Journal Issue
Share
Article

Revenue Authorities

Author(s):
Maureen Kidd, and William Joseph Crandall
Published Date:
October 2006
Share
  • ShareShare
Show Summary Details

Abbreviations and Acronyms

ATO

Australian Tax Office

CEO

Chief Executive Officer

CRA

Canada revenue Agency

DFID

Department for International Development

FAD

Fiscal Affairs Department

GDP

Gross Domestic Product

HR

Human Resources

IFIs

International Financial Institutions

IMF

International Moneraty Fund

IT

Information Technology

OECD

Organization for Economic Cooperation and Development

PEM

Public Expenditure Management

RA

Revenue Authority

SAT

Superintendencia de Administration Tributaria

SRF

Secretaria da Receita Federal

SUNAT

Superintendencia National de Administration Tributaria

URA

Uganda Revenue Authority

I. Introduction

A. Purpose

Over the course of the last 15 years or so, many countries have embarked on programs of reform for their revenue administrations (tax and customs), and some of these have decided to adopt the idea of a “revenue authority” (RA) as the delivery model for improved administration. For the purposes of this paper, an RA refers to a governance model for revenue administration where traditional ministry of finance departments (tax and usually customs administrations) are established as an organization or agency with a degree of autonomy from government and independence from standard public service policies.2 The RA has been seen by some as a possible solution to critical problems, such as poor revenue performance, low rates of compliance, ineffective staff, and corruption. It has often been argued that an RA can lead to improvements, including better accountability for results, synergies in administration across the revenue departments, and management based on professional skills and isolated from external influences.

No comprehensive assessment has been made of the RAs that exist today and of how well they have served their intended purpose. Much of the literature and many of the evaluation studies on the subject of RAs have made a direct link between overall revenue administration performance and the particular governance model of the revenue administration. In other words, they have assumed that measurable changes in performance can be attributed to the adoption of the RA governance model. The inherent challenges in this attribution have confronted virtually all those who have studied RAs and who have tried to make some assessment of RA performance.

The objective of this paper is to advance the discussion of the issues and challenges associated with the evaluation of RAs. While there may be many perspectives from which to view RAs, this paper generally looks at the issues from the perspective of revenue administration, and the almost universal desire to improve performance and compliance with the law.

As a point of departure, the paper presents a compilation of a wide range of information about RAs—from very specific descriptions of key features (e.g., legal form and character and mandate) to aspects requiring a more subjective judgment on the part of the countries involved (e.g., how RAs made practical use of their autonomy). It then looks at similarities and differences and explores trends, before concluding with a series of findings. Given its role in the provision of technical assistance to member countries in fiscal administration, the International Monetary Fund (IMF) is often asked for advice regarding revenue administration reform and strategic change. By extension, this can sometimes include the question of whether an RA should be considered as an effective tool or catalyst for improving performance revenue administration performance. The findings of this paper will be taken into account by the Fiscal Affairs Department (FAD) in its provision of advice on this subject.

B. Nature and Scope

This paper presents comprehensive information on RAs and explores issues associated with evaluating the impact of RAs on performance. Ideally, it might be possible to objectively assess the impact on performance where an RA has been established and compare that to what would have otherwise been the case had the RA governance model not been adopted. However, this is neither practical nor achievable without heavy reliance on comprehensive multi-year data to develop indices that could correlate performance, autonomy, reform measures, and the like. With such data, panel data analysis could be attempted where changes over time could be observed and causality established using econometric techniques. Performance data for countries with RAs (the target group) and countries without RAs (the control group) could be analyzed and compared and inferences made about the impact of the RA governance model. Before and after comparisons could be made.

However, there are many impediments to this kind of approach:

  • Data required would be extremely hard to obtain, and would demand a sample size that could not possibly be delivered. A very large data set would be needed, and data for the pre-reform period would be particularly difficult to obtain.
  • Broad indicators used to assess revenue administration performance can be impacted by many exogenous factors. A good example is total revenue collections, where the precise effects of the quality of the revenue administration (whether under an RA structure or not) are difficult to isolate and attribute. That is, one would have to correct (or neutralize) for the impact of factors, such as the economy in general, the quality of national statistics used, tax policy changes, regional growth trends, oil prices, population demographics, level of political commitment, resources available, and dozens of other factors in order to isolate the revenue administration impact on a change in revenue collection performance.
  • A further complication would be the requirement to exercise considerable judgment about how to isolate the impact or outcomes of the particular governance model that the tax administration is using. Even if performance data, such as revenues collected, audit coverage, debt collected, taxpayer services, and so on, could tell us something specific about the quality of the revenue administration, such data would likely tell us very little about how having a particular governance structure (such as an RA) might or might not have led to different outcomes.
  • Finally, it must be noted that the term “RA” covers a very broad spectrum of governance choices, rendering any broad “RA to normal government department” comparison tenuous at best.

A recent study by the U.K. Department for International Development (DFID) provides an excellent summary of these challenges.3 It states that there are complex measurement, data, and estimation issues around the subject of autonomous RAs, leading to three major problems with econometric or quantitative research:

  • Measuring autonomy—The different dimensions of autonomy are tricky to measure quantitatively, but might be measured by a scoring system. However, it would be complicated and possibly misleading to merge all the dimensions into one index, as the weighting of the dimensions would be necessarily subjective. It would be difficult to capture, in numbers, a rather fuzzy political concept like autonomy.
  • Data limitations—There is limited scope for a before-RA and after-RA comparison of the same indicators.
  • Establishing and isolating causality—In the absence of panel data analysis, where changes over time can be observed, it is difficult to establish causality with econometric techniques. Even with such data, many factors impact on outcomes, such as revenue, making the precise effects of administrative reforms difficult to isolate. In order to control all possible factors affecting revenue collection, researchers would need a huge dataset, which is currently unavailable.4

For any particular revenue administration outcome, one must determine what is due to exogenous factors and what is due to the effort of the revenue administration. Then, we need to know, of that portion due to the administration’s effort, what portion is exclusively attributable to the fact that there is an RA governance model at work. Without this knowledge—and it is very difficult knowledge to obtain—assessments of RAs will necessarily need to rely on qualitative inputs and subjective criteria, such as perceptions.

In summary, the goal of this paper is to bring together a body of information on RAs and to explore more closely the challenges in trying to assess the difference the RA governance structure makes, compared to not having an RA governance structure. The paper uses a combination of quantitative and qualitative information provided by RA countries themselves, including perceptions about the contribution of RAs, various descriptive documents and references, and views and opinions of knowledgeable individuals in order to conduct this task. This methodology is presented in more detail in the section below.

C. Methodology

The following description of the methodology ultimately used for this paper outlines the many sources of information used to build a database about RAs and to explore the difficulties in assessing their performance.

Avenues of inquiry included a review of existing literature, an assessment of the IMF Fiscal Affairs Department’s (FAD) experience to date with countries implementing RAs, an analysis of legislative and other reference documents, and two questionnaires for a select number of countries that have implemented or are implementing an RA. Some preliminary information was also obtained from revenue administrations that have remained traditional departments of government in the interest of providing a baseline for general comparison. Certain preliminary data prepared by the Organization for Economic Cooperation and Development (OECD) on revenue administration features of various countries—RAs and not—was also considered.

In terms of the literature review, there exists an extensive body of articles and papers on RAs (many dating from the early to mid-1990s) and a review of this literature helped frame the approach to the paper. Much that has been written to date about RAs demonstrates the challenges in trying to evaluate whether implementing an RA gives revenue administration any incremental gain in performance. Section III captures themes from six seminal papers, while a complete bibliography of the literature referenced can be found at Annex I.

FAD has worked closely with many countries that have decided to implement an RA and has provided advice through the period leading to the decision and during implementation as well. A review of experiences and recommendations provided to these countries also formed part of the information base, and specific country references are found throughout this paper.

The most important tools used were two questionnaires (see Annex II), developed and sent to various IMF member countries at different points in the research phase of the project, the first in the spring and summer of 2005, and the second in early 2006.

The design of the first questionnaire presented several challenges. It needed to be as comprehensive as possible without imposing too great a burden on responding countries. A questionnaire that was either too complex or that took too long to complete would result in a low response rate and likely would have little value to the exercise at hand. As discussed in previous sections of this paper, it would be impossible to seek a full time-series data set, so the questions were designed to elicit responses that provided the opportunity for a qualitative picture to emerge.

This questionnaire comprised four major components, each designed to solicit certain types of information that could assist in the RA review:

Basic information—this component gathers basic information about the RA including its legal status, characteristics and features, degree of autonomy, governance and accountability, organization, and scope of activities. This information was intended to present a common picture of RAs, identify critical structural similarities and differences, and assist in creating an understanding of the RA concept.

Original rationale for considering and adopting the RA—the idea behind this component was to review the considerations that went into the decision to establish the RA in the first place. In other words, what was the value proposition or case-for-action that prompted the initial decision, what were the expected benefits put forward at the time, and how was success to be measured?

Implementation of the RAthis component gathers information about the process of implementing an RA, including such matters as the duration of the process, project management issues, the use of consultants or other forms of technical assistance, costs, and so on.

Self-assessment of RA effectiveness—how countries utilizing RAs actually perceive themselves and their organizations was considered an important aspect for the study. Therefore, a component was developed to provide respondents the opportunity to self-assess the effectiveness of their RA in terms of how it contributed to tax administration modernization.

In addition to these four components, the questionnaire also requested analytical background data on various quantitative and qualitative performance measures in a time series over the life of the RA. It was anticipated that, if this data were available, it could prove useful in identifying trends and confirming other observations.

Some 21 countries were included in the population for the first questionnaire. This reflected a sizable proportion of countries that had in fact opted for the RA model and included two countries that were not RAs, as a control group. Efforts were made to ensure a geographic balance and to include an appropriate number of developed and developing countries.

Detailed replies were received from only 12 countries, and while it was clear that in many cases every effort had been made to reply to the questions posed, the resulting set of data was still incomplete. Therefore, it was decided to develop a second and shorter questionnaire to be sent to both RAs that had not replied to the first questionnaire as well as to an additional 12 countries, in an effort to enhance the overall database on RAs for the study. Notwithstanding the fact that it was an abridged questionnaire, it was designed to capture the more critical elements needed for this research.

Furthermore, it was clear from an analysis of the responses to the first questionnaire that the management of human resources (HR) posed many challenges for which the HR autonomy of the RA model was considered a possible solution. In an effort to probe more deeply into HR autonomy, and how it had in fact been used by the RAs, a detailed questionnaire on this specific topic was developed and sent to all countries that received questionnaires. This HR questionnaire focused on the areas of recruitment, remuneration, performance pay, retention, internal staffing, control over organizations and positions, training and development, and integrity (see Annex III).

Finally, the study considered two other sources of information. First, information was sought from a small number of non-RA countries in an effort to judge whether the advances attributed to RAs in human resources management, in particular, are also witnessed in more traditional departments of government. Second, the study reviewed some aggregated and preliminary data on tax administration prepared by the OECD.5

D. A Brief History of the Development of RAs

Restructuring of government has been a constant theme over the last three decades as governments have sought to deliver services more effectively and at a lower cost to citizens. In some cases, traditional government structures (e.g., a government ministry organized along hierarchical lines) have been viewed as too rigid to respond to the rapidly changing needs of the public and the challenges confronted by government in modern society. While changes in government have been described as “evolutionary rather than revolutionary,”6 a developing trend has been for government to devolve power to agencies or appointed bodies acting on their behalf.

The United Kingdom is seen by many as in the forefront of this new way of doing business and, since the 1990s, its quasi-autonomous agencies account for close to 25 percent of all government expenditure.7 Canada and New Zealand employ a similar approach, and significant areas of government whose diversity range from film classification to advertising standards are being run by agencies, rather than traditional ministries of government.

Revenue administration has not been completely immune to this trend. Governments of developed countries sought ways to deliver better service and some have turned to a form of semi-autonomous agency to help them meet goals of improved collections, better service to taxpayers, and more flexible human resource management options. Governments of developing countries share many of these goals and have additional complications. Problems related to low capacity and the need for massive administrative reforms, combined with corruption and long periods of non-performance, have made the case for a different form of government structure, compelling both to decision-makers as well as to the donor agencies interested in funding the needed reforms.

The creation of the first RA dates to the mid-1980s, while within the survey population the oldest RA is in Peru (where legislation was passed in 1988 to create an agency). There are close to 40 RAs around the world clustered largely in Africa and Latin America. Generally, revenue administrations in these regions were faced with the need for massive reform, and the creation of an RA was seen by some as a launching point for this work. Development agencies of some national governments have at times actively promoted RAs as a solution to poor performance (e.g., DFID, USAID) and international organizations have also played a role in this field. Private sector consulting firms have honed and marketed their expertise in this area and had a strong presence in Africa through the 1980s and 1990s. Countries that have opted for an RA represent a little over 20 percent of the membership of the IMF.

II. Key Definitions and Issues

A. Some Key Terminology

Throughout this paper, certain terms are used that, while often commonly understood, can in fact have a range of meanings. For clarity, the following definitions are offered and will be used in this paper.

Autonomy—The degree to which a government department or agency is able to operate independently from government, in terms of legal form and status, funding and budget flexibility, and financial and human resources policies.

Accountability—The extent to which roles and responsibilities are clear, authorities are appropriately delegated, and those so empowered to make decisions are in fact held responsible for them and their consequences. There can be both individual and organizational accountability.

Governance regime (or model)8—The institutional or structural framework that determines the responsibility, authority, and accountability for a government institution. In the context of a RA, these parameters dictate the relative autonomy of a given government organization in terms of government control and of the applicability of public service policies.

Government control—The degree of involvement by central government in decision-making within the agency, both from a program and administrative perspective.

B. Defining an RA

A revenue authority (RA) is simply a term to describe a governance regime for an organization engaged in revenue administration that provides for more autonomy than that afforded a normal department in a ministry.

Since RAs are organizations within a national government, it can be useful to look at typical national government structures along a spectrum that indicates decreasing government control and increasing autonomy (Figure 1). In the last few decades, many models have emerged along this spectrum. The standard model, furthest left, is a normal department of government within a ministry, such as finance or agriculture. Moving to the right on the spectrum are the semi-autonomous agencies, and this broad and expanding group would include the RAs (all RAs are semi-autonomous by this definition). Further along the spectrum are the autonomous agencies and regulatory bodies that have true distance from the government and the public service. This group could be said to include the central banks and other independent regulatory bodies of government. Even further along are state-owned enterprises engaged in a variety of activities from quasi-public ones, such as post offices, to traditionally private sector activities, such as oil production. The far end of the spectrum, for illustration, depicts activities previously carried out by government that have been fully privatized. Examples might include air traffic control, correctional facilities management, and (in areas more closely related to RAs) pre-shipment or arrival inspection in customs administration.

Figure 1.Typical National Government Structures

The above example illustrates various governance models in the context of a national administration. The same technique can be used to look at governance models for revenue administration, again with the continuum from left to right denoting decreasing government control and increasing independence from public service rules (Figure 2).

Figure 2.Revenue Administrations

Revenue administrations fall under the first two possibilities only—normal departments of government or semi-autonomous agencies. RAs deal predominately with indirect and direct tax administration at the national level, and usually with customs administration as well. Tax and customs laws include some of the most intrusive powers of the state, and in no known cases are they further along the autonomy spectrum than “semi-autonomous.” Their powers can never be too far removed from the control and accountability of elected government, and therefore tax and customs organizations are likely to always remain public institutions.

Figure 2 also illustrates examples of relative autonomy within the semi-autonomous group using an assessment based on results provided later in this study.

The organizational nomenclature used in the world of revenue administration varies considerably—department, directorate, commission, agency, service, board, authority. However, most countries do not attempt to define these terms—they simply use them to label their own governance model. More important is where various models (countries) are placed along the spectrum, and it is readily noted that there could be debate over this placement. Clearly there are some countries where the revenue administration is just the same as any other department of government (France and Cambodia are just two of dozens of examples). Within the semi-autonomous group, the Internal Revenue Service (IRS) of the United States could be included, given the presence of such features as external oversight of the IRS. This organization has less relative autonomy within the group. The South Africa Revenue Service (SARS) might be next, with its independence from public service rules but close control by government. A subsequent example could be Canada, with its management board and human resources autonomy, but relatively close control by government. What follows would be the broad group normally referring to themselves as RAs, such as Kenya or Peru, similar to Canada but with more autonomous budget features and somewhat less general government control.

For the purposes of this review, an RA can simply be thought of as a revenue administration operating within any governance model providing greater autonomy than that for a standard department of government.

C. Design and Scope of RAs

An understanding of RAs requires some knowledge of their design and their scope of activities. Both involve a range of options—all of which contribute to where the RA will fall on a continuum of autonomy as described in the previous section. Depending on specific policy choices made about design and scope, an RA will have more or less relative autonomy.

Little discussion exists in the literature about this range of policy choices that must be considered in the design of an RA. While Box 1 provides a summary outline of these policy choices, Annex IV offers a much more detailed discussion that will be of interest to any government considering the implementation of an RA.

Box 1.RAs: Design and Scope Issues

Design—Includes degree of autonomy or independence, and related governance framework and subsequent accountability structures.

Degree of autonomy

  • Legal form and status—legislation could establish an agency relatively similar to a normal government organization, or set up a corporate body with considerable independence from government.
  • Funding—could range from normal funding via parliamentary appropriations to direct retention of a percentage of collected revenues.
  • Budget flexibility—from limited flexibility to the complete flexibility of a one-line budget.
  • Financial policies (such as accounting, asset ownership and management, procurement)—subject to standard civil service laws and regulations or as determined by “corporate body” status.
  • Human resources—within the civil service control framework, or outside it.

Governance framework

  • Role of the minister of finance—direct supervision of the authority by the minister, or role limited to appointment of board or board chair only.
  • Role of the board—advisory, or fully empowered in legislation to take management decisions.
  • Role of director general—coordinating role only, or full responsibility for revenue operations with all vested powers from revenue laws.

Accountability

  • Reporting to the government and parliament—part of general government reporting, or special requirements specified in legislation.
  • External audit—responsibility of auditor-general, or the board selects private sector external auditor.

Scope—Refers to the scope of taxes and taxing agencies to be included ranging from direct and indirect taxes at the national level to the inclusion of customs administration and agencies or units that collect local taxes or fees.

III. Review of Selected Literature

This section deals with some of the key articles or documents in the literature related to RAs. Generally speaking, the literature focuses on the overall concept of the RA, the case for autonomy and how to measure it, and RA performance based on case studies. While there are many references to RAs in the literature, and a number of descriptive pieces, there are relatively few analytical assessments of RA performance.9 Given that this paper attempts to capture the issues and challenges in evaluating RA success, a review of some of the key papers to determine what has been said on this point is of fundamental importance.

For the purposes of this review, six papers have been selected to illustrate areas of particular interest: an assessment of revenue boards (Jenkins, 1994);10 a study of RA performance (Taliercio);11 and a practical guide to RAs (Mann);12 a review of RAs in the context of DFID projects in Africa;13 a DFID evaluation of RAs in Uganda and Zambia;14 and a political economy analysis of the RA in Uganda.15 These papers span the earliest writing on RAs to the most recent, and they represent the work of some of the leading thinkers on the subject. All grapple with the issue of how to measure RA performance.

Jenkins (1994)

This seminal article (most later studies refer to it) reflects changes in the concept of public administration that had emerged by the early 1990s—that public administration should be focused on citizens and their needs, and that different public institutions had different organizational needs. It notes further an increased emphasis on tax reform in developing countries, and the reality that tax law or tax policy is only as good as its administration.16

Jenkins’ main observation is that a critical impediment to change for tax administrations is that they were traditionally a part of the general civil service. This, he opines, has led to situations where salaries are low, qualifications of staff poor, hiring subject to political patronage, and opportunities for rent-seeking plentiful. Furthermore, public perceptions of this state of affairs were leading to diminished voluntary compliance, the key underpinning of a modern tax administration. Revenue administrations, Jenkins states, need to be more like the central banks—good pay, equipment, facilities, and so on. To respond to this, he notes, many countries have set up autonomous (or semi-autonomous) RAs.

Based on the above assessment, Jenkins called for a restructuring of revenue administrations to provide independence in the same manner as central banks: appointment of a CEO for a specified term; authority to establish operating policies; clear and transparent objectives; and a measure of financial independence. It was clear that these more independent revenue administrations should be fully responsible for their own recruitment, training, and salary structure.

