REINVENTING VAT COLLECTION: INDUSTRY VERTICAL ASSESSMENT, REVENUE INCREASE, AND PUBLIC SECTOR RELIABILITY
Mônica de Maria Santos Fornitani Pinhanez
Prior degree titles: Bachelor of Laws, São Paulo Pontifical Catholic University, 1987 Master in Public Administration, Getúlio Vargas Foundation, 1995 Master of City Planning, MIT, 1997
Submitted to the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in Economic Development and Public Policy at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
February 2008
Author ______Mônica de Maria Santos Fornitani Pinhanez Department of Urban Studies and Planning
Certified by______Professor Alice Amsden Department of Urban Studies and Planning Thesis Supervisor
Accepted by ______Professor Eran Ben-Joseph Chair, Ph.D. Committee Department of Urban Studies and Planning
© 2008 Monica Fornitani Pinhanez. All Rights Reserved The author hereby grants to MIT the permission to reproduce and to distribute publicly paper and electronic copies of the thesis document in whole or in part.
REINVENTING VAT COLLECTION: INDUSTRY VERTICAL ASSESSMENT, REVENUE INCREASE, AND PUBLIC SECTOR RELIABILITY
Mônica de Maria Santos Fornitani Pinhanez
Submitted to the Department of Urban Studies and Planning in partial fulfillment of the requirements for the degree of
Doctor of Philosophy in Economic Development and Public Policy at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY
Abstract This dissertation shows how administrative reforms of the State Tax Administration Bureaus (STABs) in Brazil between 1997 and 2005 contributed to strengthening public sector bureaucracies and institutions at the sub-national level, while increasing tax revenues and compliance. STABs’ administrative reform comprised changes in organizational structure (i.e. rationalization of procedures and processes), technological processes (i.e. computerization and on-line processes), and institutional arrangements (i.e. development of public-private sector relationships to improve tax collection and tax collector professionalization). Beyond the market-oriented reforms, these joint changes made possible to taxpayers and tax collectors to visualize the connections and functioning of the tax administration and collection process that they were not able to think about before, i.e., there is evidence for how the administrative reform facilitated the understanding of the links among the several processes in the tax collection, particularly the backward and forward linkages in each industry. The understanding of these linkages advanced a conceptual shift and the adoption of a new cognitive axis—the industry linkages that changed the way tax collectors envisioned the tax collection process. This dissertation shows that understanding these new connections led to increased efficiency, effectiveness, accountability, and transparency, providing faster and more accurate information on taxpayers and tax returns.
In addition, I show that in the case of tax administration reforms in the Brazilian STABs the conjunction of tailored organizational and technological changes influenced tax collectors’ and taxpayers’ compliance, cooperation, and adhesion to reform efforts. Technological change enabled the new understanding of the tax sector and enabled industry vertical assessment and functional specialization. The capacity to grasp the conceptual shift emerged from the joint efforts. In turn such efforts led to institutional change and the strengthening of public sector bureaucracies, particularly of the STABs.
Through a multi-level methodology, I use quantitative and statistical analysis to show that separate and joint aspects of the administrative reforms had an effect on Value Added Tax (VAT) collection. Then, I carriy out a qualitative analysis using case studies to evaluate the aspects of reform that had an impact on the changes in the STABs and how they acted together to create a shift in the conceptual understanding of taxations, of its relationship to industry structure, and of the identity of tax collectors. The choice of cases was structured to track public sector reform, development capabilities (e.g. public officials’ capacity training), and the effectiveness of tax collection over time, specifically a period of eight years, between 1997 and 2005. The focus was on the local states system that implemented tax administration reforms, specifically related to the collection of the VAT. This data, presented in this dissertation, supports the argument that success is a function of technology, organizational changes, functional specialization, and institutional arrangements, which combined, led to envision a new way to achieve greater tax collection and compliance, a shift in tax-collectors identities, and the strengthening of public sector’s institutions. Finally, the diversity of features in the chosen states range from wealth and industrialized, to rural and agricultural, to developing states. Yet, the same national policies cut across states, forcing less developed and active states to engage and catch-up with the reform plans. In this fashion, my findings can be applied to the most developed nations as well as to striving developing countries and poorer nations.
Thesis Advisor Alice Amsden Barton L. Weller Professor of Political Economy Department of Urban Studies and Planning
Thesis Committee
Thesis Reader ______Richard M. Bird Professor Emeritus of Economics International Tax Program, Rotman School of Management University of Toronto
Thesis Reader ______Jane Fountain Professor of Political Science and Public Policy Director, National Center for Digital Government University of Massachusetts Amherst
Table of Contents
Chapter 1 Introduction...... 19 1.1 What Makes a Good Administration?...... 19 1.2 A Novel Measure of Success: Quantitative Modeling of Tax Administration Reforms ...... 25 1.3 A New Paradigm for Improving Tax Administration and Increasing Public Officials’ Professionalization: Industry Vertical Assessment...... 26 1.4 Structure of the Dissertation...... 30
Chapter 2 Research Design and Methodology...... 33 2.1 Research Query ...... 33 2.2 Research Design...... 34 2.3 Questionnaire ...... 43
Chapter 3 Human and Managerial Determinants of Tax Revenue Growth: A Novel Statistical Model ...... 45 3.1 Introduction...... 45 3.2 Brief Literature Review on the Determinants of VAT Revenue...... 46 3.3 New Empirical Evidence, Indicators, and Model...... 51 3.4 Data...... 54 3.5 Correlations ...... 56 3.6 Results...... 60 3.7 Multiple Reform Variable Estimation Models ...... 66 3.8 Other Alternative Estimation Models...... 67 3.9 Conclusion...... 69 3.10 IDB Indicators...... 70
Chapter 4 Vertical Economic Assessment of the Tax Basis: The Central Strategy for Raising VAT Revenues...... 73 4.1 Introduction...... 73 4.2 Chapter Structure ...... 75 4.3 Conversations Between the New Public Administration and the Administration Reform Literature: Connections and Implications ...... 76 4.4 Tax Administration Guidelines for the Brazilian States ...... 81 4.5 Important Tax Administration Traits in Brazil ...... 85 4.6 From Territories to Industry Segmentation: Production Chain in the Tax Sector and the New Cognitive Process ...... 91 4.7 From Learning by Doing to Learning by Rationally Processing...... 100
Chapter 5 How Technological Changes Reveal Production Chains, Increase Tax Revenue, and Foster Alliances Across the Public-Private Divide...... 103 5.1 Introduction...... 103 5.2 The Technological Archetype...... 104 5.3 The Enactment Frame...... 109 5.4 e-Taxing...... 112 5.5 Computers at Work: The Information Age in the STABs ...... 114 5.6 Computerized Services ...... 121 5.7 Connecting the Dots in the States...... 124 5.8 Catching Up with the Economy: Visualizing Linkages and Chains...... 127 5.9 Resistance, Teamwork, and Alliances ...... 130 5.10 Conclusion: Information Systems in Developing Country Bureaucracies...... 133
Chapter 6 The Dilemmas of Tax Collectors (TC) and How They Learn a New Way to Increase Revenue Collection ...... 135 6.1 Introduction...... 135 6.2 The Job...... 136 6.3 The Individual ...... 141 6.4 The Professional ...... 144 6.5 New Processes and New Professionals ...... 146 6.6 Everything is Illuminated: The Rationale for a Rational Change...... 149 6.7 Accountability and Alliances...... 152 6.8 Conclusions: The New Professional and the New Identity ...... 157
Chapter 7 Conclusion...... 161 7.1 What is the Puzzle?...... 161 7.2 Axiomatic Model ...... 162 7.3 The Lasting Power of Vintage Ideas ...... 163 7.4 Empirical Results: Tax Administration, Reforms and Recommendations...... 164 7.5 Implications for Theory and Methodology ...... 167 7.6 Implications for Technology Implementation...... 167 7.7 Implications and Contributions for Public Sector Management ...... 168 7.8 Future Research...... 169 Appendix A Brazilian Tax Structure: Facts and Figures...... 171 A.1 Brazilian Regional Differences ...... 171 A.2 Brazilian Tax Revenue ...... 173 A.3 VAT in Brazil by State ...... 176 A.4 Other Information ...... 184
Appendix B List of Interviews ...... 185
Appendix C Indicators ...... 193
Appendix D State Economic Development Levels...... 199
Appendix E Time-Lagged Predictors...... 203
Acronyms ...... 185
References...... 207
List of Figures
Figure 1. Model Diagram: Technological and Organizational Change => Industry Segmentation => New Cognition Process => New Institution Image and Identity...... 29 Figure 2. Multi Level Methodology...... 35 Figure 3. Brazil - Political Division - Regions and States and States Chosen in this Research.. 38 Figure 4. Geographic Territory Tax Collection Diagram ...... 92 Figure 5. Vertical Industry Assessment...... 95 Figure 6. Letter to Tax Collectors from Segment Manager...... 99 Figure 7. Technological enactment framework (Fountain 2001)...... 111 Figure 8. Technological Enactment Framework for e-Taxing based on Fountain (2001)...... 113 Figure 9. São Paulo STABS’ On-Line Connections with Internal and Outside Agencies...... 126 Figure 10. Integrated Diagram...... 148 Figure 11. Cognition Axis Assessment Model: New Cognition Embedded in Technological and Organizational Change =>New (Tax Administration Process)=> New Ties and Identity...... 162
List of Graphs
Graph 1. Six Brazilian States’ VAT Growth in Percentage from 1997 to 2004 ...... 21 Graph 2. Brazilian States’ GDP Growth in Percentage from 1997 to 2004...... 22 Graph 3. Brazilian Tax Revenue -- Business Related Taxes...... 47 Graph 4. Best Fitted Line (VAT by GDP). Simple Regression Model...... 61 Graph 5. Best Fitted Line, controlling for AdminConnectivity (VAT by GDP). Simple Regression Model...... 65 Graph 6. VAT and Total State Revenue Growth from 1995 to 2005...... 74 Graph 7. Tax Returns ON-LINE...... 115 Graph 8. STABs’ Administrative Costs as a Percentage of Tax Revenue, Selected States..... 117 Graph 9. On-Line Budgeting in Selected States...... 123 Graph 10. STAB On-line Connectedness (for data sharing) with Other Agencies in Selected States...... 123 Graph 11. On-Line Validation in Selected States...... 124 Graph 12. VAT/GDP Ratio Predict by TCs’ Wage...... 138 Graph 13. Regional Population, GDP, and Tax Revenue...... 172 Graph 14. Participation of Social Contributions and Taxes to total Brazilian GDP...... 174 Graph 15. State Taxes as a % of Total Tax Revenue and as % GDP (2000)...... 176
List of Tables
Table 1. Brazilian Tax Revenues as a Percentage of GDP...... 20 Table 2. VAT Fiscal Effort, Potential Taxation, and Effective Taxation...... 48 Table 3. Fiscal Effort and Trend. All states, 1987 – 1999...... 50 Table 4. Summary Statistics...... 57 Table 5. Correlation Table...... 59 Table 6. Results of the Base Model Estimation and Reform Indicators...... 62 Table 7. Multiple Variables Models...... 67 Table 8. Lag Time for Predictors...... 68 Table 9. Large Taxpayers...... 84 Table 10. Official Initial Date for STAB Reforms...... 87 Table 11. States by IDB Initial Development Classification...... 89 Table 12. Political Features in the Selected States...... 118 Table 13. Computerization Processes...... 122 Table 14.STAB Tax Collectors’ Salaries by Brazilian State...... 139 Table 15. Courses Offered at the STABs (between 1997 and 1998)...... 146 Table 16. Features and Changes in Brazilian STABs: Selected States...... 165 Table 17. Brazil: Regional Disparities in Social Indicators...... 172 Table 18. Brazilian Basic Statistics by State (various years)...... 173 Table 19. Brazil Tax Assignments for Federal, State, and Local levels...... 175 Table 20. VAT Growth from 1996 to 2003 per Year...... 177 Table 21. Cumulative VAT Growth between 1997 and 2005...... 178 Table 22. GDP Growth per State (1994-2002)...... 179 Table 23. Ratio of VAT per GDP by State and Coefficients of Variation...... 180 Table 24. VAT Collected by Brazilian States, from 1995 to 2005, in R$ (thousands), in Nominal Values...... 181 Table 25. GDP by Brazilian States, from 1995 to 2005, in R$ (thousands), in Nominal Values.181 Table 26. VAT Elasticity by Brazilian States, from 1995 to 2005...... 182 Table 27. VAT Rates per State...... 183 Table 28. Selected States Features (Capital City, Territorial Area, Number of Municipalities, Population)...... 184 Table 29. Results of the Base Model Estimation, Reform Indicators and State Economic Development Levels...... 200 Table 30. Time Lagged Predictors...... 204
Acknowledgements
My gratitude goes to all the interviewees, mostly tax collectors, in Brazil who patiently gave of their time and shared their insights and often confidential data, and to those elsewhere who provided me valuable information on comparative administrative reforms. I would like to thank Professors Alice Amsden, Richard Bird, and Jane Fountain who served on my Ph.D. committee and supported me throughout my writing. Alice Amsden saw me through many years at MIT in various capacities and was a supportive advisor; she believed that I could finish this thesis and gently pushed me towards doing it. Professor Bird was a meticulous reader, always encouraging and critical of my writing—exacting in words and numbers. Jane Fountain was a critical reader of my work and research, particularly on organizations and information technology. Several other professors have been encouraging me throughout my time at MIT, among them special thanks to professors Paul Smoke, Karen Polenske, Marty Rein, Michael Piore, Richard Locke, and Judith Tendler. Special thanks go to my MIT colleagues who participated in several phases of my doctoral trajectory, commenting on colloquium and defense materials: Raquel Gomes, Sylvia Dohnert, Janice Goldman, Natasha Iskander, and Nicola Lowe, the terrific final destination group, of which it was a honor to be part of. Sumila Gulyani, Maria Elosua, and Smita Srinivas were friends who did not give up on me and were constantly trying to push me forward in one way or another. These women, my colleagues and peers -- wives, mothers, professors, and professionals are the brightest, the most beautiful, and the bravest young women I have met in any possible way. They inspired and motivated me along this journey. Audrey Kalajian and Alexandra Eurdolian, Livia Pedalini and Agnaldo Valetin, Jesse Kahn and David Fernandes, and Ian Sue Wing and Carolyn Wood were constant friends who helped me beyond the call of duty by reading, editing, revising, thinking with me and sometimes for me, but above all, giving me love, hope, and shelter in crucial moments of my work. Monica Amorim, Mansueto Almeida, Alice and Dimas Mello, Benedita Souza were friends who received me in their houses and families as one of them. A special thanks to Charles Thibault, my statistics tutor, for his tolerant and thoughtful guidance for the statistical analysis in this dissertation. There are not enough words for my gratitude. I would never be able to pursue my dreams and my studies without funding. I would like to thank CAPES, the Brazilian Federal Agency who funded my studies. The Department of Urban Studies and Planning funded me for my initial studies through research and teaching positions with Professors Karen Polenske and Judith Tendler, who supervised my initial research. I am also in debt to the International Institute for Labour Studies, at the International Labour Organization, in Geneva, for a Phelan Fellowship, which allowed me to develop my analysis of case studies as I studied public sector labor relations in other contexts. This dissertation is dedicated to my husband, Claudio, my anchor and my engine, whose patience, love, and compassion was crucial for my success; my sister Yelrihs; my mother Shirley, and my second mother, Marília, my role models of faith and strength; and my son Luca, my future. They have collectively supported this endeavor in more ways than I can describe. A special dedication goes to my father, an honest tax collector who always believed that it was possible to change the world—the individual, its society and its governments—for better.
Knowledge must come through action; you can have no test which is not fanciful, save by trial.
Sophocles (496 BC - 406 BC), Trachiniae
Chapter 1
Introduction
1.1 What Makes a Good Administration?
What makes a good tax administration? Enough revenues to fund social and developmental needs with no excess tax burden on citizens? Efficiency, credibility, and professionalism from tax collectors and officers? Is it possible to rescue crumbling public sector bureaucracies from the demise and discredit in which they have fallen? These mundane inquiries are at the core of my research. Suspicion and skepticism may have discouraged serious thoughts about these topics for many, but much research has been done in order to understand how to strengthen government institutions, to alter them to serve the people’s needs, and to adopt the best organizational practices. 1 While a few advanced developed countries’ governments have become the role models for a more efficient, credible, and sustainable public sector—achieved with a fierce hand focused on restructuring and reengineering—less developed countries have fallen behind trying to follow the blueprint of these earlier adopters of government reform. Brazil, amongst a large group of countries, is trying to catch up with the wave of good government and has been experimenting with government restructuring in several different sectors, particularly taxation.
1 Much of the literature on government reforms relies on reports of private sector experiences and is constructed around market-oriented ideologies. In the same line, the literature on new public management focuses on recommendations about how to “reinvent government” and emphasizes client-user focus, cost-effectiveness, accountability, and transparency. Good accounts of this debate can be found in Grindle (1997 and 2000), Tendler (1997), Fishlow (1989), O’Donnell (1993), Osborne and Gaebler (1993), Pereira (1997), Pereira and Spink (1999).
19 Chapter 1. Introduction
My doctoral research sought to explain how the Brazilian State Tax Administration Bureaus (STABs) were able to implement tax administrative reforms that contributed to the strengthening of public sector bureaucracies and institutions, and to the professionalization of public officials at the sub- national level, while increasing tax revenues and compliance. At the same time that the state was retracting from its leadership role of commanding the economy, and in spite of the government’s manifest inabilities to deal with mounting fiscal deficits and service demands, a set of successful tax administration (TA) reforms emerged that pushed public sector restructuring beyond the ordinary prescriptions of increasing tax rates and reducing expenditures. These reforms reconfigured the role of the state and included changes in organizational structure, technological processes, and institutional arrangements that encompassed the entire state tax system—ranging from motor vehicle and fuel taxes to taxes on services and value-added taxes(VAT)).2 Particularly, in the case of the VAT (Value-Added Tax), reforms led not only to an overall increase in revenues,3 without increasing tax rates, as shown in Table 1, but also changed tax administration rationale.4 This is especially important since the VAT is the primary state level tax, accounting for approximately 80%-90% of state revenues and 8% of the total federal tax revenues (see Appendix A).
Table 1. Brazilian Tax Revenues as a Percentage of GDP.
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 FEDERAL 16.72 17.5 18.47 20.46 20.01 19.35 19.8 20.73 22.47 22.97 23.37 22.08 21.51 22.24 23.25 23.75 STATES 7.31 7.37 6.48 8 8.34 8.21 7.95 7.89 8.17 8.69 9.02 8.4 8.4 8.6 8.74 9.14 VAT (ICMS) 7.4% 7.4% 7.0% 6.6% 6.7% 7.2% 7.7% 7.6% 7.5% 7.7% 8% n.a. MUNICÍPIOS 1.19 0.99 0.78 1 1.39 1.4 1.34 1.28 1.51 1.52 1.53 1.38 1.55 1.39 1.39 1.46 TOTAL 25.22 25.86 25.73 29.46 29.74 28.96 29.09 29.9 32.15 33.18 33.92 31.86 31.46 32.22 3338 34.23 Source: Brazilian Federal Treasury 2007 (SRF 2007).
The reform strategies, which departed dramatically from past policy and were implemented during a period of discontent with government corruption and fiscal mismanagement, make this study of the changes in state tax administration unique and worthy of attention. The strategies consisted of
2 Several reforms were carried out at the federal level prior to the reforms in the states. These reforms were equally successful. However, this dissertation analyzes only the state level. 3 Although the VAT increased by only 0.6% during the years studied, this rise in revenues can be considered a great effort from the tax administration since Brazilian states were already quite efficient in collecting VAT (Ebrill, Keen et al. 2000). 4 Although STABs’ reforms encompassed several different taxes, I focus on the VAT reform. ICMS is the state Value Added Tax in Brazil (ICMS). There is also a federal VAT, the IPI (Tax on Industrialized Products), which is not the focus of this study. Refer to Appendix A for an overview of tax shares in Brazil.
20 Chapter 1. Introduction rationalization and industry segmentation of tax collection, implementation of computerization and on- line processes, and the development of a public-private sector culture that improved tax collection. A grounded, case study approach to tax reform in Brazil is possible because VAT is collected in Brazil at the state level and not by the Federal government, as is typical in most countries (Bird and Gendron 2001). Although tax administration reforms were implemented at the same time and under very similar conditions in each of Brazil’s 27 states, these reforms were implemented differently.5 As a result, states with similar economic performance performed differently with respect to tax collection (Graph 1 and Graph 2)6.
Graph 1. Six Brazilian States’ VAT Growth in Percentage from 1997 to 2004
VAT % Growth (1995-2004)
120%
100%
80% Bahia Ceará 60% Paraíba Pernambuco 40% % Growth Rio Grande do Norte São Paulo 20%
0%
-20% 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Year
Source: IBGE (Brazilian Institute of Geography and Statistics).
Although GDP grew in the majority of the states, the VAT grew at a much more dramatic rate. However, a quick look at Graphs 1 and 2 values proves such an assumption wrong. A regression analysis
5 The Ministry of Finance and the Inter-American Development Bank (IDB) jointly coordinated the 1990s tax administration reforms in Brazil. The IDB funded the implementation of the reforms and established the criteria for their evaluation (Febres, Cruz et al. 1998; IDB 1996). 6 For the coefficient of correlation and VAT tax buoyancy, refer to Appendix A.
21 Chapter 1. Introduction of VAT and GDP growth shows that a strong correlation between state GDP and VAT growth exists. However, the impact of qualitative aspects changed in the reform process explains a significant portion of VAT growth.
The data collected regarding tax revenue collection show that the states improved their performance. However, the transformation of a tax system is not limited to improved tax collection as measured in increased revenues: it affects the government, the economy, and the society as a whole. As Bird and Casanegra de Janstcher reflect, tax administration is about equity and welfare as well:
Graph 2. Brazilian States’ GDP Growth in Percentage from 1997 to 2004.
GDP% Growth (1995-2004)
120%
100%
80% Bahia Ceará 60% Paraíba Pernambuco Rio Grande do Norte 40% % Growth São Paulo
20%
0%
96 99 02 -20% 1995 19 1997 1998 19 2000 2001 20 2003 2004 Year
Source: IBGE.
"The best tax administration is not simply one that collects the most revenue. How that revenue is raised - that is, the effect of the revenue-generation effort on equity, on the political fortunes of governments, and on the level of economic welfare – may be equally important. A poor-quality tax administration may collect large amounts from easy-to-tax sectors such as wage earners, while being unable to enforce taxes on business enterprises
22 Chapter 1. Introduction
and professionals. The level of collection is therefore a somewhat unsophisticated measure of the effectiveness of tax administration. A more accurate measure is the size of the ‘compliance gap’ - that is, the gap between actual and potential tax revenues - and how that gap varies among the different sectors of the taxpaying population” (Bird and Casanegra de Janstcher 1992).
Although several studies have measured the Brazilian STABs’ tax effort and efficiency, (Marinho and Moreira 1999; Vasconcelos, Siqueira et al. 2005), quantitative and qualitative data on the STABs’ compliance gap in not readily available. Through in-depth interviews with more than 140 public officials and consultants, however, I discovered professional, organizational, and institutional changes that differed from state to state. In this dissertation I explore the paths and strategies that these public sector bureaucracies used for organizational and individual learning to consolidate this acquired knowledge.
In addition to the differences in tax revenue shown above, some states were able to boost tax collectors’ professional levels and to bridge the public and private realm, leading to an enhanced image of the taxation agency and improving the interaction between tax collectors and taxpayers. Rio Grande do Norte, Bahia, São Paulo, and Pernambuco were all successful in achieving whereas Paraíba and Ceará were less successful in their attempts to improve their cadre of tax collectors and to integrate the business class into the public process.
For example, the STAB in São Paulo, Bahia, Rio Grande do Norte, and Pernambuco each conceived of new ways to improve tax collection. Although their choices were not identical, they were similar. These new methods that STABs implemented influenced their treatment of workers, contributing to increased employee satisfaction, and improved their relationship with the private sector. The Ceará and Paraíba STABs, however, implemented more traditional reforms. This caused to more traditional reactions, such as discontent among business groups and uncooperative behavior from tax collectors. The qualitative results of TA interventions enrich and offer details to support the overarching analysis carried out here.
My initial hypotheses regarding the array of results in Brazil’s tax administrative reforms resonates with a large body of literature that documents the evolution, successes, and pitfalls of administrative reforms among developing countries. Within this literature, there is an important debate regarding the adoption of new technologies, the role of political leadership and institutions, and public participation in the reform process. Most existing explanations of successful reforms focus on technological and organizational changes designed to echo the private sector and on market oriented
23 Chapter 1. Introduction reforms. My original hypotheses focused on the triad of administrative reforms: available technology, straightforward- and clear-cut processes, and professionalization of employees. However, I had conditional qualifiers for each one of them. Although I felt technology was important, I suspected that the key was not the quest for technology per se, but rather a government’s strategy to reallocate technology resources away from traditional tasks and toward more productive tasks such as tax systems that are more closely attuned to the economic system. Clear and simplified processes would increase compliance and transparency, and subsequently the STAB credibility. Professionalization would require relevant training and better career opportunities in order to get public officials committed to the new system. Finally, I suspected that better government and higher revenues could have been achieved even without these reforms due simply to the economic recovery in each state.
The Brazilian states’ tax reform programs provide an exceptional opportunity to learn under which conditions reforms are prone to succeed at the sub-national level. Brazil is a federal nation, composed of 27 autonomous states that share the same institutional and constitutional structure.7 In 1988, the Congress approved a new Federal Constitution, which accentuated the power and autonomy of the states, as well as their responsibilities. This process of decentralization to state and local governments places Brazil among the most decentralized countries in the developing world.8 The newly acquired role of state and local governments in service delivery awakened the concerns of local politicians and executives. Governments needed to increase revenues to cope with the rising costs in public services, without worsening the tax burden on citizens. However, despite the fact that states were granted great fiscal autonomy and liability to raise their own revenue, the ill-equipped and untrained bureaucracy, the state’s low credibility with tax-payers, rigidities in tax jurisdiction and legislation, and the lack of power of states and municipalities to borrow new money, created imbalances across states in terms of increasing local revenues and improving service delivery.9
7 Brazil is a country organized as a Federative Republic, with three levels of government: the federal government, the state government (26 states plus the Federal District), and the municipalities (approximately 5,600 municípios— municipalities). Each level of government possesses administrative autonomy. With respect to financial autonomy, each level draws up its own budget, defines obligatory taxation, and carries out public spending. The Brazilian states’ tax jurisdiction includes taxes on goods, transportation, and communications services (VAT and sales taxes), known in Brazil as ICMS; transfer of property by inheritance or as gifts (known as IT); and ownership of motor vehicles (known as IPVA). 8 For more information regarding decentralization, refer to Smoke 2003, Bahl and Linn 1992, and Bird, Ebel et al. 1995. 9 Also, an important factor in the Brazilian economy was the stabilization of the currency. Before, local governments could roll their public debts without being inquired by any authorities. However, with the stabilization plan from the Cardoso government (1995-2001), it became clear that governments would have to find clear resources for the services and public work. See (Ramos 2001).
24 Chapter 1. Introduction
Previous attempts to develop solutions for tax evasion and non-compliance were isolated, taken at each state’s own risk, and bore uncertain results.10 STABs followed generic recommendations from international organizations, such as the World Bank, the Inter-American Development Bank (IDB), the International Monetary Fund, and the Inter-American Center of Tax-Administrations (CIAT),11 but failed to achieve the expected gains in efficiency and effectiveness.12 Difficulties in implementing tax reforms and restructuring tax agencies frequently stem from regional and economic differences and diversity of state level organizations, from deficiencies in information systems, and from differences in the level of professionalization. Because of such differences, comparisons across national states are usually considered difficult, resulting mainly in individualized accounts of each experience without an encompassing analysis of common trends across case studies.13 Accounts of reforms at the sub-national level are even less common in the literature than comparative studies of national economies, due either to the lack of systematic information or to discrepancy among sources (Bird 2000, 2001; Bird and Gendron 2001; Varsano 1999). Compounding this, VAT is rarely collected at the sub-national level, so the collection and presentation of this data is unique. Brazilian states, with their differences across and within regions, present features that are common to countries in Latin America, Africa, and Asia.
1.2 A Novel Measure of Success: Quantitative Modeling of Tax Administration Reforms
One of the fundamental contributions of this work is to examine the correlation between some of the qualitative aspects of the reform with the growth in tax collection, particularly, the growth of VAT. These results have significant implications for policies aimed at constructing government reform packages, particularly in developing countries.
My approach uses the nominal value of Value-Added Tax (VAT), which is collected at the state level in Brazil and will be later examined in the chapter 3, defined as a function of economic performance (GDP) and administrative reform (REFORM). Denoting REFORM as the combined product of Administrative Efficiency, Computerization, and Training, this definition allows formulating an identity for VAT performance, as follows:
10 Interview with Yoshiaki Nakano, former head of the São Paulo STAB. 11 http://www.ciat.org. 12 For several accounts on tax administration reforms around the world, see the works by IDB (1999), by Bird and Casanegra de Jantscher (1992) for the IMF, and by Kalilzadeh-Shirazi and Shah (1991) for the World Bank. 13 In the case of Brazil, most studies focus on individual state, e.g., the case of Rio Grande do Norte, or within regions, instead of comparing reforms across states.
25 Chapter 1. Introduction
Equations
(1) VATit= β0 + β1GDP + β2Reform + β3GDP*Reform + εI
(2) δVAT/δGDP = β1 + β3Reform
Using the array of variables gathered from an IDB database described in chapter 3, this study is one of the first showing a statistically significant correlation between VAT efficiency (measured with the proxy of VAT revenue increase) and administrative reform (i.e. administrative efficiency, computerization, and training). In this sense, the model above suggests that investments in training, computerization, and administrative efficiency, led to a positive and significant raise in VAT efficiency.14
1.3 A New Paradigm for Improving Tax Administration and Increasing Public Officials’ Professionalization: Industry Vertical Assessment
Tax administration reforms occurred during mid-1990s in all 27 Brazilian states, using guidelines proposed by the IDB.15 Government leaders discussed these reforms between 1994 and 1996, and official implementation began in 1997. Tax collection grew at phenomenal rates, first abruptly and then moderately, during the first five years after the reform, leading to a relative stabilization after 2002. Therefore, the period between 1997 and 2005 presents a unique opportunity to analyze the immediate aftermath of reform and the process towards (re-) construction and consolidation of new public institutions.16
Despite the similar reforms across the states, the results differed both with respect to quantitative goals (e.g., increased tax collection) and qualitative ones (e.g., public officials’ professionalization and improvements in the taxpayer/tax-collector relationship). For example, even though the states of Paraíba and Rio Grande do Norte have similar GDP and populations, their VAT collection growth between 1995 and 2005 was quite different, 39% and 102%, respectively. Similarly, Ceará’s VAT growth in the same period was 21%, whereas Bahia’s and Pernambuco’s were 23% and 35% respectively. In the case of São
14 VAT efficiency is simply defined as a VAT/GDP ratio. I discuss tax efficiency in chapter 3. 15 Funding guidelines will be studied in chapter 4. 16 Although the reform timetable varied from state to state, the guidelines for the reforms were similar across all states. For example, Bonfim (1999) reports that the reform in Ceará’s State Tax Administration Agency started as early as 1987. In Bahia, my interviewees reported that the program started in the early 1990s as a derivation of the program BAIA AZUL (Blue Bay). In São Paulo, reforms also began in the early 1990s (interview with Dimas Mello, tax collector, 2000). However, the bulk of the reforms occurred after 1997 as a result of an agreement between the Ministry of Finance and the IDB (IDB 1997). Despite reform initiatives carried out before 1997, only the post-1997 reforms were all encompassing and implemented across all the states.
