DISCUSSION PAPER No. 4 November 2019 Governance Global Practice Thinking strategically about revenue administration reform: The creation of integrated, autonomous revenue bodies Raul Felix Junquera-Varela Rajul Awasthi Oleksii Balabushko Alma Nurshaikova This series is produced by Governance Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on Governance topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For information regarding the Governance Discussion Paper Series, please contact contact: Ayse Boybeyi, at aboybeyi@worldbank.org JEL Classification: H11, H29, H83 Key Words: Domestic Resource Mobilization, Tax Administration, Revenue Administration, Public Administration. © 2019 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved ABSTRACT Over the past two decades, two trends in tax administration reform have emerged – granting increased autonomy to tax administrations and establishing a unified integrated revenue authority (IRA) charged with administering both tax and customs laws and, in some cases, social security. This paper reviews the theory and practice of these revenue administration models and synthesizes lessons learned from the World Bank’s engagement with clients in recent years. The paper looks at several case studies across the world on the reform design as well as results of integration or increased autonomy. Semi-autonomous and integrated revenue administration models have the potential to bring crucial improvements to the efficiency and effectiveness of tax administration. The implementation shows both successes and challenges. The paper concludes with listing pre-conditions for these reforms including need for comprehensive approach of combining increased autonomy with improved administration processes and anticorruption efforts; importance of building collation of all revenue bodies prior to integration; and focus on non-revenue collecting functions of customs administration. Table of Contents Acronyms 1 1. INTRODUCTION 2 2. ORGANIZATIONAL ARRANGEMENTS OF REVENUE ADMINISTRATIONS: KEY TRENDS 3 3. SEMI-AUTONOMOUS REVENUE ADMINISTRATIONS 7 3.1 Why create semi-autonomous revenue administrations? 7 3.2 What are the levels of autonomy? 9 3.3 Design and implementation of the model: The Devil is in the details 10 3.4 Does autonomy impact tax administration performance? 12 3.5 What can be learned from country case studies? 13 Failing to keep tax administration reforms sustainable: The case of Peru’s tax agency 13 Risks of increased corruption: The case of Uganda’s SARA 15 Lessons learned from a successful SARA: The case of the South Africa Revenue Service (SARS) 16 Granting autonomy only in name: The case of Kenya 17 4. INTEGRATED REVENUE AGENCIES 17 4.1 What are the key benefits of integration? 17 4.2 What can be learned from country experiences? 18 The governance dimension: The case of Mozambique 18 Lack of prior strategic thinking and planning: The case of Kazakhstan 19 Successful planning and execution: The case of Bulgaria in the 2000s 20 5. CONCLUSIONS 23 The semi-autonomous revenue administration model 23 The integrated revenue bodies model 24 Annex 1: Institutional arrangements for tax administrations worldwide 26 References 27 Acronyms ARA autonomous revenue agency AT Revenue Authority (Mozambique) CIT corporate income tax EC European Commission GDP gross domestic product GTD General Tax Directorate (Bulgaria) HR human resources ICT information and communications technology IMF International Monetary Fund IRA integrated revenue authority IT information technology KRA Kenya Revenue Authority LTU Large Taxpayer Unit MOF Ministry of Finance NGO non-governmental organization NHIF National Health Insurance Fund (Bulgaria) NHF National Health Fund (Bulgaria) NRA National Revenue Agency (Bulgaria) NSSI National Social Security Institute (Bulgaria) OECD Organization of Economic Cooperation and Development PIT personal income tax RA revenue authority SARA semi-autonomous revenue authorities SARS South Africa Revenue Service SCC social security contribution SRC State Revenue Committee (Kazakhstan) SUNAT National Tax Administration Super intendency (Peru) TA Tax Administration URA Uganda Revenue Authority VAT value-added tax 1 1. INTRODUCTION Over the past two decades, a prominent trend in tax administration reform has been to grant increased autonomy to tax administrations to allow them greater management capacity.1 The rationale behind establishing semi-autonomous revenue bodies is threefold: to lessen political interference in the conduct of their operations, to allow greater responsibility and accountability for managers to achieve their objectives, and to build more management capacity, especially in terms of budgeting decisions and human resources (HR) policies. Such institutional arrangements have been termed unified semi-autonomous bodies (OECD 2013) or autonomous or semi-autonomous revenue authorities (SARAs). Another recent trend in countries around the world has been to establish a unified integrated revenue authority (IRA) charged with administering both tax and customs laws and, in some cases, social security. The stated rationale for this structure is that it can exploit efficiencies of scale by downsizing support functions, ensure better information flow, and improve taxpayer/trader services. Whether integration of revenue administrations has achieved its desired objectives and led to more effective revenue administration is debatable. Nevertheless, countries continue to take such decisions at the political level, and the newly created agency is often left with the difficult task of integrating responsibilities that had previously been carried out by fully independent agencies. It is unclear whether the semi-autonomous revenue agency or non-autonomous revenue agency model delivers better results, and merging tax and customs agencies has also shown mixed results. An IMF working paper surveying autonomous and non-autonomous tax administrations across the world finds no evidence that, overall, SARAs perform better than non-autonomous agencies (Crandall and Kidd 2006). Similarly, a World Bank study found that not all efforts to integrate tax and customs authority worked out well (World Bank 2010). It identified as a key challenge the responsibility for non-revenue services, such as securing borders and facilitating the flow of legitimate trade and travel, both typically customs responsibilities. This policy note reviews the theory and practice of these revenue administration models and synthesizes lessons learned from the Bank’s engagement with clients in recent years. The policy note does not argue for any one particular model; rather, it suggests a list of “dos and don’ts” for governments to consider before creating a semi-autonomous or integrated revenue authority. The policy note is organized into five sections. Following this introductory section, Section 2 provides a brief discussion of recent trends in organizational arrangements of revenue administrations across the world. Section 3 considers the rationale behind creating semi-autonomous revenue authorities, reviews the levels of autonomy used, and discusses details of design and implementation. This section also provides initial reflections on the impact of autonomous bodies on performance and presents case studies from Kenya, Peru, South Africa, and Uganda. Section 4 reviews the rationale behind the integrated revenue authority model and presents case studies from Bulgaria and Kazakhstan. Section 5 concludes the paper by drawing lessons from the case studies and providing general observations on the issues involved in revenue administration reform. 1 The majority of countries (around 60 percent) surveyed by the OECD in 2012 have adopted the “unified semi-autonomous body” for institutional set up of revenue bodies responsible for the administration of direct and indirect taxes (OECD 2013). 2 2. ORGANIZATIONAL ARRANGEMENTS OF REVENUE ADMINISTRATIONS: KEY TRENDS Revenue administrations play an important institutional role in collecting the domestic tax resources needed to fund public services while administering the tax system in a fair and impartial manner. The large number of staff and clients with whom these administrations interact; the complex network of stakeholders and other public institutions with which they are enmeshed; the range and nature of the tax laws they enforce; the multitude of core tax business processes, including IT systems, they employ—all of these elements require vesting these revenue bodies with adequate powers and autonomy to perform their vital functions effectively and efficiently. A broad consensus exists on the institutional, organizational, and operational arrangements desirable in and conducive to an efficient and effective administration of a tax system. The fiscal blueprints developed by the European Commission (EC) provide guidance on desirable features of national revenue bodies aimed at strengthening their operational capacity. One addresses internal revenue bodies, and another addresses customs bodies. Box 1. EU Fiscal Blueprints: Desirable Features for a National Revenue Body • Is there a guaranteed and adequate level of autonomy? Key indicators: Is autonomy provided for by law? Is there a statutory basis defining to whom the head reports? Is autonomy reflected in its structure and operational responsibilities? Is it able to design and implement its own operational policy? • Is there a clear description of responsibilities for bodies at the central, regional, and local level? • Are its obligations clearly translated into its mission, vision, and objectives? • Does it have its own structure and powers for effective and efficient operation? • Is it provided with adequate resources? • Does it have a stable legal framework? • Is it accountable for its operations and is subject to control and assessment? Source: EU Fiscal Blueprints. Over the past decade, the public sector has gone through significant changes to deliver services more effectively while keeping administration costs to a minimum. A key trend in this wave of institutional transformation has been granting more autonomy to government departments and agencies. As noted in Crandall (2010), a basic principle behind this move is that such autonomy can lead to better performance by removing impediments to effective and efficient management while maintaining appropriate accountability and transparency. In the context of public sector administration, autonomy usually refers to the degree to which a government department or agency operates independently from government, including legal form and status, funding and budget, and financial, human resources, and administrative practices (Crandall 2010). 