relationship between government efficiency and non-compliance measured by financial corrections is stable across models and programming periods.However, a problem arises in case of lack of a 'good fit' between EU and domestic policies and objectives, which also leads to a higher rate of financial corrections, as was the case in Hungary after the change in policy to the implementation of CP in 2012.No such policy change has occurred in Slovakia.In this context, ensuring administrative capacity, governance efficiency and a 'good fit' between EU and domestic policies and objectives remain a priority.
The findings in this study open the door to further research, primarily on (1) the effects of institutional stability and administrative capacity on EU funds management system performance in other countries, (2) the commonalities in failures leading to financial corrections across countries and (3) investigations on how the experience, systems and institutional knowledge of managing EU funds carry over between programming periods.Understanding these topics is relevant for individual countries concerning their absorption performance and EU funds management and aligning national policies with EU CP.
independent Hungarian Territorial and Regional Development Office.10.The implementation system carrying out the actual transactions was rather fragmented, with 22 intermediate bodies.
11.The Prime Minister's Office was responsible for member state-level coordination tasks, including the preparation of programming documents, functions related to programme implementation, monitoring of the use of funds, preparation of legislation and proposals for amendment, and centralized management functions related to programmes (e.g., communication, evaluation).12. Figures are from 2018.13.A total of 79% of TA allocation in Hungary and 72% in Slovakia (Polverari et al., 2020, p. 83).

Question 3 :
What was the performance of the EU fund management systems in Hungary and Slovakia in the 2000-06 and 2007-13 programming periods?

Figure 1 .
Figure 1.Correlation matrix.Note: Only the upper part of the matrix is shown.Crossed correlation coefficients are insignificant at the 10% level.Sources: AMPR reports (European Commission, various dates), worldwide governance indicators (WGI, 2020) and regional authority index (RAI)(Hooghe et al., 2016).

Table 1 .
Staff and absorption performance in Hungary and Slovakia.
Note: Employees involved in the management, implementation, control and audit of European Structural and Investment Funds (ESIF) (in 2014 and 2015 only Cohesion Policy funds in the case of Slovakia).Source: Data from European Commission (2020); data on employees from Ministry of Innovation and Technology (Hungary) and Government Office of the Slovak Republic (2015, 2016-20).

Table 2 .
Number and value of reported irregularities in Cohesion Policy and fisheries.
Note: Irregularities cover all three programming periods.Presented are irregularities detected by member states exceeding €10,000 reported to the European Commission and European Anti-Fraud Office for Hungary (HU) and Slovak Republic (SR).Source: PIF reports 2014-18(European Commission, 2014-18).

Table 3 .
Beta regression and fractional logistic regression models for financial corrections.
) excludes Hungary (HU) and Slovak Republic (SR).The standard errors in the fractional logistic model are computed using robust (heteroscedasticity consistent) covariance matrix estimators.Countries with zero financial corrections excluded in beta regression are included in the fractional logistic model.

Table 4 .
Financial corrections/contributions versus model forecasts.