26 Chapter 1. Introduction
Paulo, by far the biggest state in Brazil, VAT collection stagnated over the period evaluated (Appendix A, Table 21).
The difference between GDP growth and VAT growth among states raises many questions. Why does tax collection increase in some states more than in others? Why does tax collection vary from state to state within the same regional jurisdiction and legal framework? Under which conditions are state governments able to increase taxes without raising tax rates? What factors caused disparities among states within the same region and similar economic backgrounds? For example, why did Paraíba and Rio Grande do Norte perform differently? Although these questions have been explored at the national level (Aizenman, Kletzer et al. 2007; Keen and Lockwood 2006), there is no significant study at the sub- national level.
Findings from two years of field research suggest that to transform from a bureaucracy that barely enforces the minimum law—i.e., collecting taxes due—to an efficient and accountable public institution demands investments in both organizational and technological change. This in turn leads to a deep change in the way the tax agency and its employees, the tax-collectors, interact with the private sector, and particularly, the business sector in the VAT case. This shift in the technological, organizational and institutional environment increases a tax agency’s efficiency in tax collection. For example, upgrading from a territorially-based tax collection system to a vertically-, industry-assessed, production-chain- organized collection system was crucial in accomplishing the results achieved by the STABs in Rio Grande do Norte (RN/STAB), Pernambuco (PE/STAB), and Bahia (BA/STAB). Four factors were crucial to the success of this change: (1) the change in the rationale behind the tax collection process, i.e, the change in the way of thinking about tax collection; (2) the use of IT systems to enact this new rationale; (3) a change in public sector professionals’ sense of identity; and, (4) a much closer interaction between tax collectors and entrepreneurs in the private sector. The study of these four factors is the core of this dissertation.
Success, therefore, did not come through simple, top-down strategies. Rather, it was necessary to restructure institutional arrangements in order to engage the STABs and their tax collectors in a new tax collection process and change their relationship with taxpayers. In due course, the new practices allowed the transfer of knowledge among tax collectors and STAB management regarding production processes that affect VAT collection, which in turn enabled the dissemination of information to the private sector about taxpayers and the new taxation processes. The outcome drew respect from business sector
27 Chapter 1. Introduction associations and boosted interaction between the two sectors. Eventually, these interactions led to the formation of alliances that in turn led to best practices in tax collection.
In a similar vein to Fountain’s work on institutions (Fountain 2001), which focuses on the interaction between technological and organizational change, this paper analyzes the interaction between technological and organizational change as a systematic and conjoined process in shaping and bringing forth new institutional configurations in the public sector. However, this work is ultimately about the cognitive process that surfaced from this combination of enacted technologies and organizational changes that shaped new professional identities and philosophies. This new cognition within the sector was based on industrial vertical assessment, an assessment that led to the professionalization and specialization of tax collectors. In turn, this specialization enabled formation and consolidation of knowledge, which empowered and increased individual participation and let to strengthened institutions. This finding contradicts those who believe that specialization fragments knowledge and thwarts workers’ empowerment (Brusoni 2005; Shepard 1970; Tyler 1973) and offers an additional prescription for public sector reforms beyond the common ideas of mission, responsibility, and rewards (Grindle 1997; Lipsky 1980; Tendler 1997; Wilson 1988). Although these three ideas offer important explanations for improved workplace performance in public bureaucracies, they do not account for ongoing and joint changes in technology, organization, and personnel. This insufficiency is of particular concern with the information age’s widespread use of technology in all sectors and in all aspects of management, including planning, budgeting, and taxation.
Ultimately, I intend to show that in the case of tax administration reforms in the Brazilian states, the conjunction of tailored organizational and technological changes influenced tax collectors’ and taxpayers’ compliance, cooperation, and adhesion to reform efforts. Technological change enabled a new rationality in the tax sector and enabled industry segmentation and functional specialization. In turn, such efforts led to institutional change and to the strengthening of public sector bureaucracies, particularly of the STABs. Since institutional change involves altering rules, belief systems, and cognitive processes, this study illustrates how understanding these processes is key to creating new modes of governance, as illustrated in Figure 1.
28 Chapter 1. Introduction
Figure 1. Model Diagram: Technological and Organizational Change => Industry Segmentation => New Cognition Process => New Institution Image and Identity.
New Institutional Innovation in the Cognitive Process of Taxation Image, New Ties, New Identity
Unlike other works on reform which that either on technological, organizational, or professional change separately, this dissertation analyzes the interaction of both technological and organizational change — as a systematic joint process — in shaping and enabling new professional identities and institutional configurations in the public sector. Therefore, this work also highlights the cognitive process that emerged from the administrative processes in shaping new identities and philosophies. It establishes that vertical segmentation of tax assessment and consequent industry specialization, rather than alienating the whole process, enables knowledge and participation which, in turn, leads to individual empowerment and institutional strengthening.
I hope to contribute to the understanding of improved performance and capacity building in the public sector by extending the fundamental economic development idea conceived by Hirschman’s in his theory of backward and forward linkages (1958 and 1977). This idea is developed further in chapter 4. I add in elements from theories of professionalization and learning, drawing on the works of Dewey, Freire, and Lave (Dewey 1933; 1963; Freire 1997; 2007; Lave 1988), in order to further understand the interaction between people, technologies, and change in government reform conditions. I thus hope to contribute to three distinct areas of research: the reform of the state, the reform of tax administration, and organizational and institutional studies in public bureaucracies.
29 Chapter 1. Introduction
1.4 Structure of the Dissertation
This dissertation is structured as follows:
In Chapter 2, I describe the methodology used, the variables studied, and my strategy to accomplish this research. I examine the background for each case study (i.e. each state, viz. Paraíba and Rio Grande do Norte, Bahia, Pernambuco, and Ceará, and São Paulo). I describe the institutional and historical background of the states and present a brief comparison of their economic, political, and social conditions. I also develop the research question and the rationale for a double-edged research that includes case studies and quantitative methods, such as statistical analysis. In showing the most relevant data for considering the successful states vs. non-successful states, I suggest the path for my conclusion in subsequent chapters, a conclusion based on understanding revenue increase as correlated to concurring organizational and technological efforts.
Chapter 3 evaluates several indicators for the variable Reform and proposes a novel statistical method to explain the difference in VAT increase among Brazilian states. I describe the quantitative analysis’ dependent and independent variables, the descriptive statistics, and the model derived from statistical analysis. Finally, I demonstrate the relationship between the increase in Brazilian state VAT collection, economic development (given as state GDP), and administrative reform, using IDB indicators and data, initially produced by the STABs. Although the database used has some level of inaccuracy, this is the first time that such analysis has been pursued. I evaluate this type of data and consider other performance indicators.
Chapter 4 looks inside the STABs, examining the reforms and the prescriptions followed for changing government and creating new tax administrations. I briefly describe how the literature on administrative reform explains efficiency and institutional strengthening, particularly with respect to tax administration reform and efficiency. I examine the purpose, functioning, and evolution of tax administration reform efforts, emphasizing current trends and best practices in the field. I discuss the conditions that have been identified to date as relevant in determining the success and effectiveness of such reforms, stressing the virtues and shortcomings of the theoretical frameworks usually found in the literature. I review the guidelines for tax administration and specific tax issues, such as the Brazilian states’ tax collection structure related to VAT, and evaluate specific data on Brazil’s administrative reforms. Finally, I assess each of the state’s current tax collection organization and describe how the STABs shifted their VAT collection from a geographical model to one based on industry segmentation. This industry vertical assessment is the first part of my argument for a better reform.
30 Chapter 1. Introduction
In Chapter 5, I present the main features of the reform related to computerization. I describe and discuss the impact of computerization on the Brazilian STABs and examine how technological change has impacted public sector reforms.
In Chapter 6, I discuss the professional and institutional arrangements that both contributed to and resulted from the implementation of modernization programs at the STABs. I address the professionalization of public officials, the changes in their public image and identity, and issues related to training. This chapter analyses the learning paths of public officials, going beyond the literature conventional focus on mission and rewards, to reframe the debate about swapping reforms and learning paths in reforming the public sector. I also discuss the institutional arrangements and the relationships between and among the government agencies, private and non-governmental organizations (such industry federations, business associations, unions, and workers associations), and international agencies such as the IDB.
In the conclusion, I review my hypothesis and I propose a new model to understand and implement bureaucratic restructuring and modernization reforms. This model suggests a different way to increase the effectiveness, efficiency, and strength of public institutions: one needs to uncover the underlying rationale for each sector in order to catalyze reforms. The more reform is structured, logical, and linked to the productive reality of the sector, the clearer, stronger, and easier its implementation will be.
Finally, I place the research on tax administration reforms into a larger context—the study of public sector bureaucracies’ reform and individual learning—and I highlight a series of possible alternative paths to successful change in the public sector. I elaborate on this framework, comparing it to alternatives and attempting to contribute to unsettled questions and gaps in the technical knowledge about tax administration reforms. I also discuss what is needed to ensure a successful public sector reform. Finally, I explore some of the theoretical implications that can be derived from this research and offer policy recommendations for restructuring public sector bureaucracies and reforming public administration.
31
Chapter 2
Research Design and Methodology
2.1 Research Query
The desire to understand the key conditions under which administrative reforms strengthen bureaucracies in the public sector and translating this understanding to less developed settings to improve government motivated this research. The main research question of this study is a straightforward outcome of this unsettling ambition: “Under which conditions can administrative reforms endorse and strengthen public sector bureaucracies and institutions?” My research question is thus positioned at the intersection of several different fields of study: administrative reform and public management, bureaucracies and organizations, individual learning, human management, and personnel training. Each one of these fields unfolds in several branches: administrative reform, for example, encompasses issues ranging from technological change to organizational restructuring and service customization practices in the public sector—to cite a few, all interconnected and relevant. The challenge was to choose a research strategy that would support my study and at the same time allow tangible analysis.
My first strategy was to find a case study and an applied area of interest. In an exploratory field trip to Brazil, I found that state governments had taken on restructuring efforts in their tax administration bureaus across all 27 states in Brazil. This provided me a valuable window of opportunity to follow my interest in government reform. Moreover, the reforms had been initiated at the same time and under the identical guidelines from the IDB and the Ministry of Finance, granting a suitable set of comparable cases for empirical analysis. However, given the particularity of the field of study, tax administration, a
33 Chapter 2. Methodology
secondary set of questions formed: “How can administrative reforms in the tax sector improve tax collection, increase government credibility, and tax compliance?”
Further interaction with a selected group of cases raised more questions investigated under the same lens of whether and how reforms can strengthen public sector: “Why in some states have the reforms produced results closer to the goals and in other cases not?” From the interaction of extant theory and empirical observations emerged my research methodology. This research work has been an iterative process of back and forth questioning and debating in the search for intellectual and yet realistic explanations, that would provide me tools to learn, intervene, and replicate successful experiences in the public sector reforms. The methodology derived from and was constructed during the interactions with the officials and consultants that I interviewed and according to the availability of the data. Initially, I did not intend to undertake statistical analysis; only later on, when the indicators for all the states were kindly provided to me by IDB consultants17 and STABs’ officials, at the end of the reform efforts, I decided that there was enough data to allow it.
2.2 Research Design
This research was designed to examine how the restructuring of tax administration bureaus has strengthened public bureaucracies and institutions, increased tax collection, compliance, and credibility, and gone beyond to improve the public-private sector relationship and to promote a shift in the professional identities of tax collectors. I have done bibliographical and archival investigation, and conducted 167 open-ended interviews with officials from six state governments in Brazil and consultants in international organizations (besides several informal conversations). These interviews lasted between one and half and two hours, sometimes longer and were spread over a period of 3 years (see Appendix B). I interviewed current and former tax collectors and heads of tax agencies, representatives of business associations and public servants` associations, entrepreneurial leaders, scholars, and officials of international donor agencies IDB, the World Bank, and the IMF, and the Ministry of Finance’s leading office,18 among others.
17 Interviews with Jaime Mano, Orlando Reos, and Julieta Verleun in Brazil and in Washington, 2004-2005. 18 At the Ministry of Finance, the controlling unit was the UCP, Central Unit, which coordinated and gathered information about all STABs.
34 Chapter 2. Methodology
Figure 2. Multi Level Methodology.
COUNTRY: BRAZIL
Developed Region Regional level: Developing and South and Southern Regions Underdeveloped North São Paulo Northeastern and Central Regions
Pernambuco, Rio Grande Ceará, Bahia do Norte Paraíba STATE TAX ADMINISTRATION BUREAUS Procedures
Organizational Technological Personnel Reform Innovations Capacity Training
Results (Tax Increase) and Processes (Tax Collection)
For this dissertation, I undertook a comparative case study of the tax administration reform at the state level in Brazil. This method was appropriate for answering my research question for two reasons. First, it required me to make a theoretically driven and systematic collection of data about the same variables across cases, such as tax collection and efficiency measurements, state economy, number of tax collectors and their education, training, worker’s performance, and technological change in each STAB. I also investigated local historical facts and an array of political institutions. This gave me the opportunity to make good descriptive and accurate causal inferences. Second, the immersion into the complex and
35 Chapter 2. Methodology
untidy practice of tax collection administration made my research viable and productive, essential to understanding both particular occurrences and general trends, while contributing to the advancement of the grounded theory.
Figure 2 provides a schematic of the research design, with a focus on state governments and, ultimately, the STABs.
2.2.1 Grounded Theory Building I chose to use grounded theory for my methodological approach because it allows for a systematic inductive approach to evaluate the interactions and processes that occurred in the implementation of the tax administration reform in the STABs in Brazil. The systematic approach is ideal to engage field observations and interviews as data collecting techniques, allowing for the integration of individuals’ perspective in analyzing processes (Binz-Scharf 2003). I was concerned with both individual and collective behavior within the organization.
The outcomes of this learning and changing process can be best understood by iteratively considering theories with the data: I discovered new patterns of behaviors and practices in tax administration reform. That is what grounded theory tries to do: “to understand why and how structures, conditions, or actions might arise to ferret out generative mechanisms, to explore conditions under which these effects might vary or not, and to qualify their temporary and emergent aspect (Dougherty 2002: 851, cited in Kaplan 2004).” As Giddens says, “all so-called ‘quantitative’ data, when scrutinized, turn out to be composites of ‘qualitative’ – i.e., contextually located and indexical – interpretations produced by situated actors” (Giddens 1984, cited in Kaplan 2004). My goal is to generate a theory of the middle- range that highlights the critical events in the process that led to success implementation of reform.
2.2.2 Methods The research comprises a one country (Brazil), six states study, namely Pernambuco, Bahia, Ceará, Paraíba, Rio Grande do Norte, and São Paulo (see Figure 3). For this dissertation, a combination of (a) a study of reform processes and phases over time, (b) a set of case studies, and (c) analytical statistical study of reform indicators were used. The primary data collection tool for the case studies was structured, open-ended interviews with a list of sample questions included at the end of this chapter. Subsequently, case-analysis was used because the focus of the dissertation was the intersection of three subject areas: organizational and institutional reforms, technological innovation in the public sector, state capacity building, and successful tax administration restructuring, whose combined effects are difficult to
36 Chapter 2. Methodology
assess statistically. The efforts taken have provided reliable data for drawing a conclusive assessment of the performance and effectiveness of the restructuring endeavors in the state tax administration bureaus during the period of 1997-2005.
2.2.3 Case Study Approach The choice of cases was bilaterally informed by the tax administration reform literature and by the results observed after reform implementation. Initially, the case selection was structured to track public sector reform, development capabilities (e.g., public officials’ capacity building), and effectiveness of tax collection over time, specifically in a period of eight years, between 1997 and 2004. Although I wish I could have extended my research close to the final writing of this dissertation, for statistical scrutiny, data on state GDP was available only until 2004. Due to “red tape” and other bureaucratic obstacles, I was unable to collect data on wages and employees’ year-by-year numbers and human resources.
The emphasis was on tax administration reform and the set of technological tools and practices in public policy aimed at reform, specifically (a) government restructuring, and (b) tax administration reform and effectiveness, and (c) tax collection. Increasing tax collection and effective government restructuring are key government prerequisites to provide better services to the public and to promote economic development. Therefore, these aspects became central in my initial research as proxies for successful administrative reforms.
The focus was on the state governments that implemented tax administration reforms, specifically related to VAT, for two main reasons, viz.: although tax administration reform has been widely explored in the literature, most cases studied are national government reforms implementation. Second, while other units of analysis—national, macro-regional, and municipal would be without a doubt, interesting and insightful for any study of reforms, reforms at the national level have been more frequently studied, whereas reforms at the municipal level is understudied often due to lack of data.
Many of the reforms implemented by the Brazilian governments have been carried out across diverse sectors and government levels, such as general administrative reforms in health and education. Although more developed states have stood out as more active reformers, these reforms have been exhaustively examined, such as is the case of New Zealand, England, and the United States (Kettl 1998; Kickert 1997; Osborne and Gaebler 1993), to cite a few. Furthermore, the Brazilian states present a complex and diverse set of features—ranging from highly industrialized and developed regions to some
37 Chapter 2. Methodology
of the poorest regions in the world,19 which allows me to draw comparisons and policy recommendations for countries across the development divide.
The localized examples of good tax administration reforms can be applied to wealthy and poor nations and governments alike providing an insightful and interesting model for nations with different profiles. The diversity of features in the chosen states range from wealthy and industrialized urban settings (as is the case of São Paulo), to rural and agricultural, to developing states (as is the case of Rio Grande do Norte and Paraíba), encompassing states with a very mixed urban and rural economic reality, such as Bahia, Pernambuco, and Ceará (See Appendix A). Yet the same national guidelines informed national policies cutting across states, forcing agricultural-oriented and rural states to engage and catch-up with the reform plans designed for industrial-driven, urban areas. In this fashion, São Paulo can be compared to the most developed nations, while Bahia, Ceará, and Pernambuco fall among the striving developing countries, and Rio Grande do Norte and Paraíba would serve as a model for poorer nations.
Figure 3. Brazil - Political Division - Regions and States and States Chosen in this Research.
Therefore, for the purposes of this research design, looking at the state governments is a relevant and pertinent to unveil the reasons why modernization reforms—encompassing administrative,
19 UNDP indicators on quality of life in Appendix A.
38 Chapter 2. Methodology
technological, and institutional—are more likely to reach the goals initially set, despite their diverse context.
2.2.4 Sample Selection As for the selection of the cases, São Paulo, Ceará, Rio Grande do Norte, Paraíba, Pernambuco, and Bahia were chosen not only because they figured among the most interesting cases for tax administration reform in Brazil, according to my primary key informant interviews,20 but also because these states represent different levels of economic development and institutional capacity. Although most of my cases are located in Brazil’s northeastern region, they represent different strata of Brazil’s economic development. For example, while Pernambuco and Bahia are very developed with respect to industrial and institutional arrangements, Rio Grande do Norte and Paraíba fall among the states with the poorest institutional capacity and lowest economic development. Ceará can be compared to Bahia and Pernambuco, given its location, institutional development, and size. São Paulo, however, stands on its own, due to its economic and advanced institutional development (for general information on the states— e.g. population, state GDP, tax collectors’ wages—please see Appendix A).
Therefore, I separated the six states into three different groups, according to social-economic features collected about each state, comprised of (1) a combination of secondary data from Brazil’s leading economic think tanks (IPEA, BNDES, IBGE), international donors (IDB and IMF) on tax collection and revenue from 1995 to 2004, and (2) primary and secondary data from interviews with individuals in the six states, to wit, Bahia, Ceará, Paraíba, Pernambuco, Rio Grande do Norte, and São Paulo, considered to be relevant cases in the reform process.
Other characteristics make these states an interesting comparative study. Paraíba and Rio Grande do Norte are neighboring states with very similar economic activities. They present similar population and tax collection figures as a rate of the state GDP. However, when looking at the tax collection growth obtained in the reforms, Rio Grande do Norte stands out as one of the most successful experiences. The same comparative economic and regional conditions apply to Bahia, Ceará, and Pernambuco, respectively. This choice of states allows for greater variation across cases, since I compare the same policies and outcomes—the modernization of tax collection administration— across them, while accounting for the differences related to the administrative reform. This circumstance led to questioning the reason why STABs achieved different results when implementing reforms.
20 In the initial phase of my research, I always asked to my interviewees in the Ministry of Finance, IDB, and State Tax Administration Bureaus to identify cases of reforms that they would emulate.
39 Chapter 2. Methodology
The natural research path here was to look inside the reform. Finally, my analysis of government reform and institutional rearrangement as related to the performance of Tax Administration bureaus are located in three research areas: (1) Technological innovation; (2) Organizational processes; and (3) Professional and Institutional arrangements.
Therefore, choosing the state sample was not a straightforward exercise. Faced with a trade-off of great geographic spread for Brazilian state governments, a limited time frame and budget for data collection, this sample selection allowed an investigation of composition, capabilities, and evolution of the taxation sector as reflected in the states across different regions. No other state sample combination in close proximity could have afforded me the same sample coverage in the technology, public administration, administrative capabilities, and institutional diversity spheres.
2.2.5 Inductive Approach I depart from the premise that the organizational change and tax collectors’ professionalization enabled by computerization are the key elements leading to new arrangements between government agencies, businesses’ associations, public servants’ unions, and institutional strengthening. There was also enough starting evidence to indicate an alternative explanation based on each state economic development. At first, I collected detailed historical data about the elaboration, discussion, and implementation of the restructuring process in the 1990s in Brazil. Information was gathered about the process of tax collection, the resistance and collaboration among different sectors, and the influence that international agencies exert upon state governments and agencies. Specifically, I looked into the institutional settings with respect to the economic background and lobbying capacity of workers organizations and business associations, the consistency of the policies, and the historical trajectory of the STABs over the last decade. I also collected data on the acquisition and development of technical expertise, the computerization process, and the effectiveness of the control mechanisms to discipline tax collectors and to ensure tax compliance. Then, I showed how these conditions shaped the larger set of institutions that affected the successful restructuring of the tax agencies and influenced the design and implementation of sponsored projects, and thus determined the performance and effectiveness of public bureaucracies at promoting administrative change.
Hypothesis testing took place in two research arenas—processes and institutions. My initial hypothesis for success in increased tax collection and reform implementation emerged from the literature review and comprised: tax collection increased as a function of economic development and administrative reform imperatives, such as streamlining of the organizational collection process,
40 Chapter 2. Methodology
computerization, and professionalization of tax collectors. For the first part I looked at the economic performance through GDP. Then, I described the content and traced the process of restructuring practices and the technology used. For the second part, I studied the relationships that evolved during the last decade between business associations and professional organizations within the public sector that helped to develop the implementation of the tax administration process in the states. My secondary hypothesis was that there was an antagonistic relationship between public sector and private sector. Hence, in the states were such antagonism was curtailed or the government was strict in enforcing the tax regulations, tax collection would increase and restructuring would take place promptly.
Several instances of the implementation of the modernization process were analyzed as well as the institutional arrangements for each of the case studies. I focused on the technology implementation order, training procedures, and the organizational structure of tax collection. This allowed for comparisons within and across cases, since I applied the same analytical framework to the study of the six cases.
The analysis was not strictly equivalent in each case. This was because of (1) the limitations to gather and validate comparable data from the six states between 1997 and 2005,21 (2) the complexity of the projects and (3) the specific tasks state governments have undertaken varied in slightly different ways across states. That said, I believe that the underlying logic of the argument applies in all cases, and that most instances of restructuring and modernization share enough structural elements in common to be comparable. Therefore, even if the comparison across states has some variability, it will be a methodologically valid and theoretically illuminating step.
2.2.6 Statistical Method Approach Finally, a decisive step complements my endeavors towards supporting my findings. I chose to do a statistical, longitudinal approach to crosscheck the findings in my case studies. This analysis will be further discussed in more detail in the appropriate empirical chapters (Chapter 3 for the quantitative study of the entire population).
21 The period of eight years was defined due to specific economic changes in the national economy. Until 1994, the Brazilian economy faced several economic plans that changed the currency and adjusted prices. From 1994, however, the federal government stabilized the currency. For this reason, it is possible to have coherent results in terms of economic growth and tax collection. In this study, we examine the period between 1997 and 2005, given that although the reform plan had been discussed since 1995, implementation only started in 1997. We consider that a period of eight years is enough to evaluate the performance of reform efforts, after which the efforts fade out, or are incorporated into regular policies and procedures, or lead to a plateau in terms of tax collection increases. (Ramos 2001)
41 Chapter 2. Methodology
The formal framework of the analysis uses a series of indicators and features related to the restructuring, computerization, and effectiveness of tax administration reforms (e.g, number of computers installed, staff attendance of training courses, on-line filing, etc.). For the tax collection growth assessment, I have identified VAT revenue growth and treated it as dependent variable (Chapter 3). The independent variables are the conditions of the increase in the tax collection, the changes in the administrative process, the online procedures, and the training of public officials.
When looking at tax administration internal structure and process, the professionalization of employees, and the healthy relationship between tax collectors and taxpayers I was able to select a few factors to compose my independent variables. I only used in my analysis those independent variables which I believe have most explanatory power and for which it was feasible to gather sufficient and reliable data for testing, such as new on-line procedures in taxation processes, data on administrative efficiency, and training. A presentation of the variables is found in chapter 3. Although the study is inductive in nature, hypothesis testing was done as findings emerged. Eventually, I was able to test out my inductive conclusions through this statistical evaluation.
I use the existing indicators for tax administration performance as a proxy to test the variables that I construct for my case study. The Reform variable is then composed of several elements stratified in the list provided by the IDB, 14 different variables in all. Combined, these indicators represent the administrative efficiency, computerization, and training in all the 27 states. The methodology and details are discussed in chapter 3.
The goal of using two methods to assess the same question is to bridge the field study and the quantitative research traditions. Coupling a quantitative study of the entire set of states with a qualitative study within one STAB helps to amplify the context and the connections to explain the observable facts. Statistical analyses were used to test the casual connections between qualitative and quantitative analysis.
2.2.7 Data Collection Techniques and instruments for data collection include the following:
• Bibliographical investigation, drawing mainly on books, articles, scholarly work, and government documents about government reform in Brazil, particularly in the six states selected;
• Archival investigation, based on internal documents and from each STAB and on internal documents from some of the agencies sponsoring the projects;
42 Chapter 2. Methodology
• In-depth interviews with current and former officials of Tax Administration Agencies, international donors, representatives of businesses associations, and workers` associations.
2.2.8 Timeline The initial in-field research was carried out during the fall semester of 2001 and spring semester 2002. I then returned to Brazil in 2005 and 2006 and devoted a few months to complete the data gathering process and to finish up writing the preliminary draft.
2.3 Questionnaire
The questionnaire below was applied in most of the open-ended interviews conducted for this study.
General question about the administrative reform and tax collection procedures:
1. Do you remember the first administrative reform program to be implemented in the State Tax Administration Bureau? When was it? When was the most recent change? 2. What were the first steps taken in the last reform? How did it differ from the previous reforms? 3. When does the reform really started in this state (e.g. in Bahia)? What was the launch event? 4. Who was the Secretary (Head of the Tax Administration Bureau) at the time those reforms were implemented? What was his management style? 5. What is the functional structure of the STAB? Who does what? What are the specific functions for each public official? 6. Have you heard of tax administration reforms in other states? What were the comments and which states you think have been able to make a change in the STAB?
Questions about the tax collector professionalization: 7. Since when have entrance examinations been established in this State? 8. How many years have you worked as a tax collector? 9. Do you work internally (office paper work) or externally (street tax collection) in the Tax Administration Bureau? 10. Do you work in the state capital city or in the interior? 11. Do you have a choice about place? Have you chosen? Why? 12. Why have you chosen to be a tax collector? 13. What is the best thing about being a tax collector? 14. Please comment on the following professional aspects (after the answer is given): a. Prestige b. Wage c. Working conditions d. Freedom 15. Which strategies have been more effective during the STAB modernization process? 16. Please comment on the following modernization features: a. Computerization b. Reorganization c. Training
43 Chapter 2. Methodology
d. Service to taxpayers 17. Do you disagree with any reform policy strategy? Why? 18. Which steps in the modernization process do you think are mostly relevant/irrelevant? Could you elaborate on your answer? 19. What would you have done differently? 20. Do you think the current policies are helpful to your work? 21. Which ones are helpful and which ones are not as helpful? 22. Which previous policies were more helpful than the new ones? 23. What was your education level before being a tax collector? 24. What is your background? 25. Do you think that your previous background helps/helped you with your career as a tax collector? How? 26. Do you think the Tax Administration Bureau is doing something to take advantage of it? 27. How do you relate to the taxpayers? How do they treat you? Has the relationship changed?
Questions about unions and workers’ associations: 28. Are you affiliated with the union? 29. To what extent has the union been helpful for your career advancement? 30. Has the union opposed or advanced any of the modernization policies? 31. What has been the most interesting work done by the union up to now? Why? 32. Are you a member with the Tax Collectors’ Union or any other Workers’ Association? 33. Why? How is it helpful? 34. What do you like about them? 35. Have you taken part in any courses after you have become a tax collector? What kind of courses? Who offered such courses? 36. What was the best course that you took? Why was it a good course? 37. If you had to leave the Bureau, what would you do? 38. Are you happy with your work?
Questions for the business sector representatives: 39. What is your business? 40. How is your relationship with the Tax Administration Bureau? 41. Has any significant change occurred recently? 42. How would you describe and evaluate these changes?
44
Chapter 3
Human and Managerial Determinants of Tax Revenue Growth: A Novel Statistical Model
3.1 Introduction
This chapter completes the mainstream methodology that estimates tax collection performance only through the analysis of monetary and fiscal determinants. A diverse set of variables stemming from the reform package are used to explain revenue accretion after reform implementation: administrative efficiency as a measure of organizational change; online services and connections as a measure of computerization; and the number of courses for personnel training as a measure of human capital investment. The statistical study in this chapter analyzes the changes effected by the administrative reforms by quantifying traditionally “qualitative” aspects of government performance, and controlling for the level of economic activity. In other words, this study, in addition to the traditional economic determinants of tax revenue generation, explicitly allows state-level administrative behavior to affect tax collection effectiveness. This chapter shows the normatively positive and significant correlation between some policy reform parameters and tax revenue increase. For instance, the results show that for each percentage-point increase in the number of employees placed in training, there is a positive and significant 0.17 percentage point increase in state tax collection ability, as detailed later. For example, a state such as São Paulo, which VAT revenues approximately R$ 50 billion in 2005, could have a revenue increase of R$ 85 million (i. e, approximately US$ 43 million).