Jenkins also discusses the notion of a “revenue board,” suggesting these boards should have specific administrative responsibilities and essentially operate as a policy committee, with responsibilities such as ensuring the revenue departments enforce the laws; establishing human resources policies; ensure the sharing of information; providing resource allocation direction; recommending legislative changes; providing revenue forecasts; and setting up internal audit. He is clear that the board should not intervene in tax cases.

RA practice in the 2000s generally followed the lines described by Jenkins (although many RAs in Latin America did not choose to create a revenue board), with one exception. Jenkins suggests the CEO17 should support the board, but not have line authority over the heads of the revenue departments. This, in fact, was the model for some of the early RAs, such as Kenya. However, all RAs have now moved to the model of a fully empowered CEO (see subsequent sections of this study), especially as modern concepts about integrated and function-based organizations have led to the merging of previously separate revenue departments (except for customs).

Taliercio (2004)

Taliercio looks at the overall performance of RAs by assessing six case studies across Latin America and sub-Saharan Africa: Peru, Mexico, Venezuela, Kenya, Uganda, and South Africa. Although he concludes results are mixed, he presents two main findings: (1) performance improved most where the RAs had comparatively more autonomy; and (2) performance improved where autonomy was stable. This is based on the author’s assertion that managers with greater autonomy are able to take reforms further and faster.

Taliercio looks at two main criteria upon which to base assessments—management and performance. Management criteria for human resources include the scale of retrenchment and whether staff had to re-apply for their jobs at the time of establishment of the RA, skill, and salary levels of staff, total wages, grading structures, retention, turnover, and corruption levels. For financial management, criteria include collection costs, adequacy of funding, and investments in training and IT. Other management criteria encompassed decentralization, standards, and relations with the private sector. All these are important, Taliercio argues, since “behind any gains in performance are internal reforms to management and personnel systems, most of which would not have been possible without autonomy.”

His second area is performance in practice, covering revenue collection, compliance, taxpayer registration and services, tax revenues as a share of GDP, growth rates in arrears, VAT productivity, audit coverage, and the like. Measuring such criteria to assess performance engenders all the difficulties inherent in measuring tax administration results in general. Taliercio recognizes these difficulties, and talks of painting a picture based on case studies rather than “presenting a rigorous quantitative analysis that controls for exogenous variables.” He speaks of the problems of “data availability,” the general lack of a rigorous methodology for evaluating tax administrations (the great difficulty of measuring the marginal impact of tax administration reform on outcomes such as revenue collection), “measuring” autonomy, and sustainability of reforms.

Taliercio asserts that various reforms would not have been possible without autonomy (an example being staff improvements and reduced corruption in the Kenya Revenue Authority).

In other areas, Taliercio states that the more autonomous RAs have reasonably low collection costs (expenditures as a percent of revenues collected). While this measure may work to identify trends over time in a static tax policy environment, the comparison of this ratio between and among countries is fraught with difficulty because of the many exogenous factors affecting the ratio.

Taliercio also discusses the advantages and disadvantages of RAs. In terms of its benefits, he suggests autonomy can solve administrative and corporate governance problems and deliver fair and effective revenue administration. Problems such as low pay, low quality staff, low retention rates, and patronage-based recruitment can be addressed through autonomy in human resources matters. Political intervention can be addressed through independence from the executive branch of government. And autonomy can also provide adequate funding, such as a percentage-based cost-of-collection formula, instead of uncertain and unstable resource allocations. “Autonomy is a response to inadequate central government civil service and public expenditure management (PEM) systems …[and] .. is considered an antidote to the failure of the political system to build in effective accountability systems.”

Four potential arguments against the concept of RAs are identified: (1) creation of an RA represents an enclave approach to reform, begging the question about reform in the rest of government (Taliercio argues a compelling case can be made for tax administration because it provides the revenue for the whole of government);18 (2) there could be negative externalities throughout the rest of government (i.e., why special treatment for tax administration?); (3) there could be concern about bureaucracy run amok (i.e., insufficient accountability, which can be addressed through proper accountability mechanisms in the RA); and (4) concentration on RA reform could divert attention from the basic operational functions of tax administration.

In summary, this review is a good case study of certain countries. It discusses the performance of the tax administrations (which happen to be RAs) in those countries, with all the attendant qualifications and limitations related to the difficulty in isolating the contribution of the tax administration to various outcomes, such as revenue improvements. Aside from some qualitative observations and inferences in the human resources area, the study does not advance the difficult discussion about whether the RA governance model itself has made any difference in performance compared to what would otherwise have been the case.

Mann (2004)

Mann makes a very useful summary of the advantages and disadvantages that have been put forward regarding RAs. On the advantages, or benefits, side he lists: enhanced revenues; greater efficiency; more competent staff; de-politicization of tax administration; reduced corruption; improved services; more professional work ethic; comprehensive accounting for tax revenues; and integration of databases. Disadvantages of RAs are listed as well. Excluding those already mentioned by Taliercio, Mann adds: potential inherent conflict with the ministry of finance; and the possibility of establishing an “unnecessary” organization whose tax collection functions, given the political will and resources, could be upgraded within already existing departments of the ministry of finance.

Mann also points out that cost-benefit analyses are not usually carried out when considering the creation of an RA, something he advises should be done as a matter of importance. The paper makes a number of recommendations in terms of RA features and governance, in all cases recommending the maximum autonomy possible as best practice. The section on governance structure deals only with appointment processes for the board, its members, the CEO, and so on.

In terms of quantifiable data to assist in the basic evaluation of RAs, Mann uses VAT compliance and productivity rates and general tax-to-GDP ratios to make judgments about pre- and post-RA performance. He clearly states, however, that this type of analysis does not control for all the exogenous variables in addition to tax administration efforts that affect outcomes. The paper talks about providing performance indicators that will permit an assessment of RA performance, though not necessarily in assessing the RA governance model per se.

This paper does not address the issue of how and when the “modality” of the RA impacts revenue administration. Given the caveats necessary on the quantitative analytical tools available and the relatively small sample size, Mann’s conclusion is tempered—RAs have neither lived up to expectations, nor can they be categorized as having failed. They do, however, he says, provide a platform from which tax administration efficiencies can be generated, but their mere establishment offers no guarantee of success. Throughout the paper, discussions of the decision to establish an RA appear to apply equally to any governance model. In a general sense, the paper addresses the question of tax administration reform, rather than the modality of the RA.

Delay, Devas, and Hubbard (1999)

A DFID-commissioned study by the University of Birmingham in 1999 reviewed the experience of DFID-funded projects to support tax administration in Africa against wider international experiences to determine lessons for future work. With specific reference to RAs, the paper concluded that while RAs have much to offer (especially where civil service management inhibits proper HR management) they are not a panacea, do not prevent political interference, and are no guarantee against incompetence or malpractice. The study also notes that the urgency of the need to mobilize revenues in some countries makes the model attractive and that there may be important symbolic value in the “fresh start” which such institutional change promises.

Gray and Chapman (2001)

A further DFID evaluation of revenue projects in 2001 focused on two RAs in particular, Uganda and Zambia. While the prime purpose of the report was to assess linkage between poverty reduction and DFID support for revenue administration improvement, some comments and conclusions are worth noting in the context of RAs. Key project objectives were revenue related in both cases—to increase the revenue/GDP ratio for Uganda, and to prevent the ratio from deteriorating owing to poor macroeconomic conditions in Zambia. Both projects were considered a success in these areas although, as was the case in other studies, no direct link is made to the RA structure being the cause of the revenue performance improvements. In fact, the report suggests for Uganda a key reason for the increase in revenue/GDP was DFID support itself, since the ratio subsequently “stagnated (when) DFID’s support was scaled down.” This report also confirms general attribution problems where it noted projects had goals to increase the revenue/GDP ratio but that this ratio “is sensitive to a wide range of other influences…” Specifically regarding RAs, the report concluded that (1) the de facto degree of adoption of the RA framework varied from country to country; (2) RAs have led to more flexibility in allocation of RA budgets, but have not brought independence from the state budget; and (3) the RA framework has not led to significant easing of the problems of enforcement.

Di John and Putzel (2005)

Another important issue is raised in a paper by Di John and Putzel that focuses on the Uganda Revenue Authority. It deals with the concept of political economy analysis of an RA—i.e., a qualitative assessment of the impact of political and economic developments on the operation and autonomy of the RA. This paper states that IFIs and aid donors have developed the proposition that, in weak states, revenue collection authorities are more effective when they operate autonomously from the state as a commercial entity at arms length from the government. Reference is made to Taliercio on this point, but this is likely a little more categorical than presented by Taliercio. The report suggests that while RAs may have been instrumental in initiating reforms in some cases, sustainability is another matter altogether. In fact, the report questions whether there can ever be true autonomy where critical interdependencies exist, such as those between tax policy and tax administration. One of the main contributions of this analysis is that it identifies the importance of political and economic realities in the operation of the RA, specifically those related to: macroeconomic crises; the existence of conditions for successful reform such as political support; public communications and the perceived legitimacy of a tax and its administration; and vulnerability to shifting political attention.

Conclusion

There are several common themes that emerge from these varied papers on RAs.

With the exception of Jenkins, all the papers grapple with the problem of quantification and how to measure RA performance to assess whether the RA governance model made a difference. In no case were any of the authors able to present or to propose an analytical model that would produce measurable results for the RA. In instances where some quantifiable variables are presented, they are variables that are designed to measure the progress of reform—whether through an RA or through a traditional department of government.

Many writers on RAs tend to articulate and discuss objectives for the RA that are in fact objectives for any modernization program. This exacerbates the ability of the RA itself or an external evaluator to isolate the contribution made by the RA governance regime to overall progress on modernization.

Based on various descriptions about autonomy, and discussions of the governance model in place for central banks (and the potential benefits if such a model was carried to revenue administration), many have drawn the conclusion that autonomy is good for performance and that more autonomy is better. However, the basis on which this conclusion is reached remains unclear and largely anecdotal. In addition, there are few descriptions of any specific and practical results directly attributable to autonomy except in generally indirect terms (e.g., ability to pay higher wages, more professional staff, etc). If the autonomy itself cannot be directly linked to improved performance in some way, it is difficult to accept the “more is better” argument.

Finally, the basic problem with the literature to date is really one of measurement. It is extremely difficult to measure the contribution of tax administration generally to particular outcomes because there are so many influences on any one outcome. It is virtually impossible to measure the contribution of the RA governance model specifically to these same outcomes. In most of the literature that tries to assess RAs, the authors are really talking about any tax administration, not specifically an RA. For example, consider the simple statement “the revenue authority has increased revenues by 2 percent of GDP over a four-year period.” Some would take this as an indication of success for the RA in question. However, the statement tells us virtually nothing about the performance of the RA (as a tax administration) nor about the contribution of the RA governance model to that performance, as there are literally dozens of factors contributing to this outcome.

IV. The Questionnaire Responses and Other Data

A. Overall Response

Table 1 presents the RA questionnaire population and response rate. From the 29 RA countries that were solicited through the first and second questionnaires, 23 responses were received for an overall response rate of close to 70 percent. Information was also sought from four countries that have not opted for the RA governance model—Brazil, Chile, Ireland, and Norway. Many countries invested considerable time and effort in the preparation of responses which has greatly aided in the development of this paper.

Table 1.RA Survey Population and Response
Countries Surveyed*ResponseCountries

Surveyed *
Response
Argentina *Malawi
BoliviaMalaysia
BotswanaMauritius
Canada*Mexico*
Colombia*Peru*
EcuadorRwanda
EthiopiaSierra Leone
GhanaSingapore*
GuatemalaSpain*
GuyanaSouth Africa*
Italy*Tanzania*
JamaicaUganda*
Kenya*Venezuela
LesothoZambia
Zimbabwe

Indicates countries who responded to the first questionnaire.

Indicates countries who responded to the first questionnaire.

B. Organization and Presentation of the Data

The two questionnaires and the other information requests generated a significant amount of data. Results include factual information as well as more qualitative evaluations or assessments of the reasons for the initial decision to move to the RA model as well as the perceived effectiveness of the RA. This data was further supported by the experience gained by FAD in the provision of technical assistance and support to reform efforts in revenue administration.

A series of tables has been prepared in an effort to compile related data elements—to see an emerging picture or points of commonality or difference—and to provide a starting point for the analysis which is found at Section VI of this paper. The resulting tables include information on the legislative base of the RA, general features of the RA, the RA governance framework, scope of the RA, reasons for initially establishing the RA, assessment of RA effectiveness, and information on the practical results of HR autonomy. These tables may be found at Annexes V through XI and the information in the tables forms the basis for the findings which are discussed in the following section of this paper.

As with any exercise of this nature, the quality and completeness of the responses varied considerably. As was noted in the earlier section on the nature, limitations and scope of this paper, many options were considered to obtain the most reliable data that would form the basis for analysis. The questionnaire—in its two iterations in this case—was judged to be a realistic means to seek information. However, the responses are subject to the quality of record keeping in the organization and to the presence of corporate memory about decisions taken, in some cases many years ago. Finally, the qualitative nature of the questions inevitably leads to an element of subjectivity in the responses. Nevertheless, the data received does represent a comprehensive compilation of information on RAs. With the other information sources used (i.e., other literature, FAD experience, etc.), a reasonable basis is created for the findings and subsequent analysis.

While 23 RA countries answered the questionnaires, data from all 23 countries may not necessarily be included in all tables, because complete answers were not provided in every case.19

V. Summary of Findings

This section presents the findings from all the data compiled in Annexes V through XI as well as responses from the surveys themselves and other information, following generally to the structure of the two questionnaires.

A. Basic Information on RAs

Much of the basic information on RAs can be derived from Annexes V through VIII. The following observations or highlights are noted:

From Annex V: Legislative base for RA

  • A legislative instrument (law or decree) was used to establish the RA in all cases.
  • The earliest RA is 1988 (Peru), the most recent is 2004 (Mauritius).
  • A majority of countries (14 of 21) has introduced subsequent revisions to their enabling RA legislation, often to enable organizational or other changes (e.g., Kenya and Canada).
  • About 80 percent of the RAs (17 of 21) in the survey were described as having separate legal status (established as a “body corporate” or having “legal personality”). There appears to be common terminology around the world from the perspective of administrative law in that these entities are considered legal persons in their own right (can sue and be sued, own assets, and so on).
  • There is a great deal of commonality in the essential mandates of the RAs, according to their legal basis. Almost all have the equivalent mandate of assessing and collecting tax and administering and enforcing the revenue laws (a mandate notably shared by all tax and customs administrations, regardless of governance model). In addition, almost all have a further mandate to provide advice on tax laws to the minister of finance.

From Annex VI: General features of the RA

  • None of the respondents appears to be very far removed from government, in the sense that the minister of finance or equivalent minister is considered to have at least general supervision and oversight of the RA, with the possible exception of Guatemala where legislation clearly establishes significant independence for the RA (however, the minister of finance is named president of the board of directors in the law).
  • Seventy percent of the RAs (14 of 20) are outside the normal public service, in the sense that they have at least a minimum level of exemption from the normal public service regime (for example, the RA may be able to do its own hiring but must follow public service rules). Colombia, Ethiopia, Italy, Spain, and, to some extent, Mexico remain in effect part of the public service.
  • A slight majority of the RAs (11 of 20) has the authority to borrow, although in four of the 11 cases, the approval of the minister of finance is required.
  • Most RAs are able to own assets, in keeping with their legal status as a body corporate.
  • Different funding arrangements exist for RAs:
    • Normal budget and appropriation funding only (PEM)—40 percent or 8 of 20.
    • Normal budget and appropriation funding with an option for a percentage of revenue normally as incentive payment—30 percent or 6 of 20.
    • Funding based on a percentage of tax collection—30 percent or 6 of 20.
  • Almost all RAs appear to have the flexibility of a one-line budget and most include some kind of provision for the carryover of unused funds, especially if those funds had been earmarked for some particular project or program.
  • All RAs have arrangements in place for external audit, usually by the auditor general (17 of 20). In three cases (Botswana, Lesotho, and Zambia), the external auditor can be appointed by the Board (although in Lesotho, the government has the option of requesting the auditor general to audit).

From Annex VII: RA governance framework

  • Seventy-five percent of RAs (15 of 20) have boards (the exceptions are Colombia, Ethiopia, Peru, South Africa, and Spain). All except Zambia are empowered management boards20 with specific responsibilities and oversight functions. In the case of Zambia, the law establishes the board of directors as an operational board, not just a management board, vesting in it the powers and functions from the revenue laws normally vested in the CEO or equivalent.21
  • The boards range in size from 5 to 15 members and all have private sector representation.
  • Of the 15 RAs with boards, the chair of the board is named by the government (minister or head of state) in ten cases, the chair is specified in the legislation in four cases, and in one case the board elects its own chairperson.
  • In all cases except Zambia, the CEOs of the RAs are fully vested with the powers established in the revenue legislation and most countries specifically indicate these powers can be sub-delegated to other officers in the RA.
  • In most cases, the CEO is appointed by the government (11 of 20); in a further six cases, the appointment is made by the government on the recommendation of the board; and in three cases, the board itself makes the appointment of the CEO.

From Annex VIII: RA scope of activities

  • All RAs administer income tax and a value-added tax.
  • With two exceptions (Colombia and Singapore), all RAs administer excise taxes.
  • Almost all RAs (19) administer some other types of tax, including stamp tax (12), gaming tax (6), property tax (4), as well as a combination of various other taxes, including motor vehicle licensing and transport fees.
  • Less than 25 percent of the RAs (5 of 21) collect social security contributions.
  • Generally speaking, the RA model around the world includes customs administration within the scope of activity of the RA. Only Ethiopia, Singapore, Canada (where customs was removed from the RA in 2004 to form part of a border services organization), and Italy (which has a separate RA for customs) do not have customs as part of their RA.22

From data provided on the establishment of a RA and organizational reform (in response to the initial questionnaire)

  • The initial transition to RA does not appear to have precipitated organizational change, with the exception of the creation of a board with certain management responsibility for the organization and the creation of common service organizations. The shift to the RA model does not appear to lead to new reporting lines, organizational layers or even new type of organization, e.g., functional or other.
  • More significant change has often occurred within a few years of the RA being established. These changes included amongst others a move to a function-based organization, the integration of direct and indirect tax administration, organization based on taxpayer segmentation beginning with a large taxpayer offices, and other changes that flow from the modernization of revenue administration. In these respects, RA organizations appear to resemble non-RA organizations, except for boards and common support services (for tax and customs organizations).

B. Original Rationale for Considering and Adopting an RA

Respondents ranked prevailing reasons for the initial establishment of the RA (Annex IX: Reasons for Initially Establishing the RA). The aggregated results of these rankings is summarized in Table 2 below.

Table 2.Ranking—Reasons for Initially Establishing the RA
RankReasonAverage Ranking
1Low effectiveness of tax administration and poor levels of compliance1.8
2Need for a catalyst to launch broader revenue administration reform (modernized operations, improved automation, integrated and function-based structures, and so on)2.73
3Impediments caused by poor civil service human resources policies (recruitment, remuneration, promotion, training, discipline)2.9
4Poor communication and data exchange among the existing revenue departments (e.g., income tax, sales tax, customs)4.21
5Desire to create “islands of excellence” within the public sector4.54
6Perceptions of political/ministerial interference4.55
7High levels of corruption4.67

A number of observations can be drawn:

  • No single reason was identified by all respondents as being the main motivating factor for the establishment of the RA.
  • The highest priority reason for establishing an RA was low effectiveness of tax administration and poor levels of compliance (cited by 16 of 17 respondents and, of the 16, 11 rated it the highest priority). Box 2 describes the case of Peru, where revenue collection reached a record low in the early 1990s.
  • The next highest priority was the need for a catalyst to launch broader revenue administration reform, again with 16 of the 17 (and 7 of the 16) rating this factor either highest or second highest priority.
  • Third highest ranked was impediments caused by poor civil service human resource policies with 12 of 16 respondents rating this factor either first, second, or third in terms of priority.

Box 2.Peru: National Superintendency of Tax Administration (SUNAT)

What were the motivations for the establishment of the RA?