3 Five broad categories of institutional setups for conducting tax administration can be identified, although in practice these may be subject to a number of exceptions (OECD 2013): • A single directorate in the Ministry of Finance (MOF): Tax administration functions are the responsibility of a single organizational unit (e.g., a directorate) located within the structure of the Ministry of Finance (or its equivalent). • Multiple directorates in MOF: Tax administration functions are the responsibility of multiple organizational units (e.g., directorates) located within the Ministry of Finance (often sharing necessary support functions, such as information technology (IT) and HR). • Unified semi-autonomous body: Tax administration functions, along with necessary support functions (e.g., information technology and human resources), are carried out by a unified semi- autonomous body, the head of which reports to a government minister. • Unified semi-autonomous body with board: Tax administration functions, along with necessary support functions (e.g., IT and HR), are carried out by a unified semi-autonomous body, the head of which reports to a government minister and oversight body/board of management comprised of external officials. • Integrated revenue authority: Tax, customs, and sometimes social security administrations are merged into one agency with different levels of functional integration. Annex 1 contains a list of all surveyed OECD and selected non-OECD countries and indicates institutional arrangements currently in place. An institutional trend worldwide favors the creation of unified semi-autonomous revenue bodies to improve the administration of tax systems. According to 2013 OECD Comparative Information Series, 31 out of 52 surveyed countries have established a unified semi-autonomous body (or in the case of China, a separate ministry) responsible for tax administration (and in some cases customs administration) operations, while the rest of the countries operate with other (generally less unified and/or autonomous) models.2 In Latin America and Africa, establishing semi-autonomous revenue bodies is the predominant pattern of tax administration reform. In Latin America, the tax administrations of Bolivia, Guatemala, Mexico, Peru, and Venezuela have adopted the legal status of semi-autonomous revenue authorities. In Africa, countries such as Ghana, Kenya, Malawi, Rwanda, Tanzania, Uganda, and Zambia have granted higher levels of managerial and financial autonomy to their tax administrations.3 2 This series provides internationally comparative data on aspects of tax systems and their administration in 52 advanced and emerging economies. The information provided in this edition was obtained from a survey of revenue bodies conducted in 2012 and from research conducted by official of the OECD’s Centre for Tax Policy and Administration of the revenue bodies’ key corporate documents (e.g., strategic plans and annual performance reports), other OECD tax publications, and other relevant sources. 3 In 2010, nearly 40 revenue authorities around the world clustered largely in Africa and Latin America (Crandall 2010). Generally, revenue administrations in these regions faced the need for massive reform, and the creation of a revenue authority was seen by some as a launching point for this work. 4 In Eastern Europe, eight countries — Bulgaria, Hungary, Romania, Serbia, Slovak Republic, Slovenia, Turkey, and Ukraine — have established semi-autonomous revenue authorities to administer tax collections. In the other twenty countries, tax administration is still part of the Ministry of Finance (World Bank 2013). No Asian country has in place a semi-autonomous revenue authority (Kidd and Songwe 2009). Tax administrations in Asia have low levels of autonomy over their organizational structure. A survey conducted by the World Bank in 2009, and comprising 11 Asian countries, provides evidence to that effect. A critical element of HR management autonomy is the ability to recruit personnel. The survey covered, inter alia, autonomy to recruit, set remuneration, promote internally, discipline staff, fire or terminate employees, and set conditions of employment. According to the study, Asian countries compare fairly well with OECD and non-OECD countries in giving tax administrations authority to undertake this critical function. Some 50 percent of the Asian countries surveyed, however, indicated that, for at least some of these functions, their central agencies played a role. Box 2. Key Institutional and Organizational Trends in Revenue Administration Institutional and organizational reforms are a prominent feature of efforts in many countries to improve efficiency and effectiveness; in particular, these include: • Establishing revenue institutions with increased autonomy and, in a number of countries (e.g., Hungary, Malta, Portugal, and the Slovak Republic), integrating the administration of tax and customs operations. The majority of OECD and selected non-OECD countries surveyed by the OECD (around 60 percent) have adopted the “unified semi-autonomous body” form of institutional setup to administer the tax system. In 11 countries, a formal management/advisory board comprised of external representatives has been established as part of the overall governance framework. • Twelve OECD member countries have aligned tax and customs operations within a single agency. In the 18 non-OECD countries surveyed, this alignment took place in six countries. • The introduction of organizational structures with reduced layers of management has provided a more centralized form of national management (e.g., Estonia, Finland, and Latvia), with ongoing efforts to establish taxpayer segment-based compliance structures underway (including for large taxpayers, e.g., in Belgium, Czech Republic, and Portugal). • In some countries (e.g., Greece, Norway, and Portugal), wide-ranging programs are underway to significantly reduce the size of office networks. • Decisions to integrate the collection of tax and social security contributions (e.g., in the Czech and Slovak Republics) over the medium term are underway, with similar reforms foreshadowed for some other countries (e.g., Greece and Portugal) in the longer term. Of the 32 OECD countries with social security contribution (SCC) regimes, 13 have integrated SCC and tax collection. • The national revenue body in the majority of European countries is also responsible for administration of property taxes (and often, motor vehicle taxes); elsewhere, these taxes are generally administered by revenue bodies of subnational governments. Source: OECD 2013. 5 Over the past 20 years many countries4 have also chosen to integrate their major revenue collection agencies under one administrative umbrella. No one formula or model has been applied across the world; approaches vary depending on circumstances. Typically, however, the integrated model has been employed to resolve critical problems associated with poor revenue collection performance, low rates of voluntary compliance, poor technical skills, and high levels of corruption. Box 3. Organizational Structures of Tax Administration Organizational structures of tax administration can be based on one of three main criteria: type of taxes collected, functions performed, and taxpayer segment. Over the past decades, tax administrations evolved from a type-of-tax model to the now dominant functional model. The taxpayer segment model, which has been established by a number of advanced tax administrations, such as those of Australia, the Netherlands, and the United States, organizes taxpayer services and enforcement activities principally around taxpayer segments (e.g., large businesses, small and medium taxpayers, individuals, etc.), to better tailor tax administration processes to taxpayer characteristics. In the developing world, the South Africa Revenue Service (SARS) is currently developing holistic segmentation designs for its taxpayer and trader segments, based on a thorough understanding of their behavior and needs. In practice, most tax administrations have organizational structures that mix different features and characteristics of the available models, creating what are termed “hybrid” models. In this regard, while practically all tax administrations’ organizational structures are based on the functional model, the majority have also established a Large Taxpayer Unit (LTU), and in many of them some residue of a type-of-tax administrative orientation survives as well. In the type-of-tax model, each department or unit performs all of the functions required to administer the specific types of taxes for which it is responsible. Prior to 2005, the United Kingdom provided a good example of a tax-based organizational structure; after that time its tax administration moved toward a function-based structure. The tax-type-based model has three main disadvantages. First, the system may lead to high administrative costs and low staff productivity because administrative functions are duplicated across the tax departments. Second, the tax-type organization imposes high compliance costs on taxpayers. Third, tax-type administrations are vulnerable to collusion between taxpayers and tax officials. Source: Verhorn and Brondolo 1998; OECD 2013. Revenue administration integration can take several forms. It can refer to establishing a unified body for the collection of direct and indirect taxes, a prominent feature of revenue administrations nowadays, and The list includes Albania, Argentina, Bolivia, Botswana, Bulgaria, Canada, Columbia, Croatia, Czech Republic, Denmark, Ecua- 4 dor, Estonia, Ethiopia, Finland, Ghana, Guatemala, Guyana, Hungary, Iceland, Italy, Jamaica, Kazakhstan, Kenya, Latvia, Lesotho, Malawi, Malaysia, Mauritius, Mexico, the Netherlands, Norway, Peru, Romania, Rwanda, Sierra Leone, Singapore, Slovenia, South Africa, Spain, Swaziland, Sweden, Tanzania, Uganda, United Kingdom, Venezuela, Zambia, and Zimbabwe. 6 it is part of the general trend in tax administration away from organization based on the type of tax to one based on function. A fairly recent unification, in April 2015, in Her Majesty’s Revenue and Customs (HMRC) in the United Kingdom merged the Inland Revenue Service and the customs and excise departments, another prominent trend in revenue administration. Factors in some countries contributing to this trend include efforts to obtain economies of scale or perceived synergies with customs operations responsible for collecting value-added tax (VAT) on imports. Historical factors may lead to the separation of direct and indirect tax administration (OECD 2013). Revenue integration may encompass alignment of tax and social security contribution collections based on the similarity of core processes or the efficient use of resources. 