This study shows that organizational processes, structural arrangements, professional enhancement, and technological change are significantly correlated to increases in tax collection. These
45 Chapter 3. Statistical Models
results have significant implications for the policies aimed at constructing government reform packages, particularly in developing countries.
3.2 Brief Literature Review on the Determinants of VAT Revenue
Central to government preoccupations—in countries both developed and less developed—are the effects of budgetary deficits on the economy and the well-being of its population. The role of the state as service provider, despite contravening privatizing interests, keeps growing and depends on the success of governments to raise revenues. It is also critical to bear in mind that the economic cycle of the 1990s, the diminishing level of foreign direct investment, and the inflation “ceilings” set by international donors and agencies, all altered the scope within which developing countries’ governments could handle their finances and obligations. The VAT tool comes then as a blessing for many governments to raise revenues and has been touted as the new “money machine.” Keen and Lockwood (2006) discuss how countries with a VAT raise more revenues than do those without. However, the authors also point out that those revenues can be offset by reduced revenues from other taxes. This last comment, however, does not seem to resonate with the Brazilian tax system overall collection, since other sources of fiscal revenues related to value-added and industrial production have also increased (Graph 3).
Surprisingly, the literature and empirical work on VAT is still scant and “few papers have sought to model the revenue raised by, and compliance with the VAT, often with a view to estimating a revenue- maximizing rate” (Keen and Lockwood 2006). Keen and Lockwood model the revenue ratio r as:
rit = αVit + β‘Xit + β’vVit Xit + µit+ λt+ uit where Vit is a dummy variable if a country has VAT, Xit captures other variables affecting the tax ratio, the term µi is a country-specific effect, λt is a time effect, and the idiosyncratic error, uit, assumed to be uncorrelated.
Common-sensical comparisons across countries are usually based on the taxes internal features, such as the rates and existing reductions, the exemptions, the presence or absence of a large taxpayer unit, exports and imports, per capita income, retail economy and so on (Ebrill, Keen et al. 2000).
46 Chapter 3. Statistical Models
Graph 3. Brazilian Tax Revenue -- Business Related Taxes.
700000
600000
500000
Federal VAT 400000 BUSINESS INCOME TAX Industrialization Tax Total Import Tax 300000 VAT Total Federal Tax Revenue
200000
100000
0
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
Source: IBDG and IPEA. Nominal values (see Appendix A).
Other studies emphasize the role of monetary and fiscal determinants for tax revenue growth. Frequently, these studies refer to the performance of a tax system according to the region or country’s economic performance. For example, VAT is typically measured as a ratio of its revenue to GDP or by variables relating to the structure of the tax and competing taxes in the economy. In measuring the VAT performance in Mexico, (Tijerina-Guajardo and Pagán 2000) use a roll of variables including VAT rates, manufacturing production, government spending; and exemptions, number of VAT rates, and the rate of inflation.
47 Chapter 3. Statistical Models
In Brazil, tax efficiency has usually been measured as a function of those variables in addition to urban and rural population rates, employment, and a rate for the urbanization index (Marinho and Moreira 1999). The authors use the following equation to specify the tax capacity of Brazilian states and explain tax collection in Brazil by the equation:
ln(ICMSit)=β0 +β1 ln(GDPit)+β2 ln(UPit /Pit )+β3 ln(Xit )+β4 ln(Mit )+β5 ln(πit)+β6 ln(UDit)+vit −uit where ICMS is the state VAT collection in the period t; GDP is the GDP by state i in the period t; UP/P is the ratio of employed urban population over 10 years-old state to the total population of each state i; M is the imports in each state I, in the period t; π is the medium price index variation in the period it; UD is a proxy for urbanization, using data on residential and industrial electricity, telecommunication, and diesel consumption, in each state I, in the period t; v are errors and u are non-negative random variables, such as technical inefficiency.
In the same study of the Brazilian northeastern states, Marinho and Moreira observe that the potential tax burden is way below the real tax burden, by calculating the fiscal effort for each state. They claim that the difference between potential and real tax burden could be reduced if the state could “promote a effective tax collection and administration (Marinho and Moreira 1999).”
Interestingly, Marinho and Moreira also found that the fiscal effort (EF), measured by the difference between potential tax revenue (CP) and effective tax revenue (CE), is still low in the northeastern states (see Table 2).
Table 2. VAT Fiscal Effort, Potential Taxation, and Effective Taxation.
1991 1992 1993 1994 1995 1996 CE EF CP CE EF CP CE EF CP CE EF CP CE EF CP CE EF CP MA 3,49 68,11 5,13 3,85 76,39 5,04 3,31 66,08 5,02 3,73 61,36 6,07 3,84 55,51 6,92 4,18 54,82 7,63 PI 4,38 84,40 5,18 4,69 98,70 4,75 4,36 91,12 4,78 4,06 65,55 6,20 5,12 74,27 6,89 5,63 70,97 7,94 CE 5,37 83,72 6,42 4,94 76,97 6,42 4,80 74,89 6,41 5,41 68,55 7,89 5,90 73,75 8,00 6,18 61,58 10,03 RN 3,77 65,52 5,75 3,41 59,88 5,69 3,24 59,19 5,47 3,73 51,93 7,18 4,59 56,31 8,15 4,91 54,43 9,03 PB 4,83 77,64 6,22 4,40 76,14 5,77 4,05 71,13 5,69 4,19 59,49 7,05 5,29 63,17 8,37 5,66 63,78 8,87 PE 6,59 99,99 6,59 6,17 93,57 6,60 5,80 90,27 6,43 6,27 79,74 7,87 6,93 78,74 8,80 7,47 73,48 10,16 AL 4,68 96,29 4,86 4,39 88,25 4,98 4,07 89,55 4,55 4,42 72,32 6,11 5,18 76,51 6,78 5,38 73,64 7,30 SE 5,85 98,99 5,91 5,16 96,61 5,34 5,41 97,18 5,56 6,75 98,39 6,86 7,37 92,13 8,00 7,31 90,53 8,08 BA 5,95 88,67 6,71 6,30 96,26 6,54 6,03 92,70 6,50 6,41 79,64 8,05 6,79 74,32 9,14 7,16 72,03 9,93 Source: Marinho and Moreira 1999.
Observing Table 2, one can notice that revenue potential is higher than revenue in all the states. For example, in the state Rio Grande do Norte the potential capacity for tax collection was 9.03% of state
48 Chapter 3. Statistical Models
GDP in 1996. However, the STAB collected only 4.91%, suggesting that there is room for improvement in the efficiency of VAT collection and administration.
The authors added that, however,
“(…) The tax burden (sic) does not tolerate any increase… and it is possible that tax payers can not afford to pay any additional increase in taxation via rate increase, tax basis extension, or new taxes’ creation. Yet, the tax collection system is not equal across states, so that it is possible to increase tax efficiency by careful regulation and tax enforcement (Marinho and Moreira 1999).” (Author’s translation)
However, Vasconcelos et al. (2005) find that some Brazilian states do have more revenues than their theoretical potential. Vasconcelos et al. (2005) analyze the disaggregated fiscal effort of all Brazilian states, using a model that includes partial value of GDP, specified for agriculture, industry, and services sectors, and the estimation models show slightly different outcomes. Table 3 summarizes their results (Vasconcelos, Siqueira et al. 2005). The third column shows the fiscal effort (EF). Accordingly, EF values greater than 1 indicate an EF above the one predicted in their model; EF values less than 1 show effort below tax potential. Negative values in the fifth column indicate the trend in the state to increase or decrease the fiscal effort. Interestingly, Bahia, Ceará, Paraíba, and Rio Grande do Norte all present positive trends. São Paulo and Pernambuco present negative trends, indicating a diminishing effort to collect taxes over the years. Although the results refer to the period immediately before the reform, including only two years after the launching of the modernization program, it is an indication of the trends in the states.
In sum, government revenues are, at the very least, determined by the absolute level of economic activity. However, because of continued political and economic circumstances, and an increasing urge to reform governments and to make them fiscal solvable as well as liable and accountable for the services they ought to provide, administrative reform takes a preeminent place in the prescriptions to prompt feasible and industrious responses to those real problems. Knowing which aspect of tax administration reform works is the first step to identify why it works.
49 Chapter 3. Statistical Models
Table 3. Fiscal Effort and Trend. All states, 1987 – 1999. State Effective Tax Potential Tax Fiscal Effort Trend between Burden Burden 1987 and 1999 (%) AC 5.221 5.363 0.971 1.13 AL 9.251 5.363 1.067 -3.13 AM 16.367 16.13 1.016 0.08 AP 5.976 4.256 1.417 3.65 BA 12.918 12.702 1.013 0.15 CE 14.363 14.272 0.999 0.65 DF 8.582 10.237 0.996 7.63 ES 18.738 17.513 1.046 0.54 GO 17.744 16.626 1.033 -132 MA 9.325 9.101 1.012 1.33 MG 13.269 12.660 1.024 -1.25 MS 15.524 13.359 1.157 -9.68 MT 20.668 19.765 1.044 -3.82 PA 7.502 7.420 1.002 -2.10 PB 12.506 12.325 1.013 0.52 PE 13.819 13.681 1.016 -1.11 PI 12.802 13.161 0.971 -0.42 PR 10.070 8.610 1.175 -9.51 RJ 12.029 12.300 1.021 4.34 RN 9.743 9.579 1.005 0.94 RO 10.743 11.640 0.962 2.45 RR 8.334 8.074 1.034 -4.77 RS 14.494 13.724 1.025 -1.42 SC 14.704 14.035 1.027 -1.67 SE 11.471 11.530 0.993 -0.65 SP 14.621 14.225 1.023 -1.61 TO 20.795 22.170 0.967 5,41 Source : Vasconcelos (Vasconcelos, Siqueira et al. 2005).
When describing and understanding the revenue performance of a given tax, its nature or determinants, it is important to differentiate between factors that affect a given tax system, so that we are able to address particular obstacles or bottlenecks. In this chapter I analyze state Value-Added Tax (ICMS) revenue growth using a statistical methodology to ascertain some of the possible factors correlated to the success of the Brazilian states tax revenue increase over the past decade. A theoretical model is developed that draws from my fieldwork research and the framework described in the introduction. Specific reform components featured are integrated during the tax modernization period. This system is particularly related to the evolution of VAT collection and modernization effects. I use GDP as a proxy for the whole economy, and therefore, did not worry about including variables such as other taxes, imports, exports, and industrialization measures.
50 Chapter 3. Statistical Models
3.3 New Empirical Evidence, Indicators, and Model
In the 1990s, a series of indicators were created to measure fiscal reform (Febres, Cruz et al. 1998; Febres, Cruz et al. 1998; IDB 1996). These indicators range from the ratio of tax collection per GDP to the STAB administrative costs to employee training. The IDB and the IMF created and used these indicators to evaluate the reform implementation. The IDB and the STABs were kind enough to provide the data for the indicators (see Appendix C).
Fourteen disaggregated time series data are used, ranging from on-line tax file returns to numbers of arrears to number of employees’ trained. Given the novelty of this study, the uniqueness of the dataset, and the unorthodoxy of the theoretical model that is being tested, this approach should be of academic and methodological interest to policy makers, government officials, and tax analysts in developing countries lacking data or other information to implement government’s reforms.22
Although the role of indicators in determining tax revenue has been well explored in recent studies23, there have been few or no studies that apply regression analysis to sort out the reform predictors contributing to success, particularly in the case of VAT collection (PNAFE 2004). Alongside substantive fiscal reforms that deal with the nature of taxes and their rates, tax administration reforms gained place as an important tool to improve government. Both quantitative and qualitative measurements were created, though despite such an endeavor, government leaders, tax consultants, and donors alike still rely on the quantitative measurements of tax collection and compliance to account for success. They casually demote service quality, public sector consciousness, citizen participation, and fiscal education to a secondary plane, mostly because it is more difficult to obtain data on these features. The model described here does not provide an evaluation of how accurate the indicators are. Nonetheless, the timing as well as the variety of cases covered in this study is long and rich enough to allow for some inferences.
The emphasis will be on computerization, personnel training, and increased administrative efficiency. Given the novel approach to the data and the uniqueness of the dataset itself, the empirical results are merely an embryonic peak at the possibilities of tax reform evaluation.
22 Often, VAT performance is measured as the ratio of VAT revenue to GDP, ratio that in fact is commonly taken as an indicator of the tax efficiency (Silvani 1997, Marinho and Moreira 1999, Ebrill, Keen et al. 2000). 23 Several studies and projects by the World Bank, the IMF, the IDB and other organizations in the development world realm have not only given incentives to governments that are able to improve their indicators, but also studied them as indicators to measure performance (Febres, Cruz, et al. 1998; Cruz 2002).
51 Chapter 3. Statistical Models
In this paper, tax “efficiency” can easily be substituted by the term “collection effectiveness”. Indeed, if the VAT is a simple function of output, deviations of actual revenue from potential revenue can be interpreted as administrative inefficiency. In other words, if the tax base produces a certain level of potential VAT, then differences between what a state can collect and what it does collect can only be attributed to the way in which it goes about the business of collecting taxes. So, increases in efficiency can be viewed as percent deviations of tax effort relative to the mean tax effort.
First, since the amount of VAT is theoretically a percentage of GDP, the base model should equate VAT to GDP in a linear fashion. This relationship is captured via a linear regression equation. Previous literature has clearly shown how taxes such as the VAT are based on economic fundamentals.
Then, in order to capture the effects of administrative reforms, the form of the regression equation must somehow allow the level of key variables to affect the amount of taxes collected. Interacting the Reform metric with GDP allows the level of the reform variable to affect how effectively a state can extract VAT by unit of GDP. The regression equation, combining both the linear VAT-GDP relationship and the policy reform interaction term, is shown in Equation 1 below:
Equation 1
VATit= β0 + β1GDP + β2Reform + β3GDP*Reform + εi
Again, it is by interacting the reform variable with the GDP variable that the reform variable is able to enter into the GDP’s marginal effect:
Equation 2
δVAT/δGDP = β1 + β3Reform
So, β1 can be interpreted as the average tax effort or effective VAT rate, and β3 can be viewed as the additional tax effort allowed by the implementation of certain reforms.
The Reform variable captures some measure of tax policy, and only one measure of computerization, personnel training, or administrative efficiency is included at a time in the regression equation.
Therefore, the null hypothesis is that different features of reforms do not affect tax revenue: the null hypothesis is that the slope parameter β3 is equal to zero.
A simple model was selected for several reasons. The first is the nature of a Value Added Tax. A VAT is supposed to be intertemporally stable and solely based on economic output, measured in this
52 Chapter 3. Statistical Models
case with GDP. Indeed, simple linear regressions results of nominal VAT on nominal GDP produces a slope coefficient of 0.079, a slope coefficient standard error of 0.47%, and a regression R-square of 0.92. In other words, states are able to apply an average VAT rate of 7.9%, and the simple model is able to explain 92% of the variation in VAT collection.
To capture the effects of administrative reform on VAT collection ‘efficiency’, 14 regressions were run, with one predictor at a time. This allowed each reform policy variable’s effect to be measured.
There are three policy reform themes explored in this thesis. Each category is composed of several variables. Each variable captures some aspect of the policy theme. So clearly, variables from the same policy theme cannot be included in the same equation, because they would compete for statistical significance.
Once individual policy reform parameters have been estimated and their significance determined, we control for Administrative Costs. Administrative Costs are correlated with some of the policy reform variables, and controlling for costs can be viewed as proxy for controlling for the other larger scale policy reforms since these policies require additional expenditures.
The regression model corrects for autocorrelation in the error term. This is particularly important, since it is well known that positive autocorrelation biases coefficient standard errors downward, producing significant coefficients where there should not be. In other words, if serial correlation were not taken into account, we would find statistically significant tax policy reform impact where none was present. Autocorrelation is known to affect macroeconomic time series such as GDP.
Regressions are also run using ‘robust’ standards errors. This also increases the average coefficient standard error. Combined, the robust standard errors corrected for serial correlation are the “most restrictive” on coefficient significance in the sense that both robust standard errors and standard errors corrected for serial correlation are larger than those that are not corrected for serial correlation or are restricted in some way, making coefficient significance more difficult to attain.
The dependent and main independent (GDP) variables were not transformed into logs. First, there is no theoretical justification for this: VAT collected should be a relatively straightforward percentage applied to GDP. When the relationship is non-linear, taking logs allows the data to fit linear regression equations. In this case, the data already fit a linear model, both theoretically and empirically – there is no need to “linearize” a non-linear relationship.
53 Chapter 3. Statistical Models
Finally, the model was kept simple for a few other reasons. One could model each state’s VAT rate using dummy variables and a set of dummies interacted with the GDP term. However, there are not enough observations in the data set to allow this. There are about 200 observations, and unrestricting the VAT rates by state would add about 54 additional parameters, reducing the number of degrees of freedom to less than 150. Practically speaking, coefficient significance is very sensitive to the number of degrees of freedom when the degrees of freedom drop below 200.
3.4 Data
The data are panel formatted. This dataset covers the years 1997-2004, and covers all 27 Brazilian states (Appendix C).
3.4.1 Dependent Variable The dependent variable, VAT, is the yearly amount of Value Added Taxes (VAT) collected by each state, according to the Brazilian Ministry of Finance and the States’ Tax Administration Bureaus. It is given in nominal value of Reais, the Brazilian currency.
3.4.2 Independent Variables There are several classes of reform policy measures, each of which is captured through a variety of metrics.
3.4.3 Economic Variables Since value added taxes are linearly related to the level of economic activity, nominal GDP, GDP, is included in the regression equation as a control variable. This data is collected from the Brazilian Bureau of Geography and Statistics (IBGE). GDP is also in nominal Reais.
Notice that both the dependent and independent variables are in nominal terms; the results of a simple regression equation does not change if both sides are divided by the deflator.
3.4.4 Administrative Efficiency The first measure of administrative efficiency, AdminCost, represents operational costs as a percent of revenues collected. The informal reform objective was for each state to keep administrative costs below or around 4%.
54 Chapter 3. Statistical Models
The second measure of administrative efficiency, TaxCompliance, is the percentage of registered taxpayers who file tax returns. Although one would think that the ideal level would be 100%, many taxpayers do not need to file a return given their firms’ low income. Note that the Brazilian informal sector is quite large, so this measure only captures registered firms. Note also that several other forms of compliance can be measured, for example the percent of taxes due that are actually paid.
The STAB’s level of outreach and communication was measured in DebtNotification. It measures the percent of taxpayers in debt who received a notification of their arrears.
The STAB’s ability to enforce the payment of taxes dues was captured in DebtCollection, the value of the debt collected that year as a percent of total outstanding debt. This variable measures the level of government proactiveness as a percentage of what is available for it to take action upon.
The STAB’s inability to enforce the payment of taxes is measured in DebtOutstanding. This variable is the total outstanding debt measured as a percent of yearly tax revenues. For example, in the case of Pernambuco State, the goal was to have only 25% of unpaid debt.
DebtInstallments measures tax payments in arrears received in installments as a percent of yearly tax revenues. It measures level to which the STAB negotiates outstanding debt.
The level of outstanding debt under dispute, DebtDisputed, is the amount of VAT under dispute in the administrative agency as a percent of yearly revenue collected. A lower value indicates STAB enforcement and tax assessment accuracy.
After the tax assessment is disputed, it is up to a judge to render a final ruling. The amount of taxes to be paid under the judicial ruling, as a percent of yearly taxes collected, is captured in DebtJudged. In other words, these are taxes assessed, appealed and justified as appropriate as a percent of total tax revenue.
3.4.5 Computerization The first measure of computerization, OnlineBudgeting, is the percent of state expenditures submitted to the central state authority electronically.
The second measure of computerization, OnlineFiling, measures the percent of firms that file their tax returns electronically. It is calculated by dividing the number of taxpayers who file VAT returns online by the number of active taxpayers registered on December 31 of the preceding year. Given that the
55 Chapter 3. Statistical Models
objective of the reforms was a 100% electronic filing rate, this variable is dichotomized. The variable is equal to 1 for those states who achieved a 95% electronic filing rate otherwise 0 is assigned.
The third measure of computerization, OnlineConnectivity, uses the percentage of other state, municipal, or federal agencies, such as Federal Agencies and the Ministry of Finance, or even private agencies, such as the Firms’ Registries, connected in some on-line fashion to the STAB. These entities contribute to, work with, or share information with the STAB. Ideally, all STAB partners would be electronically connected, given that all STABs are fully computerized. However, several STABs were only recently computerized and did not have their system online running at full steam. In many instances, tax collection and budgeting remains a function carried out manually by public tax accountants. Because a large number of STABs were computerized at 100%, creating a highly skewed independent predictor, this variable was dichotomized. The dummy was set to 1 when at least 95% STAB partners connected to it electronically in some way, 0 otherwise. It is noteworthy that in some states partner connectivity reaches 100% at the end of the sampling period. This predictor can also be evaluated as a measure of administration efficiency. However, this indicator does not show the internal use of computers by the tax collectors in their individual operations and auditing.
The next measure of computerization, OnlineValidation, measures the percentage of tax returns that are immediately validated online, instead of by manual auditing. This variable captures the efficiency and expeditiousness of the electronic tax filing system. Because this independent variable was highly skewed, it was dichotomized. It took on the value of 1 when the validation rate exceeded 95%, and assumed a 0 value otherwise.
3.4.6 Training Training and professionalization of the STAB’s employees, Training, is the number of employees registered in different training courses as a percent of the STAB workforce. This value may exceed 100%, since employees may be enrolled in more than one class at any given time or over the course of a year.
3.5 Correlations
The data on Table 4 provides summary statistics of the policy reform measures. Outliers were not eliminated because these data reflects the circumstances of the Brazilian states that had been created
56 Chapter 3. Statistical Models
in 1980s, just before the modernization project. These states had not yet developed a well-structured tax collection system, and outliers can be viewed as “start-up costs.”
Table 4. Summary Statistics. Variable Obs Mean Std. Dev. Min Max
Measures of Administrative Efficiency: AdminCost 191 3.52 2.59 0.00 20.0 TaxCompliance 197 67.40 25.30 4.25 100.00 DebtNotifications 161 1.11 2.53 0.00 22.60 DebtCollected 163 0.14 0.25 0.00 1.73 DebtInstallments 180 0.07 0.09 0.00 0.75 DebtOutstanding 177 0.57 0.72 0.00 4.42 DebtDisputed 171 0.25 0.30 0.00 1.75 DebtJudged 165 0.59 0.60 0.00 3.62 LargeTaxpayers 198 0.08 0.15 0.00 1.00
Measures of Computerization: OnlineBudgeting 187 0.86 0.35 0.00 1.00 OnlineFiling 197 0.54 0.50 0.00 1.00 OnlineValidation 199 0.48 0.50 0.00 1.00 AdminConnectivity 207 0.71 0.46 0.00 1.00
Measures of Human Capital: Training 203 1.00 0.97 0.00 5.90
Note: Data from IDB, author’s own calculations.
The data on Table 5 refer to the correlation among the policy variables. It displays lows levels of correlation, with exception of a few clusters. DebtJudged, DebtOutstanding, DebtDisputed, DebtInstallments, and DebtCollected are generally positive correlated among themselves. Correlation between these variables is to be expected. DebtOutstanding positive correlation with DebtDisputed and DebtJudged stands from the fact that the more debt from taxpayers is identifies, the more likely it is that taxpayers will dispute it in court and more higher the value to be eventually collected by the STAB. The positive and significance in the correlation between DebtOutstanding and OnlineBudgeting is somewhat surprisingly. It probably stems from the fact that when STABs become more connected online with other agencies, even within the government, the more able to identify stop-filers and tax evaders. In this sense, OnlineBudgeting and OnlineValidation works as a proxy for the STAB improved capacity to reach-out taxpayers, their assets, and accounts. In the same vein, it goes almost without saying that the positive
57 Chapter 3. Statistical Models
correlation between DebtOutstanding and DebtInstallments is expected because the more remaining debts and arrears there are, the more likely the taxpayer will try to negotiate it and pay in installments, therefore, the significant correlation.
With respect to the DebtCollected, it is significantly, positively correlated with OnlineFiling, reflecting that the more taxpayers file their returns online, the more the STAB is able to collect. It goes without saying that computerization shows not only the STAB capacity to identify, validate, and collect due taxes faster, but also to actively identify default taxpayers. Inverse, but logic, is the significant correlation between DebtCollected and DebtOutstanding, indicating the more the STAB is able to collect, the less debt it accumulates.
DebtNotification is correlated with OnlineFiling given that the more the STAB is capable to collect on-line and the more efficient the system is, the faster TCs will notify the taxpayers of irregularities and mistakes. In the same line, DebtJudged is positively and significantly correlated with all variables indicating computerization. It is relevant to say that computerization of a STAB facilities actually correspond to capacity building in terms of being able to expedite information and finalize judgments about tax frauds, misunderstandings, and mistakes, either by the administration or by the taxpayers.
Related to TaxCompliance, it is positively and significantly correlated with Training, signaling where tax collectors are better educated, there is higher tax compliance. With respect to DebtNotification, it shows that it pays to have an aggressive attitude in terms of letting taxpayers know that there is a mistake in their return. The more they are let know, the higher the compliance.
Surprisingly, with respect to AdminCost, there is only significance in correlation with DebtDisputed and OnlineFiling. In the former case, the administrative cost of keeping a Tax Court System and of dealing lawyers imposes an extra burden in the STAB. In the latter case, it indicates that the computerization of STABs actually lowers the total administrative expenditure once in place, as predicted by reformers.
58
Table 5. Correlation Table.
AdminCost TaxCompliance DebtNotificationsDebtCollected DebtInstallments DebtOutstanding DebtDisputed AdminCost 1 TaxCompliance -0.044 1 DebtNotifications -0.136 0.220* 1 DebtCollected -0.069 0.058 0.157 1 DebtInstallments -0.138 -0.057 0.031 -0.072 1 DebtOutstanding -0.053 -0.138 -0.051 -0.292* 0.353* 1 DebtDisputed 0.183* 0.112 -0.115 -0.157 -0.127 0.196* 1 DebtJudged -0.091 0.013 0.067 0.042 0.344* 0.600* -0.090 LargeTaxpayers 0.099 -0.050 -0.102 -0.018 0.015 -0.096 -0.045 OnlineBudgeting -0.013 0.025 0.113 0.042 0.128 0.171* 0.059 OnlineFiling -0.15* -0.018 0.178* 0.161* -0.067 0.149 -0.122 OnlineValidation -0.004 0.129 0.145 -0.001 0.134 0.048 0.115 AdminConnectivity -0.012 -0.061 0.065 -0.020 0.208* 0.090 -0.100 Training 0.056 0.255* 0.030 -0.024 -0.076 -0.080 0.076
DebtJudgedLargeTaxpayers OnlineBudgeting OnlineFiling OnlineValidationAdminConnectivity Training DebtJudged 1 LargeTaxpayers -0.157 1 OnlineBudgeting 0.270* -0.088 1 OnlineFiling 0.299* -0.201* 0.141 1 OnlineValidation 0.294* -0.179* 0.184* 0.410* 1 AdminConnectivity 0.175* -0.155* 0.238* 0.201* 0.309* 1 Training 0.050 -0.019 0.058 0.049 0.046 -0.013 1
Chapter 3. Statistical Model
With respect to the computerization cluster, it makes sense that all variables are positively and significantly correlated. The computerization package is usually the easier one to be launch and the first to be implemented in the reform process—despite the relative difficult for writing new software programs (as discussed in chapter 5). The higher the number of agencies connected online with the STAB, the easier the budget online, online validation is a step forward from online filing. Connectivity also shows how prepared the STAB is to computerize other activities, such as filing and validation.
Finally, LargeTaxpayers is negatively correlated with the computerization variables in a small degree (4%, 3.2%, and 2.4%). It suggests that once there is increased visibility, a sharper fiscal imbalance is more apparent in the states. It is expected because computerization supposedly unveils smaller taxpayers that were defaulting and fall below the threshold of large taxpayers responsible for 50% of the tax revenue, while the number of large taxpayers is well known even before computerization.
3.6 Results
Table 6 presents the regressions results for the base model and iterative substitution of the policy reform variables.
Simple regressions results closely match expectations, as shown in column base model in Table 6. 2 GDP explains approximately 92% of the variation of VAT collection, as the R is 0.92. The value of β1 is
0.079, or 7.9%, which is interpreted at the average effectual VAT rate. The standard error of β1 is quite small at 0.0047, or 6% of the coefficient’s value.
Graph 4 shows the best-fitted line for the growth of VAT by GDP using a simple regression model.
60 Chapter 3. Statistical Model
Graph 4. Best Fitted Line (VAT by GDP). Simple Regression Model. 50,000 40,000 30,000 20,000 10,000 0
0 200,000 400,000 600,000 GDP (1000's nominal Reais)
VAT (1000's nominal Reais) Fitted values
Eight of the fourteen policy variables are statistic significant in the regression equations. AdminCost, DebtOutstading, DebtDisputed, DebtJudged, OnlineBudgeting, OnlineValidation, AdminConnectivity, and Training, are all significant at the 5% level. TaxCompliance, DebtNotification, DebtNotification, Debt Collected, DebtInstallments, Large Taxpayers, and OnlineFiling are not statistically significant within this relatively simple regression framework.
Table 6 shows the regression results of the core model and 14 successive models, each one using a Reform predictor.
61
Table 6. Results of the Base Model Estimation and Reform Indicators. Model Base Model AdminCost TaxCompliance DebtNotification DebtCollected DebtInstallments DebtOutstanding DebtDisputed
Constant -283,572 -280,170 -106,001 103,002 -129,090 -105,483 -224,917* -176,719** (189,845) (173,493) (142,198) (103,912) (121,976) (133,261) (135,510) (88,671)
β - GDP 0.0793*** 0.0761*** 0.0711*** 0.0687*** 0.0742*** 0.0745*** 0.0789*** 0.0722*** (0.0047) (0.0037) (0.0042) (0.0031) (0.0034) (0.0034) (0.0038) (0.0025)
β -Reform -38,200** -1,557 -59,751 -144,678 343,167 88,118* -164,088 (15,608) (1,193) (60,653) (280,605) (485,564) (46,645) (141,551)
β -Reform*GDP 0.0025** 0.000084 0.000752 0.012 -0.0019 -0.0037** 0.024*** (0.0010) (0.000054) (0.000995) (0.011) (0.0206) (0.0019) (0.008)
N 216 191 197 161 163 180 177 171 R2 0.922 0.929 0.952 0.954 0.970 0.945 0.955 0.975
Notes: * : Significantly different from zero at the 10% level **: Significantly different from zero at the 5% level ***: Significantly different from zero at the 1% level 1: Coefficient standard errors are shown underneath in parentheses. 2: Coefficient standard errors were corrected to account for first-order serial correlation, and are robust to arbitrary forms of heteroscedasticity.