The SUNAT was established in early 1991. The authorities’ main motivation for setting up the agency were the chaotic economic and political situation in the late 1980s and the very low tax revenue/GDP ratio (tax revenue collection reached a record low 5.8 percent of GDP in the first semester of 1991). Inflation reached 7,000 percent in 1990. The guerrilla movements made it very difficult to enforce the tax laws, and the average salary of tax department staff was US$ 50/month.1/ Staff were frequently on strike (in one year tax administration staff were on strike for a total of 5 months). The creation of the SUNAT and its sister organization, the customs administration (Superintendencia Nacional de Aduanas, SUNAD) took place during a period of major political and economic changes. From its early days, the SUNAT followed a modernization strategy suggested by FAD. The strategy was based on simplifying the tax system, strengthening VAT administration, setting up a large taxpayer unit, and transferring the responsibility for processing tax returns and payments to the banking system. The most senior government authorities (including the president) fully supported the SUNAT.

What is the nature of the RA autonomy?

According to the law that established the SUNAT, it is a decentralized public institution of the economic and financial sector. This law endows it with autonomy to manage its budget and human resources, including the ability to determine personnel salaries on the basis of staff evaluations independently of the rest of the public sector. When the SUNAT was first established, the number of staff was reduced from 3,100 to 800. New staff were hired on a merit-based system. Average salaries rose from US$ 50.00 in 1990 to US$ 1,000.00 in 1991. The SUNAT’s budget is not a line item of the general government budget, but is a separate budget approved by congress that is, in principle, equivalent to 2 percent of total estimated annual tax collection. However, the agency has not always been allocated two percent of total tax collection.

The SUNAT is now subject to a budget law which bars it from increasing salaries or introducing new salary scales, which it has not been able to do since1998. Because salaries of staff hired after 1998 are frozen, their salaries are approximately half of those of old staff. Thus, salaries require readjustment and vacancies need to be filled.

What are the main results of the RA introduction?

Compared to the period before 1990, the SUNAT now has a much stronger and more skilled team of officials—a group of professional staff who have been recruited from the top graduates of Peru’s best universities. The agency has modern office buildings and modern IT systems. A modern web site is available to the public with useful information, and more than 60 percent of tax returns are filed electronically.

The SUNAT went through a difficult period in 1999-2001, when it lost the political support that had been key to important achievements it attained during the previous decade. Audit activities were drastically reduced and the quality of database information on taxpayers deteriorated.

This trend has been reversed, although continued efforts are needed to strengthen enforcement and more investment is needed in information technology. Since the SUNAT was established, tax revenue collection increased from 5.9 percent of GDP in the first semester of 1991 to 14 percent of GDP in 2005. It is difficult to gauge what percentage of this improvement can be attributed to improvements in tax administration associated with the establishment of the SUNAT, especially because some tax rates have increased and new taxes have been introduced. However, it is clear that tax administration effectiveness has improved and the level of noncompliance has fallen—from over 50-60 percent in the early 1990s to about 38 percent in 2005. Thus, the SUNAT is a valuable asset that needs to be preserved and strengthened.

1/ This was the average public sector salary at the time.

  • The remaining four reasons (poor data exchange among existing revenue departments, desire to create islands of excellence, perceptions of interference, and corruption) clustered between an average rating of 4.21 and 4.67, suggesting perhaps not a lot to differentiate amongst them.
  • High levels of corruption, often thought to be a key reason for establishment of RAs, was cited by the fewest number of countries (13) and ranked last on the list.
  • Six of the 17 respondents indicated there were other reasons as well as those listed in the questionnaire that were a priority consideration in the initial establishment of the RA. These included such reasons as improved general management, improved credibility, autonomy and flexibility over resources, better facilitation of organizational change, improved relationship with provinces, improved customs efficiency, and a better ability to combat tax evasion. It is fair to say that when countries identified these additional reasons, they rated them quite highly as a priority (usually first or second); however, because of the infrequent citing of each reason, they were not included in the tabulations.

C. Implementation of the RA

The first questionnaire asked countries to provide information on the implementation process as to: (1) duration of the process; (2) comments on the various stages followed; (3) project management processes or techniques used; (4) use of consultants; and (5) implementation costs. From the 12 replies received, some general observations follow:

  • Answers to questions about implementation were incomplete. This may in part be because the actual implementation of the RA is now somewhat dated in many cases (most were established in the early to mid-1990s) or that the implementation of a new organization, management, procedures, and so on was left to the new board of directors. Some governments may have seen implementation of the RA as more of a legislative and technical exercise rather than as a project management issue.
  • Respondents focus on legislative issues and little attention is given to the implementation process, e.g., critical stages, project management, and techniques and costs. Only three countries provided some information about the stages in the process and only two countries have responded to the question about project management and techniques.
  • None of the countries has given information about the costs of the transition to RA. All countries have, however, responded to the question on duration on the process,23 with the transition from being a ministerial department to RA likely to take two to three years.
  • Nine of the 12 responding countries have used consultants to assist in the establishment of the RA. The extent and areas of consultation are related to research work (reviewing the current administration), building the case for change, and the preparation of legislative issues (development of a strategic decision basis for the RA).

D. Self-Assessment of RA Effectiveness

Both questionnaires requested information dealing with the self-assessment of RA effectiveness. This question identified and defined 16 common revenue administration reform areas and asked respondents to rate the impact the RA governance model had on achieving progress in the particular reform, where applicable. Furthermore, respondents were asked to provide additional information about the nature of the impact of the RA model that would substantiate the quantitative assessment.

From Annex X (Self-Assessment of RA Effectiveness), a ranking of the perceived impact of the RA governance model on specific areas of revenue administration reform has been derived and is set out in Table 3.

Table 3.Ranking—Perceived Impact of RA Model on Reform
RankTypical Revenue Administration Reform AreasAverage Degree of Impact
1Integrate tax administration8.29
2Increase tax revenues8.11
3Provide a catalyst or momentum for reform8.00
4Improve automation7.79
5Improve services to the public7.53
6Improve remuneration7.53
7Improve human resources regime7.47
8Improve compliance management7.42
9Improve public confidence7.32
10Simplify laws, regulations, procedures7.06
11Improve budget provision6.89
12Improve anti-corruption measures6.83
13Introduce risk management6.76
14Improve valuation, classification, origin6.21
15Introduce taxpayer segmentation5.94
16Introduce self-assessment5.92

Some of the key observations from Table 3 include the following:

  • Countries ranked integrated tax administration,24 increased tax revenues, and providing a catalyst for reform as the areas of reform where the RA had the strongest impact. However, respondents did not describe how the RA model had impacted so strongly in these areas.
  • Improving the HR regime was the seventh highest ranked reform area where the RA had an impact.
  • The impact on improving anti-corruption measures placed 12th in the ranking.
  • It is noted that, in no cases, were respondents able to describe a direct link between the governance model and the particular revenue administration reform area. Where narrative comments were provided (less than half the cases), respondents tended to describe outcomes (such as “revenue increased”) and did not draw any conclusions or provide any specific examples about how the governance regime of the RA made a contribution to the reform in question.

Narrative information was also requested in the first questionnaire about the specific objectives initially established for the RA, processes for evaluating those objectives, including parliamentary review, and the results of any evaluations of the RA model that individual countries had undertaken.

Specific objectives identified are largely identical to what was described as main benefits and critical success factors. In addition, countries indicated that quantitative and qualitative objectives were set for the RA but no methods for evaluation were identified. Some countries have commented that the introduction of an RA worked as a catalyst to improve results and develop a modern tax administration, and positive results were realized in modernization of operations, HR and organizational structure. No substantiating qualitative or quantitative data was provided. To illustrate the rationale behind the creation of one RA and the key results achieved, Box 3 highlights the experience of the Uganda Revenue Authority.

The tax administration of almost every country reports its business results in their yearly report to government25 and mentions this as their key vehicle to evaluate the RA. Different objectives are described in these reports, and the annual report is used for evaluation and measuring overall performance, looking at broad performance indicators. Only one country (Canada) undertook a formal review of the RA as a requirement of the legislation that established the RA.

Box 3.Uganda Revenue Authority1/

What were the motivations for the establishment of the RA?

In the early 1990s, administration of national taxes and duties was the responsibility of four departments in the ministry of finance. Reasons for poor performance of these departments included: (1) low staff morale and productivity—partly due to low pay and shortage of resources; (2) corruption; (3) ineffective collection of tax; (4) weak management of revenue administration; (5) lack of a tax-paying culture—partly because taxpayers viewed the tax system to be unfair.

The argument for the RA was that, by moving away from civil service terms and conditions of service and management practices, may of these problems would be overcome. With higher salaries, staff would not need to seek alternate sources of income and coupled with stricter discipline, corruption would reduce, morale and productivity would increase as would revenue collections.

What is the nature of the RA autonomy?

The URA is a body corporate. It has human resources and budget autonomy. The URA is able to recruit and salaries and the rate of retention have improved. On funding, it receives a budget appropriation like any department of government but the minister of finance may authorize the retention of a percentage of the revenue collected. The URA has autonomy in setting all financial policies, with the exception of procurement.

What are the main results of the RA introduction?

Established in 1991, the URA was one of the first African RAs. Results were impressive in early years, reflected by strong revenue growth in real terms that leveled off by the late 1990s. Many of the previous administration and taxpayer compliance problems gradually returned, including serious problems of corruption and inefficiency. The URA had become a fragmented organization with unclear accountabilities when a major modernization initiative was launched in late 2004. The latest reform strategy has exploited the flexibility afforded by the RA model to competitively appoint an entirely new management team and workforce, structurally reorganize by integrating tax administration with a focus on segmentation, and begin streamlining and automating operations. Thus, the RA model has been the URA’s vehicle to both success and failure over the past 15 years.

1/ Source: (1) Planning for a Revenue Authority, Final Report, Coopers and Lybrand, June 1991; (2) questionnaire response provided by the URA; and (3) IMF staff.

Also requested in the first questionnaire was information concerning the customs/tax relationship, the existence of a statement of taxpayer rights and obligations, information about harmonized tax procedures and information concerning the rate of turnover of RA chairpersons and CEOs. Responses in these latter areas were not complete or consistent, and the information is not being further pursued in this study.

Finally, it is noted that the first questionnaire also requested quantitative information about revenue administration performance. The quality and completeness of responses varies considerably; however, some of these data has been tabulated and included for reference as Annex I.

Given the limited data provided, any substantive assessment of trends is not practical. It may be of general interest to examine the figures reported for tax-to-GDP revenue and to assess growth over the life of the RA. Of the 12 countries that provided data, one country reported static tax-to-GDP revenue over the life of the RA while four reported a decline in this figure. Of the remaining seven countries reporting an increase over the life of the RA, five countries saw growth of less than 2 percent while the remaining two countries (Argentina and Colombia) had growth over the life of the RA of 3.5 percent and 6.2 percent. This assessment is obviously unsophisticated and excludes the multitude of exogenous factors that can influence tax revenues other than the form of the administration.

E. Human Resources

As noted in the discussion of this paper’s methodology, the management of human resources was often cited as a key constraint in the ability of the revenue administration to operate effectively. As a result, it is one of the key reasons cited underpinning the decision to implement a RA. The questionnaires had a particular focus on HR and the specific question posed on HR is included at Annex III.

Annex XI (Practical Results of HR Autonomy) summarizes information collected from all RA country respondents to the two questionnaires. Based on the responses, the following observations can be made.

Recruitment

  • While respondents generally did not address the question of whether ability to recruit from the outside has improved with RA status, almost all countries (19 of 21) confirm they have the ability to recruit from outside. Countries did not draw a specific link to the RA governance structure; however, it was clear in the responses that some of these countries were clearly not able to recruit on their own from the outside prior to having RA status.
  • Slightly more than half the countries (11 of 21) indicated an improvement in outside recruiting times. Close to 30 percent indicated increased staffing times and the balance either did not comment or did not know.
  • All countries but one have the final say in acceptance of external recruits. This is as expected for administrations with full HR autonomy.

Remuneration

  • A clear majority of countries (17 of 21) indicated salaries had improved as a result of the movement to RA. Some of these countries provided quantitative data on the nominal change but only three provided comparative data, using either changes at the central bank (Uganda), changes in the public service (Canada), or comparisons with the public service minimum wage (Guyana) as benchmarks.

Performance pay

  • Twelve countries (57 percent) indicated they have performance pay plans of one kind or another. Most of these countries indicated they believe the RA framework had contributed to improving these plans but only one gave specific examples (Canada) of how the RA structure had led to a change in its management performance pay.
  • Many of the countries that indicated they had performance pay plans, in fact, had bonus plans related to the achievement of revenue targets, rather than plans for individual performance pay. A number of countries indicated that the development of performance pay plans was under consideration.

Retention

  • In terms of retaining staff and reducing attrition rates, more than half the countries (12) did not know or were unable to say whether this was being achieved in their RA. For the most part, these countries attributed this inability to a host of factors, including downsizing, budget constraints, and various program changes, all of which had an impact on retention and attrition.
  • Five countries of 21 (24 percent) stated clearly that their attrition rates were not declining in their RA; the remaining four countries indicated that attrition rates and employee retention were improving.

Internal staffing

  • Twelve of 21 countries either did not know or did not indicate whether timeframes for internal staffing had been improved. Of the remaining nine, eight indicated improved timeframes for internal staffing.
  • A third of the countries indicated that there were fewer restrictions for internal staffing under their RA status, with most of the remainder either not knowing or not stating anything about restrictions.
  • Regarding the basis for promotion, 17 of 21 countries indicated merit was a factor, 11 of 21 indicated competencies were a basis for staffing, five countries indicated that seniority was still a factor.

Organization/positions

  • The majority of countries has authority to determine its own organizational structure (15 of 21). Of the remainder, three do not, and three either did not know or did not comment.
  • There is a parallel position regarding the authority to classify positions.
  • The basis for classifying positions is job function for most of the countries, although a few also use job title.

Training and development

  • Only seven of 21 countries indicated that the number of training days or training days per employee had increased under the RA governance regime. Three countries said that the amount of training had actually decreased. Fully 11 countries did not know or did not say.
  • Seventeen of 21 countries indicated that their RA has its own training institute.
  • Seven of 21 countries indicated they used specific measures to quantify training and development improvements attributable to the RA. The remainder were either silent or indicated they did not have such measures.

Staff discipline and integrity

All countries reported developments in this area—including codes of conduct, asset and interest disclosures, ethics officers, internal affairs units, integrity action plans, increased disciplinary dismissals, and so on. No country specifically attributed any of these developments to the existence of the RA governance structure.

F. Comparison with Countries that Do Not Have an RA

As noted in the very first section of this paper, this research is not attempting to make a comparison of performance or reform and modernization between RA and non-RA countries. However, in examining the difficulties in evaluating the impact of the RA model on performance, it can be useful to look at the behavior of a smaller group of countries that has not opted for the RA model. An attempt was made to obtain relevant information from certain non-RA countries to use as a soft benchmark for any initial conclusions that could be drawn about evaluation issues.

Three countries provided responses to questions about reform and modernization (six countries had been solicited). These replies have contributed generally to our information about RAs but clearly do not represent a reasonable sample size for drawing any specific conclusions.

Box 4.Brazil: Federal Tax Administration

The Brazilian Federal Tax Administration (Secretaria da Receita Federal, SRF) was established in 1968 as a traditional Ministry of Finance (MOF) department. Since its creation, it has been organized by function and has administered domestic taxes (income, property and consumption taxes) and customs. Because it is part of the MOF structure, the SRF is required to follow government-wide legislation on human resources and budgeting, although it has been able to enjoy some degree of differentiation within the public service.

Human Resources policy:

Recruitment. Employees are admitted on the basis of a two-stage exam, a written test (in areas such as tax and constitutional law, accounting, economics) and a 3-month in-house course for those who pass the test. Even though a public exam is required for all civil servants in Brazil, there is a separate exam for admission to the SRF. The SRF defines the requirements for the exams, prepares the tests and conducts the course. This has resulted in the recruitment of personnel of good quality. However, the number of openings and the administration of the exam must be approved by the Ministry of Planning, Budget and Management (MPBM).

Remuneration. The SRF is required to apply salary scales that are previously approved by the MPBM. However, there is a special pay scale that is higher than the regular civil service pay scale that applies only to the SRF and other departments that conduct tax audits (such as the Social Security Administration). In addition to this basic wage, SRF staff also receive a bonus linked to performance (both individual performance and the collective attainment of revenue targets). This system has helped differentiate the SRF remuneration system from the rest of the civil service, and has enabled the department to recruit and retain quality staff. One problem the SRF faces is that the ‘additional’ salaries for managerial positions must be allocated based on civil-service wide rules. Because of this, the ‘additional’ salaries for line managers that do not always reflect the increased degree of responsibility, especially at the executive levels.

Lack of Political Influence

Because SRF staff are hired on the basis of an open competitive exam and the SRF has traditionally had a well-established corps of employees, the service has been able to create a culture of professionalism, and has eliminated the practice of political appointments within the organization over the last decade. This may also have helped to assure stability for commissioners.

Since 1995 the SRF has undertaken a modernization program which has focused on information technology and automation. The first electronic filing program was introduced in 1991, with 3 percent of taxpayers choosing this medium to file their returns. In 1997, the SRF introduced electronic filing over the Internet. Currently, all corporate tax returns and 97 percent of personal tax returns are filed electronically. All taxes and systems work according to a self-assessment regime. The SRF web site has received many awards for its taxpayer services and education programs.

During this period, the tax to GDP ratio grew continuously, reaching approximately 36 percent in 2004. Collections administered by the SRF account for nearly half of this amount.

The research also included numerous discussions with IMF staff, particularly concerning the issue of revenue administration reform generally. It is very clear from this information source that revenue administration reform and modernization is being pursued with equal comprehensiveness and rigor in many non-RA countries as it is in many RA countries. The kinds of revenue administration reform issues outlined in the questionnaire dealing with perceived effectiveness of RAs (see the questionnaire at Annex II) are exactly the same reform areas being pursued by non-RAs. Box 4 describes federal tax administration in Brazil where, although part of the civil service, some differences in HR policies are permitted and broader reform results have been realized.

Some preliminary data was available from OECD survey research (as noted in Section I-C of this paper) from 45 OECD and non-OECD countries and dealt with a variety of revenue administration program and structural issues. Of particular interest to this paper is data relating to certain elements of HR and financial autonomy, namely the ability to establish and design the internal structure of the organization, to allocate and re-allocate budgeted funds, and to determine the levels and mix of staff. The following table compares the RA and non-RA populations in the OECD research.

Table 4.RA/Non-RA Comparison on Ability to Design Internal Structure, Allocate/Re-Allocate Budget, and Determine Level and Mix of Staff(In percent)
Non-RA

(population = 14)
RA

(population = 7)
Ability to establish and design internal structure7986
Ability to allocate and re-allocate budget86100
Ability to determine the level and mix of staff86100

From this limited sample, RAs and traditional departments share a roughly similar ability to establish and design their own internal structures but diverge (although perhaps not significantly) when it comes to budget allocation and the level and mix of staff.

VI. Analysis of Major Issues

The previous section presented the findings of this study, organized largely around the headings from the survey questionnaires. However, to provide a focus on the most important issues, this section presents an analytical assessment along the following lines: (A) the decision to implement an RA; (B) perceptions of success and the link to reform; (C) features of a modern RA; (D) human resources autonomy; and (E) RAs and non-RAs.

A. The Decision to Implement an RA

Low effectiveness of the tax administration and the need for a catalyst for reform lead as reasons to implement an RA, while corruption and concerns about political interference rank last. Little to no cost-benefit analysis is undertaken, disadvantages of the RA are often not considered, and evaluations are seldom undertaken after implementation, which can be a lengthy process.

The two highest priority reasons for initially establishing the RA (low effectiveness and the need for a catalyst for reform) address issues that also tend to be high priorities for all revenue administrations, RAs or not. These two reasons received the highest scores from the survey questionnaires. This is not a surprising outcome given the importance of both reasons and given the broad movement toward reform in almost all countries. The third highest ranked reason namely removing impediments caused by poor civil service HR regimes is obviously an issue of much significance to the RAs (HR is dealt with separately in this analysis section). The remaining reasons from Table 2 are clustered closely together in terms of average ranking, and there is not much to differentiate among them. This inter-mingling of the reasons for which RAs are established with the reform objectives of traditional departments underlines the complication of any evaluation of the performance of RAs or comparison of such performance to that of a traditional department.