3. SEMI-AUTONOMOUS REVENUE ADMINISTRATIONS 3.1 Why create semi-autonomous revenue administrations? It is good practice to provide revenue administrations with a level of autonomy that allows them to adequately deliver their functions. As stated in the World Bank Group’s Handbook for Tax Simplification: Performance, once it is properly defined and adequately measured, requires management flexibility to allow the Tax Administration (TA) to focus on achieving its goals and to link incentives to performance delivery. This implies that senior TA officials should have the freedom to choose an appropriate organizational structure and assign responsibilities to the organizational units. Furthermore, managers should have the freedom to acquire and deploy physical resources and the manpower necessary to carry out their assigned responsibilities effectively, subject only to external oversight to ensure no malfeasance or abuse of powers. (James 2009) Governments establish semi-autonomous revenue bodies for one or more of three reasons. The structure reduces political interference in revenue administration operations; it provides more responsibility and accountability for managers to achieve their objectives; and it increases management capacity, especially in terms of budgeting decisions and HR policies. Establishing revenue authorities mirrors the broader development in public sector administration sometimes termed the “executive agency” model.5 The “revenue authority” model favors removing tax administration (and sometimes customs administration) from the internal structure of ministries of finance to improve service delivery and better control tax compliance. These executive agencies administer taxes guided by business principles of efficiency and effectiveness. These governance structures allow tax administrations to free themselves from rigid bureaucratic structures and to apply private sector management principles. The autonomy granted to the revenue bodies must be balanced against a high degree of accountability for the outcomes. It should be stressed that this autonomy does not mean that semi-autonomous bodies are independent from the government and cease to be part of the public sector. On the contrary, external controls and oversight on the part of the minister of finance and external audit bodies must be strengthened. 5 An extensive literature analyzes the Revenue Authority Model, including Jenkins 1994; Taliercio 2004; and Crandall and Kidd 2006. Jenkins points to Autonomous Central Banks as a reference point for building autonomous revenue bodies in terms of HR policies, budget resources, and lack of political interference. 7 It is worthwhile emphasizing that establishing a semi-autonomous revenue body will not translate in the short or medium term into increased revenues and improved operational capacity of the tax administration. Establishment must be complemented with improvements in core business processes, organizational reforms, and investment in information and communications technology (ICT). Political commitment and the willingness to implement the necessary changes is essential, because comprehensive tax administration reforms are complex and require considerable time to mature and yield outcomes. Many countries that have established this governance model view it as a catalyst for tax reform and improved performance. As noted by Crandall and Kidd (2006), “There is a strong perception held by those countries that have adopted the revenue authority concept that this particular governance model has made a significant contribution to reform and improved performance.” Taliercio (2004) emphasizes four ways that autonomy helps to attain the intended outcome of a competent, effective, and fair tax administration. (1) Autonomy can be seen as a way of unburdening public administration from undue political influence that undermines public service provision. (2) Autonomy can remove obstacles from sclerotic civil service systems that impede the development of a competent and effective tax administration, since tax administrations are not able to recruit the professional, specialized employees that they need. (3) Autonomy is also seen as a remedy for weak corporate governance and may empower managers to gain the control over personnel, funding, budgeting, and procurement that allows them to manage the tax agency effectively and efficiently. (4) Autonomy can provide the agency with financial autonomy and stability, which in many cases means granting it a fixed percentage of the tax collected.6 Box 4. Rationale for the Revenue Authority Model The arguments for the executive agency model relate primarily to effectiveness and efficiency. 1) As a single purpose agency, it can focus its efforts on its single task: administering the tax system and raising revenue. 2) As an autonomous organization, it can manage its affairs in a businesslike way, free of political interference in day-to-day operations. 3) Freed from the constraints of the civil service system, the autonomous organization can recruit, retain (or dismiss), and motivate staff to a higher level of performance. Source: Delay, Devas, and Hubbard 1998. 6 Taliercio (2004) also mentions some arguments that have been leveled against the idea of autonomous revenue agencies (ARA): “Some argue against autonomous tax agencies on the grounds of reform strategy. Thus the ARA reform is considered a variant of the ‘enclave’ approach to reform, which is sometimes disdained for its lack of comprehensiveness. The anti-enclave argu- ment, posits, perhaps implicitly, that it is better either to reform the government as a whole or leave things the way they are. A key assumption in this line of thinking is that partial reforms might actually make more comprehensive reforms more difficult at a later date.” Whether this argument is compelling or not depends on the potential for global reform and the goals of the reformers. Taliercio also highlights that this argument loses some of its sway given the uniqueness of the tax administration. 8 3.2 What are the levels of autonomy? The nature and extent of autonomy varies greatly among countries. The 2006 IMF Survey shows a considerable range in the levels of autonomy enjoyed by autonomous revenue agencies. Some tax administrations have limited managerial and financial autonomy, such as Mexico’s, while other countries have fully autonomous revenue authorities, such as those in Canada, Kenya, and Peru, which are characterized by independence from public service rules, including budgeting and human resources, despite falling under close control by the national government. Systems of governance across countries create significant differences in the levels of autonomy granted to tax administrations. The range of powers given to a national revenue body depends heavily on factors including the system of government in place, the state of development of the country’s public sector administration practices, and the institutional model adopted for tax administration (OECD 2013). Some of the usual powers delegated to autonomous revenue bodies are summarized in Box 5. Box 5. Typical Powers of Autonomous Revenue Bodies • Budget expenditure management: Discretion to allocate/adjust budgeted administrative funds across administrative functions to meet newly emerging priorities. • Organization and planning: Responsibility for (1) determining the internal organizational structure of the revenue body; and (2) formulating the revenue body’s strategic and operational plans. • Performance standards: Discretion to set its own administration performance standards (e.g., for taxpayer service delivery). • Personnel recruitment, development, and remuneration: Effective use of these powers should enable the revenue body to make more effective use of its human resources. • Information technology: Authority to administer its own in-house IT systems or to outsource the provision of such services to private contractors. • Tax law interpretation: The authority to provide interpretations, in the form of both public and private rulings, of how to interpret tax laws, subject only to review by judicial bodies. • Enforcement: The authority to exercise, without referral to another body, certain enforcement powers associated with administration of the laws (e.g., to obtain information from taxpayers and third parties and to impose liens over property in respect of unpaid debts). • Penalties and interest: The authority to impose administrative sanctions for acts of non-compliance. Source: Crandall 2010. Table 1 below examines the extent to which selected tax administrations have flexibility/ autonomy in relation to revenue bodies’ authority to: (i) design and implement their own internal organizational structure; (ii) allocate budgeted funds to meet new/changed priorities; (iii) determine the levels and mix of staff; (iv) influence/negotiate staff remunerations levels; (v) make tax rulings; (vi) set service standards; (vii) influence staff recruitment criteria; and (viii) hire and dismiss staff. 9 Table 1. Levels of flexibility/autonomy of selected revenue bodies Influence Design Fix levels/ Hire and Negotiate Make tax Allocate Set service staff Country internal mix of dismiss staff pay rulings budget standards recruitment structure staff staff levels criteria Canada x x x x x x x x Czech x x x x x x x x Republic Belgium x x x Denmark x x x x x x x x Estonia x x x x x x x Finland x x x x x x x x Germany x x x x x Hungary x x x x x x x Ireland x x x x x x x Slovak x x x x x x x x Republic Slovenia x x x x x x x x Turkey x x x x x Bulgaria x x x x x x x Romania x x x x x Russia x x x x x x x Source: OECD 2013. Based on responses from the surveyed countries, the OECD observes that the powers least frequently devolved are: (1) authority to design their internal structure (16 countries); (2) budget allocation discretion (11 countries); (3) ability to set levels and mix of staff within overall budget limits (16 countries); and (4) authority to influence/negotiate staff remuneration levels (23 countries). These responses corroborate that the degree of autonomy of revenue bodies varies significantly across countries. 3.3 Design and implementation of the model: The Devil is in the details When a country decides to establish the SARA governance model, it must make a series of policy choices concerning degree of autonomy, funding arrangements, and accountability. Based on the results of the 2006 IMF Survey, Crandall builds a framework that encompasses degree of autonomy, governance framework, accountability, and scope. The framework offers governments a range of possibilities from which to choose. Box 6 provides an overview of this framework. 