Table 6 (cont). Results of the Base Model Estimation and Reform Indicators. Model DebtJudged LargeTaxpayers OnlineBudgeting OnlineFiling OnlineValidation† AdminConnectivity Training
Constant 4,966 -292,548* -213,471 -302,426* 35,506 -230,344 -33,898 (89,739) (177,268) (176,418) (167,697) (52,308) (187,710) (76,141)
β - GDP 0.0680*** 0.0794*** 0.0755*** 0.0783*** 0.0665*** 0.0764*** 0.0715*** (0.0040) (0.0048) (0.0045) (0.0047) (0.0028) (0.0049) (0.0019)
Β - Reform -146,713** 303,676 -62,976 30,196 -220,409** -60,623 -62,606** (69,773) (283,438) (51,572) (54,840) (95,785) (45,792) (25,833)
Β - Reform*GDP 0.0073*** -0.021 0.0038*** 0.00096 0.0118*** 0.0034*** 0.0017*** (0.0028) (0.022) (0.0011) (0.00217) (0.0034) (0.0008) (0.0004)
N 165 198 187 197 199 207 203 R2 0.987 0.915 0.929 0.922 0.987 0.913 0.982
Notes: * : Significantly different from zero at the 10% level **: Significantly different from zero at the 5% level ***: Significantly different from zero at the 1% level 1: Coefficient standard errors are shown underneath in parentheses. 2: Coefficient standard errors were corrected to account for first-order serial correlation, and are robust to arbitrary forms of heteroscedasticity. †: OnlineValidation: Because the estimated serial correlation coefficient was greater than 1, the Cochrane-Orcutt procedure produces an R-squared that is incorrect. This model was run without correcting for serial correlation, but the standard errors are robust to arbitrary forms of heteroscedasticity. Source: Regressions done by the author.
Chapter 3. Statistical Model
For each point-percentage increase in administrative cost, AdminCost, there is 0.25 percentage- point increase in the effective VAT. AdminCost is significantly different from 0 at the 5% level. This indicates that the more the STAB spends on tax administration, the more it can collect, what is quite logical. Practically speaking, a 0.25 percentage-point leads to a 3.3% increase in tax collection efficiency. Most states have administrative costs representing in between 2% and 4% of VAT collected and there is now a stated guideline of 4%, so total collection efficiency gains are limited.
DebtOutstanding is negatively related to VAT collections and this relationship is statistically significant at the 5% level. The result makes sense because the more registered taxpayers have outstanding debt; the less VAT can be collected.
DebtDisputed is positively and significantly related to VAT collection at the 1% level. It means that the more efficient the STAB is in identifying taxpayers and trying to collect from them, the more taxpayers will appeal. This is a proxy for the STAB efficiency in the sense that it shows the aggressive attitude of the state going after default taxpayers and evaders. A one-percentage point increase in DebtDisputed increases the effective VAT by 2.4 percentage points. This seems like a large effect. However, the mean value of DebtDisputed is 0.25%, with a standard deviation of 0.3%, which means a state would never fully effectuate a one percentage-point increase in this variable.
By the same token, the relationship between DebtJudged and VAT collection is positive and significant at the 1% level. A one-percentage point increase in the amount of DebtJudged leads to a 0.73 percentage point increase in VAT collected. Presumably, not all debts judged are immediately collected and not all taxpayers pay the full amount or pay on time, or in one installment, which explains why this coefficient is less than 1.
With respect to computerization, the models indicate that all predictors have a significant effect on VAT collection, except for OnlineFiling. In the case of OnlineBudgeting, the results show that the predictor has a positive impact on VAT collection, and that this impact is statistically significant at the 1% level. There is a 0.38 percentage point raise in VAT for each percentage point OnlineBudgeting. Therefore it adds 5% VAT efficiency. Considering that the mean value of OnlineBudgeting of 0.86 is quite close to 1, and that most of the states have reached full capacity in terms of computerization of online budgeting process, the gains from pushing this further in Brazil are limited.
OnlineValidation, too, has an important effect on VAT efficiency. For every on-point percentage increase in online validation, there is a 1.18 percentage-point increase in the effective VAT rate – is an 18% increase in VAT efficiency. This is significant at the 1% level. Given the great variation in the
64 Chapter 3. Statistical Model
observations, the impact in the collection will not be as important for all the states. However, for states with full capacity online validation, it is quite significant. For São Paulo State, which operates full validation on line, the revenue increment can reach up to US$ 4 billion.
AdminConnectivity, a reform predictor that is also positively and significantly correlated to tax revenue at the 1% level. There is an effect of 0.34 percentage-point increase in VAT revenue for each one-point percentage increase in investment in connectivity. This is an 4.5% increase in VAT efficiency. Accordingly, this means a potential great impact for STABs’ revenue collection, given that the trend is to have all the agencies connected on line. Graph 5 shows the upward shift in tax collection when AdminConnectivity is added. Despite small, such a shift brings a large impact on the state revenue.
Graph 5. Best Fitted Line, controlling for AdminConnectivity (VAT by GDP). Simple Regression Model. 50,000 40,000 AdminConnectivity = 1 30,000
20,000 AdminConnectivity = 0 10,000 0
0 200,000 400,000 600,000 GDP (1000's nominal Reais)
VAT (1000's nominal Reais) Fitted Values, AdminConnectivity = 1 Fitted Values, AdminConnectivity = 0
Finally, Training shows that efforts in personnel capacity pay off. For each one point-percentage increase in Training, there is a 0.17 percentage-point increase in VAT collection, or a 2.4% increase in the ratio of VAT collection to GDP. This effect is positive and significant and denotes the value of investing in better-educated professionals.
65 Chapter 3. Statistical Model
3.7 Multiple Reform Variable Estimation Models
The relatively simply framework described above was enhanced by controlling for administrative costs, AdminCost. Only those variables that were statistically significant in the simple framework are carried forward here. The results of the models that yielded an effect are presented in Table 7. Results are robust to the specification change.
AdminCost is included because it is correlated with other measures of administrative performance, most notably OnlineFiling, OnlineBudgeting, OnlineFiling OnlineValidation, AdminConnectivity, and DebtDisputed. In other words AdminCost can be viewed as controlling for a generalized set of reform parameters.
Overall, holding both GDP and AdminCost constant, it is possible to verify that DebtOutstanding still has a negative and statistically significant effect on VAT collections, and that DebtJudged, OnlineBudgeting, AdminConnectivity, and Training have all positive and statistically significant effects on VAT revenue. In most of the multiple reform variable models presented, the results are significant at the 1% level. However, DebtOutstanding and DebtDisputed’s level of significance goes down from 1% to 5%.
Quantitatively, the marginal effects for DebtOutstanding, DebtDisputed, DebtJudged, and Training are the same compared to the earlier simpler models.
However, the marginal effect for OnlineBudgeting has increased from 0.38 percentage points to 0.57 percentage points as compared to the simpler framework. This coefficient difference is statistically significant at the 5% level. This represents an 8% efficiency gain, as compared to 5% previously.
The marginal effect for AdminConnectivity has also jumped from 0.34 percentage points to 0.49 percentage points. This represents a 6.9% efficiency gain, up from 4.5%.
AdminCost is not statistically significant in the regression equations for DebtDisputed, DebtJudged, or Training. AdminCost is positively and significant related to VAT collections for regressions containing DebtOutstanding, OnlineBudgeting, and AdminConnectivity. The magnitude of AdminCost’s marginal effect is slightly larger in this framework than in the simpler framework – around 0.30 percentage points compared to 0.25 percentage points.
66 Chapter 3. Statistical Model
Table 7. Multiple Variables Models.
DebtOutstanding DebtDisputed DebtJudged OnlineBudgeting AdminConnectivity Training
Constant -212,148 -176,242* 17,381 -116,445 -174,686 -26,794 (146,013) (101,986) (91,795) (172,584) (161,512) (86,595)
GDP 0.0751*** 0.0708*** 0.0673*** 0.0693*** 0.0713*** 0.0726*** (0.0032) (0.0026) (0.0038) (0.0031) (0.0034) (0.0017)
AdminCost -40,229** 940 1,927 -47,014*** -40,516*** 16,479 (16,198) (19,110) (13,280) (15,029) (13,510) (16,720)
GDP * AdminCost 0.0026*** 0.0008 0.0003 0.0032*** 0.0030*** -0.0010 (0.0010) (0.0012) (0.0010) (0.0007) (0.0008) (0.0011)
DebtOutstanding 102,789** (46,197)
GDP * DebtOutstanding -0.0038** (0.0016)
DebtDisputed -102,998 (220,470)
GDP * DebtDisputed 0.0220** (0.0108)
DebtJudged -137,695** (67,921)
GDP * DebtJudged 0.0074*** (0.0027)
OnlineBudgeting -143,546** (71,123)
GDP * OnlineBudgeting 0.0057*** (0.0010)
AdminConnectivity -118,261** (54,679)
GDP * AdminConnectivity 0.0049*** (0.0010)
Training -73,761** (29,326)
GDP * Training 0.0021*** (0.0005)
N 161 160 155 170 187 182 r2 0.957 0.983 0.990 0.949 0.936 0.983
Source : Author’s Elaboration
3.8 Other Alternative Estimation Models
3.8.1 States Grouped by Economic Development The base model was unrestricted to account for each state’s relative level of economic development. States were divided into three groups of economic development. The regression was
67 Chapter 3. Statistical Model
modified to include two dummies – for the least and second least developed states (δ1 and δ2) – the interaction of these dummies with GDP, and the interaction of these dummies with both GDP and the reform policy measure. The objective was to see whether the policy reform marginal effects were higher in less developed states. Most of the coefficients on the GDP-Reform-state development dummy were insignificant, so we cannot reject the null hypothesis that policy reform affects all states equally, regardless of their level of economic development. Results are not presented here since none were of economic or statistical significance. (Appendix D).
3.8.2 Time Lagged Model The model was also unrestricted to predict whether fiscal year constraints would affect the performance of predictors. The regression was then modified to include t (i-1) and t (i-2), representing a lag of results lasting one or two years from the implementation of reform guideline. Predictors referring to computerization were not lagged, given that once in place, immediate results were seen (Appendix E).
Table 8. Lag Time for Predictors. Measures of Computerization: OnlineBudgeting Right away OnlineFiling Right away OnlineValidation Right away AdminConnectivity Right away
Measures of Administrative Efficiency: AdminCost 1 year TaxCompliance Not lagged DebtNotification Two year at least DebtCollection 1 year DebtInstallments 1-2 years DebtOutstanding 1 year DebtDisputed 3 years DebtJudged 1-2 years
Measures of Human Capital: Training Immediate Source: Author’s elaboration.
68 Chapter 3. Statistical Model
3.9 Conclusion
The evidence in this chapter supports that qualitative factors are important determinants to tax growth in the Brazilian States. The positive relationship between VAT growth and computerization is documented by using three indicators, AdminConnectivity, OnlineBudgeting (and OnlineValidation). The relationship is strong for AdminConnectivity and OnlineBudgeting, which is consistent with the expectation that computerization improves tax collection with better flow of information between the STABs and other government and non-government agencies. Although budgeting indicates efficiency of a financial function of the STAB, it could also be considered a proxy for computerization. The results show that a highly computerized administration leads to a sensible increase in tax collection.
Using DebtJudged, DebtDisputed, and DebtOutstanding as measures for the renewed and more efficient procedures in the tax administration, the positive relationship between them and tax collections signal the importance of taxation institutions.
Also quite important, is the significance level Reform variables denoting a restructuring action towards old debts. DebtJudged and DebtDisputed not only had a positive and significant impact on VAT efficiency, but also raised the regression R-square to 99% and 98%, respectively. It points to the fact that the STABs need to work out their policy to collect outstanding debts and to make their bureaucracy responsive to judge and collect them. In a few states, as in Rio Grande do Norte, the STAB has concentrated its efforts in collecting past debts. No pending file was left unturned and all the debt was been collected. Therefore, it is a reform tool to be further explored.
The implication of having a positive and significant Training variable points out the importance of investment in human resources.
Finally, the consistently positive and significant AdminCost variable shows that reforms to curb administrative costs are also important. Indeed, in Brazil, the explicit policy target is that administrative costs should be less than 4% of taxes collected. Given this, it may be appropriate to examine policy measures within a cost-benefit framework, or examine how it may be appropriate to relax this budget constraint for large information technology capital investments.
This study does not exhaust the elements that might affect VAT increase. For example, it may be worthwhile to investigate whether tax collectors wage raises would affect tax efficiency.
69 Chapter 3. Statistical Model
Clearly, econometric analysis shows how restructuring government administration can help solve some of the most pressing public finance issues of developing economies: how to raise more revenues, how to address the debt crisis, and how to improve state government solvency.
3.10 IDB Indicators
Below is a detailed description of the variables used to measure change in the Tax Bureau after the administrative reform, elaborated by the IDB (Febres, Cruz et al. 1998; IDB 1996; 1999).
1. STAB Administrative Cost (AdminCost) STAB expenditure STAB revenue STAB total administrative present cost in a given year (divided by) the STAB total revenue received by the Bureau during the same year in Brazilian currency (R$ = Reais). The goal is to be 0. 04 (4%). 2. STAB on-line connectiviness with other agencies) # Agencies online (AdminConnectivity Total # agencies Number of state agencies connected with the STAB for budgeting divided by the total number of state agencies involved in the state budgeting (including financial funds management). Goal: 1 (100%). 3. Taxpayers Compliance (TaxCompliance) Total VAT Returns Total Tax-payers registered Total number of firms that filed VAT tax returns divided by total number of taxpayers registered at the STAB. Goal; 1 (90%) 4. Value budgeted on-line (OnlineBudget) Online State Budgeting Total State Budgeting Value budgeted on-line through the integrated system divided by the total value budgeted by the entire state, in Reais during the fiscal year. Goal: 1 (100%). 5. On-line tax returns (OnlineFiling) # Online VAT returns Total tax returns Number of firms’ VAT returns filed online divided by the total number of VAT returns in each year per state. Goal: 1 (100%). 6. Electronic validation (OnlineValidation) Total revenue validated online Total revenue Collected value in Reais validated on-line by the bank or collecting agency and immediately sent to the state treasury agency divided by total revenue tax revenue, including legal additional values (e. g. , fines). Goal: 1 (100%). 7. # Administrative Lawsuits (DebtNotification) # Filed actions Total # Debts Total number of individual administrative notifications filed to recovered debt divided by total number of total notifications of debt until December 31. Goal: 3 (300%). 8. Revenue from previous debts (DebtCollection) Revenue from debts Total debt Value of debts paid in the current fiscal year divided by the total amount of debts from previous years added new debts incurred during the current fiscal year. Goal: 0. 2 (20%). 9. Total Debt Revenue Unpaid (DebtOutstanding) Outstanding Debt
70 Chapter 3. Statistical Model
STAB Revenue Total amount of debts from previous years added new debts incurred during the current fiscal year divided by total STAB revenue. Goal 0. 25 (25%) 10. Revenue negotiated (DebtInstallments) Total installments Total revenue Total amount in Reais to be paid in installments (except those under dispute) divided by total STAB revenue. Goal: 0. 07 (0. 7%). 11. Revenue under dispute (DebtDisputed) Value under dispute Total revenue Total amount in Reais under dispute in 1st and 2nd Appeal Court divided by total revenue. Goal: 0. 2 (20%) 12. Revenue to be received judicially (DebtJudged) Value to be received Total Revenue Total amount to be received through judicial collection divided by the total STAB revenue in Reais. Goal: 1. 4 (140%). 13. Training and Professionalization (Training) Trained Tax Collectors Total Tax Collectors Number of public officials who taken part in training courses divided by total number of public officials in the STAB. Goal 2 (200%). 14. Tax concentration by large taxpayers (LargeTaxpayer) Tax payers paying more than 50% Total # taxpayers registered
Total number of taxpayers who are responsible for 50% of tax collection in each state until December 31st of each given year divided by total number of tax collectors registered by the STAB. Goal: 0. 0006 (0. 06%)
71
Chapter 4
Vertical Economic Assessment of the Tax Basis: The Central Strategy for Raising VAT Revenues
4.1 Introduction
Brazilian states tax revenues increased by more than 31% between 1995 and 2005, largely driven by VAT (Graph 6, below); VAT (ICMS) revenue alone grew by approximately 20%. This rise in revenues is particularly remarkable during the period after the administrative reform began in 1997, especially considering the preceding economic and fiscal reforms (Ramos 2001). 24 Even though Brazilian sub-national states are responsible for collecting motor vehicle, inheritance, and other taxes, state revenue relies primarily on the collection of VAT. The VAT accounts for up to 91% of state revenues and respond to approximately 9% of Brazilian taxes’ share of GDP.25 Following guidelines designed by the Brazilian Ministry of finance, international donors, such as the IDB, and recommendations of the Inter-American Center for Tax Administration (CIAT) and tax economists, STAB modernization programs included changes in organizational, technological, and institutional aspects of the tax administration (Bird 1991; Bird and Casanegra de Janstcher 1992; CIAT 1996; IDB 1996).
24 The previous period of Brazilian political economy was ridden by economic plans that skewed any analysis of tax collection behavior. Since multiple economic plans were enacted and devaluation occurred, inflation skyrocketed, and as lack of data exists, an accurate analysis of tax collection performance before 1995 is not possible (Ramos 2001). 25 For example, the VAT corresponds to 88% of total state tax revenue in São Paulo, 90% in BA, 88% in CE, 85% in PB, 87% in PE, and 85% in RN.
73 Chapter 4. Vertical Assessment
Graph 6. VAT and Total State Revenue Growth from 1995 to 2005.
200000000
180000000
160000000
140000000
120000000 Total State Tax Revenue 100000000 State VAT Revenue 80000000
60000000
40000000
20000000
0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
Source: IBGE. Several factors contributed to improvements in tax collection, including organizational restructuring, computerization, and the increasing professionalization of tax collectors. In the previous chapter, I analyzed a few components of the organizational aspect of tax collection, using several indicators collected by the STABs. These indicators related exclusively to the administrative aspects, such as administrative costs, tax compliance, debt collection, and percentage of large taxpayers responsible for 50% of the tax revenue. The independent variables that presented a significant effect on VAT collection were administrative cost, disputed debt, judged debt, and outstanding debt.
Although the statistical models suggests which factors are more connected to administrative efficiency, do not sufficiently aspects that are intrinsically built-in aspects to the administrative changes described by tax collectors in Brazil. As Fainstein notes:
“A correlational analysis could show that certain input variables—e.g. presence of a paid organizer, level of property investment—are connected to certain output variables— e.g. expenditure on low-cost housing, growth in the urban economy—but it could tell us little about why these outputs occurred. The typical quantitative paper that links policies
74 Chapter 4. Vertical Assessment
to outcomes usually has a section at the end speculating on reasons for the relationship, but unless it is accompanied by case studies, it can never support these speculations with much evidence. Even correlations over time lack a genuine historical perspective precisely because they are snapshots of static moments rather than the tracing of processes. Thus, while there are descriptive aspects of comparison that require quantitative analysis—e.g. levels of inequality and their changes over time, rates of growth and unemployment, extent of segregation, etc.—the numbers can never explain their own existence. Modeling is one approach to overcoming this problem—for example, Fisher and Peters use econometrics to determine whether or not a rational firm will choose a certain location as a consequence of governmental incentives. This effort tells us a great deal about the rationality of offering such incentives, but it cannot tell us what are the actual decisional processes of the firm.” (Fainstein 2007)
Therefore this chapter elaborates on one intrinsic organizational aspects of the administrative reform that contributed to improvement in tax administration in a broader sense—beyond tax revenue— but that is not captured by statistical analysis: the vertical industry segmentation of tax basis.26 I also detail how this change in tax collection is eventually linked to the indicators shown in my estimation results.
4.2 Chapter Structure
I begin this chapter by reviewing the literature on public administration management, since this research speaks to, and is informed by, public bureaucracies managers, policy makers, and public officials.
Then I show how the literature on tax administration discusses efficiency and institutional strengthening, particularly with respect to tax administration reform and VAT efficiency. I start by examining the purpose, functioning, and evolution of tax administration reform efforts, emphasizing the best practices and current trends. I elaborate on the conditions that have been identified as relevant for determining the success and effectiveness of such reforms, stressing the virtues and shortcomings of the theoretical frameworks usually used in these analyses. I present the Brazilian state tax collection
26 For this dissertation I use vertical industry segmentation, economic segmentation, and industry specialization interchangeably.
75 Chapter 4. Vertical Assessment administrative structure, especially as it relates to VAT. Then I investigate which tax administration reform elements are behind the Brazilian STABs’ success.
Finally, I discuss the main findings of my field research based primarily on the comparison of restructuring and tax administration reform efforts in six Brazilian STABs: São Paulo (SP), Rio Grande do Norte (RN), Paraíba (PB), Pernambuco (PE), Ceará (CE), and Bahia (BA). These findings provide the basis for presenting an informed tax collection model based on a novel process adopted in some STABs: vertical economic sector assessment. This assessment methodology is organized according to a supply- chain theory and is based on a knowledge of backward and forward industrial linkages in the local economy.27 Finally, I detail the theoretical implications derived from these findings and offer policy recommendations for the restructuring of tax administration bureaus and public sector bureaucracies to achieve successful tax collection reforms.
4.3 Conversations Between the New Public Administration and the Administration Reform Literature: Connections and Implications
The guidelines for the tax administration reforms carried out worldwide stem from the confluence of two bodies of literature: the tax administration literature per se (Bird 1991; Bird 1991; Bird and Casanegra de Janstcher 1992; Silvani 1993; Silvani and Baer 1997; Ueng and Yang 2000), standing on its own right, and, more loosely, the new management techniques within the public sector (Ackroyd, Kirkpatrick et al. 2007; Kelman 1981; 1987; 1990; Kettl 1992; 1996; Kettl and DiIulio 1995; Osborne and Gaebler 1993). Despite streaming from diverse origins, the public finance and public administration literature, both sets of literature complement each other. However, the literature regarding tax administration reform has established itself as its own domain.
The background for this dissertation lies in the amalgamation of these two sets of literature. In the next two sub-sections, I discuss the current understanding of successful about public sector and tax administration reforms. I analyze the key principles and views from both areas of study and examine the case of the PNAFE (State Tax Administration Bureaus Modernization Program) in Brazil, a well- structured tax administration reform program. I illustrate how these theoretical tenets are insufficient to understand, replicate, or determine successful reforms. Yet as the cases of the Brazilian STABs will show, unexpected elements of change emerge to explain success.
27 The contributions of supply-chain and value chain theory as well as backward and forward linkages will be further addressed later on in this chapter.
76 Chapter 4. Vertical Assessment
Another important body of literature for this study is that on best practices in developing countries (Grindle 1997; 2000; Tendler 1993; 1997; Tendler and Freedheim 1994; 1994). This literature describes “best practice” policies and strategies in developing countries that have been carried around the world. Seminal works on these practices analyze experiences in good governance that have led to accountability, credibility, transparency, and the strengthening of public sector bureaucracies.
4.3.1 New Public Management (NPM) Theory The literature examining state administrative reform is intrinsically related to public administration theory. Orthodox public administration theory has traditionally focused on state affairs and on the internal division of labor (and power) explain government processes and performance (White 1930). A more contemporary view of public administration embraces a scientific study of bureaucracy and the role of governments in promoting the public interest by adopting best practices. This includes the elimination of waste, standardization of procedures, the improvement of techniques and the adjustment of the public workers to their jobs (Ackroyd, Kirkpatrick et al. 2007; Bendix 1968; Blundell and Murdock 1997; Marini 1968; Waldo 1968).
Following this trend, the paradigm of the New Public Management (NPM), which focused on these same elements, but with a different twist, came to dominate the public sector scenario in the 1980s and the 1990s. NPM suggests that external forces, such as the market, should limit public administration processes, since the market can redefine the role and size of government. NPM states that public bureaucracy must be efficient, accountable, and resonate with private sector managerial prescriptions. In this sense, efficiency and accountability should be measured according to market standards. Anglo- American countries such as New Zealand, Australia, the United Kingdom, and the United States provided the model, which became hegemonic in public administration and set the guidelines for government performance. Accordingly, governments better serve societies when they step back and steer rather than row the economy, public affairs, and provision of services (Osborne and Gaebler 1993).
NPM launched a new gospel to public sector management. It combines strategies and tools to achieve high degrees of efficiency, effectiveness, and responsiveness in the governing process (Minogue, Polidano et al. 1998; Osborne and Gaebler 1993; Osborne and Plastrik 1991).28 Its precepts are hands-on
28 Osborne and Gaebler’s precepts are: steering rather than rowing, empowering rather than serving, injecting competition into service delivery, transformation to rule driven organizations, funding outcomes rather than inputs, meeting customer and not bureaucratic needs, earning rather than spending, prevention rather than cure, participation rather than hierarchy, and obtaining change via the market (Osborne 1993; Blundell 1997).
77 Chapter 4. Vertical Assessment professional management, explicit standards and measures of performance, output control, decentralization, competition, private management practice, and parsimony and discipline in resource use (IDB 1996). When translated into policy, these guidelines normally result in the following practical instruments: downsizing, decentralization, privatization, deregulation, civil service system reform, devolving of managing responsibilities, shifting from input control to output and outcome measures, tighter performance specification, greater use of contracting out government services, and an emphasis on client-driven approaches (Dunford, Bramble et al. 1998; Lam 1997).
“Although these conditions were not fully applied in the public sector, NPM offered a strong parallel to them: the need for close tracking of performance, the importance of feedback from clients, the expectation of a sustained effort at improvement, and a commitment to worker participation.” (Montgomery 1981)
The reforms that took place in the United States during the Clinton-Gore years derived their tenets from NPM and from business practices, inseminated with total quality management canons. The literature on best practices in the private sector considers standard output production as the ultimate outcome to measure performance and the customer as the judge of managerial performance. As for worker participation, there is a strong theoretical emphasis on workers’ commitment and involvement, combined with some antipathy toward corporative and associative practices among workers, such as unionization (Johnson, Saes et al. 1996; Johnston 1994; Nunberg 1990; 1992; Nunberg and Nellis 1990).29
For many years now, the focus of reform studies has been on national states. National states have been struggling in the public arena to implement administrative reforms, among them tax administration reforms, in order to make their governments more efficient and less costly, putting into effect principles based on accountability, legitimacy, and transparency. Those reforms were usually based on a paradigm of market-based rules and “reinventing government” principles, such as a move towards efficiency, effectiveness, and customer-driven relationship.
Therefore, from the perspective of the well-intentioned public managers, a new generation of practitioners, planners, and policy makers under the spell of the NPM School, reform in the public sector must follow the precepts of the private sector, combined with accountability, transparency, and fairness.
29 Tendler (1997) comments on public sector workers and the perceptions of their unions.
78 Chapter 4. Vertical Assessment
While this theory advances important premises for the national government reform, it does not capture the specificities of reforms at the local and state levels, such as alternative paths to reform, uneven state capacities, and innovative and unexpected institutional arrangements at the local level that may in fact catalyze and/or implement efforts to change local government (Claver, 1999; Fong, 1999; Andrews, 1999; Grindle, 1997; Laurence E. Lynn, 1996; Pallot, 1996; Reschenthaler, 2002; Saint-Martin, 2001; Wollmann, 2001; Hood 1991). Traditional NPM does not elaborate on alternative policy strategies, sequencing, or trajectories. Nor does it capture how the peculiarities of each sector have an impact on the paths chosen and on their results, hindering or advancing reform efforts. In this sense, this study presents a novel perspective on strategies that can lead to successful reform in the tax sector.
4.3.2 Tax Administration Reform The literature on tax administration reform resembles the theories on new public management, it is centered on bureaucratic transparency, result orientation, and customer/user satisfaction, applied to limitations and specificities of tax administration, (i.e. clear rules and efforts to reduce corruption, increasing revenues, and increasing voluntary compliance). In practice, tax administration reforms aim primarily to increase tax collection, enforce compliance, and enhance effectiveness and efficiency (Bird 1991; Bird and Casanegra de Janstcher 1992).30 Tax experts and consultants, in addition to politicians, try to create a well-distributed set of taxes that target all income brackets, creating a redistributive effect in society, and allowing for some degree of equality (Bird 2003; Crandall and Bodin 2005; Varsano 1999). They also aim to simplify tax collection procedures in order to increase efficiency, and, most importantly, to increase revenues. However, neither set of literatures has been able to provide a complete recipe for reform, be it in tax administration bureaus or in any other area, to work unequivocally (Varsano 1999).
The challenge to reach efficiency and effectiveness in tax administration has been long sought. Donors and international organizations such as the International Monetary Fund and the IDB advocate recommendations such as harnessing appropriate technology, promoting compliance, penalize non- compliance, educate the taxpayer and promote preventive policies (Silvani 1993). An extensive literature on tax administration on the best practices in tax administration exists and highlights common-sense
30 Crandall and Bodin (2005) lists as goals of such reforms: improved revenue performance, more equitable distribution of the tax burden across the community, more consistency and fairness for business and individuals, greater ability to implement fiscal reforms, reduced compliance costs for taxpayers, increased number of registered taxpayers, reduction in tax evasion and fraud, improved management of tax arrears, improved services to taxpayers, and greater transparency and integrity in the administration’s operations.
79 Chapter 4. Vertical Assessment aspects of tax management such as tax simplification and taxpayer education (Bird 1991; Bird, Bagchi et al. 1995; Das-Gupta 2001; Silvani 1993; Silvani and Baer 1997; UN 1997).
However, as the economy grows more complex, as markets expand across the globe, as the service sector continues to expand, and as people and goods cross borders for better opportunities, it becomes harder and harder to achieve a simplified, reliable, and efficient tax system. To create a system that would be able to cope with tax evasion, non-compliance, and non-payment—even deal with public bureaucratic inefficiency—is a dream goal for bureaucrats and policy makers. In addition, fiscal reforms usually demand long-term mandates because policy makers depend on public discussion and approval by legislators who do not always share the same views on how to increase state revenue. Political stability and government continuity are considered crucial factors for implementing long-term reforms. Policy makers thus often look to improve revenue collection and public resources by optimizing and improving of tax administration systems. Ultimately, be the least troubled way to achieve tax collection maximization appears to be to diminish the gap between potential tax revenue and the tax actually collected appears to (Ebrill, Keen et al. 2000).
Increasing the efficiency of tax collection and reforming the tax system to adapt to the world economy seems increasingly difficult as the economy grows more complex. It requires commitment and political will from politicians and engagement from bureaucrats. It also requires outreach to the taxpayers, who are usually scared of potential draconian advances by tax collectors, and a change in the public sector’s mentality by imbuing public officials—particularly tax collectors—with a sense of mission usually uncommon (although highly valued) in the public sphere.
Tax administration reforms require political, technical, and organizational changes (Bird and Casanegra de Janstcher 1992). Much emphasis has been placed on technological modernization (computerization) and on organizational restructuring, placing workforce capacity and institutional strengthening as secondary to reform. Workers—their ability, engagement, and motivation—are in fact an important element in the success of the reform equation. Workers must be engaged and well trained; otherwise, they will oppose change and hinder reform. Nonetheless government officials typically perceive workers as opponents of reform, believing that they will behave in a rent seeking fashion, unmotivated and unproductive.