Typical rationales for RAs are intuitive and are not based on cost-benefit analysis. The impetus for the RA seems to stem from a belief (based on the experience of other countries or the literature or advice from experts) that a significant or even radical change, such as a move to RA, is required for reform to succeed or performance to improve. This is supported by the findings from the first questionnaire wherein respondents did not provide analysis or experience that would lead to an assessment that the RA, compared to any other form of governance, would solidify chances of improved performance or progress on reform.

Instead, government decisions appear to have been driven by the need for reform generally and were not supported by any specific assessment of the contribution the RA model could make versus the model of a traditional ministry or department.26

In addition, questionnaire respondents indicated little attention had been given to the implementation process as the focus was more likely to be on legislative issues required to establish the new model. No country provided information about the costs of the transition to an RA model, although, from the responses, it can be determined that the whole transition is likely to take a two- to three-year period. Finally, cost-benefit analyses were not undertaken in any case involving the establishment of an RA. This general weakness has been pointed out by a number of the academic papers on the topic of RAs and again demonstrates the challenge of evaluation.

The advantages and disadvantages of RAs have been well articulated. However, only the advantages seem to have been considered at the time of the decision to establish the RA. While the trend to move to RA has gained importance over the last decade (although it may no longer be in the ascendancy), much has been said in the literature about some of the negative aspects of creating RAs. Taliercio (2004) and Mann (2004)27 have both commented in some detail in this regard and their main lines of thought are captured in the literature review of this paper. Despite the clear articulation of the many negative aspects of RA, an analysis of these factors does not appear to have formed part of the initial decision-making process of any government that responded to the survey.

High levels of corruption is ranked lowest as a rationale for establishing an RA. Perceptions of political interference and high levels of corruption were ranked second last and last, respectively, despite some very common perceptions that these were important considerations in adopting the RA model in the first place. This is anomalous with popular thinking of the reasoning behind RAs and the major preoccupation of many governments with addressing corruption. One reason for this may be that it is difficult for countries to provide a frank answer to this question. In that sense, the down-playing of the role of corruption as an original rationale may be somewhat revisionist and not a true reflection of it as a motivating factor in the creation of an RA.

Post evaluations do not seem to be a feature of RAs. Generally, countries were not able to report any specific evaluation contemplated at the time of the creation of the RA, in anticipation of a future need to assess its effectiveness.28 Nor have there been very many subsequent assessments of RAs (there are some assessments of specific input-related data, such as data on HR improvements and performance data related to operations such as enforced collections). However, as has been mentioned repeatedly, linking any of this to the RA governance model has been difficult. When asked about evaluation of the RA, most countries included mention of the generic and normal reporting to government without identifying how this reporting isolated the impact of the RA on overall operations. Performance attribution difficulties remain a major obstacle in assessing any RA.

Implementation of RA can take a long time. Few respondents to the more detailed questionnaire were able to provide much detail on the implementation process. From those who did respond, and from the experience of FAD, it is concluded that many used consultants to assist in the development of the plan and, in most cases, the overall process took anywhere from two to three years. Little information is available on how the implementation of the RA was managed or the exact costs attached to this change but what is clear is that considerable resources were consumed in RA implementation. Furthermore, this level of effort clearly has opportunity cost in terms of what is not done. It is impossible to determine the extent to which the energy and time and money invested by the organization in a shift in governance of this nature may have been a sound investment, given the paucity of information available about any impact the new model had on performance of the organization.

Box 5.The Canada Revenue Agency1/

In 1996, the Government of Canada announced in the Speech from the Throne and the budget that it intended to convert Revenue Canada from a department of government into what it then called a “revenue commission.” In taking this step, the government wanted the new organization to achieve three objectives:

  • Provide better service to Canadians.
  • Become a more efficient and effective organization.
  • Establish a closer partnership with the provinces and territories.

An Act to Establish the Canada Customs and Revenue Agency was introduced into parliament in June 1998 and assented to in April 1999. The Canada Customs and Revenue Agency came into existence on November 1 of that year.

The legislation contained five major elements: mandate and governance of the Agency; accountabilities; partnership responsibilities; human resource authorities; and administrative authorities.

According to the evaluation undertaken by the CRA, the manner in which the Agency implemented the legislation has resulted in meeting two of the original objectives that were set for it: a more efficient and effective administration and a closer relationship with the provinces and territories. Importantly, however, the fundamental transformation of the Agency’s human resource and administrative regimes also served as the essential ingredient in enabling the Agency to meet its objective of improving service to Canadians.

1/ Taken from The Canada Revenue Agency: the First Five Years, Setting the Foundation for Tax and Benefit Administration in the 21stCentury, November 2005.

B. Perceptions of Success and the Link to Reform

The role of perception in any assessment of RA success should not be underestimated. RAs are seen as making a strong contribution to reform—although no direct linkage was made by respondents to the RA model itself.

Perceptions of RAs themselves and others are that the RA model has made a significant contribution to revenue administration reform. Strong perceptions exist on the part of RAs themselves, some donor countries, international organizations, and academics that RAs have generally made a contribution (and sometimes a significant one) to revenue administration reform. In the response to the survey question that asked for a self-assessment of how effective the RA model was in contributing to specific aspects of reform, respondents assigned very high scores (an average of 7.2 out of 10), across a range of 16 different reform areas. Furthermore, of those 16 reforms, there was an average frequency of response rate of more than 88 percent, indicating a very strongly held belief that the RA governance model was a significant contributor to reform in almost every area. Box 6 describes the case of Guatemala, where an RA has been in place since 1998 and where HR autonomy is seen as a contributing factor to success in tax administration reform.

Box 6.Guatemala: Superintendency of Tax Administration Superintendencia de Administración Tributaria (SAT)

What were the motivations for the establishment of the RA?

In early 1997, the government introduced a program of reforms to achieve the objectives of the Peace Agreement, which was signed in December 1996. As part of this program, the authorities adopted a package of reforms to strengthen its tax system and administration. The reform program’s objectives included raising the tax/GDP ratio from around 8 percent of GDP (a figure that had not been surpassed for many years) to 12 percent through reforms of the tax system and its administration. Soon thereafter, and based on the successful experience of the SUNAT in Peru, the authorities decided to establish the SAT. The main motivation for establishing a relatively more autonomous agency was to provide greater flexibility to the tax administration in order to create a corps of professional, appropriately paid and motivated tax administrators, based on a professional career system. The SAT was established on February 21, 1998, and took over the powers and responsibilities of the former Inland Revenue and Customs Departments.

What is the nature of the RA autonomy?

The SAT is a decentralized public institution of the economic and financial sector, and is a body corporate. It has human resources and budget autonomy. The SAT is able to recruit, establish salary scales and define career paths. The SAT has autonomy in setting all financial policies, with the exception of procurement.

What are the main results of the RA introduction?

The SAT is a modern tax administration that has the basic requirements to efficiently and effectively manage the taxes assigned to it. It has systems to monitor delinquent taxpayers and current accounts; is expanding services provided to taxpayers and has qualified staff, good facilities and computer equipment. The organizational structure is also sound and there is a large taxpayer unit that monitors about 75 percent of collections. Tax collections rose from 8.7 percent of GDP in 1998 to 10 percent in 2005. Even though it has many of the desirable features of a modern tax administration, the SAT has not always been able to act effectively during its existence due to managerial instability, changes in the agency’s strategic focus, legal problems, lack of a consistent use of information in the operational areas, and some cases of corruption. Since its creation, there was also a period (2000-2001), in which there were attempts to remove the SAT’s autonomous status and convert it back into a traditional government department in the ministry of finance.

Source: IMF staff.

Perceived reform success was not linked by respondents to the existence of an RA governance model. The three highest-ranked revenue administration reform areas impacted by the RA structure were: integrating tax administration; increasing tax revenues; and providing a catalyst for reform. It is noted that these are also goals for reform in a non-RA environment. Furthermore, narrative explanations for these and the other ratings in the questionnaire did not draw any link to the governance model.

The creation of the RA per se, with the attendant expectations of increased management control and autonomy, better resourcing and staffing, and more control over its own destiny, was generally expected to result almost automatically in overall improvements in revenue administration. However, the contribution of the RA governance model to performance improvement in revenue administration has proved to be most elusive to establish.

The lack of data available makes it difficult to assess the impact of the RA governance model on performance other than by assessing perceptions. Through the more detailed questionnaire, respondents were asked for data covering a number of aspects of revenue administration performance, as well as for their own assessment of how the RA governance model has contributed to improved performance. The responses demonstrated that most administrations do not collect or analyze data that could permit specific conclusions to be drawn.29

C. Features of a Modern RA

Revenue authorities remain firmly entrenched in government and there is much homogeneity in their mandates and corporate character. Management boards are a feature of many RAs and, while the board seems to have evolved to a common model, the influence of the board is difficult to assess. The CEO generally maintains management control of the RA. Organizational structures did not change dramatically with the implementation of the RA. Customs is included within most RAs but managed as a separate entity from tax administration. Budget autonomy and certainty of funding vary greatly between RAs.

Revenue authorities remain an integral part of government. Regardless of the degree of autonomy that may or may not be enjoyed by RAs, they remain key institutions of government. The powers to tax and to control the movement of people and goods at a country’s borders are among the most pervasive and potentially intrusive aspects of the state’s involvement in the lives of its citizens. Government leaders rarely agree to cede oversight and control of (and accountability for) such core functions of government.30

From the review of RA legislation, RAs continue to operate under the regular oversight and often direct supervision of a minister of government, to administer laws that clearly establish authorities and accountability to the government, to report in a formal manner to the legislature on plans and activities, and to receive core funding as a decision of that legislature. The RA governance model appears to have done little to shift this fundamental relationship between the tax and customs administrations and the executive branch of government. While autonomy from government is frequently noted as an important benefit stemming from the creation of an RA, many features of RA legislation, discussed below, suggest that this autonomy may be more limited than thought.31

In addition, such autonomy mostly pertains to management issues related to revenue administration, not operational or technical administration of pertinent revenue legislation. There does not appear to be a significant difference in this operational/technical autonomy between RA and non-RA countries.

Management boards in RAs have evolved to a common model—a board empowered to exercise management and oversight roles, but excluded from taxpayer operations. Management boards are an integral part of 75 percent of the RAs surveyed and all but one are empowered management boards that are prohibited from direct involvement in taxpayer cases (this was not so for some of the early RA boards, which were directly involved in individual tax cases). These boards exercise oversight over the classic management disciplines, e.g., human resources, planning, financial management, asset management, and so on, and generally play a role in the selection of officials for key senior positions. Excluding the boards from involvement in tax cases, and thus from access to confidential taxpayer information, paves the way for private sector representation (which is the case for all 15 boards). The presence of these boards is one of the more distinctive characteristics of an RA.

The real influence of the management board is difficult to assess. While the surveys did not seek specific information on the outcomes related to the active presence of the board in the organization, it is argued by some that management boards with private sector representation can inject a more business-oriented approach to the workings of the RA and thus bring more rigor to financial and HR matters. As these private sector members are usually appointed by the government, the extent to which they will operate within the same philosophical parameters as that government remains unclear.

Some boards do have some explicit legislative responsibility for operational matters (Tanzania—to review and approve operational policy; Mexico—to provide opinions to the minister on a range of operational issues; and, of course, Zambia). These cases represent a stronger operational oversight than the norm. Conversely, South Africa passed legislation that abolished the advisory board in favor of the creation, as needed, of advisory councils that meet only as required (two advisory councils have been established), which may imply at least a dissatisfaction with the advisory board model and a weaker involvement of the advisory councils in the workings of the RA.

The authority of the Board varies but the daily operation of the RA rests squarely with the CEO. Notwithstanding the empowered board concept and the fact that their responsibilities are often extensive, a little over half the boards have no role in the appointment of the CEO (11 of 20). In six cases, the board recommends the appointment of the CEO to the executive branch in government and in three cases the board itself makes the decision on the appointment.

Furthermore, except in the three cases referred to earlier in board involvement in operational issues, no board has any case-specific involvement in the operations of the RA and the entire responsibility for the daily management of operations, including the handling of taxpayer cases, continues to rest with the CEO. A CEO with these operational authorities vested directly from the revenue legislation is in many senses similar to the head of a tax administration in a traditional department of government. Any real assessment of the differences in their roles has been difficult to ascertain.

The general mandate and corporate character of RAs have evolved to a common model. Of the survey respondents, 80 percent have separate legal status. However, all share the same general mandate—the administration of tax laws and, in most cases, customs. With the implementation of the RA, there appears to have been little change in the core mandate of the organization. RAs continue to be responsible for direct taxation and indirect taxation at a national level (Annex VIII).

RA organizational structures did not change dramatically in the transition from government department to RA. With the exception of the creation of the board and the creation of common support services, in most cases (based on the detailed questionnaire), there was little real change to the organizational structure with RA implementation. Once the RA had been in place for a few years, some countries further refined the organizational structure, often to implement a functional model of revenue administration or to implement a taxpayer-focused organization. It is a matter of open debate whether these organizational changes were driven by the fact of the existence of a maturing RA or were rather a natural organizational evolution that would have been witnessed regardless of the governance model. These kinds of organizational changes do represent the core of expert advice (including that of the IMF) to revenue administrations seeking to improve performance and are in many cases implemented in the absence of an RA. Once again, the problem of attribution of change or improvement to the existence of the RA remains.

The scope of most RAs includes customs, and customs organizations within an RA tend to remain separate from the tax administration. In most countries, the customs administration is included within the overall organizational framework for revenue administration, as a part of the RA. However, in the great majority of countries, customs administration is the responsibility of a completely separate organization and operates with close linkages to the tax administration. Customs operations are not truly integrated with those of the tax administration. This is likely because of the inherent “real-time” nature of customs operations compared to tax administration, and because customs must carry out its operations where it has physical control over its transaction environment (border crossings, international airports, secure warehouses, and so on). Furthermore, there appears to be a widening non-fiscal role for customs (security-related) that will tend to keep it apart. These different features of customs operations tend to preclude the true integration of business processes and client services with tax operations and, thus, reduce the scope for effectiveness and efficiency gains, such as are possible with the integration of direct and indirect tax operations.

Budget autonomy and certainty in terms of funding is not a characteristic shared by all RAs. Perhaps surprisingly, most of the RAs that responded to the questionnaire do not enjoy a significant level of autonomy from the ministry of finance in terms of funding. Seventy percent (14 of 20) receive basic funding through the normal government appropriation and budget process. However, close to half of this group also has the option to receive a portion of revenue collected as an incentive payment (often to augment staff salaries), but this is at the discretion of the minister of finance. It could be argued that this same option is open to any revenue administration, RA or not.32

Many RAs must operate within the broader budget and spending framework established by the government, following normal rules for public accounts, and often for asset management and procurement. Most RAs are afforded some flexibility to carry forward unspent funds to the next financial year, but there are often limits on the carry-forward provision, e.g., length of time, total amount, specific project slippage only, and so on. All RAs in the study can own assets, but again many countries reported that these assets are subject to broader government rules on asset management. About half of the RAs (11) reported the ability to borrow, but the approval of the minister of finance is still required in four of these cases.

The key difference between most RAs and a government department in the areas of budgets and spending appears to be the ability to own assets, some ability to carry forward unspent funds (a feature increasingly being applied to normal departments in many countries), and the possibility of additional incentive funding provided by the minister of finance based on a percentage of collections arrangement. This level of autonomy is somewhat less than might have been expected based on conventional wisdom about RAs. However, it is clear that some of the countries that are operating on a true cost-of-collection formula are convinced their budgets are larger than would have otherwise been the case. Anecdotal evidence also indicates that, for countries funded on a percentage of collection basis, the full allocation is often not made available to the RA, given broader fiscal constraints facing the government.

D. Human Resources Autonomy

HR autonomy is accepted as desirable with little consideration of the alternatives. HR autonomy is considerable but quantifying the impact of this autonomy on performance has proven challenging—as has quantifying the impact of the RA on the development of performance pay plans. While HR autonomy is stronger than financial autonomy for RAs, it is still circumscribed by other factors. Respondents were generally unable to measure adequately attrition and retention rates and, where measurement was possible, were unable to attribute changes. Training and development and the establishment of integrity programs were equally difficult to link to the existence of the RA.

Human resources (HR) autonomy has generally been accepted as a desirable feature without a great deal of discussion or consideration of alternatives. Most of the literature written about RAs cites independence from the HR regime of the central civil service as one of the key motivators to create an RA. Revenue administrations are often described as burdened with the inefficiencies of these central regimes and, if given real autonomy (or even partial autonomy), they could offer more competitive salaries, attract and retain better staff, deal more effectively with the problems of corruption, and generally exercise better management control over their own human resources.

In the first, more detailed, questionnaire, few respondents described the exact nature of the constraints of the civil service regime. There was no reference to options other than withdrawal from the civil service regime that may have been considered by the government, for example addressing the HR situation across government rather than limiting solutions only to tax and customs administrations.

Most RA countries reported considerable autonomy in human resources, and large performance gains related to that autonomy. However, quantifying these improvements has proven difficult. For instance, 19 of 21 reported full authority to recruit externally, with final say in selecting individuals recruited in this manner and 17 of 21 reported full authority to determine organizational structure and to classify positions in the organization. These and other authorities that derive from autonomy have led to some significant HR performance gains, including: improved recruiting times (11/21), increased remuneration (17/21), and better timeframes for internal staffing (12/21).

However, in fewer than half of these cases of HR improvement were the respondents able to provide quantifiable measures of results. Furthermore, in most cases where results had been quantified, respondents simply reported a change in a particular measure (for example, remuneration “has increased by x percent” over some time period, internal staffing times “are getting close to a target of x days”) without relating this result to some external benchmark, such as average central bank wages, or average internal staffing times before they became an RA. One exception was remuneration—where three of the countries provided comparisons with outside benchmarks that demonstrated a relatively greater improvement in remuneration inside their RA. One of these countries also had quantifiable data regarding attrition and retention rates.

Improvements in HR performance are, of course, an indirect measure of improvement in revenue administration performance. However, it is generally accepted that there is a direct link between the two. It is disappointing, then, that there was not more quantification of the improvements in HR performance stemming from the autonomy offered by RA status. It can only be concluded that countries just assume improvements such a salary increases or reductions in recruiting times are attributable to their RA status, and see no particular need to measure the improvements or to compare them against outside benchmarks.

Information gathered was unable to shed any light as to how the RA governance structure may have contributed to improvements in performance pay plans. The ability to link remuneration and performance is widely considered an important aspect of HR management, and the RA model is often proposed as being able to facilitate this development. However, even though slightly more than half the RA countries reported having a performance pay plan (for about half of these the plan consists of a bonus scheme related to the achievement of revenue targets), only one was able to demonstrate how RA status had led to improvements in the development of performance pay plans. A number of others indicated that the introduction of performance pay was under active consideration. It is not clear from this rudimentary analysis that the RAs are any more advanced in this area than non-RAs.

Increasing retention of key employees and reducing attrition rates are critical objectives of increased HR autonomy. However, most respondents indicated that there were too many factors at play affecting these issues to permit meaningful measurement or attribution of results. These are key HR objectives for any revenue administration, and HR autonomy has usually been seen as a critical means to achieve these outcomes. However, it appears that this assumption may have been too simplistic. Many countries reported difficulties in measuring impacts in these areas, and in attributing any changes in attrition and retention rates to their own HR policies. Various external factors are seen as having a much more profound impact: the general economic situation; retirement and pension-related issues; downsizing initiatives; perceptions about public sector employment. As a result, the RAs surveyed generally had very little measurable and quantifiable data in these areas, and very little information about trends.

Training and development also proved difficult to relate to the RA governance model. Only 7 of 21 RAs surveyed indicated that the number of training days per employee had increased, and fully half did not report on this question or did not know. The remainder reported that training days per employee had gone down. In these latter cases, respondents pointed out this was usually due to exogenous factors, such as budget or staff reductions. None of the information provided can assist in answering the main question about the impact of the RA governance framework on training and development.

Many activities are being taken by RA countries regarding staff discipline and integrity; however, all specific initiatives cited in the survey questionnaires are also being undertaken by countries in a non-RA governance environment. There does not appear to be any particular set of initiatives RAs have launched in developing their integrity plans that could not also be applied in a non-RA environment. While an adequate level of remuneration is only one factor in a modern approach to integrity, it appears to be the main area where the RA governance model could make a difference.