10 Box 6. Design Considerations for a Semi-Autonomous Revenue Authority 1. Degree of autonomy a. Legal form and status: From an agency relatively close to a normal government organization to a corporate body with considerable independence. b. Funding: From normal funding via parliamentary appropriations to direct retention of a percentage of collected revenues. c. Budget flexibility: From limited flexibility to the complete flexibility of a one-line budget. d. Financial policies (such as accounting, asset ownership, and management procurement): From a situation in which the revenue authority (RA) is subject to standard civil service laws and regulations to one in which it is determined by a “corporate body” status (i.e., not part of the government’s accounting entity) e. Human resources: From being within the civil service control framework to being outside it. f. Operational autonomy: From a situation in which the minister has day-to-day authority to one in which the minister has no involvement in operational decisions. 2. Governance framework a. Role of the minister of finance: From direct supervision of the authority by the minister to a more limited role, such as appointment of the board or CEO only and limited broad strategic and tax policy directive powers. b. Role of the board: From no board at all to one with advisory powers only to being fully empowered through legislation to take management decisions. c. Role of the commissioner general: From a coordinating role only to full responsibility for revenue operations with all vested powers from revenue laws. 3. Accountability a. Reporting to the government and parliament: From being part of normal general government reporting to the need to follow special requirements specified in legislation. b. External audit: From being a legislated responsibility of the auditor-general to the RA or its board selecting the external auditor as it sees fit. 4. Scope Usually, the RA has authority for the administration and enforcement of all direct and indirect taxes at the national level as well as customs administration. The RA may also include the collection of local taxes or fees, social taxes or levies, and collection of social contributions. Source: Crandall 2010. Implementation poses serious challenges to countries, since the effort can span as many as two years and even more. Crandall (2010) lists a number of critical questions that governments should ask themselves before establishing a SARA. 1. Are the problems facing revenue administration known? 2. Is a reform and modernization strategy in place? 3. Is there a risk that creating the SARA will overwhelm other reform efforts? 4. Are the benefits and downsides of revenue authorities well understood? 11 5. Are the experiences of other countries relevant? 6. Is a lengthy time frame acceptable? 7. Are the requisite skills and other resources available? 8. Is the government prepared to deal with possible upheaval in labor relations due to the move to a revenue authority structure? 9. Are there any reasonable and practical alternatives? 10. Are the conditions for success and sustainability present? 3.4 Does autonomy impact tax administration performance? As noted above, a goal for establishing an autonomous revenue body is to improve the effectiveness and efficiency of the tax administration. Proving causality between the creation of an autonomous revenue authority and these desirable outcomes is difficult, however. Studies conducted to evaluate the success of the SARA model, including a survey carried out by the IMF in 2006, have not been able to clearly assess the model’s impact on tax administration performance. The OECD emphasizes some key points drawn from existing research: • A number of practical issues arise concerning the quantification of any benefits resulting from the introduction of the model. These include: (1) measurement: the difficulty inherent in gauging the impacts, in quantitative terms, of a concept such as autonomy; (2) data: limitations on obtaining relevant data items before and after implementation; and (3) attribution/causality: ability to attribute any observed benefits to specific initiatives (including the model itself). • Effective implementation of the model requires various types of support. These include, for example, good relationships with the MOF, strong leadership by senior management, and human resource policies for achieving good performance and addressing poor performance. • As noted by Crandall and Kidd (2006), improved effectiveness and efficiency is likely to flow most directly from an ongoing commitment to the reform of structures, systems, and processes. In particular, well-designed programs of service and enforcement, sound allocation of resources, and effective management are needed. Implementation of the new governance structure is, at best, a first step in this direction. • Countries that have adopted the revenue authority concept feel strongly that this particular governance model has made a significant contribution to reform and improved performance. Taliercio (2003) argues that SARA reforms represent an attempt by politicians to initiate a credible commitment to taxpayers that tax administration will be more competent, effective, and fair. According to Taliercio, the factor that enables politicians to make their commitment credible is the level of autonomy given to the revenue authority. Politicians are interested in making a credible commitment because they believe it will increase tax compliance. To the extent that revenue authority reform increases tax compliance, the payoff to politicians is greater tax collection. This, he argues, motivates politicians to install revenue authorities. 12 Taliercio probes this hypothesis by analyzing the results of a survey of large corporate taxpayers and professional tax consultants in four Latin American countries: Bolivia, Mexico, Peru, and Venezuela. The statistical results support the hypothesis that autonomy has an independent effect on perceptions of the political commitment to reform, even controlling for other variables such as specific service quality, the wider institutional context, and fixed country effects. This indicates that autonomy is valuable in itself, and not only as a means to better services. Rather, taxpayer ratings of organizational performance are positively affected by higher ratings of autonomy, suggesting that autonomy signals the government’s commitment to reform. In a study from 2004, Taliercio gathers evidence from case studies to assess the extent to which autonomy improves performance, analyzing greater autonomy in HR management and financial resource management in Kenya, Mexico, Peru, South Africa, Uganda, and Venezuela. Increased revenue performance results both from exogenous factors and from institutional and operational improvements in tax administration. It is extremely difficult to determine the impact on revenue performance resulting from establishing an autonomous revenue authority. Therefore, when it comes to evaluating how a SARA model affects tax administration effectiveness, we must rely more on perceptions than on quantitative criteria. 3.5 What can be learned from country case studies? Failing to keep tax administration reforms sustainable: The case of Peru’s tax agency 7 For Peru, as in many cases of tax reform, fiscal and economic crisis provided the catalyst. The country’s strategy for tax administration reform involved separating the National Tax Administration Super intendency (SUNAT) from the Ministry of Finance and creating an agency the independence of which was modelled on that of the Central Bank. Before the reform, Peru’s tax administration had been riddled with corruption and on the verge of collapse. The tax agency neither recruited experienced professionals nor provided training. Salaries were low, yet the wage bill was high due to overstaffing. Most general directors were political appointees whose tenure averaged less than two years. Moreover, the agency did not have adequate resources to finance its operations, let alone to invest in infrastructure. As a result, tax collections declined, a trend exacerbated by the economic crisis of the late 1980s. Tax revenue dropped from 14 percent of GDP in 1978 to 9 percent in 1988. Frequent organizational restructuring did nothing to improve performance. In this context, serious attention was given to radically reforming — even recreating — the tax administration. Building on the legal framework established by President Alberto Fujimori, the Peruvian government decided to pioneer the concept of a semiautonomous revenue authority, the first in Latin America. The reforms were remarkably successful: by 1997 internal tax revenue had recovered to 13 percent of GDP — despite an extremely difficult political and economic environment — and 90 percent of large corporate taxpayers surveyed believed that taxpayer services had improved. The reforms had several key elements: granting SUNAT meaningful administrative and financial autonomy; implementing radical 7 World Bank 2001; Taliercio 2003 and 2004; and World Bank Group 2009. 13 personnel reform; investing in infrastructure and information technology; and generating public support. The administrative reforms rested on three key pillars. First, SUNAT’s personnel system was exempted from civil service regulations and allowed to operate along private sector lines. Second, SUNAT’s superintendent became a presidential appointee, reporting directly to the president on important matters. Third, SUNAT’s budget was set at 2 percent of collections, and the funds were automatically deposited in the agency’s accounts. These features were intended to empower management while insulating it from political interference, to imbue staff with professionalism and integrity, and to ensure the timely and predictable provision of budget resources. The reforms also forged a new relationship between taxpayers and the tax agency. SUNAT pledged to treat all taxpayers fairly and committed itself to improving services. At the same time, the agency made clear its intention to enforce compliance with the tax code. SUNAT communicated its “new deal” to taxpayers through a sophisticated public relations campaign. The agency’s new managers also understood the importance of earning public support for reforms, and this support was cultivated through better services. While establishing SUNAT is a success story, the failure to maintain these reforms is very instructive for other tax modernization efforts. First, SUNAT lost the direct backing of President Fujimori when, to shore up his support among the poor, he switched his focus from fiscal austerity to executive-led populist spending programs. Second, the leadership changed four times between 1992 and 1998. Such changes proved