Therefore, tax growth should not be the sole indicator of success: professionalization of the work force, institutionalization of custom-driven practices, and superior governance should also be indicators. The impacts of reform on public servants’ professionalization, on organizational structure, and
80 Chapter 4. Vertical Assessment
institutional arrangements and norms should be considered as important as increased tax collection. In an effort to acknowledge the importance of these distinct aspects of reforms, the IDB developed a set of indicators that includes tax efficiency, computerization, and training policies. Even in the IDB’s list, however, quantitative indicators are leading figures. These indicators do not allow for a critical measure of these strategies’ quality.31
Reforms should not merely target those quantitative aspects of the administrative machine. Yet, as a rule, indicators such as increased tax collection and state tax collection as a ratio of State GDP (Gross Domestic Product) continue to be the emblematic signs for successful tax administration reforms. And indeed, growth in tax revenue usually suggests that the administration of tax collection is running better, that taxpayers are complying, and that overall control is more effective. Of course, as a result of increasing revenue, governments are able to readdress public needs and invest in services.
As mentioned earlier, despite the numerous accounts of tax administration reforms at the national level, reports of sub-national level tax agencies reforms are scarce. This dissertation proposes to examine the impact of reform at the sub-national state level.
4.4 Tax Administration Guidelines for the Brazilian States
In the wake of the theories presented above, the IDB and the Brazilian Ministry of Finance approved a loan for up to US$500 million in 1996 to create and implement a national tax administration program for the Brazilian states. The states’ tax administration reforms strictly followed the IDB’s prescription in spite of lurking risks such as unsatisfactory execution of projects, a lack of institutional capacity in the states, states’ inability to meet technical criteria, and inadequate preparation by the states during the execution of the project (IDB 1996).
The stated goals of the program were “to increase administrative efficiency, effectiveness, rationality and transparency” in consonance with the theories above presented. And in order to achieve these goals, specific fiscal modernization projects that had the following objectives:
“improve the legal, operating, technological and management mechanism of the tax administration agencies in Brazil;
strengthen and integrate the financial administration and consolidate audit and internal control;
31 See IDB (1996) and Cruz et al. (1999) for the IDB indicators. See also Chapter 3 and Appendixes C.
81 Chapter 4. Vertical Assessment
ensure effective control of the tax payment by taxpayers through the use of new techniques and methods in the tax collection and supervision;
expedite legal actions for collection of delinquent tax debts and strengthen integration between the tax administration and the agencies responsible for judicial collection.” (IDB 1996)
The program offered technical assistance, comprising two main elements: (1) strengthening the Federal Ministry of Finance in the supervision, integration, and coordination of the national fiscal affairs, and (2) establishing a fiscal administration to finance modernization projects in tax and financial administration (budgeting, cash management, accounting, auditing, and internal control).
The executing agency was the Federal Ministry of Finance through its Executive Secretariat, which created a program coordination unit (PCU), responsible for technical assistance and coordination.32 The states’ PCUs were responsible for the actual modernization projects.
The loan agreement established detailed criteria for implementation and evaluation. With respect to the fiscal administration component, it was designed to finance institutional strengthening activities for the states’ bureaus of finance and planning. The tax administration subcomponent included the following activities:
“economic and tax relates studies; state tax legislation; single taxpayer list; tax collection; tax inspection and audit; tax collection at administrative and judicial level; and integrated tax administration system.” (IDB 1999)
The financial subcomponents included:
“studies and analyses of public spending; integrated financial administration system; budget, cash management; and governmental accounting; control and management of public spending; and institutional and regulatory framework for internal control.” (IDB 1999).
Political will seems to be the most important determinant in any successful tax administration reform. Taxpayers compliance requires ironclad political will on the part of authorities and government leaders to implement reforms as well as public consensus from political supporters (Bangura 2000). The second most important requirement is human resources: implementation out the work (Silvani 1993).
32 Interviews with and data provided by Verleun, in 2005.
82 Chapter 4. Vertical Assessment
In an analysis of reforms in several countries, Silvani points out “a feeling of urgency can catalyze effectiveness in the taxation sector.” Reforms to increase tax collection also include the simplification of tax administration systems, improvements in auditing, computerization, organization development, taxpayer education, and enforcement. Simplification of tax systems includes the simplification of tax returns, the creation of different tax brackets, the adoption of flat tax rates, the elimination of subsidiary schedules, and the streamlining of multiple taxpayers requirements (Varsano 1999).
Auditing is crucial to improving voluntary compliance and to spotting tax evasive behavior. This includes both cross–checking information with other governmental sources and detecting unreported sales or fake purchases. Auditing activities should not diminish with reform, and non-compliance penalties must be imposed. However, many tax agencies’ administrators try to reduce auditing staff in order to reduce administrative costs as part of reform efforts towards geared towards reducing government expenditures (Interviews with tax administrators in Ceará, Brazil).
Also recommendation is to differentiate taxpayers. In most countries, large firms account for between 75 and 95 percent of the tax revenues collected. It is now widespread to organize teams to audit and monitor the largest taxpayers. It has also been widely recognized that taxpayers in different tax brackets (usually identified by firm size in the case of VAT) should be treated differently with regards to tax collection and audits. However, this does not mean that the remaining taxpayers should be ignored (Febres, Cruz et al. 1998).
Indeed, in the quest for tax system equity, “it is the tax bureau that will ultimately improve or worsen, since they are directly responsible for ensuring that all who must pay taxes do so” (Febres, Cruz et al. 1998).
According to data provided by the STABs, Brazil follows this pattern of large taxpayers, who at most times represent less than 1 percent of the total number of taxpayers, account on average for more that 50 percent of total taxes collected in Table 9.
83 Chapter 4. Vertical Assessment
Table 9. Large Taxpayers. Large Taxpayers Taxpayers paying > 50% VAT Revenue Total Taxpayers 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd 0.03% 0.11% 0.09% 0.09% 0.20% 0.17% AC AL 0.15% 0.11% 0.10% 0.05% 0.06% 0.05% 0.03% 0.03% AL AM 0.04% 0.06% 0.05% 0.05% 0.04% 0.04% 0.03% 0.03% AM AP nd 0.08% nd nd 0.14% nd 0.26% 0.26% AP BA 0.02% 0.02% 0.01% 0.00% 0.00% 0.01% 0.00% 0.00% BA CE 0.05% 0.02% 0.02% 0.03% 0.02% 0.02% 0.01% 0.01% CE DF 0.04% 0.05% 0.01% nd nd 0.10% 0.65% 0.30% DF ES nd 0.02% 0.03% 0.03% 0.04% 0.04% 0.03% 0.03% ES GO 0.01% 0.03% 0.02% 0.08% 0.01% 0.02% 0.00% 0.00% GO MA 0.02% 0.02% nd nd 0.02% 0.03% 0.04% 0.03% MA MG 0.02% nd 0.00% 0.00% 0.01% 0.00% 0.00% 0.01% MG MS 0.08% 0.06% nd 0.00% 0.13% 0.67% 0.01% 0.01% MS MT 0.04% 0.02% 0.01% 0.01% 0.01% 0.02% 0.02% 0.02% MT PA 0.10% 0.04% 0.05% 0.33% 0.22% 0.63% 0.02% 0.05% PA PB 0.15% 0.15% 0.20% 0.30% 0.10% 0.10% 0.06% 0.06% PB PE 0.04% 0.06% 0.05% 0.05% 0.02% 0.02% 0.02% 0.03% PE PI 0.01% 0.06% 0.08% 0.04% 0.04% 0.02% 0.04% 0.04% PI PR 0.01% 0.03% 0.01% 0.01% 0.01% 0.01% 0.00% 0.00% PR RJ 0.01% 0.01% 0.00% 0.01% 0.00% 0.01% 0.00% 0.00% RJ RN nd 0.03% 0.06% 0.05% 0.08% 0.08% 0.10% 0.03% RN RO nd 0.04% 0.08% 0.10% 0.25% nd 0.05% 0.00% RO RR 0.14% nd nd 0.09% 0.09% 0.30% 0.21% 0.21% RR RS 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% 0.00% RS SC 0.04% 0.03% 0.02% 0.01% 0.01% 0.01% 0.00% 0.00% SC SE 0.50% 0.50% 0.50% 0.48% 0.97% 0.07% 0.06% 0.05% SE SP 0.02% 0.02% 0.01% 0.01% 0.01% 0.01% 0.01% 0.00% SP TO 1.00% 0.13% 0.10% 0.40% 0.10% 0.10% 0.02% 0.05% TO Source: IDB
Table 9 shows the concentration of revenue per type of taxpayer. This indicator allows the analysis of income concentration. Using this data, it is possible to verify the relationship that may exist between the numbers of taxpayers who pay their VAT taxes and their contribution to state VAT revenue (Febres, Cruz et al. 1998; Febres, Cruz et al. 1998). Interestingly, although 0.33% of large taxpayers in Brazil account for 92% of STAB revenue (Silvani and Baer 1997; SRF 2002), the management of large taxpayers was not a significant variable when regressed in comparison to VAT (refer to chapter 3).33
33 In Brazil, 550 firms/taxpayers contribute to 44% of the total Brazilian tax revenue (including all taxes and all levels), from a universe of 2.9 million taxpayers (SRF 2002).
84 Chapter 4. Vertical Assessment
However, it could be that strict treatment of the remaining taxpayers was not accounted for. As Silvani and Baer (1997) elucidate,
“Arrangements to monitor large taxpayers should be accompanied by special programs for controlling medium-size and small taxpayers, who represent the largest proportion of taxpayers. [M]onitoring them with special systems and procedures has also proven to increase their compliance.”
4.5 Important Tax Administration Traits in Brazil
The Brazilian STABs are powerhouse in the sub-national governments’ fiscal apparatuses. They establish state governments’ general fiscal goals, communicate public policy trends, and control the state government budgeting for each fiscal year. STABs also supervise the performance of other public agencies in their fulfillment of these strategic goals and objectives. In order to keep track of state budgets, STABs request that public agencies provide planning and budgetary reports, such as budget projects and spending and performance reporting. As such, the STABs act as budgetary, accounting, and general financial manager. They perform accounting controller rolls for public spending policy and supervise the execution of authorized appropriations by adhering to legal and normative budgetary frameworks. In addition to their budgetary functions, STABs are responsible for state taxation, which is the scope of this dissertation.
Although several aspects of tax administration reforms are important, I concentrate on those that differ across the states analyzed. First, however, I describe some general aspects of tax collection and administration that are pervasive across all the Brazilian states.
4.5.1 Brazilian States’ Tax System: VAT Assignment and Rates The current assignment of revenue sources in Brazil departs in several respects from the traditional prescriptions found in fiscal federalism literature, especially regarding indirect taxes. The federal government collects cur the personal and corporate income taxes, which are shared with the states and municipalities. The states collect a broad-based value-added-type tax (the state VAT known locally as ICMS), a tax on motor vehicles, and estate and gift taxes. The VAT has traditionally suffered from a number of shortcomings, some of which were eliminated in 1966 through a major tax reform, specifically:
85 Chapter 4. Vertical Assessment
‘The VAT is levied at different rates on different categories of goods: reduced rates on “essential necessities,” 17 percent on most goods,34 and 25 percent on luxury goods. These rates are levied on a tax-inclusive base, resulting in higher rates on a net-of-tax basis. All interstate transactions are taxed on an origin basis. The federal senate sets the rate on these transactions at 12 percent, except for exports from the states in the South and Southeast to states in the other regions, which are taxed at 7 percent. The difference in rates on interstate and intrastate transactions complicates administration and provides an incentive to misclassify transactions, to reduce the tax liability.” (Ter-Minassian, n.d.)
Until the 1966 reform, exports of non-manufactured goods were not zero-rated, and credit was not allowed for the purchase of capital goods. Rates were also not uniform across the country. Exemptions and other preferential treatments for selected sectors have eroded the base of the VAT.35
A selective federal VAT (IPI) and a municipal service taxes (ISS, a service tax) are superimposed on top of the VAT, giving rise to cascading and distorted tax burdens across sectors and localities. This cascading of taxes is reinforced by the fact that sizeable contributions earmarked for the social security system (COFINS and PIS/PASEP) are levied on turnover.
4.5.2 Staging and Timing of the Reform In order to assess the success of the STAB’s reform efforts, it is imperative to differentiate the initial administration conditions among states. Given the lack of data prior to the 1997 reform, it is quite difficult to assess the conditions of the STABs’ administrations at the commencement of reform. On the one hand, starting points differed based on experience: São Paulo STAB, which oversees the largest part of the Brazilian economy, had operating since 1936,36 whereas STABs in states such as Tocantins, Acre, Roraima, and Rondônia were not even created until recently (in the case of Tocantins, it became a state only in 1989). On the other hand states with less capacity also benefited from the lessons learned by the most developed states, making a complete assertion of their initial developmental condition in 1997. When the initial disparities among states were regressed in comparison to the level of VAT performance,
34 Brazilian States utilize several different VAT rates. However, some standardization exists with respect to rates across states. For more information on VAT rates, please refer to Appendix A. For intra-state general rate is 17%or 18%, for inter-state transactions, rates are either 12% (for transactions with the southern states) or 7% (for transactions with remaining states). 35Tax preferences under the VAT have to be approved unanimously by a committee of the Secretaries of Finance of the states (CONFAZ), chaired by the Deputy Minister of Finance. The unanimity rule has helped moderate, but far from eliminated, the proliferation of preferences under the VAT. 36 Interview with Benedicto Franco, 2001, SINAFRESP.
86 Chapter 4. Vertical Assessment
these differences proved not significant (refer to comments on alternatives regression models in the previous chapter).
The STABs’ restructuring efforts centered around two main areas, one focused on tax administration and the other around budgeting. The initial dates for the programs are shown in Table 10, below.
Table 10. Official Initial Date for STAB Reforms. State Tax Administration Reform Budgeting Administration Reform AC April 1997 AL January 1997 January 1997 AM January 2000 January 2000 AP March 1997 April 1997 BA February 1997 December 1999 CE January 1996 July 1996 DF November 1997 ES January 1997 May 1997 GO January 1997 MA December 1996 MG January 1997 January 1997 MS March 1996 March 1996 MT February 1996 October 1998 PA August 1996 December 1996 PB January 1996 January 1998 PE July 1997 July 1997 PI January 1997 January 1998 PR January 1997 April 1997 RJ January 1997 January 1997 RN March 1996 RO October 1997 RR July 1997 July 1997 RS January 1997 January 1997 SC January 1997 October 1997 SE January 1997 SP June 1998 March 1994 TO April 1997 Source: PNAFE 2004 (PNAFE).
Yet, official dates do not reflect the development stage in each state. There was a lot of eagerness to assume the leadership in terms of reform ideals:
87 Chapter 4. Vertical Assessment
“Our state started reform in the late 1987, with Tasso Jeressaiti. It was a truly entrepreneur government and very tuned up with updated management strategies. We were the first to start tax administration reforms.” Ednilton Soarez, 2001, prior head of the STAB/CE.
“São Paulo is and has always been the locomotive of the Brazilian economy. Of course, it was ahead in the reform process.” Yoshiaro Nakano, 2001, prior head of the STAB/SP.
“Bahia has been at the head of process since the project Baía Azul in the early 1990s. Although the Baía Azul project focused on the cleanliness of the beach in the state’s capital city, we needed to resolve our finances because we were running out of funding and international organizations refused to lend more money, given the state’s fiscal crisis. Before continuing with the Baía Azul project, we had to “clean” our finances.” Dimas de Melo, 2001, STAB/BA.
“São Paulo, Bahia, and Ceará were way ahead in the reform process, which make sense since they are the most developed states in Brazil. However, the increase in state debt left them without resources to implement the reform. Their STABs decided to request a loan from the IDB, whose tax administration reform projects were widespread in Latin America. When they sent their loan requests to be approved by the Senate (international loan requests must be approved by the Federal government and by the Senate), the first step toward the implementation of the reforms, the representatives of the other states barred them. All the states wanted to participate on the program. So we had to wait for all the states to get ready and put together projects so that we could present one whole project. Only then did we get approval for the loan to fund the reform.” Mauro Bogea, PCU, Ministry of Finance.
The conflicting testimony above demonstrates how many state leaders’ considered their states to have spearheaded the reform movement and the competitive spirit among state’s leadership. This testimony also suggests that it is difficult to categorize the commencement of reforms in each state.
For the sake of equity and fairness, the IDB classified states’ administrative capacity using diagnostics provided by the World Bank. This accounted for the diversity among the states. Three groups were created (see Table 11): one made up of states with the least self-financing capacity (group 1, which included Acre, Amapá, Roraima, Rondônia, Tocantins, Sergipe, the Federal District, Piauí, Alagoas, Mato Grosso do Sul, Rio Grande do Norte, Paraíba, and Mato Grosso; which together accounted
88 Chapter 4. Vertical Assessment
for 13.5% of all state tax revenues in 1995); a second made up of states that were already making efforts at modernization but that still displayed structural deficiencies in connection with database integration, internal control, and management (group 2, which included Amazonas, Maranhão, Para, Espírito Santo, Goiás, Ceará, Pernambuco, and Santa Catarina; which together accounted for 21% of all state tax revenues in 1995); and a third group made up of states that had already started the modernization process in various areas of the public sector, including fiscal matters (group 3, which included São Paulo, Paraná, Bahia, Rio Grande do Sul, Rio de Janeiro, and Minas Gerais, which together accounted for 65.5% of all state tax revenues in 1995). I studied six states from the groups above and using this classification.
Table 11. States by IDB Initial Development Classification. Self-financing Capacity States Group 1 Poor AC, AP, RO, RD, TO, SE, DF, PI, AL, MG, RN, PB, MS Group 2 Developing AM, MA, PA, ES, GO, CE, PE, SC Group 3 Developed SP, PR, BA, RS, RJ, MG
4.5.3 Parallel Study Committee: The PROMOFAZ (Modernization Program of the STAB) “We were apart from the system and we were a part of the system. We worked with a parallel structure that allowed us to work separated from the original STAB structure and that created for us a protected zone within which we could examine the Bureau. We could work and rethink the traditional structure, and still work inside it and be a part of it. It was parallel in the sense that, structurally, we had the same structure, but we were there to invent the new organization from inside. But at the same time, we kept our positions in the original STAB and continued working day-by-day. So we were somewhat within and somewhat outside. We were insulated from being attacked by the structure and by our colleagues, who might think that we were plotting against them, and we had autonomy. That is how our group worked.” Nelson Machado, Tax collector, STAB/São Paulo, 2001.
An internal unit studied reform in each STAB. In São Paulo, this group was called PROMOFAZ (Programa de Modernização Tributária). In other states, it had similar names, such as PROMOAT, PROMOCAT, and PROMOSEFAZ.
89 Chapter 4. Vertical Assessment
The PROMOFAZ governance structure at each tax agency was simplified by virtue of its status as a core remodeling structure. This allowed the agencies to eliminate layers of bureaucracy while adopting a somewhat different layering of complexity, based upon consensual decision-making, working committees, and a structure that organized around ties that follow a flat, rather than, hierarchical organizational chart.
The importance of this unity was described by a tax collector/auditor who said that he was
“proud of working for the committee, and that despite the increase in the workload, I liked the fact that I kept my regular work. It allowed me to keep in touch with my peers, to not be perceived as privileged or a traitor, and to have a down-to-earth perspective on the reform. Nelson Fagundes, Tax collector, São Paulo, 2001.”
The prominent role of tax collectors in setting the tone and direction for their agencies is evident in the strategic influence of this committee. The unit reviewed directions and new suggestions, and it discussed the use of questionnaires and studies from consulting firms (Azevedo 1998; Machado 1998). Because the members of this parallel structure were tax collectors themselves, the internal animosity against them was low.
4.5.4 Tax Administration Models Initially, tax administration functions were not specialized. STABs’ structures were quite political and they were organized vertically. The tax collector could be assigned either an internal or external job according to the superior’s orders. As management capabilities improved and as scientific functionalism was disseminated, STABs began to organize employees by function: collecting, auditing, accounting, etc. With the growth of local economies and industrialization, STABs needed to select firms. In this phase, firms were selected by size. STABs selected larger firms first, since they could contribute more to revenues and had more prominence (Vehorn and Brondolo 1999).
Vehorn and Brondolo (1999) found that the tax administration agencies’ were organized in four ways: (1) by type of tax (e.g., Personal Income, Property Income Tax, VAT, Automobile Tax); (2) by function (such as departments to collect arrears, audit taxpayers, process tax returns and payments, etc.); (3) by type of taxpayer (large-, medium-, and small-sized enterprises); and, (4) by vertical (geographical) hierarchy (headquarters, regional, and local offices). STABS throughout Brazil used a mixture of features from different models from size, functions, type of tax, and geographic territory. The organizational change from a mechanical, territorial, and functional model to an economy-rationalized model that I
90 Chapter 4. Vertical Assessment
found in Brazil is distinctive of the ones described in the tax administration literature, and therefore worthy attention. This innovation happened beyond the patronage of IDB or the Ministry of Finance and is peculiar to a few Brazilian states. This is the focus of the next section and the core of this dissertation.
4.6 From Territories to Industry Segmentation: Production Chain in the Tax Sector and the New Cognitive Process
The tax reforms of the 1990s produced a radical change in the way that VAT is collected in Brazil. Previously, tax collection had been performed according to the territorial allocation of tax collectors. Each tax collector would be responsible for a geographic area, ranging from a few blocks to a district or a small town. In this territorial distributional approach, tax collectors would audit all resident enterprises, which might include garment retailers, butchers, and chemical industries. The STAB’s local offices targeted dissimilar industries operating in a single area or block. This organizational structure was implemented in the early 1930s, reflecting the spatial growth of firms and businesses around the cities. Over time, with urban sprawl and industrial growth, neighborhoods became more diverse and industries dispersed. Consequently, tax collection did not follow any rational order with respect to industry or products.
Auditing and surveillance followed a case-by-case procedure. By and large, the tax collector would audit and check tax forms, receipts, and inventories (without any third party information on the entry and exit of commodities) and would audit company books, relying largely on information provided by business owners, retailers, and manufacturers.
“Before, we parachuted into a firm. So an audit would happen every five years. So suppose the last time we visited the firm was in 1996. Now it is 2001. The problems are other ones. It does not mean that what happened in 1996 is still irregular now. Then, we would notify the taxpayer and leave. It did not guarantee that the tax was paid nor would it guarantee revenue for the state. No one really knew what would happen afterwards and the information might never reach any one else. We were totally disconnected and isolated.” (Interview with tax collector in RN STAB)
STABs were physical comprised of a web of tax offices or stations, where the tax collector would spend most of the time at the street level. There, the manager would assign firms and areas for the tax collectors, who would audit firms in their specific region. This system allowed little or no visibility for economic changes within the industries from the street level to the senior management. Furthermore, the
91 Chapter 4. Vertical Assessment territorialization of tax collection did not allow for a broad understanding of economic or industrial performance, or for the use of “input–output” economic or production-chain rationale in the taxation process. Consequently, tax collectors ignored any forward and backward linkages in each industry and did not apply economic concepts, such as a production chain perspective, when estimating future tax revenues from industries (see Figure 4).
Figure 4. Geographic Territory Tax Collection Diagram
Centralized Vertical Coordination
Individualized Action
Territorial Auditing and Assessment Axiomatic Rationale: Backward and forward linkages
The territorial approach to tax collection prevailed in all Brazilian states for decades until the recent breakthrough of the industry segmentation oriented approach. Pressured by the growth and increasing complexity of the national economy, in addition to the increasingly decentralization of the production process (outsourcing and contracting) and to the rise in the number of subsidiaries, local STABs offices became powerless to control the circulation of taxable products and the returns filled by firm owners.
The traditional tax collection system focused on the traditional modus operandi of tax administration; that is to say, tax evasion, as a final goal in itself. This is ultimately a closed system resulting in auditing and accounting for each individual taxpayer’s evasion. This is a reductionist vision of the system, since this strategy is based on a review of taxpayers who have already evaded taxes or who are involved in the evasion already detected for other taxpayers (Nascimento Júnior 2000).
92 Chapter 4. Vertical Assessment
This reductionism supports traditional tax collection because it is based on the idea that fighting individual tax evasion is a means to fighting global tax evasion. In this sense, the tax administrations believed that they were fighting systemic tax evasion in a very deterministic way. Instead of fighting tax evasion by understanding the complex way the economy works, this philosophy let tax administrations to focus on fighting tax evasion by hiring more repressive tax collectors to act upon taxpayers individually.
A new, more systemic approach to the tax collection, using the General System Theory was needed to replace this Cartesian, vertical, tax administration. This new approach account not only for the regular “on the spot” auditing of transactions (at firms or on the road), but also had to verify the “extended” parts of the productive process that were occurring in firms’ subsidiaries, branches, contract suppliers, and so on. However, such a system demanded careful and accurate information on taxpayers and their transactions within and outside of each state, as well as a detailed analysis of the production processes. Three factors were essential for this new system. First, the way tax collectors audit companies had to change. Second, it required an information system with the ability to show different taxpayers in the same sector, connect different phases of production, and detect taxpayers’ defaults. Only a powerful informational system could provide such complete information. Third, tax collectors needed to be trained and specialized in each sector, not just generally knowledgeable and prepared to analyze these facts.
Clovis Panzarini, Chief of Staff for the São Paulo STAB and a tax collector for two decades, confirmed that although such specialization had hung on for a while, the complexity of firms imposed a new reality and new needs on the way tax collectors work.
“Firms not only produce goods and book profits in different states, but also in different countries. Pieces and bits of manufactured products are produced all over the states, either by contractors or by branches of the parent company. Furthermore, suppliers of parts or raw material reside in all four corners of the country and, therefore, the accountability of these firms my raise questions and difficulties for tax collectors.”
It is well known that firms act in their own interest and attempt to evade their tax obligations by utilizing tax havens. However, the tools to control this type of evasion were not available for the STABs’ tax collectors. To develop these tools required more than a simple change in the mechanics of tax collection. It demanded a paradigm shift and a new way of thinking about the collection process.
“The new systemic model makes viable a new and broader system concerned with understanding each aspect of the tax phenomenon, departing from a study of just a few of the multiple levels of relationship and complexity. Taking into account the environment
93 Chapter 4. Vertical Assessment
with its diverse forms, interactions, and objectives, it allows an understanding of how taxpayers work to survive and compete.. (Nascimento Júnior 2000)
The first part of the new model resembles a simplified version of the production chain approach. Instead of auditing companies in a territorial basis, STABs decided to collect taxes according to industrial sectors. Rather than on the entire production chain, from the primary input to the final output, the model tries to identify specific industries and economic sectors, according to their importance in terms of tax collection (i.e. focusing on large firms and more visible sectors). Each state aligned its industry specialization process according to its main industries. STABs map out how much each sector collects, evaluates the performance of each individual firm and audits only the ones that collected taxes below the average in the sector (see Figure 5).
The value chain structure includes all firms in a vertical chain and can be characterized in terms of input suppliers, producers, processors wholesalers and retailers. Understanding the relationship between firms at different levels of the value chain is critical. Transactions among firms are vertically related and reflect the movement and delivery of a given service or product to the end market. This approach mirrors the intrinsically embedded principle of the VAT (USAID 2006).
This system resembles the backward and forward linkages model proposed by Hirschman (1958). Hirschman attempted to relate the shape of economic development, including its social and political components, to the linkage constellation of specific economic activities. He originally presented linkages as a tool of analysis and for industrialization policy. Overall, linkages denoted interdependence and, as such, the generalizable linkages approach became particularly useful for interpreting development experiences, identifying commodities and manufacturing linkages, and analyzing growth. In 1977, Hirschman proposed a more generalized concept, involving consumption and fiscal linkages. Eventually, he included a whole ideological system under the linkages conceptualization:
“the generalization can be interested as a switch in interpreting and explaining the complex interaction between technology, ideology, institutions and development.” (Hirschman 1987)
“The trade off between “operationality” and “sweepingness,” the switch from strategy to interpretation, the original emphasis was on tradables and specifically industrial sectors whose production was to be stimulated in a strategy of industrializations.” (Syrquin 1992)
94 Chapter 4. Vertical Assessment
Figure 5. Vertical Industry Assessment.
STAB
Axiomatic Rationale: Information backward and Flow forward linkages
teamwork sweeping the whole industry
For example, the Pernambuco STAB divided the economy into 12 major sectors: beverages, restaurants, vehicles, garments, cereals, and so on. In the restaurant sector, several large restaurants were not paying taxes regularly. With the new systems, the STAB was able to identify which ones were delinquent and follow up regarding possible cases of tax evasion. Similarly, the São Paulo STAB based its audit decisions according to the most important industries. In Rio Grande do Norte, however, the auditing process was more encompassing because the state government decided to include all economic sectors and firms in its audits. This comprehensive approach was probably possible because of Rio Grande do Norte economy had smaller and fewer firms than other states, which facilitated the auditing process. As an RN tax collector put it:
“Of course we think that the economic segmentation is interesting. We do it somewhat, but we have only 27,000 firms, whereas São Paulo has one million. Our goal is not to reach just sectors, but the whole economy.” (Ludemilson Araújo Lopes, TC and chief-of- staff, STAB/RN).
In Pernambuco, the PROMOFAZ’s tax administration procedures and modernization project was called REMAF (Revisão do Modelo de Ação Fiscal—Fiscal Action Revision Model). This program was created in 1997-98 within auditing department.
With the new paradigm, the focus switched from punitive to preventive. Previously, tax collectosr or auditor would go to firms and fine them (auto de infração). The new focus was on prevention, avoiding the accumulation of credit that is difficult to collect, and expensive and time consuming to sue. With the new focus, firm owners had a harder time evading taxes.
95 Chapter 4. Vertical Assessment
The Bahia and Goiás’ STABs were considered the role models for STABs in the group of poor and developing states. Goiás focused on monitoring the tax collectors’ follow-up efforts with taxpayers and Bahia focused on managing market segments. They provided good models for the other STABs. As Celso T. Ferreira, tax collector, from Bahia, explained:
“Our model brought together the two models. Today our model has two pillars: taxpayer monitoring and segmentation. We have 89,000 in our firm registry (cadastro de empresas). 5,000 are responsible for the more than 95% of the tax collection.”
According to Ferreira, segmentation allowed for the implementation of another tool in taxation, the ‘substituição tributária’, or tax credit. The firm at the first stage of production pays the tax at the beginning of the transaction. If other firms in the chain do not repay that firm, it generates a tax credit for the firm that paid the tax. That firm may demand a credit with the STAB. In this manner, collection is ex ante and it simplifies tax collectors’ work.
Segmentation demanded monitoring. However, STABs were not staffed for monitoring all firms in each segment. Segmentation gave the STAB an intelligent mechanism to monitor the companies:
“We did not have a mechanism to monitor all 5,000 firms. We were monitoring all of them at the same way. The only important thing was to check the papers and documents of the firm in locus or in the trucks on the roads. We were looking for a more intelligent mechanism that would allow us to analyze and identify the universe of these firms.” Tax collector at the Pernambuco STAB.
For the monitoring model, the objective was to scrutinize the functioning of the firm, its economic background, and supply linkages, and to then classify them into bands. These bands would identify the extent to which each firm could possibly be evading taxes:
“We had to create an indicator to measure the level of risk of evasion for each firm, and then ascribe appropriate actions for that level.” Tax collector in Pernambuco.