While most of the RAs operate outside the public service regime, their real autonomy can be circumscribed by other factors. For instance, while the RA may have complete autonomy over salary levels and hiring, its operational budget often continues to be provided by the government and the overall salary cost must be met from within this budget. The government determines the level of funds to be made available to the RA and this immediately dictates limits on job structures, salary ranges, and economic increases including those made as a result of collective bargaining. It can also be noted that an RA remains very much a public institution, and must have transparent human resources policies that reflect its public nature. As a result, RAs often end up putting in place human resources regimes based on the same principles that governed the regime they left.

E. RAs and Non-RAs

Many countries that have traditional departments of government seem to enjoy many of the flexibilities that in the past were only possible through RAs.

It may be that non-RA countries in a modern governance environment enjoy many of the flexibilities often considered to be only available in the RA governance structure. As discussed in Section V-F of this paper, it was difficult to obtain specific information from non-RA countries to establish a soft benchmark for comparisons between the governance models. Of interest, though, is the data derived from the preliminary OECD research. These data demonstrates that, for three aspects of management authorities (establish and design the internal structure, allocate and re-allocate the budget, and determine the level and mix of staff), the differences between RAs and non-RAs are indeed slight.

While this can in no way be considered comprehensive, this pattern seems to raise the question of whether many of the flexibilities and much of the autonomy sought by RAs may, in fact, be available to traditional departments of government, as part of more general initiatives in modern management and comptrollership. Box 7 illustrates this point by providing a description of the Australian Tax Office (a traditional department of government) and how its staff are remunerated.

Box 7.Australian Taxation Office Staff Remuneration

The Australian Taxation Office (ATO) enjoys a degree of flexibility in determining the remuneration levels of its 20,000 staff. Essentially, the ATO’s system of remuneration operates within a four-tiered structure:

  1. ATO head and immediate deputies (Commissioner of Taxation and three Second Commissioners) are statutory office holders whose remuneration package is set by an independent review tribunal responsible for establishing remuneration levels of politicians, judiciary, and other statutory officers.
  2. Senior Executive Service (SES) staff (the ATO’s senior leadership team comprising around 200 officers) are remunerated according to individual Workplace Agreements (WAs). A WA is a specific package of pay and conditions negotiated by the ATO head, within parameters set by government, with each individual SES officer. In practice, SES officers with broadly equivalent roles and responsibilities are placed in groups (or salary bands) for remuneration purposes, but some variation can be expected from individual to individual to take account of particular circumstances. SES officers are either permanent (open-ended) appointees, or subject to fixed term contracts (typically two to five years). Fixed-term contracts are increasingly being used in relation to high-level information technology, legal, and human resource management positions. Remuneration levels of SES staff on fixed-term contracts are generally higher than those appointed to open-ended positions. All SES officers participate in a performance bonus scheme that offers bonus payments from 5 percent of base pay (for fully effective performance) to 15 percent (for exceptional performance).
  3. Executive-level staff (including senior technical operatives, senior team leaders, and managers—about 1600 are subject to pay and conditions established in a specific executive-level staff agreement. This agreement is negotiated, within certain government-wide parameters, by the ATO head with the executive-level staff as a group; there is no union involvement because the staff are considered to be part of the executive management structure. An objective of the agreement is to provide benefits and incentives that appropriately recognize the leadership role performed in the ATO by the executive-level staff. Pay rises for the Executive level staff must be offset by productivity improvements. Executive-level staff participate in a performance pay scheme similar to that applying to SES officers.
  4. Lower-level staff (ranging from the lowest-level operative staff through to mid-level team leaders—about 19,000 are subject to pay and conditions established in a General Employees Agency Agreement (GEAA). The GEAA is a formal agreement between the ATO head and staff (and their unions), and is negotiated within policy parameters set by government. Any pay rise included in the GEAA must be offset by productivity improvements (i.e., the ATO does not receive specific budget increases from government to meet the cost of pay rises it negotiates with its staff and their union).

For both tier 3 and tier 4 staff, Workplace Agreements may be used to offer individual remuneration packages for particular staff, especially where there are attraction or retention issues.

VII. Conclusions

As was stated in the introduction, this paper has attempted to compile a substantive body of information on the key features of revenue authorities and more importantly, to discuss and assess the many issues related to the evaluation of RAs. Observations are woven throughout the various sections of the paper that confirm that any solid evaluation of RA performance and then attribution of that performance to the RA governance model is particularly challenging.

The limitations of the replies to the questionnaires used as one of the main sources of information for this paper have also been discussed. While others have used case studies and various analytical models to review RA performance, this paper has relied on questionnaire data and the experience of the IMF’s Fiscal Affairs Department in the delivery of technical assistance to countries both with and without an RA. The following conclusions, drawn from the findings based on the questionnaire along with in-house knowledge and experience, represent a starting point. By their nature and the nature of the data, the conclusions are not definitive (and caution should be used in interpreting them in any other way). Ongoing discussion among tax administration professionals, academics, and the donor community could draw on the possible application of that performance, success and potential attribution of that success to different governance models—from traditional departments of government to revenue authorities.

Defining an RA

1. A generally accepted or approved definition of an RA does not exist.

Revenue authority and other similar terms, such as semi-autonomous RA, are used by various writers and organizations, often without being defined. This paper uses a specific and arbitrary definition that covers a particular range on a spectrum.

RA and performance

2. The proposition that RAs have led to better revenue administration performance compared to what would have been the case had they remained a traditional department of government has neither been proved nor disproved.

This is an important conclusion of this paper even though it may appear for some to be indecisive. No administration surveyed appears to have measured in any systematic way the impact of the RA governance model. Until they do, decisions whether to adopt an RA structure will have to be made based on reasons other than an objective assessment of its impact on performance. With a few exceptions, basic data that might have assisted in ascertaining a link between the existence of an RA and improved performance are neither captured nor used by the RAs themselves.

3. There is a great deal of difficulty in attributing changes in performance (outcomes) to a particular governance framework.

Many performance indicators in revenue administration have literally dozens of possible contributing factors. The paper repeatedly makes the observation that linking changes in performance to features of the governance structure has proved elusive. Establishing such a link may, in fact, be impractical and almost impossible, with the result that other more intuitive or case-specific approaches may continue to be seen as the best direction for further research.

Reform and modernization

4. Clearly, there is no basis to conclude that reform and modernization of revenue administration is more likely within an RA governance framework than without.

Many RA and non-RA countries have been launching similar reform programs, and undertaking similar initiatives irrespective of their governance framework. As one developed country representative indicated to the research team, “we could have done most of these initiatives without RA status anyway.”

5. Amongst countries which have an RA, many see the existence or creation of the RA governance model as a catalyst to promote and extend reform.

This conclusion is based on the results of the questionnaires and on FAD experience in the delivery of technical assistance (especially in Latin America). The need for a catalyst was rated strongly as a reason for establishing an RA in the first place and relatively high scores were assigned in the self-assessment of the impact of the RA model on reform.

Autonomy

6. The general case has not been made that the degree of autonomy (more autonomy or less autonomy) can be specifically related to improvements in performance or the ability to undertake reform and modernization.

According to some case-study research, autonomy matters and more autonomy matters more. This has not been borne out as a general conclusion by any of the findings of this study, not because this is necessarily a faulty premise but primarily because no specific examples to support such a general conclusion were available.

7. RAs remain closely linked to government in terms of the role of the minister and accountability to the executive and to parliament.

It is unlikely that RAs would ever move too far from the accountability of elected officials, given the nature of the powers assigned under the revenue laws (intrusive and pervasive). Furthermore, perceptions of political and ministerial interference ranked very low as a reason for establishing an RA. In modern revenue administrations, RA or not, there are clear understandings about political involvement in the technical administration of the revenue laws. RAs tend to have similar provisions to prevent the involvement of their boards in these same issues.

8. RAs frequently have a great deal of administrative autonomy flowing from their “corporate” status, primarily in areas such as finance, budget, and asset control. However, the comparative advantages of this autonomy may be less significant than thought.

RAs are gaining relatively greater administrative autonomy by virtue of their legislative status. However, in terms of financial flexibility (the availability of a one-line budget, carryover of unspent funds, and so on), it is noted that such features are becoming increasingly prevalent in non-RA revenue administrations as well. With respect to budget autonomy, this is often characterized as one of the main features of an RA (i.e., an established relationship between the revenue administration budget and the amount of [forecast] revenue collected). On closer examination, RA legislation is somewhat limited in this area in many cases. Furthermore, there is some evidence that in actual practice in many cases broad public service-wide decisions of the treasury may in fact take precedence over budget “guarantees” made to the RA, including those sometimes included in legislation.

9. HR autonomy is almost always accepted as a desirable feature for an RA on the grounds that it will address poor performance. However, alternative solutions may not be given sufficient consideration.

There appears to be little consideration in many cases of alternatives to HR autonomy as the solution for poor HR performance in a country’s civil service. Options to address deficiencies by reforming civil service institutions more broadly may not receive sufficient discussion when an RA model is under consideration. However, broad civil service HR reforms are occurring in many countries and have clearly lead to some significant improvements in a number of non-RA countries.

10. RA countries have received a great deal of HR autonomy and have used it to improve HR performance. However, quantifying performance improvements, and linking them to the HR autonomy granted, has clearly proved difficult.

It was expected that quantifying and linking performance improvement to HR autonomy would have been easier than quantifying and linking such improvement to the more general RA governance model, since there are many specific areas in HR that lend themselves to a comparison with a non-RA benchmark. While there was more quantification in the HR area than generally, it was not extensive.

Cost benefit

11. Cost-benefit analyses were not and are not being undertaken with respect to the implementation of RAs, nor has an attempt been made to assess the opportunity cost associated with setting up the RAs (including costs associated with the expenditure of management time and energy).

In view of the inability to unequivocally demonstrate a link between the RA governance model and better performance, it would have been expected that this kind of cost-benefit assessment would have played a much more significant role in any decision to establish an RA. Further, academic research is often cited only as a positive reinforcement for the establishment of an RA, despite the fact that much of this research clearly indicates that there are disadvantages to the establishment of RAs.

Perceptions

12. The case for establishing RAs in the first place has been largely intuitive.

This is a broad conclusion about perceptions. Most countries clearly believe they had a sound basis for establishing their RA. On closer inspection, though, this was usually a well-articulated list of problems or inefficiencies in revenue administration that they believed needed to be addressed. An intuitive leap was made that the RA was the best solution to solve those problems and address those deficiencies. The basis of this leap appears to have been: recommendations of bilateral donors and international organizations; practice in neighboring countries; and/or generic or qualitative evaluations of success in other areas and locations. In some countries, the leap may well have been the right decision given the political context and the need for a dramatic catalyst for change, particularly in areas where other reform initiatives had failed.

13. Notwithstanding the lack of a demonstrated basis for establishing an RA, there is a strong perception held by those countries that have adopted the RA concept that this particular governance model has made a significant contribution to reform and improved performance.

This conclusion is eminently clear from the research. There is little doubt that the RA administrations truly believe and promote the notion that their RA is making a difference.

Summary

This paper is a good example of the difficulties in being conclusive on a subject as complex and with so many inter-dependencies as a governance framework for revenue administration. It is hoped, though, that the findings and conclusions do advance the discussion of revenue authorities, particularly in this era of widespread reform and modernization.

Based on this review and the reforms under way in many countries around the world, the following advice should be offered to countries considering the establishment of a RA model as a path to more effective revenue administration:

  1. Establishing an RA should not be viewed as a panacea—creating a RA is clearly expensive, may take a long time and require significant effort, and does not actually improve tax administration effectiveness.
  2. Before considering a choice of governance model, revenue administrations should clearly identify and articulate problems and deficiencies, and consider strategies for reform and modernization based on international best practice.
  3. Only after a proper diagnosis is undertaken should a full assessment be made of the extent to which the RA governance model might satisfy the problems and reform strategies identified.
  4. Whatever the governance model, continuing political commitment is of the utmost importance in establishing and maintaining a professional and effective revenue administration. Experience of many developing countries shows that after an encouraging start, the performance of RAs has frequently been undermined by political changes.
  5. The RA model alone does not lead to improved effectiveness and taxpayer compliance—its establishment must be coupled with a serious commitment and plan for reform. In fact, modernization of tax administration is ultimately the result of improvements in organizational structures, systems, and processes, including well designed programs of services and enforcement, sound allocation of resources, and effective management. A move to implement a new governance model for the revenue administration can only be, at best, a first step.

This paper is a contribution to an ongoing discussion rather than a definitive assessment of the value of RAs. A comprehensive data set on the features of RAs has been created, representing about 75 percent of the countries that have chosen this governance model. These data, and the findings and conclusions in this paper, should be of use for further research on revenue authorities. Some areas of further research could include the following:

  • Developing an “index” to determine relative autonomy or degree of autonomy of RAs.
  • Correlating such autonomy to external measures, such as changes in the Transparency International corruption index or other measurable data.
  • Using political economy analysis with detailed case studies to further explore some of the specific attribution problems identified by this paper.
  • Undertaking comparative studies with governance changes in other domains (such as the central banks).
  • Assessing “actual” autonomy versus the autonomy prescribed by law.
  • Reviewing the actual operations and contributions of RA boards to determine effectiveness.
  • Assessing the nature of customs operations within the RA, and examining the extent of customs-tax integration actually occurring.

ANNEX I

Bibliography for the Literature Review

    CastroPatricio and JamesWalsh2003The Organization of Customs Administration” in Changing Customs: Challenges and Strategies for the Reform of Customs Administrationed. by MichaelKeen(Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    CunninghamBert2004Re-Engineering and Re-Structuring the Civil ServiceCustoms and Excise Mauritius.

    DelaySimonNickDevas and MichaelHubbard1999Reforming Revenue Administration: Lessons from Experience—a study for the International Development DepartmentSchool of Public Policy, University of Birmingham (May).

    • Search Google Scholar
    • Export Citation

    Di JohnJonathan and JamesPutzel2005The Case of Ugandapaper for Conference on Institutional Change for Growth and Poverty Reduction in Low-Income CountriesJuly 6–7, 2005 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    FjeldstadOdd-Helge2005Revenue Issues and Corruption: An Overview of IssuesWorkshop on Corruption in Public Financial Management: Donor PerspectivesFebruary16–17, 2005Michelsen InstituteBergen, Norway.

    • Search Google Scholar
    • Export Citation

    GrayJohn and EmmaChapman2001Evaluation of Revenue Projects Synthesis Report Volume IUK Department for International Development.

    • Search Google Scholar
    • Export Citation

    HindriksJeanMichaelKeen and AbhinayMuthoo1999Corruption, Extortion and EvasionJournal of Public EconomicsVol. 74No. 3 pp. 395430.

    • Search Google Scholar
    • Export Citation

    JenkinsGlenn P.1994Modernization of Tax Administrations: Revenue Boards and Privatization as Instruments for ChangeBulletin for International Fiscal DocumentationVol. 48No. 2 (February).

    • Search Google Scholar
    • Export Citation

    ManasanRosario2003Tax Administration Reform: (Semi)-Autonomous Revenue Authority Anyone?Discussion Paper Series No. 2003–05 (revised) Philippine Institute for Development Studies.

    • Search Google Scholar
    • Export Citation

    MannArthur J.2004Are Semi-Autonomous Revenue Authorities the Answer to Tax Administration Problems in Developing Countries?—A Practical Guideresearch paper for the project Fiscal Reform in Support of Trade Liberalization (USAID funded) (August).

    • Search Google Scholar
    • Export Citation

    McCartenWilliam2005The Role of Organizational Design in the Revenue Strategies of Developing CountriesWorld Bank Institute.

    • Search Google Scholar
    • Export Citation

    MeyerLaurence2000The Politics of Monetary Policy: Balancing Independence and AccountabilityU.S. Federal Reserve Systemremarks at the University of Wisconsin (October).

    • Search Google Scholar
    • Export Citation

    SilvaniCarlos and KatherineBaer1997Designing a Tax Administration Reform Strategy: Experience and GuidelinesIMF Working Paper No. 97/30 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    TaliercioRobert Jr. 2004Designing Performance: The Semiautonomous Revenue Authority Model in Africa and Latin AmericaWorld Bank Policy Research Working Paper 3423 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation

    WallerMichael1996Danish Tax and Customs AdministrationInternational Training Division, Department for Personnel and Organizational DevelopmentDenmark.

    • Search Google Scholar
    • Export Citation

    WallerMichael1997Perspectives for Integrating the Administration of Internal Taxes, Customs Duties and Social Security Contributions: Effects in a Decentralized StateCanada for CIAT 1997 Technical Conference.

    • Search Google Scholar
    • Export Citation

    WallerMichael2001Review of Integrity Management in the Independent Revenue Authorities of Ghana, Tanzania and UgandaTransparency International: the Coalition Against Corruption Working Paper (June).

    • Search Google Scholar
    • Export Citation

    WallerMichael2002Evaluation of Revenue Projects- Synthesis ReportEvaluation Department, DFID.

    WallerMichael2002Options for the Establishment of a Central Revenue Authority in MozambiqueWorkshop Discussion Paper Ministry of Planning and Finance-Government of Mozambique IMF DFID.

    • Search Google Scholar
    • Export Citation

    WallerMichael2003Consultancy Support to the Implementation of the Mauritius Revenue AuthorityPrice Waterhouse Coopers.

    WallerMichael2003Supporting Institutional Reforms in Tax and Customs: Integrating Tax and Customs Administrationspaper prepared for World Bank Group by PLS Ramboll and supported by the Danish Governance Trust Fund.

    • Search Google Scholar
    • Export Citation

    WallerMichael2005Revenue Authorities and Taxation in Sub-Saharan Africa: A Concise Review of Recent Literature for the InvestmentCompetition and Enabling Environment Team, DFID.

    • Search Google Scholar
    • Export Citation

ANNEX II

International Monetary Fund

Fiscal Affairs Department

Survey on Revenue Authorities November 2005

Introduction

Over the course of the last decade, many governments have struggled with poor performance of the tax and customs departments and have sought new ways to approach the business of revenue administration. A revenue authority33 has been seen by some as a possible solution to critical problems, e.g., poor revenue performance, low rates of compliance, ineffective staff, and corruption. It has been argued that a revenue authority can lead to improvements including better accountability for results, synergies in administration across the revenue departments, and management based on professional skills and isolated from external influences.

Given its role in the provision of technical assistance to member countries in fiscal administration, IMF advice is often sought on programs of reform and strategic change. By extension, this includes the question of whether a revenue authority is an effective tool or catalyst for improving performance. However, to date, no comprehensive assessment has been made of the revenue authorities that exist today and of how well they have served their intended purpose. The purpose of this survey, therefore, is to seek information that will allow an assessment of the relative success or failure of the revenue authority model in improving revenue administration performance.

This information, along with academic papers written on the subject of revenue authorities as well as internal information in the Fiscal Affairs Department, will assist in the preparation of an analysis and assessment by the IMF of revenue authorities and an articulation of the conditions and issues affecting revenue performance.

This questionnaire is divided in to five sections:

  • Basic information on the revenue authority
  • Original rationale for considering and adopting the revenue authority
  • Implementation of the revenue authority
  • Self-assessment of revenue authority effectiveness
  • Quantitative and qualitative indicators of revenue authority effectiveness

The assistance of your administration is sought in completing this questionnaire. We anticipate that the evaluation paper that will be prepared as a result of your input will be available by the end of June 2005 and a copy will be sent to you.

If you have any queries about the contents of this questionnaire, or if you require any explanations, please contact Ms. Maureen Kidd at MKIDD@imf.org (English/French) or Ms. Katherine Baer at KBAER@imf.org (English/Spanish). For all other issues please contact Ms. Carla Cullati at CCULLATI@imf.org.

I. Basic Information on the Revenue Authority (RA)

1. Please provide a brief history of the RA, including date of implementation, any major changes in the revenue authority over time, current status e.g., no changes contemplated or structural improvements being considered, etc.

2. Please provide a copy of:

  • the initial legislative authority (enabling law) and any significant revisions made subsequently
  • any existing formal regulations relating to revenue authority governance, and a short description of the importance of each
  • high-level organization charts of the RA: (a) immediately prior to the implementation of the RA; (b) the initial organization chart of the RA; and (c) the current organization chart. The charts should be dated and should show all organizational units (departments, divisions, sections, etc.) that report directly to the CEO as well as the approximate staff and budget levels for each
  • excerpts describing the RA from any public documents such as corporate plans, annual performance reports, etc.