destabilizing to the esprit de corps in an organization that had not yet developed a tradition of procedures and norms. Third, the degree of elite-level conflict between the Ministry of Finance and SUNAT increased as Fujimori developed links with parts of the business community that had traditionally influenced the state via the finance ministry. Peru’s SARA experience highlights that autonomy is not simply a technical matter but rather depends on contingent political support. Moreover, it underlines the important role that political strategies play in the evolution of institutional innovations. Tax administration reforms in Peru provide useful lessons for other countries. Positive efforts and results included: • The immediate efficacy of SUNAT as a semiautonomous revenue authority was due to a combination of political leadership and managerial expertise. • A comprehensive and coordinated approach to both tax policy and tax administration reform contributed to SUNAT’s success. • SUNAT’s organizational design allowed managerial flexibility on a host of important issues, from personnel to financial administration. • In spite of resistance to radical personnel reforms, SUNAT instituted an organization based on high ethical and professional standards. • SUNAT’s strategy of coupling its fight against tax evasion with better taxpayer services enabled it to develop a constituency of supporters in the business community. • Peru’s experience also highlights pitfalls to avoid for other countries engaging in tax administration reform. • The relationship between the tax agency and the ministry of finance should be carefully managed. 14 • Sustainability merits careful attention. Although it initially achieved success, SUNAT subsequently faced several problems that limited its success, including a deteriorating tax policy, political interference, and weaknesses in the institutional environment. Risks of increased corruption: The case of Uganda’s SARA8 Uganda’s experience with SARAs follows a pattern similar to Peru’s. As did President Fujimori in Peru, Uganda’s President Museveni took power on an anti-corruption, anti–political party platform in the midst of a collapsing state and following the economic decline that occurred during the Amin and second Obote regimes. In 1991, the state, with the financial support of Britain, established the Uganda Revenue Authority (URA). While internal staff relations and hiring followed new public management principles, salaries were increased by eight times the public sector average; poorly performing staff were either fired or moved to other ministries. The URA’s autonomy vis-à-vis donors and the Ministry of Finance increased when it was put in charge of determining tax revenue targets. Initial results were encouraging, as tax collection as a percentage of GDP increased from 7.0 percent in 1991 to 12 percent in 1999. Thereafter, cases of corruption occurred and tax collection worsened. By 2001, tax collection had declined to 10.7 percent of GDP. To the extent that tax capacity reflects legitimacy, it is reasonable to conclude that both the Ugandan revenue authority and the Museveni regime may have experienced a decline in legitimacy in the post-2000 period. The extent to which this decline in legitimacy matters for the government’s political survival, of course, depends on the emergence of a sufficiently powerful opposition. Several political problems emerged that compromised the URA’s functioning. First, while its senior managers received substantial pay increases, the wage differential with respect to staff at lower levels grew, creating discontent. The wage differential was also a factor in the growth of corruption in the URA. Second, while the URA board’s internal autonomy vis-à-vis its employees increased, its external autonomy was limited. Relations with the finance ministry soured after the first three years, as the URA failed to meet target objectives. Finally, the encroachment of clientelism affected recruitment and promotion within the URA, and growing interference by well-connected members of the executive branch, the finance ministry, and the parliament contributed to the widespread perception that, after 1995, corruption within the URA increased substantially. Likewise, relations with President Museveni, initially very strong, soured. This tension had at least two sources. First, the semi-military operations introduced to prevent smuggling and tax evasion became unpopular. Second, the introduction of VAT, which initially had the president’s support, proved very unpopular in a country where tax payments were notoriously low. Tax revolts in 1996 over the introduction of VAT induced President Museveni to make the URA a scapegoat. Uganda’s experience with SARAs provides useful lessons for other countries. These include: • Macroeconomic and particularly fiscal and inflationary crises provided opportunities for leaders to gain leverage over the reform process. • Centralized public authority and executive support is essential for successful tax reform. • The creation of a (semi)-autonomous revenue authority can, paradoxically, increase the political attention it receives. 8 World Bank Group 2009. 15 • The pressure on governments to meet revenue targets (regardless of whether they are experimenting with SARAs) can create tension between the state and interest groups that can undermine the tax authority’s legitimacy. • The extent and nature of political commitment to reform are crucial to success. • Given the SARAs’ increased discretionary authority, procedures should be implemented to reduce the possibility that an environment enabling corrupt behavior will develop. • It is important to keep in mind that quite small, practical changes to tax design and improvements in tax administration can be important in encouraging taxpayers to mobilize. • Organizational changes and capacity development investments can have significant impacts, even in unfavorable governance environments. (See, for example, the cases of Mozambique and Rwanda below.) Lessons learned from a successful SARA: The case of the South Africa Revenue Service (SARS)9 In the mid-1990s, South Africa reformed its tax administration and granted the revenue authority greater autonomy. Tax collection was both very successful and very progressive. In the period between 1997 and 2002, the tax take as a percentage of GDP in South Africa averaged over 25 percent, compared with the middle-income-country average of 15 percent of GDP. Income tax collection averaged 14.6 percent of GDP, compared with the Latin American average of 3.9 and the East Asian average of 6.9 percent at that time. While this impressive revenue collection coincided with increased autonomy, several other historical and political factors also underlay this success. First, the highly successful income and overall tax collection capacity of the South Africa state was not a new feature of the reformed tax authority. From 1960 to 2000, South Africa’s tax collection as a percentage of GDP, particularly income tax collection, was consistently the highest among middle-income countries. The factors supporting this high level of income tax collection capacity have received considerable analysis. First, a high degree of cooperation existed between the state and upper-income groups, which supported state-led reforms. This challenges the idea that effective tax collection depends simply on instituting an autonomous revenue agency. Second, the introduction of computerization in the 1960s greatly enhanced the ability of the Department of Inland Revenue to calculate and issue assessments, to record payments, to monitor large taxpayers, and to maintain control over tax payments more generally. Third, introducing a withholding pay-as-you-earn system also greatly enhanced tax collection. Several key factors contribute to the continued success of SARS’s tax collection capacity. These include: • The state exhibited a high degree of administrative cooperation, particularly between SARS, the Finance Ministry, and the Central Bank. • The state has historically maintained a cooperative relationship with upper-income groups, including large firms, which helps reduce the transaction costs of collecting income taxes. • The reforms were introduced with substantial consultation with representatives from the state, political 9 World Bank Group 2009. 16 parties, business chambers, labor unions, and national and international legal experts. • South Africa’s enhanced revenue collection performance in that period followed a sustained campaign by SARS to emphasize that task-effective tax collection is central to the “protection of the national economy.” • SARS was formed within the context of a strong national political party, the African National Congress. Two lessons emerge from this. First, “windows of opportunity” for successful tax reform emerge in moments when a dominant political party is accorded a high degree of legitimacy. Second, it is much more difficult to sustain reforms in contexts with fragmented political structures or anti–political party leaders or in situations of substantial multi-party contestation and/or political polarization among parties. Granting autonomy only in name: The case of Kenya10 Kenya demonstrates a case of failure due to lack of investment in improving core operational processes of tax administration. Kenya’s creation of a SARA was not accompanied by corresponding changes in business processes to improve compliance and delivery of services to taxpayers. Moreover, a tax-based organizational structure was maintained. Verhorn and Brondolo (1998) stress that the formation of the Kenya Revenue Authority (KRA) in 1995 initially did little to strengthen tax administration because KRA’s three revenue departments (Income Tax, VAT, and Customs and Excise) continued to operate independently from each other; no changes were made to improve compliance and taxpayer services; and the organization itself lacked many of the key features of a true Revenue Board, such as autonomy over staffing arrangements and an independent budget. 