“Monitoring is a system that generates a series of indicators. Initially, 180 firms were classified and tested to implement the operational factor. Then, we created segments to manage… Monitoring creates consistency and normality. It classifies not the sector, but the firms. We are always comparing the firms with similar firms in the same economic segment.” Tax collector in Pernambuco, 2001.
96 Chapter 4. Vertical Assessment
The study of firms and industrial sectors also included an evaluation of past patterns of relationship with the STAB and the frequency with each firm in the sector was assessed. The regularity of audit actions, the problems that the sector presented, and the frequency of evasions in the past were all factored into a formula to establish the risk a given of the segment or sector.
“The first thing to be done was to start working with what we already know. We created matrix structures that crossed several indicators: quantitative and qualitative. We looked at the inputs and outputs of each firm and at the more important sectors. We classified them in terms of better or worse according to their methods of updating their fiscal books. Afterwards, we set to reach the rest of them, gradually, by monitoring the entire sector.” Tax collector in Pernambuco, 2001.
For each segment, the STAB created a study group that would analyze the production chain from the bottom up. These tax collectors could identify an industrial area and then chart information about each industry. They would also establish “best practices” for approaching each sector. For example, the pharmaceuticals, restaurant, beverages sectors would all be approached in different ways. They then created a matrix with all of this data, including market indicators, connections with the industry associations or firms associations, to total sector collection, fiscal debts, total product input and output, aggregate values transactions, IBGE data, formulas, and so on. Additional indicators were also created to measure the average performance of each industrial segment (for example, the number of users per restaurant type and informal sector competition in the adjacent area).
Firm economic performance was classified by levels, or bands, from 1 to 5. For example, Band (“Faixa”) 4 should be visited every week.
“We adjust the model for each segment. But we developed it from the fuel/combustibles sector. We update the model every 5 years. This way, I can know from my desk in the tax station in the border with Piauí state, what is happening with a particular taxpayer, even if I have never met them.” Tax collector in Bahia.
The following describes the minutiae carried out to map the restaurant segment.
“The restaurant segment is quite large and is known for evading taxes. Just think how many times you go to a restaurant and leave without getting a receipt for the total check, much less the tax receipt. Of course, the restaurant owners deny any claim of evasion and it was very difficult to get hold of their suppliers and inventories. First of all,
97 Chapter 4. Vertical Assessment
inventories are not well kept and documented, and second, many suppliers are from the informal sector. Thus, the restaurant industry was very murky to assess.
What did we do? In order to assess the potential of the sector and how much evasion existed, we set up a survey in several malls in the city. We hired a group of students to stand up in front of each restaurant in the mall. Usually, mall restaurants are quite busy and have a high turnover. The students had to check every single customer who left the restaurant and ask how much they had spent and whether they had a receipt. At the end of our survey, we knew the average number of customers and average revenue that the restaurants in the mall should be making and consequently, what they should be paying in taxes and how much evasion existed.
Not only did we get the information we wanted, but we had the director of the restaurant association begging us to talk about it and to set up a policy that would not only catch them, but also their competition and the informal sector.” Pernambuco, STAB, Segmentation Coordination, 2001.
Once it completed its segmentation analysis, the Pernambuco STAB created directories for each segments. The directories had seven sections: (1) legislation (regulations and norms); (2) firm auditing history (fiscalização de estabelecimentos); (3) transport and commodities circulation policies and audits (mercadorias em trânsito) that were carried out by tax collectors in roadside tax stations or in motorized vehicles who would stop trucks and check the products they were transporting (postos fiscais and volantes), (4) tax revenue (the bank offices were responsible for receiving the tax payments), and (5, 6, 7) procedures for the three regional tax stations (located in Caruaru, Petrolina, and Arco Verde) in rural areas.
The focus became different, and planning became part of the process. Everyone was involved in the process and could access information. According to a tax collector in Bahia, “all the information is available within the Bureau for all the tax collectors to consult whenever they need it.”
The results were quite remarkable. “We had a highly positive experience with taxing combustibles/fuels. It is an important source of tax revenue in PE. The team led, defined our fiscal action, drafted agreements, and everything was approved by the Director.”
Figure 6 shows the results for the wholesale food segment when reporting to his superiors and colleagues in the Pernambuco STAB.
98 Chapter 4. Vertical Assessment
Figure 6. Letter to Tax Collectors from Segment Manager.
To Friend Tax Collectors,
After a long year working in the Mills and Cereal Sector, I feel obliged to inform the results obtained in the sector during 2000. These results were only possible due to the efforts of each one of you carrying out your work. I hope that the information in the below table may shed light on the diatribe about the subject. Non-monitored firms Monitored Firms Rev % Growth Rev % Growth
Month 1999 2000 1999 2000 1 896,623 1135430 27% 3712390 5193705 40% 2 902,714 922785 2% 2524742 3274777 30% 3 929714 1045195 12% 3047729 3060388 0% 4 798637 1062973 33% 3493589 3138013 -10% 5 631975 3507641 455% 3169906 3252207 3% 6 816055 3359886 312% 3238678 3173135 -2% 7 620918 2590368 317% 3272025 3589572 10% 8 685030 2589850 401% 3465936 3439847 -1% 9 798467 1919901 224% 3206415 3090615 -4% 10 998989 2233792 92% 3360011 3137639 -7% 11 1044171 2233792 114% 3214082 2317459 -28% 12 1169258 2815168 141% 3342731 3442247 3% Total 10291933 26611833 159% 39048236 41109605 3% • Current values IGP-DI (General Price Index) • Monitored firms N=57 • In the revenue above, no fine, penalty, or past debt payments were included. Sincerely, Francisco C. Pereira Manager Wholesale Food Segment.
99 Chapter 4. Vertical Assessment
Although other STABs emphasized in interviews that their restructuring transformation lead to the formation of working teams, tax collectors at the Pernambuco STAB stated that teamwork was not a central in their work:
“We have managers or representatives (a preposto). The manager is the person from outside who can see the entire segment, reach out the industry as a whole, and establish plans that will be consistent through out the state. In the interior, there will be a single procedure for the entire industry. This manager must peruse several directories and understand whole economy.” Segment manager in Pernambuco STAB.
Despite this claim, teams are integral to tax collectors work in teams in the STABs that have adopted the segmentation model. The managers and directors of each segments meet together. They then review their plans bimonthly so that they can switch tax collectors around and supply them to groups that have more need. Although the individual remains central, planning action is done in teams. This strategy leads to coordination among segments.
“The resources are in the directors’ hand; that is why we meet all together. But there is also a functional specialization. We did not want to assign teams to each manager because 20 tax collectors for one group are sometimes too many for a single group and sometimes too few. So we have to manage according to what is needed. It is thus tailored according to the current need of each segment. We still have individual action, but what has changed is the resource allocation.” Segment manager in Pernambuco STAB.
Coordination and planning are key aspects of the work described above. However, the system preserves the individuality of the tax collector. I evaluate this in chapter 6.
4.7 From Learning by Doing to Learning by Rationally Processing
This chapter began by discussing the theories of reform—both in public management and in tax administration— and showing how they agree that organizations that work efficiently and deliver their services effectively tend to have clear goals and strategies and tend to be more open to public scrutiny. Most of readings on reform and restructuring look for best practices that can be replicated.
This chapter claims that practitioners, public officials, and politicians must pay attention to the inner structure of each sector. This is an aspect that has been overlooked—the sector rationale, the
100 Chapter 4. Vertical Assessment
linkages that one is able to make and dots to connect with one’s world. The STABs reorganized their structures to reflect the outside economy and industrial with its backward and forward linkages.
This new approach was only possible due to a core innovation in the process of tax collection: the economic segmentation of VAT collection activity. The new system focused on predicting the performance of each industry and on identifying outlier firms in each sector. The STABs mapped all the firms in each industrial sector, analyzed their respective tax performance, and identified the ones that were collecting fewer taxes, and investigated why these individual firms were not collecting as much as other firms.
This innovation were had two consequences: first, tax collectors were able to visualize the whole economy as well as its parts; and second, tax collectors could talk to owners in an informed way and use arguments based on accurate information about each firm’s transactions rather than suppositions. This accuracy, in turn, increased the tax collectors’ efficiency and credibility. Business owners began to approach STAB officials with a different attitude. They started respecting and responding to the STABs’ demands more promptly. Eventually, STABs understood firms’ production patterns as part of complex manufacturing and economic systems with inputs and outputs from different places and branches at several locations both within and outside the state boundaries. The tax collection production chain furthered the idea of input and output control.
The new process also allowed STABs to better control tax collection and determine more precisely which companies and businesses should be investigated. Before, these decisions depended on the will of the street level tax collectors. Change in both organizational and technological procedures raised the STABs’ and tax collectors’ credibility. These changes centralized tax collection information: increased tax collectors’ knowledge of procedures, tax collection status, increased the knowledge of tax collectors, how the economy functions, and led to coordination of tax collectors’ activities. Tax collectors began to work in teams to evaluate performance of each industry and to audit each sector. This organizational model replaced the territorialized, individual process that had been previously used.
However, the implementation of information technology systems was of utmost importance in developing the new model of VAT collection. Information technology proved to be a critical toll for learning about the STABs’ efforts to monitor firms’ performance and to reorganize the work process. Computerization is addressed in the next chapter.
101
Chapter 5
How Technological Changes Reveal Production Chains, Increase Tax Revenue, and Foster Alliances Across the Public-Private Divide
"I have to confess I love your project. It really makes a substantive contribution. You install computers, the (people) learn, you do away with (collectors), you reduce the state payroll, and, best of all, you eliminate strikes... A perfect case of governmental reengineering."
A Minister of Finance, Costa Rica
5.1 Introduction
In this chapter, I describe how information technology (IT) and information systems (IS) interact with organizational changes, leading to institutional innovation in public sector bureaucracies.37 I look at the role that computerization plays in improving state governments and public bureaucracy performance in Brazil. In the Brazilian STABs, computerization has led to a faster and more accurate information collection than the one previously in place, which, in turn, has led to greater accountability and
37 Throughout this chapter, I use Information Technology, Information System, and computerization interchangeably, referring both to hardware and software programs, as well as to information systems such as the Internet.
103 Chapter 5. New Technology transparency in tax collection. These technological changes have not only increased revenue collection, but have, combined with the organizational restructuring previously studied, led to increased professionalization in the public sector, improved workplace conditions, and produced new institutional arrangements, such as transparency and credibility (chapter 4). These new institutional arrangements include the revamped relationships among public sector agencies, private organizations, and professional associations, such as public employee unions and accountant’ associations.
My research reveals that computerization, encompassing both large and small equipment systems and including both hardware and software, has achieved the following results:
• increased tax revenue; • better allocation of human resources; • increased interaction between tax collection agencies and other governmental and non- governmental agencies; • the elimination of intermediaries (despachantes) in the tax collection process; • dissemination of information both within and outside the STAB, which has fostered transparency and credibility for the STAB; and, • a changed relationship between tax collection agencies and unions and business associations.
In addition to the organizational change (Chapter 4) and professionalization (Chapter 6), technological change was one of the pillars of the STABs’ modernization program. “Although technologies allow public organizations to change, they are not enough to ensure that change to take place” (Metcalfe and Richards 1990). The technological upgrade of the entire system was of utmost importance insofar as it disseminated information both vertically and horizontally—between upper and lower levels of administration and among the lower-level bureaucrats—empowering both the top and lower-level bureaucracy. In the Brazilian case, technological change took place immediately and consecutively with the reform.
5.2 The Technological Archetype
Private firms often adopt IT reforms in response to competition but such pressures were largely absent for public organizations until recently. Public bureaucracies in developing countries have few incentives to reform and their solutions may or may not include computerization. Bureaucracies in developing countries lack resources and face strong pressures to divert resources away from specific
104 Chapter 5. New Technology
tasks. IT resources are meager and organizational constraints, such as unprepared staff, are formidable (Peterson 1997). Obsolete equipment and administrative procedures, compounded with tax evasion and corruption, create an immense logjam for the flow of information on taxpayers’ assets and incomes. As such, automation and the consequent rationalization of information flows have been central features of tax reform efforts. International donor organizations played a seminal role in demanding computerization as a tool to implement governmental reforms (Tettey 2000).
The introduction of computers in traditional public organizations changes how organizations operate (Reich 1992). Despite its large, upfront installation cost, computerization lowers the costs of organizational communication and administration. Computerization also forces standardized procedures based on manuals and regulations to give way to non-hierarchical forms of communication and flexible procedures. Flexibilization allows choice and simplification and eases service delivery (e.g., documents, tax return validation, notifications, etc.) to costumers (i.e., taxpayers). In this sense, the bureaucratic principle of meeting goals as a measure of minimum success corroborates the concern for delivering services to customers fast and efficiently (Barzelay and Armajani 1992).
One of the main uses of IT in the public sector is to access information to formulate accurate policy designs. Efficiency-equity is one of the problems this information can solve. Major problems in tax administration include poor taxpayer identification, inadequate classification of taxation activities, document designs that are unsuitable for computer centralized database storage, and insufficient central computer capacity. In the tax administration arena, IT helps governments to identify taxpayers, monitor their transactions, receive reports from third parties, and crosscheck information received from taxpayers. The main role of IT is therefore as an organizational tool for the government (Das-Gupta and Mookherjee 1998). Hence, part of the solution pushed by consultants and international organizations has been to create a national database to store and process all tax information in a single place, accessible to users in any place, i.e. STABs.
In any organization, computerization can serve a range of functions. It is an integral part of administration, as computer generated outputs are in high demand. Computerization usually involves a centralized computer system to carry out an organization’s primary administrative functions. In tax administration, these functions include maintaining a registry that records taxpayers and receipt of tax payments, issuing form letters and notifications, and assessing taxes.38 This means that the computer
38 Heeks identifies four types of formal information: information to support internal management (including both personnel information and budget and accounting information for financial management); information to support
105 Chapter 5. New Technology system is in the front line of administrative activity (Glenday 1996; Hadler 2000).
Accordingly, computerization advances the battle against the four central problems of TAs, usually related to the taxpayers databases: (1) non-registered taxpayers who ought be in the system; (2) formally registered taxpayers who do not file tax returns; (3) taxpayers who file tax returns, but do not pay; and (4) taxpayers who file incorrect information and pay less than they ought to (Bird, Bagchi et al. 1995).
Finally, with the advent of the Internet, widespread connection evolved.
During the last decade, many countries have attempted to transform their governmental structures and improve the quality of their services. In terms of systems, e-government has emerged as a powerful tool for administrative reform and for change in any sector. Brazil is actively participating in the IT revolution and has designed and implemented several projects that can be identified as e-government (Fernandes and Afonso 2001; UNDP 2001). In general, e-government initiatives can be divided into four categories: e- services, e-management, e-democracy, and e-policy. The implementation of these systems provides better and more democratic information that, in turn, increases the efficiency of the state bureaucracy.” (Gil-García and Cid 2003)
5.2.1 Functionality In Brazil, the past two decades have seen an unprecedented growth in the use of information technology across the public sector, particularly in the management of fiscal systems. As a rule, computerization is a suitable process to deal with tax administration since this field deals with large amounts of data and calculations. Tax administration reforms have utilized equipment and tools that increase taxpayer identification, compliance, and effectiveness. In addition to the belief that compliance and effectiveness rely on enforcement as a primary instrument to guarantee taxpayer compliance, reform has been aided by IT to support enforcement and to advance new organizational methods.
Computers facilitated the reorganization of the VAT collection process in two ways: (1) they facilitated the collection of complete, fast, and accurate information on taxpayers, their assets, and tax returns; and (2) they allowed coordination and control of the processes within tax administration public administration and regulation (including records of entities, people, and institutions); information to support public services (e.g. education, health, and transportation); and information made publicly available for dissemination of policies, laws, and regulations or for demographic studies) (Heeks 1998; 1999). All this information can be used for everything from day-to-day operations to long-term policy analysis and planning.
106 Chapter 5. New Technology
(Barbone, Das-Gupta et al. N.D.; Bhatnagar 2000; Das-Gupta and Mookherjee 1998; DEAT 1999; Glenday 1996; Hadler 2000; Peled 2000; Peterson 1996; 1997; Silvani and Baer 1997; Southern 2001; Stanback and Jr. 1987; Tettey 2000).
More than mere formal changes in the format of tax collection, computerization affects the procedures, processes, and mindsets of the whole STAB organization. The change in the rationale of the VAT collection described in chapter 4, for example, would have been greatly jeopardized without computers. Some of the innovations during the reform period involved the translation of existing procedures into more sophisticated processes using digitalized formats. However, the bulk of the reform lay in the real and substantial implications that digitalization provided for information assessment and data integration, the deep organizational change that this enabled, and the subsequent professionalization of public officials.
Nonetheless, simple actions, such as organizing and updating centralized databases, helped to loosen several tax administration bottlenecks related to implementation problems, lack of resources, and bureaucratic limitations. These problems are not peculiar to Brazilian tax administration and have long been reported in the literature (Bhatnagar 2000; Bird 1991; Bird, Bagchi et al. 1995; Bird and Casanegra de Janstcher 1992; CIAT 1996; Das-Gupta 2001; Febres, Cruz et al. 1998; Glenday 1996; Hadler 2000; IDB 1996; Jenkins 1994; Lasheras and Ros; Management 2001; Peterson 1996; 1997; Silvani 1993; Silvani and Baer 1997; Surrey 1973; Tanzi and Pellechio 2000; UN 1997; Vehorn and Brondolo 1999; WorldBank 2001).
5.2.2 Complexity Another consideration in execution of IT is the complexity of tax administration. This complexity arises from the tax system’s broad scope and deep involvement in the economy. Successful computerization of tax administration requires high levels of management and administrative resources. TAs not only deal with huge databases, but they also have to operate in all regions of a country and to interact with all economic entities-individuals, corporations, partnerships, cooperatives, clubs, and other organizations (Peterson 1996; 1997). IT reform cannot be isolated from the wider constraints of the sector to be reformed. Successful IT reform requires interventions exactly where the constraints might be. Acting selectively, building institutional capacity in a sector, and limiting the diversion of resources demand a great effort and coordination from the STAB. At the same time, IT needs to be integrated into the whole system and learned by the community of users.
107 Chapter 5. New Technology
IT systems affect the coordination of information flows and the work of the bureaucracy by altering the relationship between information and physical factors, such as distance, time, and memory, thus making collaborative problem-solving and highly coherent organization possible. Fountain (2001) explains that through computerization, users are able to manage excessively specialized situations and compress horizontal and vertical flows of information and decision-making instantaneously.
“Not only are we now able to retrieve information while we are still auditing the firm, but we can also get information about the firm production, the product destination, the firm suppliers, and the market. We can also collect information about the taxpayers’ assets and transactions available from other public agencies, such police department, judicial courts, and so on. Prior to carrying personal laptop computers to the audits, we could not access this kind of information. The information about the taxpayer, even whether the firm was legally registered or not, was only available under request by post mail sent by the head of the STAB to the agency that had the information. For example, if we needed an information about the permit for the firm’s activity, it could take an entire year before we received the information from the official Firms’ registry office, if they ever did.” Dimas Mello, tax collector, Bahia STAB.
IT also pushes towards data standardization and sharing across agencies. As (Fountain 2001) adds:
“shared databases are not possible without standardized data, and once developed, they create a platform for further integration efforts.”
Information technology is a means to deliver superior reforms in a more efficient and accountable way. IT not only increases the capacity for organizational control, it also facilitates decentralization, resource management, and privatization (Heeks 1998).
“Before the computerization, if a firm owner would tell us information about his assets and his company, or if the tax collector responsible for that particular information was not available, we had to wait for the tax collector to return. Our working day was wasted. And the tax collector had time to manipulate his files. Now we have laptop computers and we can check the taxpayer in locu.” RN STAB, tax collector.
Since this approach seems so evident and has already been prescribed and implemented in tax administrations around the world, why had it not worked efficiently in Brazilian states before? IT was
108 Chapter 5. New Technology
used by Brazil’s public sector since the 1960s, but implementation results have been different. The explanation lies in how IT has interacted with the different organizational and institutional arrangements that are in place at various times. As Fountain (2001:6) proposes, “the challenge is to understand the institutional arrangements in which these changes are embedded.” I thus examine not only the institutional arrangements, but also how and why they have changed.
5.3 The Enactment Frame
Currently, there are some social theories that recognize the interplay between social or organizational structures and IT (Orlikowski and Baroudi 1991; Orlikowski 2000; Orlikowski and Yates 2002). However, most of the initial studies about IT were about its deterministic command. These early research efforts made the assumption that IT could transform social and organizational entities independently of institutional and cultural characteristics. Some theoretical approaches to IT that take its transformational power for granted. Orlikowski and Iacono (Orlikowski and and Iacono 2001) identified this view as the “tool view of technology.”
Technological determinism has reigned as an explanation for successful restructuring results. This idea assumes that technology by itself compels individuals, social arrangements, and institutions to conform. Those who ascribe to technological determinism believe that technology is the autonomous motor to changes within an organization. The adoption of a new technology inevitably “creates predictable (and usually modern, human error freer) structures, work, routines, information flow, and performance [providing] the technological fix.” (Orlikowski 1996)
Some theoretical traditions propose a different way alternative ways to understand the relations between IT and social structures. According to (Orlikowski and Iacono 2001), these approaches refer to technology either as an embedded system or as structure. Two examples of the structuration perspective are the structurational model of technology (Nunberg and Nellis 1990; Orlikowski and Baroudi 1991; Orlikowski 2000) and the adaptive structuration theory (DeSanctis and Pool. 1994).
Institutional theory has been very useful in understanding organizational settings for IT/IS (Powell and DiMaggio 1991; Scott, Ruef et al. 2000). Scholars from various disciplines, such as economics (North 1990; Rutherford 1999), sociology (Brinton and Nee 1998), and political science (March and Olsen 1989) have developed institutional frameworks to understand diverse social phenomena. Based on this institutionalism tradition, the technology enactment theory attempts to explain
109 Chapter 5. New Technology the effects of organizational forms and institutional arrangements on the information technology used by government agencies.
Fountain (2001) suggests that the technology enactment framework focuses on the relations among information technology, organizations, embeddedness, and institutions. The assumption that a direct relationship exists between technology and structure has produced contradictory findings because it omits important elements in this interaction between IT and the governmental structures, such as those due to cognitive, cultural, structural, and political embeddedness. To address this, Fountain calls for a new concept:
“technology enactment, i.e., the tendency of some organizational actors to implement new ways that indeed strengthen institutionalized socio-structural mechanisms even when such actors tend to enact technology to preserve ongoing social or network relationships and to maintain performance programs.” (Fountain 2001).
In the tradition of institutional analysis, Fountain establishes that an enacted technology could be understood as the perception, design, and use of objective technologies, such as the Internet and different variable pieces of hardware and software. Therefore, enacted technology encompasses both the technological features of the current system and the way in which users take advantage of those technology characteristics (see Figure 7). Similarly, objective technology includes all other features that could potentially have been selected as part of the enacted technology, but for some reason is a not part of the information technologies in use.
Social informatics (Kling 2000; 2001; Kling and Scacchi 1982) and the technology enactment theory (Fountain 2001) are based on embedded arrangements. According to the enacted technology framework, there are objective information technologies are in some way modified by organizational and inter-organizational factors. In other words,
“[T]he embeddedness of government actors in cognitive, cultural, social, and institutional structures influences the design, perceptions, and uses of the Internet and related IT.” (Fountain 2001)
Using related theoretical constructions, these theories argue that there is a dynamic interaction between social structures and information technologies. In this view, IT has the potential to change social and organizational structures, but at the same time IT is affected by these structures in its design, implementation, and use.
110 Chapter 5. New Technology
Figure 7. Technological enactment framework (Fountain 2001).
IT Features
Enacted Outcomes Organizational IT Characteristics
Institutional Arrangements and Regulations
Likewise, the relation between IT and social or organizational structures is bi-directional (DeSanctis and Pool. 1994; Kling 2000; Orlikowski 2000). IT has the capability to change government organizations, but at the same time, IT is affected by several organizational and institutional constraints (DeSanctis and Pool. 1994; Fountain 2001; Gil-García and Cid 2003; Kling 2000; Orlikowski 2000) 39.
Were we to accept that IT alone made the reform process possible, we would need to accept technological determinism as the explanatory force for successful tax administration reforms in the Brazilian STABs. However, the effects of information technology on institutional arrangements, while decisive, were not limited exclusively by the logic of the technological determinism. IT has enabled governmental agencies to learn about, and to cope with, local economic predicaments and has allowed STABs to create a new structure for tax collection. IT has also driven business associations and unions to rekindle their relationships with government agencies and it has enhanced respect for public officials and accountability in the public sector. But it has relied on the individuals within these organizations to be the driving force for change implementation. These individuals have implemented software systems that have affected the entire system of tax collection.
39 (Gil-García and Cid 2003) make a similar argument with respect to e-budgeting in Mexico.
111 Chapter 5. New Technology
5.4 e-Taxing
Traditionally, the taxation process has been conceptualized as a technical process (Chapter 4). According to this technical view, the taxation process comprises the following steps: regulation, control, audit, execution, and evaluation. Administratively, each step involves several regulations, procedures, and transactions, which are accomplished by official, documented reporting-oriented processes.
Following the technology enactment view, e-taxing enactment refers to the perception, design, and use of information technologies at any stage of the tax process. Similar to any other use of IT in government, the technological characteristics of e-taxing projects are affected by the organizational context in which they are developed.
The organizational forms contain structural characteristics, such as centralization, formalization, and communication channels (Fountain 1994). Organizational and networking factors have a direct effect on the enacted technology. In addition, institutional arrangements (represented by laws, regulations, and other cognitive, cultural, or socio-structural constraints found in government contexts) affect organizational forms and therefore indirectly impact the enacted technology. The resulting enacted technology produces certain organizational outcomes. These outcomes produce modifications in the technology itself, but may also in the long run, lead to transformations in organizational forms or even in institutional arrangements. Thus, the technology enactment framework recognizes the complex relations between information technology and social structures.
The e-taxing enactment framework proposed in this paper is very similar to the generic conceptual model developed by Fountain (2001) (see Figure 8). However, institutional arrangements have direct effects on the enacted e-taxation, whereas the effects of institutions are indirect in the original framework (institutions only affect the enacted technology through their direct effects on organizational characteristics).
In addition, although we are able to evaluate most of the bi-directional effects of IT applications on organizational and institutional characteristics of the organization, these observations may not be all- inclusive. The arrows moving from each step indicate both parts of the mechanism and process. Organizational and technological characteristics act as gears and enablers in the e-taxing process. The result of this gearing action is the enacted e-taxing, with objective and systemic features.
112 Chapter 5. New Technology
Figure 8. Technological Enactment Framework for e-Taxing based on Fountain (2001).
New Rationale; Organization: Vertical Industry STAB Assessment
Institutional Settings
Institutions Regulation s Organization STAB Norms
RATIONALE Objective Technology
Enacted Technology => Process
113 Chapter 5. New Technology
5.5 Computers at Work: The Information Age in the STABs
As occurred in the rest of the world, an impending and unavoidable need for IT restructuring reached the Brazilian public sector in the 1990s. Although it resembles a new paradigm for the public sector, IT implementation has its roots in the post-industrial society and in the information age touted by authors such as Toffler and Bell (Heeks, 1998: 3). In Brazil the information age in the public administration started in the 1960s with the advent of mainframe computers.40
“At that time, we had mammoth machines for which we needed the entire floor, one room for each machine. Just to run one state budget, it would take us the entire semester and it was basically summing up the accounts.” Darcy Almeida, Brasília, D.F., retired government officer.
Because of the high inflation rates in the 1980s and 1990s and the concomitant development of the financial banking system, Brazil built up one of the most developed bank systems around the world. Taxpayers have been able to pay taxes in thousands of bank branches and other financial institutions for several years (Haldenwang 2003; Ruediger n.d.). This network receives more than 90% of all tax returns, thereby reducing the burden on the tax officers to physically collect tax payments. Between 1980s and 1990s, the number of returns filed by taxpayers increased from about one million to approximately 25 million overall in Brazil.41 Today, almost all of these taxpayers file their returns on-line and pay at the bank branches or lottery houses (Graph 7).42
Computerization in the STABs started in 1995, both with respect to technological aspects (updating of the physical infrastructure and installing new equipment), and in terms of organizational changes (creating connections with other agencies and installing new software to monitor the firms’ economic performance and track down stop-filers and non-filers) (Silvani and Baer 1997). Initially, nine states were involved: São Paulo, Bahia, Rio de Janeiro, Paraná, Pernambuco, Rio Grande do Norte, Paraíba, Maranhão, and Ceará (Interview with Dimas Melo in BA, STAB). The joint effort reflected needs and demands in each of these states. Some, such as São Paulo and Bahia, had already equipment to facilitate this reform. However, even wealthy states had to improve their technological infrastructure in
40 Although Brazil has never been a major global player in the computer industry, attempts to develop its computer industry have included the closure of the market to protect domestic companies in the 1970s and the 1980s. For more information on the computer industry in Brazil and for information on the bank system in Brazil (Evans, Frischtak et al. 1992). 41 Communication from Pedro Luís Bezerra, Federal Treasury, Ministry of Finance. 42 Health and education systems began to be updated by the late 1980s. In India, Singapore, the Philippines, and Spain similar processes had been implemented starting in 1975 (Lasheras and Ros N.D.).
114 Chapter 5. New Technology
order to develop their collection capabilities. Of the 27 STABs, only São Paulo, Paraná, and Bahia had websites, and those were only static pages in 1997 (SEFAZ and FAZESP 1999).
Graph 7. Tax Returns ON-LINE.
Online Filing - Selected States
100% 90% 80% CE RN 70% PE 60% Percentage of BA 50% Online Filing PB 40% SP 30% 1997 20% 1999 10% 2001 0% Year SP 2003 PB BA PE RN CE States
Source: IADB 2005.
Peterson discusses how the inertia of bureaucracies in developing countries coupled with the delays intrinsic to the implementation of information systems is deadly. “On average, it takes three to five years to institutionalize information systems in private firms in the United States and Europe.” 43 Because of the resource shortages that developing bureaucracies face, particularly related to skilled staffs and the comments on the literature by donors and international organizations, it could be expected that STABs would take much longer to implement computer systems than in those developed countries or in the private sector. However, this was not the case. Results from computerization were seen right away, as statement from several taxpayers and tax collectors and by the statistical analysis in chapter 3.
43 Also, refer to the case of Singapore (Bhatnagar 2000).
115 Chapter 5. New Technology
“In most places, agencies purchased a whole lot of personal computers and made then available to tax officers. However, computers would remain untouched at the top of the desks accumulating dust.” Lobato (SP STAB) and Dimas (BA STAB).
How then were STABs able to computerize tax agencies so quickly? Several aspects commonly mentioned in the literature of administrative reform were observed in the case of the STABs visited. These elements are political will, resources availability, and incremental sequencing
5.5.1 Resources, Timing, and Political Will One of the first practices of administrative reform, even before improving facilities, is often to supply tax agencies with the latest technology: in the case laptop computers. This appears to have occurred at the Brazilian STABS.