3. Please provide a description of the key features of the RA:

1. Degree of autonomy

  • a description of the legal form and status of the RA
  • a description of funding arrangements (provisions in law and actual practice)
  • a description of budget flexibility of the RA (e.g., the controls, if any, on the approved budget)
  • a description of the financial policies of the RA (such as accounting, asset management, procurement- for example, do the civil service policies apply or are decisions left to the RA?)
  • a description of the HR regime, including whether HR policies fall within the civil service control framework or outside it. Please also attach a short description of current policies on recruitment, promotion, discipline, remuneration and training.Governance framework
  • What is the role of the minister of finance in the RA—according to the law and in practice?
  • What is the role of the board (if any)—is the board advisory or fully empowered to take decisions through legislation? Please describe.
  • What is the role of the Director General (or CEO)—is it a coordinating role only or does the CEO have full responsibility for revenue operations with all vested powers from revenue laws? To whom does the CEO report? Please describe.Accountability
  • How does the RA report to the government and parliament? Please describe any specific measures, in law and in practice.
  • Please describe the specific arrangements for external and internal audit, in law and in practice.Scope
  • Please list all the taxes administered by the RA.
  • Please describe the scope of operations of the RA (i.e., tax administration, customs administration, collection of social contributions, collection of other fees and taxes, etc.)

II. Original Rationale for Considering and Adopting the RA

4. Please rank in priority order the possible reasons listed below for the initial establishment of a revenue authority, as well as any other reasons you may identify. Do not assess any listed if not applicable for your country.

Ranking
Low effectiveness of tax administration and poor levels of compliance
Impediments caused by poor civil service human resources policies (recruitment, remuneration, promotion, training, discipline)
Perceptions of political/ministerial interference
Need for a catalyst to launch broader revenue administration reform (modernized operations, improved automation, integrated and function-based structures, and so on)
Poor communication and data exchange among the existing revenue departments (e.g., income tax, sales tax, customs)
High levels of corruption
Desire to create “islands of excellence” within the public sector
Other reasons (itemize and rank)

5. What were the main benefits of a revenue authority that were presented to the government (i.e., the case for action)?

6. Were any critical success factors or conditions for success established at the outset, or subsequently?

7. What were the reactions or observations from parliament/congress, if any, during passage of the initial legislation?

III. Implementation of the Revenue Authority

8. Please provide a brief general description of the process followed to implement the RA in your country. Of particular interest will be the duration of the process, comments on the various stages followed, project management processes or techniques used, whether consultants were engaged, information on the costs of implementation, and any other information about implementation that may be of interest.

IV. Self-Assessment of RA Effectiveness

9. Please describe any specific objectives initially set for the RA, if any.

10. Please describe any processes for evaluating those objectives, e.g., parliamentary review, as well as any criteria established for evaluation.

11. Please provide the results of any evaluation undertaken and any subsequent modifications as a result (evaluations of the RA or of the RA contribution to other objectives).

12. Most countries in the world have launched revenue administration reforms over the last ten years or so. This applies to those countries who have adopted a revenue authority model (the minority), and to those who have not (the majority). The essential difference between a revenue authority and a more standard revenue administration is the governance model (degree and nature of autonomy etc). The purpose of this table is to identify and quantify the specific impact of the revenue authority governance model in achieving revenue administration reforms. For each of the “revenue administration reform areas” in column 1, respondents are asked to quantify in column 2 the degree of impact of the revenue authority model, using a scale of 1 to 10 (10 being high impact, 1 low and 0 no impact). If the revenue authority governance model had no impact on a particular reform (say introducing taxpayer segmentation in tax administration), then it should be assigned a 0. If this area is not part of the reform program, not applicable (n/a) should be indicated. In column 3, respondents are asked to provide specific information about the impact of the revenue authority model that would substantiate the quantitative assessment. Respondents could also indicate whether the impact is direct or indirect and explain why. In order to complete the questionnaire, brief definitions of the typical revenue administration reform areas are as follows.

  • Increase tax revenues—self-explanatory.
  • Introduce self-assessment—legislative changes to move from administrative assessment of taxes (including trade taxes) to a program of self-assessment.
  • Integrate tax administration—refers to moving from a type of tax structure (for example, separate administration of income tax and VAT) to integrated tax administration along functional lines
  • Introduce taxpayer segmentation—refers to the development of specific administrative regimes based on taxpayer size, usually beginning with a large taxpayer unit and subsequently designing approaches for medium and smaller taxpayers.
  • Improve taxpayer and trader compliance management—refers generally to improved operational practices including audit, management of arrears, filing compliance, etc.
  • Improve services to the public—self-explanatory.
  • Introduce risk management—refers to efforts to adopt risk management and risk profiling practices in all aspects of revenue administration but particularly in audit and post-assessment compliance.
  • Improve automation—self-explanatory.
  • Simplify laws, regulations, and procedures—self-explanatory.
  • Improve valuation, classification and origin—applies where customs administration is part of the revenue authority.
  • Improve human resources regime—this refers to improved recruitment, staff development and training.
  • Improve remuneration—self-explanatory.
  • Improve budget provision for the operation of the revenue administration—self-explanatory.
  • Improve anti-corruption measures—self-explanatory.
  • Provide a catalyst or momentum for reform—did the fact of the creation of the RA provide any focus or catalyst for the government to launch a revenue administration reform program?
  • Improve public confidence—identify and quantify the extent to which the adoption of the RA governance model may have improved the confidence of the public in the revenue institution.

13. For countries where customs is included in the RA, did the establishment of the RA affect the relationship between the tax and customs administration? Please illustrate with examples.

14. Do you have a public statement of taxpayer rights and obligations? If so, when was it published? Please provide a copy.

15. Have tax administration procedures been harmonized among tax types (i.e., penalties and interest, timeframes for submitting returns, procedures for audit and dispute resolution, etc.)? When was this accomplished?

16. Please provide information on changes in the management of the RA, specifically, how many chairpersons of the board and CEOs have there been since the establishment of the RA? What is the average length of time in each position?

V. Quantitative and Qualitative Indicators of RA Performance

The difficulties in providing a time-series of performance data that can be used for international comparisons are well understood. So too are the difficulties in relating such data to something as specific as the features of a revenue authority. Nevertheless, the information asked for below will serve to provide a useful analytical background.

Ideally, it would be preferable to have all requested data in an annual time series that begins 2 or 3 years prior to the establishment of the RA through to end 2004. The idea, of course, is to assess the extent to which there may be trends that coincide with the establishment of the RA or with other specific events. In cases where data may not be available over this time period, respondents are asked to provide whatever time series data they have that could show such trends.

17. Please complete the attached table.

18. Table of quantitative data

1995199619971998199920002001200220032004
Tax revenues (percent of GDP)
Total
Corporate income tax
Personal income tax
Excises
Customs
VAT
Other
Arrears (at year end)
Nominal
Percent of annual collections
Registration (# of taxpayers)
Corporate income tax
Personal income tax
VAT
Excises
Income tax returns
Total returns (all taxes)
Average time required to process
Returns filed electronically
Customs (if included in RA)
Total customs
Declarations filed electronically
Declarations subject to
physical inspection
Average release time
Administrative costs
Total expenditures (not including donor financing)
Donor funded expenditures
Dispute resolution
Objections (internal)
Average completion time
Appeals (quasi-judicial)
Average completion time
Human resources
Staffing levels at year end
1. Management
2. Officers
3. Support staff
New recruitment
Average salary (entire organization)
Average salary (central bank)
Percent staff with university degree
Average training days per person
Staff discipline cases (dismissal,

demotion, suspension, etc.)

International Monetary Fund

Fiscal Affairs Department

Survey on Revenue Authorities January 2006

Short questionnaire

In 2005, FAD sent detailed questionnaires to more than 20 member countries who had adopted the revenue authority34 concept for revenue administration. A total of 14 countries responded. Information from the survey questionnaires has been compiled and a discussion paper on findings to date has been prepared.

In order to better reflect the more than 30 IMF member countries that use a revenue authority model, a revised and considerably shortened questionnaire has been developed. It is being sent to all those countries who did not respond to the 2005 survey, as well as to an additional 12 member countries.

This revised questionnaire is divided into three sections:

  • Basic information on the revenue authority
  • Original rationale for adopting the revenue authority
  • Self-assessment of revenue authority effectiveness
  • Practical results of human resources autonomy

The assistance of your administration is sought in completing this questionnaire. We anticipate that the evaluation of this additional information will be completed by April of 2006, and a final paper on revenue authorities available by mid-year..

If you have any queries about the contents of this questionnaire, or if you require any explanations, please contact Ms. Maureen Kidd at MKIDD@imf.org (English/French) or Ms. Katherine Baer at KBAER@imf.org (English/Spanish). For all other issues please contact Ms. Carla Cullati at CCULLATI@imf.org.

I. Basic Information on the Revenue Authority (RA)

2. Please provide a copy of:

  • the initial legislative authority (enabling law and relevant regulations or decrees) and any significant revisions made subsequently
  • high-level organization charts of the RA: (a) immediately prior to the implementation of the RA; and (b) the current organization chart.
  • descriptions the RA from any public documents such as corporate plans, annual performance reports, etc.

3. Please list all the taxes administered by the RA and describe the scope of operations of the RA (i.e. tax administration, customs administration, collection of social contributions, collection of other fees and taxes, etc.)

II. Original Rationale for Considering and Adopting the RA

1. Please rank in priority order the possible reasons for the initial establishment of the revenue authority. Use the reasons suggested, or others you may wish to identify.

Ranking
Low effectiveness of tax administration and poor levels of compliance.
Impediments caused by poor civil service human resources policies (recruitment, remuneration, promotion, training, discipline).
Perceptions of political/ministerial interference.
Need for a catalyst to launch broader revenue administration reform (modernized operations, improved automation, integrated and function-based structures, etc.)
Poor communication and data exchange among the existing revenue departments (e.g., income tax, sales tax, customs).
High levels of corruption.
Desire to crate “islands of excellence” within the public sector.
Other reasons (itemize and rank).

III. Self-Assessment of RA Effectiveness

1. The essential difference between a revenue authority and a more standard revenue administration is the governance model (degree and nature of autonomy etc). The purpose of this table is to identify and quantify if possible the specific impact of the revenue authority governance model in achieving revenue administration reforms. For each of the “revenue administration reform areas” in column 1, respondents are asked to quantify in column 2 the degree of impact of the revenue authority model, using a scale of 1 to 10 (10 being high impact, 1 low and 0 no impact). If the revenue authority governance model had no impact on a particular reform (say introducing taxpayer segmentation in tax administration), then it should be assigned a 0. If this area is not part of the reform program, not applicable (n/a) should be indicated.

  • Increase tax revenues—self-explanatory
  • Introduce self-assessment—legislative changes to move from administrative assessment of taxes (including trade taxes) to self-assessment.
  • Integrate tax administration—moving away from a type of tax structure (for example, separate administration of income tax and VAT)
  • Introduce taxpayer segmentation—development of specific administrative regimes based on taxpayer size.
  • Improve compliance management—operational practices including audit, management of arrears, filing compliance, etc.
  • Improve services to the public—self-explanatory
  • Introduce risk management—adopting risk management and risk profiling practices particularly in audit and post-assessment compliance.
  • Improve automation—self-explanatory
  • Simplify laws, regulations, and procedures—self-explanatory
  • Improve valuation, classification and origin—applies where customs administration is part of the revenue authority
  • Improve human resources regime—this refers to improved recruitment, staff development and training
  • Improve remuneration—self-explanatory
  • Improve budget provision for the operation of the revenue administration—self-explanatory
  • Improve anti-corruption measures—self-explanatory
  • Provide a catalyst or momentum for reform—using the fact of the creation of the RA to provide a catalyst for revenue administration reform.
  • Improve public confidence—the extent to which the adoption of the RA governance model may have improved public confidence in the institution.
Typical Revenue Administration Reform AreasDegree of Impact
1.Increase tax revenues
2.Introduce self-assessment
3.Integrate tax administration
4.Introduce taxpayer segmentation
5.Improve compliance management
6.Improve services to the public
7.Introduce risk management
8.Improve automation
9.Simplify laws, regs., procedures
10.Improve valuation, class., origin
11.Improve human resources regime
12.Improve remuneration
13.Improve budget provision
14.Improve anti-corruption measures
15.Provide a catalyst or momentum for reform
16.Improve public confidence

IV. Practical Results of Human Resources Autonomy

Increased autonomy in human resources is a major feature of most revenue authorities. It is important to try to understand what has been accomplished with this increased autonomy. Therefore, respondents are requested to provide any quantitative or qualitative information they can related to the following areas:

  • Recruitment—Has the ability to recruit staff from outside the revenue authority improved? Does it take less time? Does the RA have more or final say in who gets hired, compared to the situation before the RA?
  • Remuneration—Have salary levels improved? Can such improvements be quantified in any way (for example, average wage in the RA now compared to average wage in the central bank now, versus the same comparison before the RA was established?
  • Performance pay—Is there a better basis for performance pay (bonuses) than there was prior to the RA?
  • Retention—Are employees staying longer with the organization? Have attrition rates declined? Can these issues be quantified?
  • Internal staffing (promotions)—Have time frames for internal staffing improved? Are there fewer restrictions than before the RA? What are promotions based on? Merit? Competencies? Seniority?
  • Organization and positions—Does the RA have full authority in these areas? On what basis are positions classified? (by title, by job description, etc.
  • Staff training and development—Has the number of training days per employee per year increased? Has a specific training institute been established? Are there any specific measures used to quantify training and development improvements attributable to the RA?
  • Staff discipline and integrity—Are there specific developments in this area that have been made possible by the establishment of the RA?

ANNEX III

Question on Human Resources Autonomy—Sent to Countries that Had Answered the November 2005 Questionnaire

Increased autonomy in human resources is a major feature of most revenue authorities (RAs). It is important to try to understand what has been accomplished with this increased autonomy. Therefore, respondents are requested to provide any quantitative or qualitative information they can related to the following areas:

  1. Recruitment—Has the ability to recruit staff from outside the RA improved? Does it take less time? Does the RA have more or a final say in who gets hired, compared to the situation before the RA?
  2. Remuneration—Have salary levels improved? Can such improvements be quantified in any way (for example, average wage in the RA now compared to average wage in the central bank now, versus the same comparison before the RA was established?
  3. Performance pay—Is there a better basis for performance pay (bonuses) than there was prior to the RA?
  4. Retention—Are employees staying longer with the organization? Have attrition rates declined? Can these issues be quantified?
  5. Internal staffing (promotions)—Have timeframes for internal staffing improved? Are there fewer restrictions than before the RA? What are promotions based on? Merit? Competencies? Seniority?
  6. Organization and positions—Does the RA have full authority in these areas? On what basis are positions classified? (by title, job description, etc.)
  7. Staff training and development—Has the number of training days per employee per year increased? Has a specific training institute been established? Are there any specific measures used to quantify training and development improvements attributable to the RA?
  8. Staff discipline and integrity—Are there specific developments in this area that have been made possible by the establishment of the RA?

ANNEX IV

Revenue Authorities—Design and Scope Considerations35

In its simplest expression, a revenue authority (RA) takes one or more of the traditional revenue departments from the ministry of finance and amalgamates them together in a separate agency that is given some degree of autonomy from the government and from public sector policies. Beyond this common definition, RAs vary greatly from country to country, depending on the design and scope features discussed in this report. It is important to note that all features carry a range of possibilities, and the sum of them all determines the relative autonomy enjoyed by the particular organization.

A. Overall Design of the Revenue Authority

The design of the RA is a reflection of the policy choices made by the government. Overall design includes the degree of autonomy from the government and the public service, the governance framework that regulates the RA, and mechanisms for maintaining accountability to the government (including the legislature).

Degree of autonomy

A first important design issue concerns the degree of autonomy that is to be given to the RA. In revenue administration, a major element of autonomy is the degree of separation from the public service and its policies. In this sense, an RA can vary from a close-to-normal line department to a significantly (but not completely) autonomous corporate body. Much depends on the legal form of the RA and the provisions governing its funding and financial management policies, human resources regime, and organizational arrangements.

Legal form

Corporate character. Some RAs have a “corporate character” or status embodied in the separate enabling legislation that permits a significant level of autonomy to be designed. For example, a separate legal character can imply the right to own assets, a feature which suggests more rather than less managerial autonomy. Also, the legal status will sometimes determine the application of public administration law and policy to the RA. That is, the extent to which public service laws, regulations and policies apply, in whole or in part. RAs that lack a separate corporate character enjoy far less autonomy and are potentially subject to far more government regulations and control.

Legal mandate. The legislation establishing an RA provides a framework for the roles, functions, powers, and accountabilities of the organization itself and of the other elements of its governance framework. The general mandate and composition of the RA should be clear and unequivocal. The legislation will sometimes identify the relevant components or departments and specify that the RA will be initially comprised of these organizational units, their employees, assets, etc.

Parliamentary review. Because RA legislation can sometimes be experimental or controversial, some countries have included provisions built into the legislation for a formal parliamentary review after a certain period of time.

Assets and borrowing. It is common to find that the enabling legislation confers a corporate character to the RA. Such status offers the possibility of greater autonomy including the right to own assets, borrow or invest funds, and be exempted from certain public service requirements. Clearly, an agency that owns its assets (e.g., buildings) can optimize the use of these assets in a more autonomous mode than otherwise. Alternatively, a government could decide that asset optimization at the broader government level is more important, in which case the RA would be provided with less autonomy in this area. The RA legislation should also deal with the issue of the right to borrow or invest funds, although in many cases borrowing is not an unfettered right as prior minister of finance approval is required.

Application of public service-wide legislation. The legislation should further specify the extent to which any existing public service legislation should continue to apply, in whole or in part, to the RA. For example, a RA could be exempt from some or all public service legislation related to human resources. Other public service legislation, such as laws related to collective bargaining and to public finance and accounting, would also need careful review for the same reason.

Organizational autonomy. Organizational autonomy is a major consideration. It involves the degree of independence in establishing organizational arrangements. Where such autonomy is provided, the RA will be free to decide how to organize its staff both at headquarters and in field offices, when to add or close offices in different locations, or when to otherwise adjust the structure of the organization in any way deemed desirable. While organizational autonomy is a normal prerogative of an institution with a “corporate character,” the public sector environment presents certain sensitivities that require attention. For example, the government clearly has a direct interest in such matters as physical presence throughout the country, the location of customs operations, and regional balance.

Need for governance framework. The above legislative features carry with them a number of consequential implications. For example, a corporate body such as described above normally needs a (management) board to supply effective governance, and the autonomy suggested by this model necessitates a set of transparent mechanisms to provide accountability to the executive and the legislature. These issues and others, which will also need to be clearly set out in the legislation, such as funding, are discussed in sections below.

Funding

Normal appropriations. There are two basic means for providing ongoing funding for an RA. The first is the provision of a standard parliamentary appropriation using the normal public expenditure management (PEM) and budget decision processes of government. This in effect would be the same as for any other department of government and is the method providing the least autonomy in the funding area.

Collection fee. The second is the inclusion of a percentage-of-collection funding formula, or guarantee, in the RA legislation. This can also be referred to as a collection fee. The main arguments in favor of a legislated percentage-of-collection funding formula are that it can insulate the RA from the vagaries of a suboptimal budgeting process (if that is the case), it can provide greater certainty and reliability, and it can be structured to provide incentives for improved performance.

There are, of course, many arguments against the use of such a formula, including the main argument that there are many factors that impact revenue collections, not just the efforts of the revenue administration. Clearly, these would include such items as the general performance of the economy and tax policy changes. In addition, concerns could be raised that the formula, by linking funding to revenue collection, could create perverse incentives for the RA to focus excessively on collections and ignore other important tax administration functions, such as service to taxpayers.

In the case where a greater degree of autonomy is desired, the legislation sometimes establishes the principle of a percentage-of-collection fee, set as a percentage of total estimated net revenues. Estimated revenues are considered a better measure than actual revenues since they can be determined beforehand and hence be used in establishing budgets before the fiscal year begins.