4. INTEGRATED REVENUE AGENCIES 4.1 What are the key benefits of integration? The integrated revenue authority (IRA) model has typically been employed to resolve critical problems associated with poor revenue collection performance, low rates of voluntary compliance, poor technical skills, and high levels of corruption. The IRA model is believed to foster increased efficiency and professionalism, improved morale and integrity, and economies of scale in providing administrative services, including personnel, financial management, training and staff development, and ICT. Frequently, the IRA model was also used to address remuneration challenges that could not be tackled under normal civil service pay and condition rules. The model has proved useful in reducing costs and improving organizational performance, particularly in human resource management and training and back office support services, such as office administration, procurement, asset management (including office accommodation and regional presence), common corporate financial systems, taxpayer recruitment and registration, audit, legal support, risk management, compliance improvement, and fraud investigation. In terms of IT systems supporting revenue administration, global experience suggests that application developments must be performed by dedicated staff with extensive tax or customs 10 Verhorn and Brondolo 1998. 17 experience, according to their roles. Activities that lend themselves to a common or integrated approach include operation of the primary data center and disaster recovery center, installing and maintaining servers and associated hardware, maintaining the system network, procuring licenses, and maintaining system service performance. Physically, the equipment used to run the tax and customs systems (i.e., servers, network, storage, backups, and disaster recovery) can be housed in the same physical space. In general, experience suggests it is more productive to integrate hardware than software; in the case of software, limited sharing is likely, as deep knowledge and content-specific skills are necessary. Operation of the tax and customs systems, including monitoring and administration, will likely remain separate, as they have different availability requirements (i.e., customs operates a 24x7 system and needs operational system availability every day — usually throughout the day — whereas tax revenue, other than web services, can be restricted to extended business hours). It is possible that some front-end systems and client interfaces using the same data paths may be consolidated on one platform. The integrated model has also helped strengthen the audit function of tax and customs administrations by providing access to data. Integration of the tax and customs risk management systems has the potential to increase the performance of risk-based audits. 4.2 What can be learned from country experiences? The governance dimension: The case of Mozambique11 Beginning in 1994, Mozambique carried out a complete reform of its tax system and administration, including merging the tax and customs administrations into a revenue authority. The general objectives of the reform, implemented in phases, were to modernize the tax system and rectify flaws in tax and customs administration, thus creating modern institutions with the capacity to administer new taxes effectively and efficiently manner while improving revenue collection. The Government of Mozambique identified, categorized, and implemented the reforms in two phases between 1994 and 2007. First, the reform of indirect taxes, prior to 1999, included restructuring the administrations concerned (customs and, for domestic taxes, the VAT unit). Between 2000 and 2007 direct taxation and the associated administration at the tax department were reformed, culminating at the end of that period in the Mozambique Revenue Authority (AT), with jurisdiction over both customs and domestic taxes. An evaluation of Mozambique’s reforms indicates beneficial effects in several key indicators, such as collections, tax evasion and noncompliance, tax equity, economic efficiency and GDP growth, corruption or openness, and business climate. The beneficial effects continued immediately after the creation of the AT. Mozambique provides a good example of the need to build capacity in the tax administration as a precondition for establishing a successful IRA. The tax administration of Mozambique had very limited institutional capacity, inadequate infrastructure, antiquated organizational structure, a total lack of information technology support, outdated procedures and manuals, a weak audit regimen, and 11 World Bank Group 2009 and Castro, Junquera, Schenone, and Teixeira 2009. 18 a rudimentary and incomplete taxpayer register. Moreover, customs had higher capacity than did the tax administration. In light of this, the authorities decided to strengthen capacity at the tax administration prior to establishing the AT model. Customs and the internal revenue service would be merged and put on an equal footing before the Revenue Authority was created. Lack of prior strategic thinking and planning: The case of Kazakhstan Kazakhstan created its State Revenue Committee (SRC) on August 6, 2014, without undertaking the detailed preparatory and planning work required. The newly established SRC was given 100 days to develop its integration strategy and produce quick wins. The expectation was that by integrating the tax administration with the customs service, the former would rapidly help the latter improve its performance. Kazakhstan’s decision to merge tax, customs, and an economic investigation service was driven by the need for higher efficiency in public revenue collection and improved service delivery to taxpayers and traders. At the time, the Government was concerned with the extreme divergence of the country’s ranking on the Doing Business Indicator: Kazakhstan ranked 17th on the Ease of Paying Taxes but 185th in Trading Across Borders in the 2015 Doing Business Report. Corruption in customs had also been an inherent systemic issue, which, as perceived by the Government, required radical changes in organizational mindset. Thus, tax officials were put in charge of the integration process, better characterized as a take-over than a merger. The new SRC management turned to the Bank for technical assistance through reimbursable advisory services. The Bank had an ongoing engagement in Kazakhstan through a Tax Administration Reform Project, a Customs Development Project, and a series of tax policy advisory reports and was well positioned to provide such support. The Bank team responded quickly, organizing a large conference in December 2014 to present international experiences, discuss key issues associated with effective revenue integration with focus groups, and help the client formulate objectives. The conference warned that integration was not a one-time event, but rather a phased process that could take from three to five years to produce tangible results. Experts from Canada, the Netherlands, South Africa, and Spain all stressed differences between tax and customs, as well as the importance of a sound change management strategy. The Bank also supported SRC in defining the scope of integration in terms of business processes and potential efficiency gains and in drafting a strategy document for an integrated revenue agency. Despite the Bank’s emphasis on the need for phased integration, SRC decided to move on with a “big bang” approach. The table below depicts the order of the SRC’s integration steps. In less than three months a new organizational structure was adopted with large back up functions — strategic planning, human resources, financial management, legal support — fully integrated at the central level. Most of the merged units were staffed by former tax employees with little understanding of customs specifics. 19 Immediate “Big Bang” Steps Medium-Term Steps Long-Term Actions (implemented in 100 days) (implemented in 36 months) (3 to 10 years, still ongoing) • Changes in vertical (central/ • Business processes • IT integration: subnational) and horizontal reengineering o Single account with (functional) organizational • Single risk management connections to second-tier structure database banks • Introduction of joint audits, o Unified risk management • End-to-end monitoring of appeals, and debt collection system goods (from the moment of services o Unified IT system for tax entry into the country until and customs inspections • Single website and e-mail sale) and debt collection domains • Unified registration of o E-audit system taxpayers and traders o Customs e-declarations • Unified forms for tax and through Information customs inspections and System “Taxpayer Cabinet” appeals • End-to-end monitoring of • Elimination of duplicate goods (continued) document requirements • Merger of tax and customs legislation into a single Revenue Code (cancelled in 2016) The SRC’s ambition for its revenue integration went far beyond organizational changes; it sought identify potential synergies in operational areas and IT systems. Positive early results of functional integration included the introduction of a single risk management database providing 300 prudent taxpayers with a “Green Corridor” for imports. Another good sign for taxpayers and traders was the introduction of a unified registration, as well as improved information exchange between tax and customs offices, thus reducing the number of documents required to be presented in the form of hard- copy originals. Businesses reported additional confusion and administrative burdens at the implementation level, however (Sange 2015). Little attention was paid to communicating the changes to the wider constituency or to building internal capacity. Customs officials were left out of the reform process, lowering motivation and satisfaction levels, as evidenced by the staff survey.12 After only five years of operation, it is still too early to assess the impact of the integrated model. Kazakhstan was ranked 105th in 2019 on Doing Business’s Trading Across Borders, but this may not be fully attributable to the integration. Successful planning and execution: The case of Bulgaria in the 2000s Bulgaria successfully integrated its social contribution collection processes with its tax administration, forming an integrated revenue agency. The integration was well planned, the strategy for implementation was well thought out, and the government was committed to a process that would take several years. Notably, the customs administration was not part of Bulgaria’s integration effort. 