Information systems require large and timely investments in material and people to ensure that the systems can handle information quickly. In Brazil, government expenditures on IT have grown rapidly. The government created a special fund that receives 1% of all the firms’ gross revenue (Fernandes and Afonso 2001). Although the systems were expensive in terms of personnel and equipment, Graph 7 shows that the STAB’s administrative costs as a percentage of the tax revenue did not change significantly. However, these administrative costs may not include the IDB’s matching loans for the modernization project, given that not only the reform costs are estimated, but the total administrative cost. Training resources were also made amply available to the states, as will be discussed in the next chapter.
Timing and sequencing also played an important role in the STAB’s implementation of computers. Both are frequently brought up in the literature on reforms. Some scholars believe that the management constraint of bureaucracies in developing countries means that IT reform should be incremental, not comprehensive (Peterson 1996; 1997). Good timing for the modernization project had two dimensions: political and technological, but did not follow a predicted incremental path.
Earlier on, in the early 1980s, the few states that were able to renovate and incorporate technological equipment were not fully capable of involving the tax officers. Tax collectors continued carrying on their actions and methods as they had always done before and looked down on computers, despite training. What happened this time around?
116 Chapter 5. New Technology
Graph 8. STABs’ Administrative Costs as a Percentage of Tax Revenue, Selected States.
Administration Cost
7%
6%
5% BA CE 4% PB PE 3% RN SP 2%
1% Administrative Cost as a % of Tax Revenue as a % of Cost Administrative 0%
7 8 9 0 1 2 3 4 199 199 199 200 200 200 200 200 Year
Source: IDB and author’s estimation.
One aspect of successful reforms commonly discussed in the literature is political backing (Kettl and Milward 1996; Kickert 1997; Lasheras and Ros). Political will was strong in most of the states visited. In Bahia, São Paulo, Rio Grande do Norte, and Ceará, governors had been in power for at least two mandates (approximately 8 years). The head of these governments were known for their entrepreneurial attitudes and for their commitment to reforms (Tendler 1997; Tendler and Freedheim 1994). Continuity was an additional benefit to the implementation of the reforms. The heads of the STABs were also highly qualified economists, scholars, and technocrats, an emergent kind of “market- oriented social technocrats.” Table 12 shows that the six selected states had political stability and continuity.
117 Chapter 5. New Technology
Table 12. Political Features in the Selected States. # of Internal Business Political Party Head STAB State Years Opposition Class BA Paulo G. Souto - Liberal Party Mascarenhas 7 Oscillating Strong CE Ciro Gomes & Tasso Jereissati Ednilton + 10 Very strong + elitist PSDB Soarez PB Jose Taguino Maranhao PSB Inactive n.d. PE Jorge Jatoba PSB 4 Strong Strong RN Garibaldo Alves Fo. PMDB Ubirajara 8 Conciliatory Strong SP Mario Covas PSDB Y. Nakano ~ 8 Conciliatory Strong Source: Compiled by the author from interviews and STABs’ websites.
Table 12 is an attempt to capture the political environment and the relationship between the STABs, their own workers (Internal Opposition), and the outside community (the Business Class) in order to show internal and external resistance to computerization and other reforms. There was no apparent internal opposition to the reform in either SP or RN, whereas opposition did exist in both CE and PE. Tax Collectors’ Association in CE was fiercely opposed to most change instituted by the reform, its president was very upset by the government “undemocratic and against the personnel” practices, and promoted strikes (interview with CE TCs Association’s president and anonymous TCs during strike rally). The internal opposition in PE, however, focused on wages. The TCs in BA oscillated with respect to reforms due to elections, so it was not possible to capture their position. Finally, in TCs’ Associations in PB were quite inactive.
Therefore, the two factors that seemed to have been crucial were the iron hand of the state imposed the obligation and the political timing. Yoshiaki Nakano, then head of the São Paulo STAB, assured that the process would be followed:
“(As for computers), we did not give them an option. Either they accepted (to work on computers) or they were out. And the policy implementation came from both sides affecting both tax collectors and tax payers at the same time.” Nakano 2000 São Paulo STAB 2001.
When asked about the computerization of the offices, a TC said:
118 Chapter 5. New Technology
“I have no choice, either I use it or I leave. Plus, I have a computer at home, everybody has computers, and I already used it anyway for private things.” Interview, SP/STAB 2001.
The imposition was two-pronged in what it affected tax collectors and taxpayers alike and immediately. The demand for turning in the tax returns in electronic media was mandatory in the first year of the program and after the second year, firms started filing returns through the web. When asked why they kept filing their VAT tax returns on paper forms, Nakano from São Paulo, STAB, commented:
“On the one hand, we established that in the first year, all the tax returns must be done in electronic format. The Tax Agency would only accept diskettes. No more paper. We trained the accountants and since they actually already do most of their work on computers, they “trained” the firms for us.
On the other hand, you did not even have to force the tax collectors to change. Even the ones who were not very comfortable with the change were trained. They did not have much of an option, actually. The ones who refused to use computers or to learn, we gave them the option of leaving the bureau.” (Interview with Nakano 2000)
Interestingly, the São Paulo and Rio Grande do Norte Tax Collectors’ Associations’ presidents had a fascinating reaction and looked at me in astonishment when I asked whether they did not oppose the computerization of offices. The president of the São Paulo workers’ association dismissed the fact as given, saying:
“The reform is unavoidable. It is here to stay. Of course we will oppose the government when our rights are threatened, but computers? Every single tax collector has at least one computer at home, or his or her kids have one it. Computers are an inescapable reality, we are not fighting it” David, President Tax Collectors’ Association.
When asked whether there would be dismissals or lay-offs because of the resistance to use computers, he said:
“We offer computer courses even to the retired tax collectors. But a few will of course leave. Not because of the computers, but because they have been around for too long.”
The availability of new technologies in information processing and the possibility of cross- checking information should raise the collection of tax revenues. The regression models presented in
119 Chapter 5. New Technology chapter 3 show the impact of computerization elements as a share of the GDP. In the model using the variable Reform as OnlineConnectiviness and OnlineValidation the relationship with VAT revenue growth was strong and significant. Using these variables as proxies for the computerization development of the STABs, we also verify the importance in the public finance process as well.
Conversely to what might have been expected, computerization facilitated the transition between the old bureaucratic structure and the new one. The former organizational model that was used at the STABs was highly centralized. Modernization brought information dissemination and decentralization. But, at the same time, the reverse situation was also true: The fact that a bureaucratic structure was in place facilitated the implementation and consolidation of technological innovation given that the process was rationalized, hierarchical, and systematically structured.
Ironically, computerization also reinforced certain characteristics of the old bureaucratic structure despite bringing about dissemination of information and strengthening the public institution. Computerization may have yielded a more “pure” bureaucracy in the terms advanced by Weber (Weber 1958). Therefore, it seems that computerization affected key institutional arrangements, helping to reshape an old public sector institution—the STAB—, while enabling the consolidation of a new process of tax collection. In opposition to some expectations, computerization has also increased centralized coordination, given that it gave more visibility for the managers at the center of the movements going on at the street-level of tax collection. As such, it gave them more control over the whole process and a grip on the implementation of restructuring policies. In the process of vertical assessment of economic segmentation, as it will be explained later in this chapter, computerization was the key tool to communicate order, information, and the new rationale between the top management and the tax collectors who worked the streets.
Computerization has been centralized and has been combined with the traditional administration. Computers allow decentralized handling and recording of information because the computer system rests on an extensive communications network. Computerization has been implemented fast and has reduced the taxpayers' cost of compliance. Computerization has enhanced coordination. While the introduction of uniform procedures may improve coordination, computerization has probably not increased compliance measured by tax revenues collected and has not affected people's perception of tax fairness.
120 Chapter 5. New Technology
5.6 Computerized Services
One of the central aspects of the tax administration reform that occurred in the Brazilian states is the computerization of all steps in the tax collection process. From the e-filing of tax returns to electronic validation and identification of potential taxpayers and internal audits, every aspect of the taxation process has been converted into a digitalized form with electronic input and automated connections. Data is collected and entered either through websites or from floppy disks submitted to the tax agencies. When I asked interviewees what led to the significant increase in tax collection across the Brazilian states, the very first answer was computerization and reorganization.
Although computerization occurred concurrently with other reforms, it had a deep impact on the administrative life of tax collectors and taxpayers. The sheer reduction of paperwork was a striking achievement for both taxpayers and tax collectors. Before tax administration reform, taxpayers were expected to submit 32 different paper forms. With computerization, they could submit information on all of their firms’ transactions over the Internet, using a single form (interview with Nakano, 2001). All of this information was sent to databases accessible by the STABs, and paper forms were eliminated (Guaragna 2000). These databases also became available to other agencies, allowing agencies to share information ranging from firm locations to the names of suppliers and transporters.
The computerization process did not stop at the filing of tax returns. It included processes such as the electronic validation of tax returns to allow the government to collect taxpayer data and information. A description of computerized processes is shown in Table 13.
While it was relatively quick and easy to provide computers and implement information systems, the physical and organizational restructuring processes took longer to implement. As a result, although these organizational processes were planned to be implemented concurrent with computerization, it ended up occurring at a much later date in practice.
Graph 9, Graph 10, and Graph 11 show that on-line budgeting, linkages (agencies that are connected on-line and share data with the STAB), and validation have been completely adopted in most states.
121 Chapter 5. New Technology
Table 13. Computerization Processes. On-Line Filing of Tax Returns Requiring that all taxpayers file tax returns on-line. In place in São Paulo since 2000. Electronic Validation Creating standardized legal evidence that Tax Returns were submitted. Electronic Document Requisition Using the Internet to request legal documents. Taxpayer Record Maintenance Automated maintenance of taxpayer businesses information and payments to/from other firms. Electronic Payment Systems Utilizing ATM and Internet kiosks (e.g. Net-kiosks) to collect payments from taxpayers. Taxpayer Information System Database showing taxpayers’ assets, previous tax returns, transaction history, and criminal records. Automated Assessment Using software to automate assessment of tax returns. Taxpayer Accounting Record Electronic collection of firms’ accounting information. Collection Data Mining for Auditing and Creating automated links with other sources to obtain information Investigation on firms during audits. Non Filer Management Systems to identify firms defaulting on their tax obligations, using data from firms that do file and historical information. Identification of Potential Systems to identify new taxpayers using data from other firms. Taxpayers Management Information Systems Integrated systems that share information about taxpayers’ commercial activities with other agencies and public sector organizations. Internal Audit Systems Automated systems to checked tax collectors’ work. State Budgeting Synchronized and on-line budgeting for all the governmental agencies in each State. Source: STABs, IDB.
122 Chapter 5. New Technology
Graph 9. On-Line Budgeting in Selected States.
On-Line Budgeting
120%
100% BA
80% CE PB PE 60% RN SP 40%
% of Agencies Budget OL Budget % of Agencies 20%
0%
7 8 9 3 4 99 99 00 199 1 1 2000 2001 2002 200 2
Source: IDB.
Graph 10. STAB On-line Connectedness (for data sharing) with Other Agencies in Selected States.
STAB Linkages With Other Agencies
120%
100% BA 80% CE PB PE 60% RN SP 40%
20% % of Agencies Linked On-Lne % of Agencies
0%
97 98 99 00 01 02 03 04 19 19 19 20 20 20 20 20
Source: IDB.
123 Chapter 5. New Technology
Graph 11. On-Line Validation in Selected States.
On-Line Validation
120%
100%
BA 80% CE PB 60%
line PE RN 40% SP
20%
Percentage of REturns validated on 0%
1997 1998 1999 2000 2001 2002 2003 2004 Year
Source: IDB.
All of these actions led to an integrated electronic collection system, which in turn fed into a permanently maintained taxpayer record and asset information system. This system also allowed the tax agencies to run automated assessments, collect taxpayer accounting data, and mine data for firm audits and investigations. In addition, the STABs implemented systems to manage archives of non-filers that could be used to quickly retrieve information and identify potential taxpayers (SEFAZ 2001). All of this data was integrated into consolidated STAB management information systems. Given the significant correlation found between DebtJudged, DebtOutstanding, and online validation and budgeting (chapter 3), it appears that computerization improved taxpayer data collection and led to better collection practices, even of previous taxpayer debts.
5.7 Connecting the Dots in the States
Several trends occurred across the public sector with respect to IT implementation. The idea of taking advantage of a central computer sending and receiving information from peripheral units had been in existence since the 1960s, when most of the agencies used mainframe computers. However, the flow
124 Chapter 5. New Technology
of information only came full circle when the peripheral equipment and information became available to the lower level tax collectors.
Along with the standardization of procedures, computerization allowed for the timely processing of information from a growing number of taxpayers who assess their own taxes. One of the most important programs created in the states was the SIAFEM (Integrated Financial Management System for the Public Sector), which was implemented in all states with the support of the federal government through the Ministry of Finance. In 1998, the program covered all budgeting and financial transactions at 903 units in São Paulo state. The units were distributed among the state’s legislative, judiciary, executive, and independent agencies. The operations involved 2,500 decentralized workstations and 6,000 users (tax collectors and other public officials) (Nascimento Júnior 2000; Soboll 1998).
According to those who coordinated the implementation of this technology,:
”[t]he impacts of this modernization process brought about changes in working methodologies and structures, banished bureaucracy and cultural patterns and established a new paradigm for information transparency and the management of public expenditures [and tax collection] in São Paulo state. SIAFEM/SP is also stimulating the improvement of further software facilities such as the Integrated Financial and Management System (SIAFISICO), the Graphic Data Entries System (SIAFACIL), and the Public Expenditure Data Management System (SIGEO).” (Soboll 1998)
This connectivity not only impacted the physical structure of the STABs, but also affected their organizational configurations. Changing organizational processes involved teamwork, which according to Soboll, was “linked to team building and individual empowerment.”
“We even changed the floor plan of the STAB. We knocked down walls, eliminated paper files, and shortened the distance between offices. We saved 431,000 square feet and consolidated the STAB on to only one floor so that we could reestablish visual, physical, and verbal connections among the the internal auditing staff.” (Soboll, interview STAB/SP)
125 Chapter 5. New Technology
Figure 9. São Paulo STABS’ On-Line Connections with Internal and Outside Agencies.
POPULINA
GUARANI INDIAPORÃ MESÓPOLIS D'OESTE MIRA ESTRELA STA. RIFAINA ALBERTINA RIOLÂNDIA Secretaria da Fazenda STA. CLARA OUROESTE PARANAPUÃ CARDOSO PAULO DE FARIA D'OESTE TURMALINA MACEDÔNIA IGARAPAVA STA. RITA DOLCINÓPOLIS ARAMINA D'OESTE PONTES RUBINÉIA ASPÁSIA GESTAL VITÓRIA ORINDIÚVA BRASIL PEDRANÓPOLIS MIGUELÓPOLIS TRÊS URÂNIA BURITIZAL PEDREGULHO FRONTEIRAS SANTA ÁLVARES SALETE FERNANDÓPOLIS FLORENCE COLÔMBIA STA. FÉ SANTANA JALES ESTRELA PARISI AMÉRICO GUAÍRA SEFAZNET DO SUL DA PALE ST INA D'OESTE DE CAMPOS IT UVERAVA PONTE PENSA MERIDIANO JERIQUARA SÃO ICEM CRISTAIS NOVA CA NAÃ FRANCISCO SÃO JOÃO PAULIS TA PAULISTA DAS ILHA SOLTEIRA VA LENTI M IPUA DUAS PONTES RIBEIRÃO PA LMEI RA DIRCE GENTIL GUARACI REIS CORRENTE APARECIDA D'OESTE PONTALINDA COSMORAMA D'OESTE VOTUPORANGA GUARÁ SÃ O JO ÃO NOVA GRANADA MARINÓPOLIS BARRETOS FRANCA SUZANÁPOLIS DE IRACEMA ALTAIR SÃO JOAQUIM TANABI MIRASSOLÂNDIA DA BARRA SÃO JOSÉ MAGDA DA BELA VISTA PE REI RA BARRET O ONDA VERDE JABORANDI ITAPURA SEBASTIANÓPOLIS GUZOLÂNDIA GENERAL FLOREAL RESTINGA DO SUL IPIGUÁ ITIRAPUÃ SUD MENUCCI SALGADO MORRO AGUDO Fazenda AURIFLAMA NHANDEARA OLÍMPIA ORLÃNDIA NUPORANGA MONTE BALSÁMO COLINA POLONI APRAZÍVEL GUAPIAÇU PATROCÍNIO NOVA SÃO JOSÉ TERRA ROXA PAULISTA CASTILHO GASTÃO MACAUBAL DO RIO PRETO SEVERÍNIA MIRASSOL ANDRADINA VIDIGAL SALES OLIVEIRA MONÇÕES SANTO ANTONIO NOVA LUZITÂNIA NEVES PAULISTA CAJOBI BATATAIS UNIÃO VIRADOURO DO ARACANGUA NIPOÃ CEDRAL MONTE AZUL PAULISTA BADY TABAPUÃ PAULI STA LOURDES UCHÔA CASTILHO TURIUBA JACI BASSITT EMBAÚBA BEBEDOURO JARDINÓPOLIS Velocidade atual das Linhas PLANALTO NOVAIS PONTAL PARAÍ SO PITANGUE IRAS ALTINÓPOLIS MURUTINGA BURITAMA NOVA IBIRÁ CATIGUÁ DO SUL ALIANÇA TAQUARAL BRODÓSQUI PALMARES JOSÉ POTIRENDABA PIRANGI ARAÇATUBA BONIFÁCIO PAULIS TA STO. ANTONIO CATANDUVA TAIA ÇU TAIUVA NOVA MIRANDÓPOLIS BREJO ZACARIAS SERTÃOZINHO DA ALEGRIA INDEPENDÊNCIA ALEGRE ELISIÁRIO MENDONÇA VISTA ALEGRE URUPÊS ARIRANHA DO ALTO RIBEIRÃO PAULI CÉI A LAVÍNIA SERRANA SÃO JOÃO GUARAÇAI PINDORAMA JABOTICABAL BARRINHA PRETO DO PAU D'ALHO GUARARAPES BIRIGUI SANTA CRUZ CÁSSIA DOS IRAPUÃ DUMONT MARAPOAMA FERNANDO MONTE ALTO DA ESPERANÇA COQUEIROS MONTE VALPARAÍSO UBARANA BARBOSA PRESTES CASTELO ADOLFO STA. CAJURU NOVA SERRA AZUL COROADOS SALES ITAJOBI ADÉLIA CANDIDO PRADÓPOLIS GUATAPORANGA BE NTO RODRIGUES CRAVINHOS STA. DE ABREU GLICÉRIO GUARIBA MERCEDES BILAC TAQUARITINGA TUPI PAULISTA JUNQUEIRÓPOLIS RUBIÁCEA AVANHANDAVA STA. PENÁPOLIS GUATAPARÁ SABINO NOVO HORIZONTE ERNESTINA SÃO SIMÃO GABRIEL TAPIRATIBA PACAEMBU STA. ROSA MOCOCA PANORAMA MONTEIRO MOTUCA ITÁPOLIS DOBRADA DE VITERBO OURO IRAPURU BRAÚNA LUÍS ANTONIO CACONDE VERDE ADAMANTINA DRACENA PI ACAT U CLEMENTINA GUAIÇARA SA LMO RÃO RINCÃO TAMBAÚ SÃO JOSÉ FLÓRIDA ALTO ALEGRE BORBOREMA PROMISSÃO MATÃO DO RIO PARDO PAULIS TA LINS STA. LUCIA SA NT ÓPOLI S LUISIÂNIA STA. RI TA DO DIVINOLÂNDIA LUCÉLIA RINÓPOLIS DO AGUAPEÍ AMÉRICO PASSA QUATRO 64Kbps CAIUÁ TABATINGA FLORA RICA OSVALDO PO NGAI BRASILIENSE ITOBI CRUZ SÃO SEBASTIÃO PRESIDENTE ARCO ÍRIS CAFELÃNDIA NOVA EUROPA INÚBIA GETULINA IBITINGA CASA BRANCA DA GRAMA VENCESLAU EMILIANÓPOLIS MARIÁPOLIS PAULIS TA QUEIRÓS URU IACRI GAVIÃO PEIXTO ARARAQUARA PORTO PRACINHA STA. CRUZ VARGEM GRANDE FERREIRA ST O. PARAPUÃ DAS PALMEIRAS DO SUL < 64Kbps EXPEDITO SAGRES GUAIMBÉ BALBINOS IACANGA DESCALVADO ÁGUAS PIQUEROBI GUARANTÃ BOA ESPERANÇA DA PRATA REGINÓPOLIS HERCULÂNDIA ITAJU DO SUL CAIABU TUPÃ SÃO CARLOS STO. ALFREDO BASTOS JÚLIO MESQUITA IBATÉ ANASTÁCIO PIRAJUI SÃO JOÃO MARCONDES POMPÉIA TRABIJU PIRASSUNUNGA DA BOA VISTA AG UAÍ PRESIDENTE AREALVA BARIRI RIBEIRÃO BONITO PRESIDENTE ÁLVARO DE CARVALHO PRESIDENTE BE RNARDE S MARTINÓPOLIS BOCAINA STA. CRUZ QUINTANA DOURADO ANALÂNDIA STO. ANTONIO EPITÁCIO PRUDENTE PRESIDENTE DA CONCEICÃO ALVES BORACÉIA DO JARDIM MARABÁ PAULISTA INDIANA ORIENTE LEME AVAÍ QUATÁ GARCA MOJI-GUAÇU ESPIRITO SANTO ALVARES MARÍLIA RANCHARIA VERA CRUZ CORUMBATAI ESTIVA DO PINHAL MACHADO ITAPUI GERBI JOÃO BORÁ BAURU BROTAS REGENTE FEIJO GÁLIA PEDERNEIRAS ITIRAPINA RAMALHO OSCAR BRESSANE JAÚ LUTÉCIA CONCHAL FERNÃO DOIS CÓRREGOS ARARAS RIO CLARO ANHUMAS DUARTINA TEODORO SAMPAIO MIRANTE DO TARABAI LUPÉRCIO PARANAPANEMA MINEIROS IPEUNA OCAUÇU PIRATININGA TORRINHA ITAPIRA DO TIETÊ MOGI MIRIM ECHAPORÃ ALVINLÃNDIA BARRA ÁGUAS DE CABRÁLIA STA. CORDEIRÓPOLIS PARAGUAÇU PAULI STA LUCIANÓPOLIS BONITA ENGENHEIRO LINDÓIA ESTRELA PAULISTA MACATUBA GERTRUDES LAVRINHAS CHARQUEADA COELHO DO NORTE IGARAÇU LINDÓIA ROSANA EUCLIDES DA QUELUZ TACIBA UBIRAJARA DO TIETÊ STA. MARIA SÃO PEDRO IRACEMÁPOLIS ARTUR NOGUEIRA CUNHA PAULISTA SA NDOVA LI NA PA ULISTÂNIA AGUDOS CRUZEIRO CAMPOS NOVOS DA SERRA S. ANTONIO SERRA NEGRA LIMEIRA SOCORRO AS SIS PLATINA PAULI STA AGUAS DE DA POSSE PIQUETE PIRAPOZINHO NANTES HOLAMBRA NARA NDIBA MARACAI LENCÓIS SÃO PEDRO ESPIRITO SANTO AREIÓPOLIS COSMÓPOLIS SÃO PEDRO BOREBI PAULI STAS MONTE ALEGRE IEPÊ DO TURVO CACHOEIRA DO TURVO AM PARO DO SUL CAMPOS SÃO MANUEL JAGUARIUNA SÃ O BE NTO PAULISTA AREIAS SÃO JOSÉ DO RIBEIRÃO AMERICANA DO JORDÃO PIRACICABA DO SAPUCAÍ CANAS BARREIRO ARAPEÍ CRUZÁLIA TARUMÃ DO SUL PAULÍNI A PEDREIRA SILVEIRAS CANDIDO MOTA PINHALZINHO BA NANAL NOVA ODESSA PEDRA BELA GUARATINGUETÁ ANHEMBI S. BARBARA PEDRINHAS IBIRAREMA ÁGUAS DE SANTA IARAS PRATÂNIA PALMI TAL SANTA CRUZ RIO DAS D'OESTE SANTO ANTONIO POTIM PAULISTA SUMARÉ CAMPINAS TUIUTI DO BÁRBARA PEDRAS DO PINHAL LORENA FLORÍNIA SALTO GRANDE RIO PARDO SALTINHO HORTOLÂNDIA PINDAMONHANGABA MORUNGABA BOTUCATU VARGEM MOMBUCA MONTE MOR OURINHOS ÓLEO MONTEIRO ROSEIRA CONCHAS JOANÓPOLIS TREMEMBÉ LARANJAL CAPIVARI BRAGANÇA PAULISTA LOBATO APARECIDA PA ULISTA VALI NHO S CANITAR BERNARDINO ITATIBA DE CAMPOS MANDURI RAFARD CERQUEIRA ELIAS FAUSTO VINHEDO CHAVANTES CESAR TIETÊ PIRACAIA IPAUÇU AVARÉ PARDINHO JUMIRIM LOUVE IRA LAGOINHA CUNHA PEREIRAS INDAIATUBA ITUPEVA ATIB AIA BOFETE JARINU SÃO JOSÉ TAUBATÉ ARANDU ITATINGA BOM JESUS CAÇAPAVA PIRAJU PORANGABA CERQUILHO SALTO JUNDIAÍ IGARATA DOS CAMPOS TIMBURI DOS PERDÕES CESÁRIO LANGE PORTO FELIZ VARZEA CAMPO LIMPO NAZARÉ SARUTAIÁ TORRE PAULI STA PAULISTA PA ULISTA DE SÃO LUIS DO BO IT UVA FRANCISCO PEDRA QUADRA REDENÇÃO PARAITINGA CABREÚVA MORATO MAIRIPORÃ JAMBEIRO ITU SANTA ISABEL DA SERRA FRANCO DA ROCHA TEJUPÁ CAJAMAR JACAREÍ GUAREÍ TAT UI FARTURA PIRAPORA DO CAIEIRAS IPERÓ BOM JESUS ARUJÁ UBATUBA ANGATUBA GUARULHOS PARANAPANEMA CAP ELA ARAÇARÍGUAMA GUARAREMA NATIVIDADE DA SERRA SANTANA SANTA BRANCA TAG UA I DO ALTO SOROCABA ITAQUAQUECETUBA ITAÍ ALUMÍNIO DE PARNAÍBA SÃO PARAIBUNA TAQUARITUBA ARAÇOIABA BARUERI MAIRINQUE OSASCO PO Á ALAMBARI DA SERRA PAULO SÃO ROQUE JANDIRA VOTORANTIM CARAPICUIBA FERRAZ DE CAMPINA DO SALESÓPOLIS BARÃO DE CORONEL TABOÃO VASCONCELOS MOGI DAS ANTONINA MACEDO MONTE ALEGRE DA SERRASÃO CAETANO SUSANO CRUZES SA LTO VARGEM GRANDE BIRITIBA-MIRIM ITAPETININGA SARAPUI PAULISTA DO SUL DE PIRAPORA EMBU MAUÁ RIBEIRÃO CARAGUATATUBA ITAPORA NGA COTIA DIADEMA PIRES ITAPECIRICA RIO GRANDE BURI DA SERRA DA SERRA BERTIOGA PIEDADE SÃO SEBASTIÃO IBIÚNA SANTO ANDRÉ RIVERSUL SÃ O LO URENCO SÃO BERNARDO PILA R DO SUL DA SERRA DO CAMPO ITABERÁ SANTOS CUBATÃO ILHA BELA SÃ O MI GUEL EMBU-GUAÇU ARCANJO TAQUARIVAI SÃO VICE NTE JUQUITIBA CAPÃO BONITO GUARUJÁ ITARARÉ ITAPEVA TAPIRAI PRAIA GRANDE
MONGAGUÁ ITANHAÉM PEDRO DE TOLEDO RIBEIRÃO NOVA CAMPINA GUAPIARA GRANDE JUQUIÁ MIRACATU PE RUÍB E RIBEIRÃO BRANCO SETE BARRAS
BOM SUCESSO ITARIRI DE ITARARÉ
BARRA APIAÍ DO CHAPÉU ELDORA DO REGISTRO IGUAPE
IPORANGA ITAPIRAPUA
RIBEIRA ITAOCA
PARIQUERA-AÇU
CAJATI
JACUPIRANGA
ILHA COMPRIDA BARRA DO TURVO
CANANÉIA
Source: Harumi presentation.
Figure 9 shows the connectivity between São Paulo STAB and local stations across the state. On the top part, it shows that STAB’s connections with local agencies were atomized. On the bottom part, the picture depicts a centralized system, where the central STAB is connected to all the agencies across the state. The connectiviness increased over the last five years. The connection among local stations was further spread and eventually reached to other states. Similarly, before the system connecting RN, PB,
126 Chapter 5. New Technology
and PE was on-line, tax collectors would share diskettes with state information daily in order to keep their systems up-to-date with respect to firms’ transactions (interview in STAB/PB and RN).
5.8 Catching Up with the Economy: Visualizing Linkages and Chains
Although the STABs were enthusiastic about computerization, the sequencing of implementation varied. As explained in the previous chapter, the private sector organizations evolved at a quicker pace than the public sector, and government officers were not able to keep pace with the innovation and complexity caused by technology advancement in firms. Not only did the number of firms and industries to audit increase, but firm structures became increasingly complex, interconnected, and decentralized. Although firms kept management functions at their headquarters, manufacturing and assembly operations were increasingly moved to other locations and states. Suppliers and service providers also relocated. This economic shift made auditing and reconciling tax returns and payment across states more difficult and complicated.
“The production unit reached a level of complexity that was irreconcilable with the state’s organizational and technological needs. The system in place served to increase the inequality between the private and public sector. Therefore, we need to change the system.” (Lobato, SP STAB, interview 2001)
As the complexity of the production structure evolved, the tax collection process needed to shift from a territorial structure to one based on economic segmentation. The territorial model had worked in the past because the authoritarian political system bred during the political dictatorship of the 1960s and 1970s, guaranteed the enforcement of “intrusive” government policies such as auditing. In addition, the network of firms had been simpler, making it easier for tax officers to penetrate the firms’ transaction networks.
As the then coordinator of the segmentation strategy in the São Paulo STAB put it,
“[t]he new organizational model implied that we needed a system to disseminate knowledge that tapped into external resources and connections, that influenced tax collectors’ planning and actions, and that had more territorial flexibility.” (Glauco, tax collector, interview presentation STAB/SP, 1999/2000)
In addition, state government lacked the information technology available to firms. Although most of Tax Agencies already used computer systems to process taxpayer data and to consolidate
127 Chapter 5. New Technology government revenue and expenditures, most of them were running programs on old mainframe computers inherited from the 1960s.
The new modus operandi of economic specialization demanded systems that were capable of providing reliable and accurate information and that would provide instant information for auditors and tax collectors in the field.
The new, computerized processes allowed the central STAB agencies to control tax collection and to determine more precisely which companies and businesses should be investigated. Before, these decisions depended very much at the will of the lower level tax collector. This combination of new organizational and technological procedures raised the credibility of both the STABs and the tax collectors. It centralized tax collection information, it facilitated the coordination of tax collectors’ activities, and it increased tax collectors’ knowledge of collection procedures, collection status, and the economy.