Financial management policies

As far as financial management policies are concerned, there are two key issues: (1) the degree of application of legislation or public service regulations or rules in the areas of financial administration and accounting, procurement, asset management, and the like; and (2) the degree of budget flexibility granted to the RA (i.e., the movement of funds from one budget category to another). Revenue authorities typically obtain considerable autonomy in areas such as budget flexibility, procurement, and asset management. One issue to consider is the extent to which the RA will remain part of the government accounting entity, and thus the extent to which accounting and financial administration policies would continue to apply.

A thorough review needs to be made of the body of law, regulation, and policy in these areas that applies to normal government departments. A first order of consideration is the extent to which any of these ought to continue to apply in the case of the more autonomous RA, or at least the extent to which principles in these laws ought to be included in the RA legislation.

Accountability of the board. Where these policies become the responsibility of the board, it is the board rather than the government that will be accountable for ensuring the policies are developed and implemented, and that the public interest is respected. In these cases, the expectation is that the RA would use its autonomy in these areas to adopt practices that would allow it to be more efficient and effective, by tailoring policies to its own business environment, and by maintaining flexibility to incorporate quickly the best of other public and private sector practices.

Budget flexibility. In terms of budget flexibility, normally the treasury or the budget office in the finance ministry sets the rules and procedures for all government organizations. These concern primarily the ability to move funds across budget lines (such as capital versus recurrent expenditures, salary versus non-salary expenditures, and so on). The finance ministry must operate with a whole-of-government perspective, and may introduce controls not particularly suited to an operational set of functions such as would be performed by an RA. Where a government opts for a more autonomous RA, the board—rather than the government—usually determines policies in this area for the RA.

One reason for granting such autonomy may be that the RA needs to be able to respond quickly to changing demands in areas such as enforcement, or service, and be able to make trade-offs between budget lines in a manner not possible in the normal government context. In this case, the board would need to set out new operating guidelines for the RA for budget flexibility, which continue to respect principles of responsible management and prudent stewardship of public funds.

Financial administration and accounting. This refers to the fiduciary responsibilities of a public institution: accounting practices, payment, chart of accounts, accounts receivable, invoicing, contracts, etc. As the RA often remains part of the government’s accounting entity for purposes of public accounts, it may want to retain identical policies and practices with the public service in many of these areas. If these become a board responsibility, in the more autonomous model, the board will want to ensure that principles of responsible and transparent financial management are fully adhered to.

Procurement. The purchase of goods and services in the public sector is founded on the principles of fair and open competition, value for money, and transparency. Public sector procurement must also reflect trade agreements and other complex policies. If these become a board responsibility, then the board will have to ensure that RA policy on procurement complies with these principles, while at the same time permitting quicker response and more flexibility than any outmoded civil service system it has replaced.

Asset management. An RA needs clear policies on asset management, including such areas as life-cycle management, lease versus purchase, space optimization, asset disposal, and the like. All public sector departments must have these policies to a certain extent, normally dictated by the ministry of finance. For an RA with the right to own assets, these policies normally become the responsibility of the board.

Capacity of the RA. Two important implications follow from the large degree of financial autonomy enjoyed by a more autonomous RA. First, a substantial amount of development work will need to be undertaken to design the full range of financial management policies that are needed to support the operations of the RA. Second, the RA will need to recruit sufficiently skilled managers and officers in the financial area in line with the larger role that will be played by the organization’s finance department compared to that of the finance departments in the normal revenue departments.

Human resources regime

The human resources regime refers to all aspects of human resources, including staffing, training, compensation, performance pay, dismissal, discipline, and, in some cases, collective bargaining. As in the case of financial management, autonomy in managing human resources is another key feature that sets an RA apart from other institutional forms of revenue administration.

Board accountabilities. Many RAs are given significant authority in the human resources area, as it is sometimes viewed that management efficiency and effectiveness can be improved by getting out from under an outmoded regimen of public service human resources rules and regulations. In such cases, the legislation might specify the human resources principles that will apply to the RA (e.g., fairness, redress, competency, and so on). The board would be required to establish human resources policies that respect these principles. Depending on the degree of autonomy chosen, the board’s sphere of responsibility for human resources might include the following:

  • the overall human resources regime for the RA;
  • the salary scheme for employees and positions;
  • the performance assessment scheme, including performance-related incentives;
  • all matters relating to conditions of work, including hours of work and overtime arrangements;
  • standards of discipline, and sanctions for breaches of discipline including termination or suspension;
  • employee termination or demotion for poor performance; and
  • employment-related expenses or other terms and conditions of employment.

Governance framework

Another important design issue concerns the governance framework for an RA. Decisions need to be made on the degree to which an RA is to function independently of government control. This raises the question of the form of governance that should be put in place to ensure that the RA properly discharges its functions and remains accountable as a public institution. On this issue, RAs commonly include a comprehensive governance framework involving the minister of finance (the government), the governing board (board), and the CEO. The role and responsibilities of each of these actors vis-à-vis the RA is an important aspect of the autonomy granted.

Role of the minister of finance

Control over the RA. The legislated role of the minister of finance with respect to the RA has a major impact on the governance as well as the degree of autonomy from the executive level of government. Therefore, it is critical to set out the role of the minister in terms of the control and supervision of the RA (direct control and supervision in some cases, almost nil in others), the appointment of the chair and members of the board and the CEO, the approval of the budget, and so on. If too much authority is granted to the minister in these areas, the RA will de facto have a significantly reduced autonomy; if too little authority is granted to the minister, there is a danger that the RA may lose necessary sensitivity to its inherent public sector role.

Implications of corporate character. The role for the government in a more autonomous revenue authority will be much more limited than would be the case for a RA which was not a “body corporate” and which was in effect directly subordinated to the minister of finance. For this kind of RA, a first consideration is in regard to the board. Clearly, the government is the “shareholder” of the corporate body (the RA) and therefore needs to have a say in the appointment of those who will govern that body. There are two aspects to this: the appointment of the chair of the board and its members; and the appointment of the CEO.

Relationship to CEO. The position of CEO is one of the most important in the RA, and the CEO will in effect have a dual set of accountabilities. He or she will be subordinate to the board in terms of the management; however, he or she will also be directly accountable to the legislature and to the government for the execution of all the operational powers and functions assigned to the RA by virtue of the tax and customs laws.

Power of directive. Many government institutions that have been established as corporate bodies, including RAs, include a provision for the minister to issue a directive to that corporate body. This kind of provision allows the government as the effective shareholder to direct that some particular action be done. Any such direction requires maximum transparency, usually through publication in a country’s official gazette. The argument in favor of these kinds of mechanisms is that they maintain a certain amount of executive level authority and accountability without materially affecting the autonomous nature of the RA, since the expectation is they would be rarely used.

Role of the board

RAs normally have a board whose functions and powers form an essential part of the organization’s governance framework. Such boards can be advisory in nature, usually in cases where the minister has a strong role and autonomy is more limited, or they can be management boards with strong functions set out in legislation. Boards are almost always prohibited from involvement in the operational execution of the tax and customs laws, and from access to any information about individuals or corporations obtained as a result of the administration and enforcement of those laws. To do otherwise would place the (private sector) members of the board in an obvious and untenable potential conflict of interest situation. This section deals with a number of issues related to the board: the functions and powers of the board; the role of the chair of the board; the composition of the board; the tenure of the chair and members; and the operations of the board.

Board functions. The role and functions of the board flow directly from the legislation. Board functions, again depending on degree of autonomy, could include the following: to oversee the administration, management, and organization of the RA; to oversee the management of resources, services, property, personnel, and contracts; to approve the strategic plans and the budget of the RA; to approve the annual report; to establish policies to be followed; to establish by-laws for the functioning and operations of the board. In general, the board will have the power to execute all the authorities of RA with respect to carrying out the board’s mandate.

Board meetings. The chair will normally preside over the board’s meetings and exercise the powers and functions as prescribed by by-laws established by the board under its legislated authority.

Ex-officio members. A board has many duties and functions to perform and requires a mixture of skills and experiences in order to be effective. As a RA remains a government institution, it is often considered advisable to include certain government representatives on the board. In order to ensure autonomy at the same time, these positions are usually based on the notion of fixed ex-officio, or non-voting, appointments. This will respect the principle that all (voting) members of the board are required to act strictly in the best interests of the organization, and not represent the interests of some other constituency.

The CEO. In the context of corporate governance, there is a debate as to whether the CEO should also be a member of the board. The CEO of the RA has a critical role to play and has an important relationship with the board, as well as with the minister of finance in terms of the revenue laws. Careful consideration needs to be given to the most effective role for the CEO on the board.

Selection of board members. In the interest of ensuring sufficient capacity on the board, the legislation should clearly indicate that all members of the board must have the experience and knowledge required for discharging their functions, normally in finance, accounting, taxation, public administration, law, or some other related field

Size of the board. Considerable debate has also taken place concerning the optimum size for corporate boards. It would appear from the literature that boards of 7 to 12 members are now being considered optimal in terms of the efficient and effective functioning of corporate boards. Larger boards than this are considered unwieldy; smaller ones are felt to be too narrow and tending to lack comprehensive skills.

Role of the CEO

Powers vested in CEO. The CEO is generally responsible for supervising and managing the day-to-day operations of the RA. The management authority of the organization is embodied in the board, and in that respect the CEO, even though possibly a member of the board, is subordinated to it. However, the RA also has the mandate for the execution and enforcement of the revenue legislation, and the board will be prohibited from involvement in these areas. It is possible, then, to have an RA where the CEO only has responsibility over the areas where the board has a mandate, and where the heads of the revenue departments retain their powers and functions directly from the respective legislation.

At issue here is the extent to which all the powers and obligations related to the revenue laws (such as the power to assess taxes, make a customs determination, issue interpretations, impose or waive penalties, and so on) are actually given to the CEO through the enabling legislation of the revenue authority, who in turn delegates them to other senior officials and staff, or whether they are still given directly to the departmental heads, which serves to exclude the CEO from operations (this was a feature of some early RAs).

Accountability to the government

Although RAs are intended to have independence from the public sector, it is important that they retain accountability to the government as a public institution. After all, an RA, despite its independence, continues to perform critical public sector functions. It is thus essential to establish appropriate accountability mechanisms that reflect the desired degree of autonomy for the organization. It is generally felt that the greater the autonomy of the RA from the public service, the greater the need for unique, structured, and transparent accountability mechanisms in the legislation.

In the government context, laws assign responsibility and authority to organizations and individuals within them, and these organizations and individuals are held accountable for the effective and efficient performance of their responsibilities according to the governance framework established for them. Many of the aspects discussed earlier in this paper constitute in effect accountability mechanisms to serve this purpose.

Internal and external audit. A first element concerns the issue of auditing, both internal and external. As for internal audit, it is generally accepted that boards of RAs will have an active role in reviewing the outputs of internal audit (including internal affairs) in order to be able to exercise their management responsibilities, and that the organization should have an independent internal audit function reporting directly to the CEO (there is some current debate as to whether internal audit should report to the board). An RA must also have external audit. There are two choices for external audit—either the board appoints the external auditor, or the auditor general of the country, which reports to parliament, is named the external auditor for the RA.

Ombudsman. One element of oversight concerns the need for an ombudsman function to be prescribed in the law. If this policy choice were to be made, the ombudsman role would be limited to administrative areas, including issues of service quality and overall taxpayer treatment, and would exclude issues directly related to tax or customs administration for which there are normally redress mechanisms already in the base legislation. An ombudsman function of this type can be seen as a reasonable means of ensuring accountability in a more autonomous organization.

Reporting to parliament. Providing formal reports to parliament is another means of ensuring accountability to both the parliament and the executive. The two most common forms of reporting are through the annual corporate plan and budget (a look ahead at what the RA plans to do in the coming year) and the annual report (a look back at what was accomplished in the year past). Such documents provide valuable information to the government and the parliament, to ensure transparency.

B. Scope of Activities and Responsibilities

Another critical issue in creating an RA is the scope of taxes and taxing agencies that are to be brought under its jurisdiction. In all countries that have implemented an RA, the agencies that administer direct and indirect taxes at the central level form the backbone of the new organization. Many RAs also include the customs administration. Some even include agencies or units that collect local taxes or social security contributions, or other fees and charges.

In determining the appropriate scope of collection responsibility for an RA, an assessment needs to be made of the trade-offs involved in bringing different numbers and types of agencies under the jurisdiction of the authority. On the one hand, a broad scope of authority across multiple revenue collection agencies provides opportunities for achieving economies of scale and synergies in revenue collection. On the other hand, the task of managing an RA becomes more complex as the number of agencies and functions under its jurisdiction expands.

ANNEX V

Legislative Base for the RA
CountryLawLegal Form/CharacterEssential Mandate
ArgentinaExecutive Decree 1156/96 (1996/rev. 1997)Autonomous entity in administrative terms, with respect to its organization and operations (but must respect organizational decisions of Executive Branch)Implements tax and customs policy, including the imposition, collection, and auditing of taxes, taxes on import and export, control of international trade in goods, collection of social security contributions and related responsibilities
BotswanaBotswana Unified Revenue Service Act (2003)Body corporateAssessment of collection for tax, administration and enforcement of laws
CanadaCanada Customs and Revenue Agency Act (1999/rev. 2004)Agent of the government—body corporateAdministration and enforcement of pertinent revenue legislation
ColombiaLaw 49 (1990/rev. 1991/1992/1997/1999Has legal personalityThe administration of income and supplementary, national, stamp and sales taxes; customs duties and other national taxes for which no other state entities are responsible—as well as the management and administration of customs processing.
EthiopiaFederal Inland revenue Authority Act (1997/rev. 2003)Has legal personalityEnforce and implement tax laws—assess, collect, and execute federal taxes
GuatemalaOrganic Law of the Superintendency of Tax Administration (Decree 1- 98)Separate legal status—a decentralized state entity with functional, economic, financial, technical, and administrative autonomyPerform tax and customs administration functions stipulated in pertinent legislation
ItalyDecree 300/99 (1999/rev. 2003)Legal persona—has regulatory administrative, asset management, organizational, accounting, and financial autonomyManaging all taxes which have not been otherwise assigned
KenyaKenya Revenue Authority Act (1995/rev. 2005)Agency of the government—body corporateAdminister and enforce the revenue laws
LesothoLesotho Revenue Authority Act 2001Agency of the government—body corporateAssessment, collection, and receipt of specified revenue
MauritiusMauritius Revenue Authority Act (2004) (prev. Unified Revenue Authority Act)Agent of the state—body corporateAssessment and collection of tax; and management, operation, and enforcement of the revenue laws
MexicoTax Administration Service Act (1995/rev. 2003)Decentralized agency of the MOF—managerial autonomy for its work, technical autonomy for issuing decisions. Operates outside political considerations.Responsible for implementing tax and customs legislation
PeruEnabling Law 24849 (decree 501) (1988/rev. 2002)Decentralized public institution with legal status under public law—has economic, administrative, functional, technical and financial autonomyManage, audit and collect taxes; manage and regulate international trade in goods.
RwandaRwanda Revenue Authority Act 1997A “public” establishment and a body corporateAssess, collect, administer, and account for fiscal and customs revenue
SingaporeInland Revenue Authority of Singapore Act (1992/rev. 2005)Agent of the government—body corporateAdminister, assess, collect, and enforce payment of taxes
South AfricaSouth Africa Revenue Service Act (1997/rev. 2002)Public entity—organ of the state within the broad public administration, but outside PSThe efficient and effective collection of revenue; control over import, export, manufacture, movement, storage or use of goods
SpainArticle 103 (State Tax Administration Agency Law 31-1990) (numerous amendments)Legal entity governed by public law with full public and private authority—part of central government attached to MOF through treasuryEffective application for tax and customs systems
TanzaniaTanzania Revenue Authority Act (1996/rev. 2001)Agency of the government—body corporateAssess, collect, and account for government revenues; administer the revenue laws
UgandaUganda Revenue Authority Act (1991/rev. 1998)Agency of the government—body corporateAssess, collect, and account for government revenues
VenezuelaNational Integrated Customs and Tax Administration (SENIAT) Law 2001 (earlier version 1995)Autonomous service attached to MOF—NOT a corporate bodyApplying the customs and tax laws and exercising and managing and developing powers related to the implementation of customs and tax policies
ZambiaZambia Revenue Authority Act 1994Body corporateAssess, charge, levy, and collect revenue due to the government
ZimbabweZimbabwe Revenue Authority Act 2001Body corporate—agent of the stateAssessing and collecting and enforcing the payment of all revenues

ANNEX VI

General Features of the RA
CountryRole of MOFPublic ServiceAuthorization to BorrowAuthorization to Own AssetsHR AutonomyFundingExternal Audit
ArgentinaMinister for economy and production—general supervision and legal controlOutsideNoYesYesPercentage (approximately 2.75) of collection formula, subject to executive branch revisions, directly retained by the RA. Unused funds may be carried over to following year. Has broad budget flexibility—must comply with government accounting, asset management, and procurement regulations.National audit office
BotswanaInformation not available.OutsideYes—with MOF approvalYesYesNormal budget and parliamentary appropriation, plus any percent of revenues the Minister may determine-funds not expended are available the following year.Auditor appointed by the board
CanadaSeparate minister responsible.OutsideNoYesYesNormal budget and parliamentary appropriation –PEM system. Unused funds may be carried over to following year. Has broad budget flexibility—must comply with government accounting, asset management and procurement regulations.Auditor general
ColombiaMinister of finance and public credit—no indication of nature of role.InsideNoYesNoNational general budget appropriation + appropriations from the management fund + appropriations from the tax premium + additional investment funds + percent of revenue collected (2 percent is cited)Comptroller general
EthiopiaMinister of revenue is accountable.Inside (minister has authority)NoYesNoNormal budget and parliamentary appropriation.Auditor general
GuatemalaIndependent from minister of finance.OutsideNoYesYes2 percent of total taxes collected (domestic and trade) + any other funds approved in state budget.General comptroller of the accounts of the nation
ItalyMOF responsible—relationship based on negotiation of a framework document.Part of public serviceNoYesNoNormal budget and parliamentary appropriation with option for incentive funding based on tax revenues collected. Carryover of unused funds subject to approval of minister—follows public service procurement rules. Accounting and asset management are autonomous.Court of auditors
KenyaGeneral supervision.OutsideYesYesYes1.5 percent of estimated revenue + 3 percent of excess revenue (not directly retained). Authority to carry over unused funds. Procurement, accounting and asset management must follow central government policies.Auditor general
LesothoGeneral supervision.OutsideYes—subject to approval of ministerYesYes, but minister Has authorizati on to make regulations2 percent of estimated revenue—plus additional funds allocated by minister for exceptional performanceIndependent auditors appointed by board—government may request AG to audit
MauritiusOversight and final responsibility.OutsideYes—subject to approval of ministerYesYesNormal budget and parliamentary appropriation.?
MexicoOversight as minister; direct control as chair of board.Outside (but with controls)NoYesNoNormal budget and parliamentary appropriation-subject to civil service financial and procurement policies. Unused funds up to 25 percent may be carried over but any related payments must be authorized in federal budget.Auditor general
PeruMinister appoints National superintendent.OutsideNoYesYes2.5 percent of taxes on tariff on import; 2 percent of tax collection + other—follows public service procurement and financial policies.Comptroller general
RwandaGeneral supervision.OutsideYesYesYesNormal budget and parliamentary appropriation“A competent official organ established by law”
SingaporeGeneral accountability.OutsideYes (MOF to approve)YesYes2 tier funding: percent of expected revenue for operations + incentive or penalty element- may retain all surpluses—follow financial policies of the civil service including procurementAuditor general or other approved by MOF
South AfricaMOF is executive authority.OutsideNoYesYesGoverned by PEM No explicit fee but possibility exists by cabinet approval. Rollover of specific purpose funds beyond one year requires Treasury approval. Follows public service accounting and asset management policies. Follows many civil service procurement policies.Auditor general
SpainRA is attached to ministry of economy and finance.Part of public service (limited autonomy)YesYesYes- (limited)Normal budget and parliamentary appropriation + a percent of revenues with a cap as established in the annual budget law—follows standard public accounts, public service procurement principles and regulations.Comptroller general
TanzaniaGeneral responsibility.OutsideYesYesYes—but many aspects subject to the approval of the ministerNormal budget and parliamentary appropriation + option for retention of percent of revenue collected. Minister to decide. Prepares own regulations based on public policy for procurement, financial and asset management policies.Auditor general
UgandaGeneral supervision.OutsideYesYesYesNormal budget and parliamentary appropriation-retention of percent of revenue collected may be authorized by minister (not to exceed appropriation)—autonomy in all financial policies except procurement.Auditor general
ZambiaMinister accountable.OutsideYesYesYesNormal budget and parliamentary appropriationAppointed by the authority
ZimbabweMinister accountable.OutsideYesYesYesNormal budget and parliamentary appropriationAuditor general