12 Id 20 The integration of tax and social contribution collections in Bulgaria took place as part of a broader package of reforms that began with the Social Insurance Administration Project13 in the mid-1990s. In October 1999, the Bulgarian government decided to create an entirely new unified revenue collection agency. This move was expected to improve revenue collections of both social security and tax. The underlying concept was that a unified revenue agency would establish a single point of performance accountability and coordination for collecting taxes and social contributions. Opportunities to improve administration included streamlining registration, filing, and payment procedures for employers and the self- employed as well gaining efficiencies by eliminating duplicate functions and back office support, including IT. The Government undertook considerable consensus-building efforts, and on June 2000, the council of ministers approved the concept of a unified revenue agency with the following features, among others: • Harmonization of the tax and social contribution legislation • Development of integrated business processes and IT systems • Adoption of a single national identification number (Bulstat number) • Implementation of modern audit techniques based on risk management principles • Establishment of a modern function-based organization structure The concept paper provided a well-thought-out framework for the design of the National Revenue Agency (NRA). The organization had a clear understanding that the reform process would require phased implementation involving several years for completion. In the first, preparatory phase, the tax and social security administrations worked together to harmonize the revenue bases, conduct coordinated audits, and learn to function in a far more coordinated manner. Only in the second phase of comprehensive implementation would the unified revenue agency be set up and deep reform and modernization of the entire administration take place. While careful strategic planning for integration played a key role, implementation was also handled well. In November 2002, legislation was enacted to establish the NRA as a formal legal entity charged with collecting domestic taxes and social contributions and reporting to the Minister of Finance. In June of the following year, a loan was secured from the World Bank to provide funding to design and implement the NRA. The project14 represented the first Bank lending operation focused specifically on supporting the integration of two revenue agencies, in this case tax and social contributions collection, in a newly established national revenue agency. The NRA started operations on January 1, 2006. This marked the beginning of the integration of the former General Tax Directorate (GTD) and the collection function previously performed by the National Social Security Institute (NSSI). Subsequently, the number of field offices was reduced from 416 regional and local offices to 28 regional offices and a Large Taxpayers Unit (LTU). In the meantime, collection of contributions and fees was decentralized to local tax offices, including the withholding tax for personal income taxes. The three factors that seem to have been critical in the implementation phase were leadership, staffing, and communications. The MOF provided leadership throughout the integration process. The initial 13 Social Insurance Administration Project (P008323). 14 Revenue Administration Reform Project (P073427), 2003. 21 task force was formed in March 2000 and was chaired by the deputy minister of finance. The team comprised representatives of the GTD, the NSSI, and the National Health Insurance Fund (NHIF), and its activities were managed as a joint agency effort by a unified revenue agency commission, with a GTD deputy director as team leader. An IT directors’ board with GTD and NSSI representation was also established. Before the NRA began operation, a recruitment strategy was put in place to select NRA staff, with the old GTD and the NSSI being the most important sources of experienced employees. Prior to establishing the NRA, retraining and placement services were in place for about two years to provide opportunities for staff members who either chose not to apply to the NRA or who were not selected for it. Commonly, these employees found jobs with local governments and private companies. Overall, the NRA endeavored to create human resource management conditions that would help develop a professional workforce with modern skills and a high level of probity. This was a daunting task for the NRA, particularly because the law requires that it pay its staff members only public sector salaries. As a result, it could not compete financially with the private sector for talent. The HR team therefore developed a two-part strategy to attract highly skilled staff: first, it used psychological profiles during recruitment to select employees from the pool of GTD and NSSI staff who were the most motivated and qualified; and second, it concentrated on providing non-financial incentives, such as training opportunities, travel abroad, job certainty, and good and fair career development opportunities (including access to management positions, a recognition program that publicly recognizes high-performance staff, public relations campaigns to raise the stature of NRA staff in the community, and others). This twofold strategy appears to have worked; staff turnover ratios have been brought under control, although difficulty remains in hiring for some employee categories, particularly lawyers and IT staff. The NRA employed various strategies to improve staff communication and to raise awareness and use of NRA services in taxpayer communities. The principal activities for strengthening internal communications included periodic employee attitude surveys and direct interaction between top management and staff. The NRA executive director also visited territorial directorates regularly to interact with staff. This helped boost morale and kept staff informed through a direct feedback loop with management. Outside the NRA, national campaigns helped taxpayers and contributors pay their tax and social contributions electronically, further raised awareness about the NRA services call center, and promoted the use of generally available e-services. The NRA also implemented a number of measures to improve transparency and uniformity in applying the law. All laws, regulations, practices, and instructions used by the NRA were made available to taxpayers via an intranet and the internet. In addition, internal tax manuals dealing with the VAT, corporate income tax (CIT), and personal income tax (PIT) were also made available to the public, promoting a high level of transparency and uniformity of understanding between the NRA and its clients. To assist in designing the next round of optimization, the NRA created a consultative council representing businesses, trade unions, the accounting and auditing professions, and NGOs. The newly established National Revenue Agency and the five-year Revenue Administration Reform Project promoted significant improvements in tax policy and administration in Bulgaria, making the country the top tax system reformer in 2006–2007 of the 178 economies surveyed in the World Bank’s Doing Business report of 2008. This was a combined result of the reduced tax and social security burdens and the simplified tax-paying process. Improved revenue collection paved the way for tax reduction. The corporate income tax rate was reduced from 15 to 10 percent in 2007, the personal income 22 tax was also effectively reduced with the adoption of a flat PIT rate of 10%, and social contribution rates were cut by 7 percentage points in 2006–2007. The lessons learned from the project, while being somewhat specific to mergers of tax and social contributions collections, remain valid for other types of revenue agency mergers. 5. CONCLUSIONS Semi-autonomous and integrated revenue administration models have the potential to bring crucial improvements to the efficiency and effectiveness of tax administration. The Bank’s experience engaging with clients in this area has seen both successes and challenges, however. Review of country case studies suggest the following preconditions and factors to keep in mind when embarking on either of the two primary models: semi-autonomous revenue administration and integration of revenue administrations. The semi-autonomous revenue administration model Before embarking on the long, complex process of creating a semi-autonomous revenue administration, tax administrations should strive to meet several preconditions. A unified, modern tax administration has a much higher chance of benefiting from increased autonomy. Creating a semi-autonomous revenue administration based on a tax authority with fundamental organizational issues will only amplify its problems. First, the nation should have a unified tax administration covering all national taxes. Second, a unified tax administration should be structured and organized according to key functions, which aids in compliance management.15 Finally, the tax administration should manage its own human resource policies, allowing it to create specialized career paths and to develop sound remuneration policies for its personnel. In addition, political support and a minimum level of capacity are required to guarantee success when establishing autonomous revenue bodies. Sustainability in this process raises specific challenges, analyzed below. In some countries, for example, Mozambique, the process has coincided with merging tax and customs administrations. No firm conclusion can be drawn about whether the semi-autonomous revenue agency model delivers better results than other models. In the absence of “regime change” and ensuing reforms, the semi-autonomous revenue model per se may or may not deliver better results. It seems that greater autonomy can help deliver better results if it leads to a “regime change” that supports the delivery of needed reforms; otherwise, it seems to be ineffective (Crandall and Kidd 2006). Moreover, reforms can also be delivered in the absence of a semi-autonomous revenue agency. In Latin America, the move to greater autonomous revenue authority has been marked by “regime change,” and this has led to greater effectiveness in raising revenues. The main drivers of improved performance were reforms in management and personnel systems, which would not have been possible without HR management autonomy (Taliercio 2003). Managerial and personnel autonomy contribute to the perception of administrative fairness, and personnel management is the most important dimension of autonomy for perceptions of administrative competence (Alonso, Junquera, and Fanta 2011). 