ANNEX VII

RA Governance Framework
Board Y/N#Private

Sector

Rep.
CEO

Vested

Powers
Appoint ChairAppoint Board

Members
Appoint CEO
ArgentinaYes—advisory onlyN/AN/AYes—may sub-delegateN/AN/AExecutive branch
BotswanaYes—empowered management board—prohibited from intervention in cases7YesYes—may sub-delegateLaw names secretary for financial affairs of MOFMoFMOF on recommendation of board
CanadaYes—empowered management board—prohibited from intervention in cases15YesYes—may sub-delegateGovernmentGovernment on recommendation of provincesGovernment
ColombiaNoN/AN/AYes—may sub-delegateN/AN/AExecutive branch
EthiopiaNoNANAYes—may sub-delegateNANAGovernment on recommendation of minister
GuatemalaYes—empowered management board6YesYes—may sub-delegateMinister of public finance is president of the board of directorsPresident of the republic (from a list drawn up by a legally constituted nominating commission)President of the republic on recommendation of board
ItalyYes—empowered management board-implication board not involved in case work6 + DG As chairYesYesLaw names DGParliamentary decree on recommendation of ministerPresidential decree
KenyaYes—empowered management board-implication board not involved in case work10YesYes—may sub-delegatePresidentMOFMOF on recommendation of board
LesothoYes–empowered management board prohibited from involvement in casework8YesYesMinisterMinisterBoard subject to approval by minister
MauritiusYes—empowered management board—prohibited from intervention in cases7YesYes—may sub-delegatePresidentMinisterBoard
MexicoYes—empowered management board. May have operational casework responsibilities, e.g., forgiveness in criminal proceedings7YesYes—may sub-delegateLaw names minister as chairMOF names 3 President of Mexico names 3President of Mexico
PeruNoN/AN/AYes—may sub-delegateN/AN/APresident on recc of MOF
RwandaYes empowered management board8YesYes—may sub-delegatePM on recc of cabinetPM on recc of cabinetPM on recc of board
SingaporeYes (referred to as the authority)—considered equivalent to an empowered management board5-10YesYes—may sub-delegateMinisterMinisterBoard
South AfricaNo (advisory board prior to 2002; now ad hoc advisory ctes only)N/AN/AYes—may sub-delegateN/AN/APresident
SpainNo1/N/AN/APresident and DG—may sub-delegateN/AN/AGovernment on recommendation of the minister
TanzaniaYes—empowered management board—prohibited from involvement in tax assessments but has operational role in exemptions10YesYes—may sub-delegatePresident on recommendation of ministerMinisterPresident on recommendation of minister
UgandaYes—empowered management board—implicit prohibition regarding caseworkUp to 7YesYes—may sub-delegateAppointed by the ministerMinisterMinister on the recommendation of the board
ZambiaYes—operationally empowered board9YesBoard has vested powers—may delegate to CEOChair elected by boardMinisterPresident
ZimbabweYes—empowered management board7YesInformation not availableMinister designates a member to be chairMinister upon consultation with presidentBoard but minister if non-citizen

The RA has a president as well as a director general. The president is the state secretary of the treasury or a person of the same rank appointed by government. The president has broad responsibilities for achieving agency objectives as well as for acting as its legal representative.

The RA has a president as well as a director general. The president is the state secretary of the treasury or a person of the same rank appointed by government. The president has broad responsibilities for achieving agency objectives as well as for acting as its legal representative.

ANNEX VIII

RA Scope of Activities
Income

Tax
VATGaming

Tax
Stamp

Tax
Excise

Tax
Property

Tax
Other

Taxes
CustomsSocial

Security
ArgentinaYYNNYNYYY
BotswanaYYNNYNNYN
CanadaYYNNYNYNY
ColombiaYYNYYNYYN
EthiopiaYYYYNNYNN
GuatemalaYYNYYYYYN
ItalyYYNYYYYN 1/N
KenyaYYNYYNYYN
LesothoYYNNYNNYN
MauritiusYYYYYYYYN
MexicoYYNNYNYYN
PeruYYNNYNYYY
RwandaYYNNYNYYN
SingaporeYYYYNYYNN
South AfricaYYNYYNYYY
SpainYYNNYNYYN
TanzaniaYYYYYNYYY
UgandaYYYYYNYYN
VenezuelaYYYYYNYYN
ZambiaYYNNYNYYN
ZimbabweYYNYYNYYN

In the case of Italy, there are separate RAs for tax administration and for customs administration. For convenience and comparability, we have indicated customs is not part of the tax administration RA.

In the case of Italy, there are separate RAs for tax administration and for customs administration. For convenience and comparability, we have indicated customs is not part of the tax administration RA.

ANNEX IX

Reasons for Initially Establishing the RA1/
ABCDEFGH
Argentina41235
Botswana2364571
Colombia1Yes
Ethiopia1362547
Guatemala421536Yes
Italy12543Yes
Kenya12345
Lesotho7613452
Mauritius13452
Rwanda3271465Yes
Singapore123
South Africa123456Yes
Spain1243Yes
Tanzania1624357
Uganda126354
Venezuela1234567
Zambia1672453
Total27445041595659
Frequency15151115141213
Average ranking1.82.94.552.734.214.674.54

The following reasons were included in the questionnaire: A- low effectiveness of tax administration and poor levels of compliance; B- impediments caused by poor civil service human resources policies; C- perceptions of political/ministerial interference; D- need for a catalyst to launch broader revenue administration reform; E- poor communication and data exchange among the existing revenue departments; F- high levels of corruption; G- desire to create islands of excellence; H- other.

Generally speaking, these OTHER responses were rated highly by respondents—from highest priority to fourth highest—and include the following reasons: (1) improved relations with the provinces; (2) improved accountability for results and management; (3) improved credibility; (4) control and flexibility over human resources and assets; (5) tax evasion; (6) customs inefficiency; (7) merge tax administration; (8) better integral control of tax and customs through action plans involving tax and customs obligations; (9) part of program to strengthen national economic institutions. These responses were not included in the tabulation given the infrequency of any single OTHER response.

To arrive at a balanced weighting of the answers (i.e., to factor in the influence of the frequency), the ranks awarded by each country for each reason were totaled and divided by the frequency to arrive at an overall rating (the highest rank for an individual reason was 1).

The following reasons were included in the questionnaire: A- low effectiveness of tax administration and poor levels of compliance; B- impediments caused by poor civil service human resources policies; C- perceptions of political/ministerial interference; D- need for a catalyst to launch broader revenue administration reform; E- poor communication and data exchange among the existing revenue departments; F- high levels of corruption; G- desire to create islands of excellence; H- other.

Generally speaking, these OTHER responses were rated highly by respondents—from highest priority to fourth highest—and include the following reasons: (1) improved relations with the provinces; (2) improved accountability for results and management; (3) improved credibility; (4) control and flexibility over human resources and assets; (5) tax evasion; (6) customs inefficiency; (7) merge tax administration; (8) better integral control of tax and customs through action plans involving tax and customs obligations; (9) part of program to strengthen national economic institutions. These responses were not included in the tabulation given the infrequency of any single OTHER response.

To arrive at a balanced weighting of the answers (i.e., to factor in the influence of the frequency), the ranks awarded by each country for each reason were totaled and divided by the frequency to arrive at an overall rating (the highest rank for an individual reason was 1).

ANNEX X

Self-Assessment of RA Effectiveness1/
ABCDEFGHIJKLMNOP
Argentina78906510555878877
Botswana5284744462555n/a84
Colombia8n/a6n/a66n/a8n/an/a8307n/a6
Ethiopia855567385n/a233253
Guatemala66897578577910888
Guyana8n/a10n/a78n/a787678978
Italy88n/an/a89796n/a88n/a6n/a9
Kenya6n/a1057746547107777
Lesotho10101010641010101010101010106
Mauritiusn/an/a77n/an/an/an/an/an/a88558n/a
Mexico57n/a88108965n/an/a5589
Rwanda109108910910108998101010
Singapore9n/a9799799n/a997189
South Africa1008210101086767107910
Spain7n/a7n/a7878n/an/a988n/an/a8
Tanzania9n/an/a9898787778888
Uganda10386585766756684
Venezuela108991010591078878810
Zambia10985107610107101077107
Zimbabwe82915756558109976
Total154771419514114311514812087142143131123136139
Frequency19131716191917191714191919181719
Average8.115.928.295.947.427.536.767.797.066.217.477.536.896.838.007.32

The purpose of this table is to identify and quantify if possible the specific impact of the RA governance model in achieving revenue administration reforms. For each of the “revenue administration reform areas”(A through P on the table) respondents were asked to quantify the degree of impact of the revenue authority model: 10 being the highest impact, 0 being the lowest and n/a indicating that the reform indicator was not applicable in the particular circumstances of the respondent. These reform areas are:

A-increase tax revenues; B- introduce self-assessment; C- integrate tax administration; D- introduce taxpayer segmentation; E- improve taxpayer and trader compliance management; F- improve service to the public; G-introduce risk management; H- improve automation; I- simplify laws, regulations and procedures; J- improve valuation, classification and origin; K- improve HR regime; L- improve remuneration; M- improve budget provision; N- improve anti-corruption measures; O- provide catalyst for reform; P-improve public confidence.

The purpose of this table is to identify and quantify if possible the specific impact of the RA governance model in achieving revenue administration reforms. For each of the “revenue administration reform areas”(A through P on the table) respondents were asked to quantify the degree of impact of the revenue authority model: 10 being the highest impact, 0 being the lowest and n/a indicating that the reform indicator was not applicable in the particular circumstances of the respondent. These reform areas are:

A-increase tax revenues; B- introduce self-assessment; C- integrate tax administration; D- introduce taxpayer segmentation; E- improve taxpayer and trader compliance management; F- improve service to the public; G-introduce risk management; H- improve automation; I- simplify laws, regulations and procedures; J- improve valuation, classification and origin; K- improve HR regime; L- improve remuneration; M- improve budget provision; N- improve anti-corruption measures; O- provide catalyst for reform; P-improve public confidence.

ANNEX XI

Practical Results of HR Autonomy
ArgentinaBotswanaColombiaEthiopiaGuatemalaGuyanaItalyKenyaLesothoMauritiusMexicoPeruRwandaSingapureSouth AfricaSpainTanzaniaUgandaVenezuelaZambiaZimbabwe
Recruitment
ImprovedYYYDKDKYDKYYYYYYYYYYYYYY
Less timeDKNNYDKYYYYYYNNNYNYYDKYDK
Final sayYYYYYYYYYYYYYYYNYYYYY
Remuneration
Salaries improvedYYYYNYYYYYYNYNYDKYYYYY
QuantifiedNNYNDKYNNNYYYYNYYNYNYY
Performance payYNYNNNYYYNNNYYYYNYYNY
Retention
Staying longerDKDKNNDKDKDKDKDKDKDKNNYYYYYDKNY
Attrition downDKNNNDKDKDKDKDKDKDKNNYDKYDKYDKNY
QuantifiedNNYNNNNNNNNNNNNNNNNYN
Internal staffing
Timeframe improvedDKNYYDKYDKDKYDKYDKDKYYDKDKDKDKDKY
Less restrictionsYNYYDKYDKDKDKDKYDKDKYDKNDKYDKDKDK
Promotion basism-smcm-c-sm-sm-c-sm-cm-cDKmm-c-sm-cmmcmm-cm-cmcm
Org./Positions
Authority for org.YYYNYYDKYYYYNYYDKNYDKYYY
Authority for positionsYYYNYYDKYYYYDKYYDKNYDKYYY
Classification basisFFFTTT-FFFFFTFFFFFT-FT-FFFF
Training and development
Days increasedDKNNYDKDKDKYNDKYYYDKDKDKDKYDKDKY
InstituteNNNYYNYYYYYYYYYYYYYYY
QuantifiableDKNYNNDKYYNDKYYYDKYNYNDKDKN
Integrity measuresYYYYYYYYYYYYYYYYYYYYY
Legend: Y= yes; N= no; DK = did not respond or do not know; m = merit; c = competency; s = seniority; f = function; t = title. Brief indicators have been used in the vertical column—see Annex III for details.
Legend: Y= yes; N= no; DK = did not respond or do not know; m = merit; c = competency; s = seniority; f = function; t = title. Brief indicators have been used in the vertical column—see Annex III for details.

ANNEX XII

Quantifiable Information, 1994–2004
Tax Revenues to GDPCurrent RegistrationResourcesCustoms
19941995199619971998199920002001200220032004CITPITVATBudgetStaffDeclarationsPercent Filed

Electronically
Argentina--18.718.019.019.319.620.019.617.821.124.2226,000884,000867,000--21,883----
Botswana------------------------------------
Canada----------35.935.635.133.533.931.81,500,00023,300,0002,200,000CAN$3.1 billion39,100N/AN/A
Colombia--------10.711.011.013.213.514.2----------9,3301,450,00019
Italy--27.527.529.230.130.229.729.629.1----805,00039,800,0005,100,000EUR 2.7 billion34,50014,980,00055.3
Kenya----24.423.123.523.021.821.819.919.621.9312,0009,0009,000K Sh 3.8 billion3,275500,0000
Malawi------------------------------------
Mauritius------------------------------------
Mexico--9.38.99.810.511.310.611.311.611.410.1577,00010,000,000?MEX$10 billion32,6907,300,000100
Peru----------12.712.312.512.213.113.1------N/A7,150937,00081.3
Singapore--11.014.014.011.012.012.012.010.010.0N/A98,0001,800,00065,000S$167 million1,600N/AN/A
South Africa--22.922.223.223.624.424.123.224.123.623.71,280,0003,990,000536,000R 3,38613,1381,870,00089
Spain--16.716.417.316.616.817.717.517.917.517.61,000,00015,000,0003,000,000EUR 1,150 million25,5656,700,00096.5
Tanzania------12.011.010.510.310.911.212.0N/A243,000--84,000T Sh 44.3 million3,448----
Uganda------12.011.612.711.911.712.913.013.6----9,500U Sh 55.2 billion------
1At the time of the preparation of this paper, Maureen Kidd was a Technical Assistance Advisor with the Fiscal Affairs Department. William Crandall is a member of FAD’s panel of experts.
2A more precise definition of RA is not really possible since these governance models cover a range on a spectrum (see subsequent discussion in II.B. Defining a Revenue Authority). Other organizations and authors use slightly different terms, such as “semi-autonomous revenue authority” or “unified semi-autonomous body,” which may or may not have the same definition as used in this paper.
3Revenue Authorities and Taxation in Sub-Saharan Africa: A Concise Review of Recent Literature from the Investment, Competition and Enabling Environment Team, DFID, February 2005.
4This comment applies, of course, to any revenue administration, regardless of governance structure.
5OECD: Tax Administration in OECD and selected non-OECD Countries: Comparative Information Services (2006).
6Difficulties with Autonomous Agencies, by Manning and Matsuda. World Bank (2000), p. ..............
7The United Kingdom defines a quasi-autonomous agency as “a body which has a role in the processes of national government, but is not a government department or part of one, and which accordingly operates to a greater or lesser extent at arms length from ministers.”
8The term “governance” is often used more broadly. The World Bank uses six dimensions to measure governance: voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption.
9See bibliography (Annex I).
10Jenkins, Glenn P., 1994, “Modernization of Tax Administrations: Revenue Boards and Privatization as Instruments for Change,” Bulletin for International Fiscal Documentation (February).
11Taliercio, Robert, 2004, Designing Performance: The Semi-Autonomous Revenue Authority Model in Africa and Latin America (Washington: World Bank).
12Mann, Arthur J., 2004, Are Semi-Autonomous Revenue Authorities the Answer to Tax Administration Problems in Developing Countries?—A Practical Guide, Fiscal Reform in Support of Trade Liberalization Project (USAID) (August).
13Delay, Simon, Devas, Nick, and Michael Hubbard, 1999, “Reforming Revenue Administration: Lessons from Experience- A study for the Department of International Development,” School of Public Policy, University of Birmingham.
14Gray, John, and Emma Chapman, 2001, “Evaluation of Revenue Projects Synthesis Report Volume I,” UK Department for International Development.
15Di John, Jonathan and James Putzel, 2005, “The Case of Uganda”—paper for Conference on Institutional Change for Growth and Poverty Reduction in Low-Income Countries.”
16The notion that “tax administration is tax policy” was originally expressed by Milka Casanegra de Jantscher in “Administering the VAT,” in Value Added Taxation in Developing Countries, ed. by Malcolm Gillis, Carl S. Shoup, and Gerardo P. Sicat (Washington: World Bank, 1990).
17Many different titles are used to describe the senior civil service position responsible for the agency (e.g., director, director general, commissioner, permanent secretary, deputy minister). For the purposes of this paper, the term CEO or Chief Executive Officer is used for consistency.
18Taliercio recognizes that these problems affect the entire government but argues the problems are felt much more acutely in the tax administration as compared to most other public agencies, largely on account of the specialized skills needed in tax administration and because of its revenue-producing role. This “uniqueness” argument may not be as strong as the author suspects, especially in the context of world-wide public service reform and good governance initiatives.
19Some countries did not provide the information needed on every question, and in some cases information was not provided in a format suitable for compilation.
20An empowered management board is a board that is more than advisory, that has real powers under the law (usually for matters related to human resources, finance, and administration, and other non-operational matters). Peru and Spain have no board; South Africa previously had an advisory board but currently the law provides for the establishment of ad hoc advisory committees only; and Argentina’s is advisory only.
21In no other cases in the survey responses were operational boards identified, although the board in Tanzania does have operational responsibility in one limited area, i.e. decisions on remissions. In some of the early RAs, such as the Mauritius Unified Revenue Authority Board, umbrella-level operational boards comprised primarily of public servants were established to take specific operational decisions.
22This survey has not reviewed the nature of the integration of customs and tax administration organizationally within a RA structure. However, responses from the initial questionnaire indicated customs operations are generally separate from tax administration.
23The focus is predominantly on the legislative process—from the time the political initiative was taken until the formal mounting of the RA.
24This particular reform area was defined as integrating tax administration, for example income tax and VAT. It may be that some countries have considered this to mean the integration of all revenue departments, including customs (as this is sometimes associated with the start-up of the RA).
25Mostly ministry of finance, parliament, or other bodies, such as the executive branch (Argentina).
26Di John and Putzel (2005) note that while RAs may have been instrumental in initiating reforms, there is no evidence to indicate that this is sustainable. Gray and Chapman (2001) equally found that initial benefits of RAs tend to be eroded over time.
27Both authors consider the disadvantages can be outweighed by the advantages.
28One exception is Canada, where the requirement for a parliamentary review five years after implementation was written in to the original RA legislation. Box 5 outlines the rationale behind the decision to create the Canada Revenue Agency.
29This conclusion is consistent with overall experience in obtaining such data in developing countries.
30Even in the case of Guatemala, where the law creating the RA provides for the possible privatization of certain revenue administration functions, there remain restrictions on the extent to which the RA must remain accountable in the long term for core functions of government (assessing and auditing).
31This is supported by Di John and Putzel (2005), who note that ministries of finance continue to wield strong influence.
32Gray and Chapman (2001) found that RAs do not have independence from the state budget and that the level of budgetary funding was often a constraint to reform.
33An revenue authority, sometimes referred to as a semi-autonomous revenue authority, refers to a type of governance model for revenue administration where traditional ministry of finance departments (tax and sometimes customs administrations) are established as an organization with a degree of autonomy or independence from government and from standard public service policies. Many revenue authorities are corporate bodies, governed by a board, with varying degrees of budgetary and human resources autonomy.
34A revenue authority, sometimes referred to as a semi-autonomous revenue authority, refers to a type of governance model for revenue administration where traditional ministry of finance departments (tax and sometimes customs administrations) are established as an organization with a degree of autonomy or independence from government and from standard public service policies. Many revenue authorities are corporate bodies, governed by a board, with varying degrees of budgetary and human resources autonomy.
35Technical note prepared by William Crandall (May 2004).

Other Resources Citing This Publication