15 This is predominant organizational model nowadays. See Box 3 for an overview of tax administration organizational structures and the disadvantages associated with the tax-based model. See also lessons learned from the Kenya case. 23 Considerations of political economy play a key role when it comes to decisions about creating an autonomous revenue body, as shown in the cases of Peru and Uganda. The Peruvian experience with an Autonomous Revenue Authority demonstrates that autonomy is based on contingent political support and is not simply a technical matter. Moreover, it underlines the importance of political strategies in the evolution of institutional innovations. Establishing a semi-autonomous revenue body will not translate into increased revenues and improved operational capacity of the tax administration in the short or medium term. SARAs must be complemented with improvements in core business processes, organizational reforms, and ICT investment. Political commitment and a willingness to implement the necessary changes are also essential because of the complexity of comprehensive tax administration reforms and the length of time they require to mature and yield the desired outcomes. The following lessons drawn from the country studies are commonly borne out: • A comprehensive and coordinated approach to both tax policy and tax administration reform, as in the cases of South Africa and Peru, requires well-managed cooperation with the Ministry of Finance. • Cooperative relationships with business groups, especially large taxpayers, help reduce the transaction costs of collecting income taxes, as was the case for the South Africa Revenue Service. • Substantial consultation with all stakeholders is important, as also demonstrated by SARS. • The combination of political leadership and managerial expertise contributes to successful reforms. • Change management that encourages managerial flexibility and high ethical and professional standards is important. • Organizational changes and investment in capacity development can have significant impacts, even in unfavorable governance environments. • Coupling the fight against tax evasion with better taxpayer services improves the likelihood that reforms will succeed, as shown in Peru. • Creation of a semi-autonomous tax administration should go hand-in hand with anticorruption efforts. The integrated revenue bodies model Integration of tax and customs administration into one body does not coincide with the semi- autonomous revenue agency model. Integration can be done under an agency enjoying a limited level of autonomy. Integration of different revenue administrations frequently aims to achieve economies of scale or perceived synergies with customs operations responsible for collecting VAT on imports or with administrations collecting personal income taxes and social security contributions. While the decision to establish a semi-autonomous revenue authority or to integrate revenue administrations are not identical, many of the same considerations and prerequisites for success apply. Political support and leadership are essential, as showcased in the three case studies on integration. In addition, integration of two revenue bodies with flawed business processes will contribute to likely failure 24 of the integration. The case studies analyzed in this paper yield the following lessons for integration efforts: (1) Reform design should be grounded in clear operational issues and objectives. Integration requires long lead times, detailed planning, coordination, and communication for successful implementation. As the Kazakhstan experience shows, decision to integrate too rapidly, with only high-level goals such as “facilitating trade” and “reducing the informal economy,” will not translate into the clear steps for change required at the administrative level, leading to ambiguity and low staff morale. In Bulgaria, the integration process began in 1999; the government approved a concept in 2000; implementation took place in two phases beginning in 2002; and final integration was ongoing for several years after that. (2) A successful transition depends on buy-in from all agencies slated to merge. In Bulgaria, with the transfer of collection responsibility to the revenue administration, the social security agency would lose control over its revenue generation. Avoiding the agency’s resistance to the change required building confidence and instruments (such as protocols, memoranda of understanding, and service level agreements) that reassured stakeholders that social contributions collections would be transferred fully and immediately and that the social security agency would receive appropriate information in a timely manner. The experience in Kazakhstan was the opposite of this. Customs agency staff felt they had been kept “out of the loop.” The customs staff shared little or no ownership of the integration process with social security agency staff, who consequently always mistrusted the transition as a takeover and regarded the process as somewhat unfair. (3) Integrated work processes are important. Bulgaria’s success depended on developing integrated work processes that helped both taxpayers and tax administrations and reduced compliance and administration costs. These processes included a single national database comprising taxpayer and contributor information and unified systems for registration, processing, collection, taxpayer services, audits, and appeals. In Kazakhstan, however, while plans have been made to integrate processes with a view toward reducing compliance burdens and making administration more effective, more than twenty months after the merger these have yet to be implemented. (4) The enabling legal framework must be initiated as early as possible to ensure a clear and sufficient basis for the reform. In Bulgaria, the NRA Act helped remove difficulties in coordinating GTD and NSSI modernization efforts. To protect the interests of the NSSI and the National Health Fund (NHF), the law created the NRA advisory board, which included the heads of these agencies. Significant changes to substantive tax and social security laws were made to align legal definitions, calculation bases, taxpayer/contributor obligations and rights, enforcement powers, and penalty regimes. In Kazakhstan, the integrated agency was set up under a presidential decree, but the required amendments to the tax and customs laws to support an integrated organization were carried out much later. (5) Finally, integration of customs and tax administration require additional consideration of non-revenue functions of customs such as border security and anti-smuggling functions. As these are not typical for other revenue administration bodies, they need to be treated with caution and not be left out from the integration planning process. 25 Annex 1: Institutional arrangements for tax administrations worldwide OECD Countries Nature of Body Non-OECD Countries Nature of Body Australia USB Argentina USBB Austria SDMOF Brazil USB Belgium MDMOF Bulgaria USB Canada USBB China Other Chile USB Colombia USBB Czech Republic USB Cyprus MDMOF Denmark USB Hong Kong SAR, China SDMOF Estonia SDMOF India USB Finland USB Indonesia SDMOF France SDMOF Latvia USB Germany Other Lithuania USB Greece MDMOF Malaysia Other Hungary USB Malta MDMOF Iceland USB Romania USB Ireland USB Russian Federation USB Israel SDMOF Saudi Arabia SDMOF Italy Other Singapore USBB Japan USB South Africa USBB Republic of Korea USB Luxembourg MDMOF Mexico USBB Netherlands SDMOF New Zealand USB Norway USB Poland MDMOF Portugal SDMOF Slovak Republic USB Slovenia USB Spain USBB Sweden USBB Switzerland SDMOF Turkey Other United Kingdom USBB United States USBB *Note: USB: unified semi-autonomous body; USBB: unified semi-autonomous body with formal board or advisory group comprised of external officials; SDMOF: single directorate in Ministry of Finance; MDMOF: multiple directorates in MOF. Source: OECD 2013. 26 References Alonso, Rosa, Raul Junquera, and Enrique Fanta. 2011. Gap Analysis of the Tax Administration of the Philippines. Washington, DC: World Bank. Castro, Patricio, Raul Junquera, Osvaldo Schenone, and Antonio Teixeira. 2009. Evaluation of Reforms in Tax Policy and Administration in Mozambique and Related TA 1994–2007. Washington, DC: FAD, IMF. Crandall, William. 2010. Revenue Administration: Autonomy in Tax Administration and the Revenue Authority Model. IMF Technical Notes and Manuals. Washington, DC: FAD, IMF. Crandall, William, and Maureen Kidd. 2006. Revenue Authorities: Issues and Problems in Evaluating their Success. WP 06/240. Washington, DC: IMF. Delay, Simon, Nick Devas, and Michael Hubbard. 1998. Reforming Revenue Administration: Lessons from Experience. Report prepared for DFID. Birmingham: International Development Department, University of Birmingham. Devas, Nick, Simon Delay, and Michael Hubbard. 2001. “Revenue Authorities: Are They the Right Vehicle for Improved Tax Administration?” Public Administration and Development 21 (3): 211–222. Diaz Yubero, Fernando, and Raul Junquera. 2013. Chapter 2 in Granting Autonomy to Tax Administrations: The Case of the Spanish Revenue Agency, vol. 2. International Master on Tax Policy and Revenue Administration, Spain. Madrid: Institute for Fiscal Studies. James, Sebastian S. 2009. A Handbook for Tax Simplification. Washington, DC: World Bank. http://documents.worldbank.org/curated/en/317341468335679099/A-handbook-for-tax-simplification. Jenkins, Glenn P. 1994. “Modernization of Tax Administration: Revenue Boards and Privatization as Instruments for Change.” Development Discussion Papers 1994-01. JDI Executive Programs. Kidd, Maureen, and Vera Songwe. 2009. “Organization, Governance and Structure of Tax Administration in Selected Asian Countries.” Draft. World Bank, Washington, DC. Manasan, Rosario G. 2003. Tax Administration Reform: (Semi-)Autonomous Revenue Authority Anyone? Discussion Paper Series 2003-05. Quezon City: Philippine Institute for Development Studies. Mann, Arthur. 2004. Are Semi-Autonomous Revenue Authorities the Answer to Tax Administration Problems in Developing Countries? A Practical Guide. Washington, DC: USAID. Moore, Mick. 2013. The Changing Politics of Tax Policy Reform in Developing Countries. PREM Notes. Washington, DC: World Bank. OECD. 2013. Tax Administration 2013: Comparative Information on OECD and Other Advanced and Emerging Economies. Paris: OECD. Sange Research Center. 2015. Monitoring of the Quality of the SRC Services (prepared for Kazakhstan Taxpayers Association). 27 Taliercio, Robert, Jr. 2003. Administrative Reform as Credible Commitment: The Impact of Autonomy on Revenue Authority Performance in Latin America. Washington, DC: World Bank. ———. 2004. Designing Performance: The Semi-Autonomous Revenue Authority Model in Africa and Latin America. Washington, DC: World Bank. Verhorn, Charles L., and John Brondolo. 1998. Organizational Options for Tax Administration. Washington, DC: IMF. World Bank. 2001. Strengthening Peru’s Tax Agency. PREM Note, RAPTG. Washington, DC: World Bank. World Bank. 2010. Integration of Revenue Administration: A Comparative Study of International Experience. Washington, DC: World Bank. World Bank. 2013. Tax at Glance. World Bank: ECA Region. World Bank, Washington, DC. World Bank Group. 2009. Reforming Tax Systems for Political Stability and Sustainable Economic Growth: A Practitioners’ Guide. Foreign Investment Advisory Service, WBG, in collaboration with DFID. Washington, DC: World Bank. 28