“The system tries to increase compliance by giving appropriate replies to questions by tax administrators, auditors, and collectors about taxpayers. The system standardizes the information about the taxpayers and updates it each time it is used. It facilitates the identification of taxpayers and the amount of tax due. Through periodic crosschecks it spots tax evaders. The most important achievement of the new procedure is to include in one computerized record all the information on a taxpayer. The national database has all the information of the new tax management procedure system described above. The database contains the actual record of payment to government and reminds taxpayer when payment is due. The national database contains about 82 million taxpayers.” (Panzarini, tax collector, STAB/SP 2001)
An instrumental part of the reform then was to provide fiscal auditors with laptop computers and other equipment so that they could track the financial and fiscal life of the firm from the site where they were working and did not have to go back to headquarters to gather information. In addition to laptop computers, this equipment included mobile stations, or vans, with satellite dishes attached to their roofs. Auditors and collectors could use these vans to travel to firms or to pursue cargo trucks, and allowing them immediate access to information on firm’s production chain.
A tax collector in Rio Grande do Norte proudly reported:
128 Chapter 5. New Technology
“We were doing a road blitz and this truck stopped and had no documentation on its load. We had all the information at hand. Who he was, for what company he worked, and what kind of product the truck was carrying. Now with powerfully equipped vans and rooftop satellite dishes…we could tell everything that was happening and how to argue… that is how we are now acting.” (Interview from Rio Grande do Norte)
Although the tax collector noted that pursuing cars and merchandise trucks was part of the job before, it was far more worthwhile to pursue them with information on hand .
Once IT systems had been implemented and organizational process had been changed, it became even more important to learn about the firms’ performance. Tax collectors began working in teams to evaluate each industry’s performance and to audit each sector, replacing the previous territorialized, individual process.
Another factor to consider when examining computerization is the administrative capacity of the organization being computerized. Peterson analyzes the degree of administrative capacity to institutionalize computerization for the analysis of public sector management (Peterson 1996; 1997).44
Computerization also resulted in updated records for the largest taxpayers files (those few companies who are responsible for 50%, 70%, and sometimes 90% of state VAT tax revenues). These companies were the initial tax collection targets in states such as São Paulo, Bahia, and Pernambuco, and computerization greatly assisted efforts to monitor these firms.
In order to encourage computerization, the BID and the Ministry of Finance created benchmarks and competitive deadlines for the 27 states. The first benchmark occurred in 2000 and the last occurred in 2005. The growth in quantity and sophistication of services available to the taxpayers during this period is remarkable. By 2005, 15 of the STABs provided more than 20 out the 40 potential taxpayer services.
Information technology teams developed websites to help STABs communicate both internally (with tax collectors) and externally (with taxpayers). Services provided by these websites included updated and downloadable regulations and procedures, tax consultation services, and the ability to communicate with a tax ombudsman.
44 The frameworks of the OECD administrations are among the largest and most complex systems in the public sector and have few comparable systems in the private sector. Tax systems affect most aspects of the economy: trade, employment, investments, and consumption.
129 Chapter 5. New Technology
5.9 Resistance, Teamwork, and Alliances
Introducing IT in the public bureaucracies of developing countries that have limited resources, weak management, and unskilled staff is even more difficult when there is both individual and organizational resistance. Such resistance can range from fear that technology could take over and cause public employees to lose perks, to fear of increased governmental surveillance. Information systems are difficult to implement for at least three reasons:
1. individuals may believe that they will disrupt established operating procedures; 2. IT reform requires good management, which is scarce in the bureaucracies of developing countries; and 3. IT professionals who implement new systems are insensitive to organizational constraints (Keen 1991). Some assume that public officials will implement administrative reforms strictly according to the blueprints of the reform plans, without resistance or questioning. Others, however, believe that public officials need to be forced to join these efforts through restructuring.
As described above, computerization affects the public and private sector differently. For the public sector, it changes employees’ working conditions and transforms operations—altering sequencing, patterns, and pace. For the private sector, computerization creates new rules and procedures for taxpayers that make them more visible to law enforcement.
Computerization and new equipment may not necessarily cause organizations to adopt new practices. For example, tax collectors could be reluctant to use computers and could be afraid that computers could alter their work routines or even replace them entirely. The literature on modernization and computerization is filled with reports of organizational (union) and individual resistance to change in the public sector (Cornfield, McCammon et al. 1996; Freeman 1996; Freeman and Ichniowski 1988; Freeman and Medoff 1984).
One reform pursued has been to find ways to tax a larger number to taxpayers with a minimal number of tax collectors (Interview with Dr. Ednilton Soarez). Although the STABs were unable to formally dismiss public employees, several strategies were used to reallocate human resources without incurring further expenses. In São Paulo, for example, 500 public accountants were reassigned to other public agencies once the budget was automated. Other strategies to reduce excess staff included voluntary early retirement programs and the reassessment of employee functions and performance, in order to
130 Chapter 5. New Technology
assign old employees new functions. The STABs had to maintain the allegiance of their employees while implementing these changes, however.
Senior staff frequently do not clearly understand the role and limitations of IT in solving their needs. One way to develop this understanding is to send senior staff to courses on management of IT and IS (Peterson 1996). Unfortunately, there are no shortcuts or substitutes for experience and management of IT systems is learned mostly “by doing.” Although Keen notes that “IT reform should not burden senior staff because the technology will not reduce staff or "delay" the bureaucracy,” (Keen 1991), management of IT requires experience and understanding. Keen states that IT reform introduces new relationships within the organization. To be accepted, an IT reform has to be both unobtrusive and a complete solution.
Although few, if any tax collectors were dismissed, in any of the six states I visited the pace and onset of information technology differed. Although staff in Rio Grande do Norte was reduced between 1994 and 1998 (from 519 tax collectors to 479, from 358 “borrowed” staff from other agencies to 261, and from 180 politically appointed positions to 98), most of the tax collector positions were reduced through retirements, the borrowed staff returning to their original positions at the state bank, or the political appointments leaving when new officials were elected (interview with Ludemilson RN STAB Chief of Staff, 2001). However, he also stated that the STAB had just hired 75 tax collectors who had passed the newly prepared entrance examination:
“We do not necessarily have to cut people. Our goal is not to do without people. It is to be efficient, and if we need more staff we will open more positions and hire them. However, the former senior cadre of professionals was old, was not in tune with our new principles, and was about to retire, anyway. The change in staff was better for both sides.” (Ubirajara, head of the RN STAB)
Public bureaucracies, their unions and other associations are often considered to have a diverse repertoire of practices to stall changes in the workplace. This is especially true for restructurings: employees and their unions fear layoffs, replacement of workers, and the demise of existing interest groups and alliances (enclaves and cliques). Public officials also fear losing privileges granted by their positions. In this instance, however, policy makers and theorists were surprised to find that tax collectors did not fear or complain about the new technology.
131 Chapter 5. New Technology
Tax departments are high-intensity systems, are extremely vulnerable to fraud, and have great technical complexity. Vulnerable and complex, high-intensity systems require the deepest organizational change, which threatens staff. In addition, the development of tax department IT systems requires an understanding of taxation and accounting. As a result, it is extremely difficult to keep such organizational reforms on schedule. In spite of this, the STABs’ reforms—especially those related to computerization— were implemented in a timely fashion. The reason has already been stated: “it was unavoidable this time around.”
IT reform in tax collection takes time because it requires building new and important relationships between the tax collectors and those implementing the reform and between the tax collectors and the taxpayers. Building trust requires time. In a classic study in 1974, Gibson and Nolan, as cited in (Peterson 1996), discuss that trust building should be included in the early stages of IT adoption, even if resources are scarce. In São Paulo, Pernambuco, and Rio Grande do Norte, trust among tax collectors was established because reforms were implemented by a committee of their own peers. The joint-state committee held frequent discussions ranging from new reporting requirements to the integration of the different IT systems throughout the STABs. This structure also required that the government staff members be allowed to perform their primary work first, and then fit in the IT reforms. Because the technical staff was part of the larger group, it was not intrusive. As computerization of the STABs developed, the new technology changed the relationship between the STAB leaders and the lower-level tax collectors, fostering the development of the new tax administration model discussed in Chapter 4.
Unlike most Brazilian state and local government agencies, the STABs—especially those in São Paulo and Rio Grande do Norte—decided to hire tax collectors with backgrounds in information technology so that they could manage their own information systems and not have to rely on outside agencies.45
“[T]he alternative was to introduce the information technology from within each STAB, instead of having an outsider provider. We tried to have a vertical control of all phases of the tax collection process.”46 (Guaragna 2000).
Initially, the Tax Bureau in Rio Grande do Norte hired 10 tax collectors with computer programming backgrounds. They developed a software program to manage and accompany the “Fiscal
45 For example, STABs in São Paulo, Rio Grande do Norte, and Bahia canceled their contracts with PRODESP, the State Data Processing Company, and created their own teams of IT specialists (interviews). 46 Interview with Seco Pereira in Gestão da Informação na Administração Fiscal, MF-UCP-PNAFE (PNAFE N.D.).
132 Chapter 5. New Technology
Action,” i.e. audit, assess, etc. This software, developed for the Internet, enabled tax collectors auditing a firm to obtain a historical overview of each and every fiscal transaction related to that firm (including all notifications, permissions, taxpayer account updates, pending administrative processes, and judicial processes) over the past five years, in chronological order.
The computerization of tax procedures purged the system of undesirable intermediaries, such as non-certified accountants and despachantes, by making their roles redundant. Because the accountant unions had a vested interest in eliminating these intermediaries, they played a key role in disseminating on-line procedures such as the on-line VAT return. A true symbiosis appears to have occurred between the accountants’ unions and the state government, intensifying their relationship and establishing new roles for unions in publicizing governmental procedures and in delivering public services.
A computerized national database helped to create public awareness about the effectiveness of the fight against tax fraud. This national database allowed taxpayer returns to be compared with information from third parties, dramatically increasing the total information collected.
“Eventually, [the national database] increased visibility and credibility for the tax collectors’ actions.” (Glauco presentation to STAB/SP)
5.10 Conclusion: Information Systems in Developing Country Bureaucracies
Although the tax administration reform’s new processes might have been effective without technological updates, as Silvani suggested in an interview, they did not occur prior to the computerization of the entire system. IT reform is a solution to an organizational problem. However, an IT reform is not a computer recipe. Trust, need, help, and urgency determine whether the government decision-makers accept that a problem exists and that technology offers a solution.
The three stages of the computerization process—design, adoption, and institutionalization— were greatly supported by organizational reforms. The design stage includes determining information needs and developing a user-friendly computer program to process information and produce reports. During the adoption stage, the system is placed in operation concurrent with the existing manual systems. During the adoption stage in the computerization of the STABs, the technical staff worked closely with tax staff to ensure that the system performed as expected. This allowed for institutionalization of the technological change. According to Peterson (1996), when computer systems for government operations
133 Chapter 5. New Technology are fully managed and operated by government staff, implementation is guaranteed. Implementing an effective information system is never easy. It requires a complex process of reordering relations in an organization. The more intense a system reform is, the more widespread the organizational reordering and the greater the difficulty will be.
In sum, the STABs’ experience with the implementation of modernization reforms suggests that success is influenced by how well public agencies have managed the technological and organizational aspects of the process. This implies that a different approach—one that integrates technology, organizational aspects, and rational processes and procedures—might be preferable. The individual circumstances that each STAB faced, such as local resistance, low performance, and pressures for lower taxes from business associations, affected its management of the reform process. The circumstances under which public sector agencies computerize may hinder or reinforce their ability to reform using technological change.
The next chapter analyzes how organizational and technological change has affected the STABs’ professionalization and identity.
134
Chapter 6
The Dilemmas of Tax Collectors (TC) and How They Learn a New Way to Increase Revenue Collection
“He who has a why to live for can bear almost any how.” (Nietzsche)
“There is nothing in the world, I venture to say, that would so effectively help one to survive even the worst conditions, as the knowledge that there is a meaning in his life.” (Argyris)
6.1 Introduction
Who is the Tax Collector? What motivates a person to become a Tax Collector? How does one get to be a Tax Collector in Brazil? How does society perceive the Tax Collector? How do Tax Collectors see themselves? Does being a Tax Collector make one a public servant or the personification of an administrative evil? Several interviewees were very stern when they described how people, family, and friends perceived their profession.
By and large, Tax Collectors are perceived as the evil hand of the state; the hand that enforces one of the most undesired and functions of a state that is more and more perceived as a coercive and abusive force which charges and punishes people without giving anything back.
135 Chapter 6. Learning
From the point of view of the citizens, Tax Collectors might be the least accepted of government functionaries. While their mission is to levy and collect taxes, and to apply penalties, many people feel Tax Collectors work for their own benefit and embezzle whatever is in front of them. As such, TCs may be seen as the street-level corrupt hand of state bureaucracy.
There are certainly monetary and social reasons for individuals to become Tax Collectors, such as good wages, job security, and end-of-life full-salary pensions, or hidden, such as bribes, empowerment, economic prestige, and social status. But are there more reasons as well? Have they ever been imbued with a sense of public good in their mission?
This chapter addresses the dilemmas faced by TCs and taxpayers alike when dealing with the everyday tasks of tax collection. It describes the TCs’ working conditions and career development, as well their obligations, aspirations, and missions. At the end, it discusses the latest change in the profession brought about by administrative reforms. It also relates how professional sense-making and identity change are related to the new organizational structure, i.e., vertical industry segmentation. As seen in the previous chapter, technological shift was crucial to guaranteeing the change in the tax collection rationale. The emphasis is on the way individuals learned the new knowledge, i.e., the organizational rationale, how this new knowledge propagated among TCs, leading to individual empowerment and instilling a sense of professionalism and identity.
6.2 The Job
“I am a foot soldier in the most feared, hated, and maligned agency in the federal government. I work for the Treasury. (…) I collect taxes, but do not call me a TC. Call me what the Service calls me. Call me a revenue officer.” (Yancey 2004)
“The four protocols: Find where they are; Track what they do; Learn what they have; Execute what they fear.” (Yancey 2004)
Given that the pay is quite rewarding, Tax Collectors are quite well-off as compared to the rest of the society (Table 14). They are very well educated and wield discretionary power equivalent to that of
136 Chapter 6. Learning
the police. There is also a public resentment against the profession and the daily life of a Tax Collector can be quite rough.
In Brazil, TCs are usually called “treasury agents” or “fiscal auditors.” Although the STABs’ career structure is composed of several fiscal and administrative levels (e.g., Fiscal Agent I to VI in São Paulo), I focus exclusively on the VAT auditors’ and collectors’ occupations. These jobs include visiting, evaluating, and auditing firms; collecting due and delinquent taxes; internal and external general consulting regarding taxation; solving impasses in the Administrative Tax Court; and systems and information managing. TCs can also be assigned cases, called taxpayer delinquent accounts or investigations about firms’ economic performance. The Brazilian National Tax Code and the State VAT Regulation Code describe the occupations and minutiae rules which are quite complex.
To enter the career, candidates must take an entrance examination comprised of several different subjects (quantitative logic, rational logic, math, Portuguese, English, taxation, and accountancy). Usually, competition for such positions is quite intense. In the last entrance examination for the São Paulo STAB, 26,708 candidates tested for 350 entry-level positions.47
Besides the promise of a stable job for life and a good pension, the high salary offered is an equally, if not more, compelling attraction to become a TC. As shown in Table 14, fiscal auditors have one of the highest paid jobs in the public sector in Brazil, equivalent only to careers in the judicial system, such as judges and prosecutors. In the same table, one can observe that the STABs’ wages for TCs sits way above the minimum wage. Where typically a household must earn the equivalent of three times the minimum wage to enter the middle income, a TC makes the equivalent of nine times that the minimum wage in Ceará, fifteen times in Rio Grande do Norte, twenty-one times in Paraíba, and thirty times in Pernambuco and São Paulo. Higher wages, however, do not correlate with better economic performance reflected in the VAT. States with the highest wages are not necessarily the ones performing better. In a regression of 1998 VAT on wages, the results were completely unrelated (Graph 12).
The low r-square of “goodness of fit” shows no correlation between the salaries received by the TCs. Had the data set encompassed years before and during the reform, the inferences would be more accurate. It is, nonetheless, instructive and suggests that public officials, particularly, TCs, might respond to incentives other than money. In chapter 3, the analysis of the regression models actually shows that VAT is responsive to training. Although the two datasets are different and the dependent and
47http://www.centraldeconcursos.com.br/docs/Diversos/ICMS-estatistica.pdf and http://www.centraldeconcursos.com.br/docs/edital/Edital_de_Abertura_de_Inscricoes_FINAL__APROVADO.pdf
137 Chapter 6. Learning
independent variables here are different (tax ratio and tax collector’s wage, respectively), the lack of correlation is evident.
Graph 12. VAT/GDP Ratio Predict by TCs’ Wage.
10.00% 9.00% 8.00% 7.00% 6.00% 5.00% 4.00% y = -1E-07x + 0.0741 3.00% R2 = 1 Tax Efficiency
VAT/GDP 1998 VAT/GDP 2.00% 1.00% 0.00%
0 0 0 0 .0 .0 .0 .0 .00 .00 .00 .00 .00 0 0 0 0 0 0 0 0 ,0 ,0 ,0 ,0 R$ 0.00 0 0 0 0 1 2 3 4 R$ R$ R$ R$ R$ 50,000R$ 60,000R$ 70,000R$ 80,000R$ 90,000 Annual Tax Collector Wage 1998
Source: Author’s elaboration. Data: FENAFISCO and IBGE.
In 1998 the exchange rate for one R$ 1.00 fluctuated between US$ 1.12 and US$ 1.2. The annual minimum wage in Brazil was R$ 1,560.00 (R$ 130.00 per month as of May 1998).48 Therefore, the career’s high salary and generous fringe benefits explains the 80 candidates per opening.
Such a coveted job does not, however, come without heavy duties. The initial working conditions are not always as pleasant as an airy office in developed urban areas. Only the first-placed candidates are eligible for a position in offices of the capital city. Most incoming officers are sent to work in the street, far from home in the countryside, in a face-to-face job that is mostly despised and feared. Ultimately, they are street-level bureaucrats who literally work the streets, to collect the state cut of business’ revenues.
48Source: http://www.portalbrasil.net/salariominimo.htm and, http://www.data360.org/dsg.aspx?Data_Set_Group_Id=59 by Board of Governors of the Federal Reserve. System.
138 Chapter 6. Learning
Table 14.STAB Tax Collectors’ Salaries by Brazilian State. Annual Tax Income Collectors' Annual Tax Collector/ Per Capita Nominal Wage Minimum wage Tax Collectors' States (a) (c) (R$ 1,560) Wage/IPC (b/a) Acre R$ 2,826.57 (nd) (nd) (nd) Alagoas R$ 2,283.01 R$ 73,200.00 46.9 32.1 Amapá R$ 3,564.35 R$ 30,600.00 19.6 8.6 Amazonas R$ 5,990.04 R$ 78,000.00 50.0 13.0 Bahia R$ 3,014.41 R$ 44,928.00 28.8 14.9 Ceará R$ 2,684.16 R$ 14,750.40 9.5 5.5 Distrito Federal R$ 13,052.89 R$ 65,400.00 41.9 5.0 Espírito Santo R$ 5,996.10 R$ 40,476.00 25.9 6.8 Goiás R$ 3,671.66 R$ 36,632.16 23.5 10.0 Maranhão R$ 1,347.81 R$ 20,376.00 13.1 15.1 Mato Grosso R$ 4,958.46 R$ 57,000.00 36.5 11.5 Mato Grosso do Sul R$ 4,307.66 R$ 42,955.32 27.5 10.0 Minas Gerais R$ 5,230.31 R$ 40,560.00 26.0 7.8 Pará R$ 2,697.94 R$ 38,400.00 24.6 14.2 Paraíba R$ 2,163.93 R$ 36,000.00 23.1 16.6 Paraná R$ 6,131.02 R$ 42,360.00 27.2 6.9 Pernambuco R$ 3,295.83 R$ 52,752.00 33.8 16.0 Piauí R$ 1,624.31 R$ 60,063.84 38.5 37.0 Rio de Janeiro R$ 7,354.21 R$ 69,030.00 44.3 9.4 Rio Grande do Norte R$ 2,606.69 R$ 26,569.08 17.0 10.2 Rio Grande do Sul R$ 7,145.08 R$ 51,720.00 33.2 7.2 Rondônia R$ 3,610.80 R$ 48,000.00 30.8 13.3 Roraima R$ 2,861.47 (nd) (nd) (nd) Santa Catarina R$ 6,446.26 R$ 52,941.60 33.9 8.2 São Paulo R$ 9,182.95 R$ 52,116.00 33.4 5.7 Sergipe R$ 2,984.06 R$ 44,616.00 28.6 15.0 Tocantins R$ 1,751.21 R$ 26,400.00 16.9 15.1 BRAZIL R$ 5,647.64 Source: FENAFISCO 1998.
The conditions affecting a TC include:
• Dealing with fearful, hostile, and defensive individuals and organizations; • Working in difficult environments, such as high crime areas or distant fiscal stations; • Dealing with prominent and powerful taxpayers who could employ intimidating measures or subject the Tax Collector’s actions to news media coverage,
139 Chapter 6. Learning
• Working long shifts into the night, often without proper facilities; • Applying complex statutes, regulations, and judicial decisions to complicated situations; and, • Working under intense pressure to resolve delinquent cases within deadlines and to meet regional collection quotas.
The remote locations and poor quality of the facilities also affect the TC’s overall working conditions. Because VAT is also charged on product transportation around the state and from state-to- state, there are state border VAT patrol stations where trucks and vehicles carrying goods for commercialization are supposed to stop for control. Many of these stations are in the corner of the states, in the frontier zones, close to the desert, or in forested regions. Such areas are poor and isolated and the physical conditions of the facilities are meager. Some of the larger STABs, such as the ones in the southern states, have closed their border VAT patrol and rely exclusively on mobile stations (i.e., cars or vans that follow trucks) and firm inspections. However, the poorer states still depend on their fiscal stations. In Brazil, many STABs still depend on the TC going after trucks and vans that are looking for secondary roads and trails so that their load will not be taxed.
“Sometimes we have to follow a truck as if we were police and bandit. And they know the region so much better than we do that it is quite hard to get them.”
“Not only is the fiscal station isolated, but most truck drivers are quite intimidating. They either try to bribe you or to intimidate you. I used to carry a gun to work everyday.” Maura, TC, STAB/RN, 2001.
A female TC told me:
“In the beginning, we would travel two or three hours to get here (to the fiscal station), and then we had to work 7- hour shifts. We are supposed to stay in the station. But there was only one bathroom for both TCs and truck drivers alike and only one bedroom at the back of the station that we had to share with everybody.” Interview with tax collector, Rio Grande do Norte and Paraíba border fiscal stations, 2001.
The administrative reform carried out was quite effective in addressing the physical conditions and capabilities. Computers and digital networks were put in place, weaving information across different governmental agencies (chapter 5). The physical infrastructure was widely renovated, mainly in the poorer states, such as Rio Grande do Norte and Paraíba. In these two states, not only were the
140 Chapter 6. Learning
headquarters of the STABs entirely revamped and refurbished, but the fiscal stations were also rebuilt and mobile stations were created.
“Now we have a decent place to stay overnight. Two years ago, we worked in that cabin across the road. No light, no AC, no privacy any time during the day. Even to get here, we had to hitch with a police patrol. Now we have a nice building, with good light and AC. There are two bedrooms with bathrooms, one for women and one for men. We have computers, so that we can access the data about the load that is being transported.” RN and PB border STABs.
6.3 The Individual
The career is predominantly composed of male officers, although there seems to be no discrimination against women in the STABs’ rank-and-file. Given the precarious conditions of the stations, the danger of the street-level work, and the requirements to work away from the hometown, fewer women than men apply to the position. This trend declined quickly in the last decade, when more women joined the rank-and-file of the STABs and their contingent in the workforce grew significantly.49 Yet, several female interviewees revealed some reluctance to work the streets and expressed a preference for working at desk jobs, such as data analysis, assessment, and consulting. After the first year in the frontline, many women migrate to city or regional offices.
The high pay and the benefits attract the most well-educated and well-trained professionals in society. Economic ups and downs also contribute to the high demand for public employment, especially considering that unemployment rates range between 15% and 35% across the Brazilian states. An auditor concurred:
“After working 16 years for IBM, and having been “outsourced” to work as a contractor, I found myself making one third of my original professional wage. Applying to a stable position was the best, if not my only option.” Newspaper interviews.
In the last entrance examination for TCs in São Paulo, most of the applicants were between 30 and 50 years-old, had several years of professional experience in very diverse fields, held degrees ranging
49 I do not have accurate number on female TCs.
141 Chapter 6. Learning
from mathematics and engineering, were dentists, physicians, historians, accountants, and economists.50 Although the above observation suggests an aging cadre of professionals, it is somewhat illusory; most TCs enter the profession in their mid-30s. Among the interviewees, two graduated from ITA, the Brazilian Aeronautics Technology Institute, regarded as the pinnacle of Brazilian education; there was also a veterinarian, several economists, lawyers and accountants, several computer engineers, and an historian. Given the nature of the skills demanded by the profession, engineers, lawyers, and economists lead in terms of hiring. Most of the interviewees had two college degrees and often they held a graduate degree and specialization courses.
The person who enters the career must have a strong drive and ambition. Once inside, she or he will have to deal not only with the public criticism and lack of sympathy but also with the challenge of the job. The training for the career is also intense and demanding. Considering the nature of the job, ethics must have a central role in the profession. However, high pay, education, training, and insulation within the STAB does not guarantee a well-rounded ethical professional.
When asked whether he had faced corruption in the service, a TC told me:
“Even if you try to be correct it is quite hard. Everybody who works with you expects you to get the dough and distribute it. When I first came to the STAB, I was sent to the streets. It is actually the place that we are close to where things happen and there is more money. Each TC had a partner and we were assigned to visit the firms. Usually, we had a driver and police officer that would come along just in case we need police enforcement. One day I came down to visit a firm and the police officer said to the driver: ‘I will not go in the blitz with this guy, he never gets us anything.” Retired TC, STAB/SP.
Not only does the general public have a tarnished perception of TCs, but their own colleagues and supporting staff share the belief:
“Even my mother who praised me said: ‘From now on, my son, you will be rich. You will get the dough.’ An accolade not for the merit that I had in passing the entry examination
50 Interviews with VAT TC’s applicants in locus at the examination site. São Paulo, April 17-18 2006. Also: http://www.cursoaprovacao.com.br/cms/entrevista.php?cod=1819 http://www.editoraferreira.com.br/publique/cgi/cgilua.exe/sys/start.htm?infoid=2819&sid=7 http://clipping.planejamento.gov.br/Noticias.asp?NOTCod=235374
142 Chapter 6. Learning
and the good salary I would receive, but for the bribes and money that I would have on the side, because that is the way people think about us.” Retired TC, BLF, 2000, SP.
“The worst in the profession is not only the career pressure, the taxpayer offers, or the peer pressure to bribe, it is that even your family thinks that you are corrupt and you have to explain that you are not.” Neil Armstrong de Andrade, 2000, STAB/RN.
“Corruption is so much part of the culture that a friend of mine, who had retired some time ago, every now and then would go around the area where he used to work and ‘collect’ some ‘gifts’ from his ‘friends’ the taxpayers.” Retired TC, São Paulo.
The STAB environment was quite the breeding ground for all sorts of misdeeds and corruption. As noted by Rose-Ackerman and Aidt (Aidt 2003; Rose-Ackerman 1997), corruption is a political and economic issue and usually flags a poorly functioning state. Ineffective states can retard and misdirect economic growth, and international organizations have emphasized the need to control corruption. However, in the STAB, several aspects of the working conditions presented a specific hindrance to change.
A major problem when dealing with taxation is corruption. The World Bank (WB), recommends that governments and organizations seek “to counter this bias (corruption) with a strong focus on economic rationality. The emphasis on technical and economic training within the World Bank's professional staff is the organizational response to these pressures (Rose-Ackerman 1997).”
Very few people sought to enter the profession “in order to have a vocation consistent with ideals of service and sacrifice,” as Lipsky suggests (Lipsky 1980). To maintain a sense of relative altruistic behavior, high standards, and self-monitoring, as proposed by Lipsky (1980), is almost impossible. Power and status are a hidden aspect of the profession that is mostly perceived from inside given the prejudice that surrounds the profession. But once one inside the organization, it is clear that
“The Service offered people like us the one thing that we would find irresistible: a world of practically limitless power, nectar of the goddess to the ineffectual dreamer, for whom life was not a pursuit of happiness, but a struggle for recognition and control. (Yancey 2004)”
“I’m sure it’s occurred to you, Rick, that your job here is unique. It cannot be compared to anything else in the government or even to the private sector collectors. You’re charged with enforcing the most unpopular laws in the country. Everywhere you go,
143 Chapter 6. Learning
people hate and fear you. As a result, the tendency is to withdraw. It’s the agency’s tendency and it’s our personal tendency.” (Yancey 2004)
Lack of professionalism may result from the lack of rewards for productivity and good work, excess bureaucracy, and political influence. In most of Brazilian STABs, there is no common system to reward tax officials, except for the good salaries.51 Recognition, however, does not necessarily come with the better salaries. Professionalization and accountability might be achieved in another venue as suggested in the next sub-sections.
6.4 The Professional
Except for passing the entrance examination and holding a college degree, the professional requirements for a TC are not specified. It goes without saying that the preparations for the exams are quite challenging. Typically. not only do the candidates carry two or more degrees from the top schools in Brazil, but they also carry master and doctoral degrees. This is point of enormous pride among TCs. Consequently, TCs think of themselves as belonging to a privileged category and claim to have exclusive ownership of specialized taxation and fiscal knowledge.
Although well-qualified and with great potential to form an excellent cadre of professionals, professionalization of TCs only occurs on the job. There is no professional education prior to becoming a TC or fiscal auditor. There is no particular aptitude test for the position, no requirement of a call to public service among the candidates. The professionalization of TCs is provided on the job at the STAB, both as a necessity to train the new recruits to deal with the specificities of the new profession and as a response to the demands of administrative reforms. STABs not only created their own training programs and schools, a.k.a., the “Escolas Fazendárias” (Tax Administration Schools - TAS), but also started sending the public officials to train in other states or in the federal tax school, run by the Ministry of Finance. The more developed states, such as São Paulo, Bahia, and Rio Grande do Sul have their own TAS. Frequently, they are responsible for training TCs from other states.52
Very few TCs specifically talked about the social function of taxation. In my interviews, only TCs in Rio Grande do Norte (Interview with N.A., RN/STAB) mentioned the connection between taxation and social welfare. It is worth noting that the RN/STAB stood out as one of the best performing
51 São Paulo STAB used to pay TCs by quotas of work, measured by number of audits carried out in a month. 52 The sending STABs cover the cost of training tax collectors in other states.
144 Chapter 6. Learning
in terms of tax revenue, professional satisfaction, and good connections with the TCs’ association and the business sector.53
The first months’ on-the-job training attempts to imbue TCs with a professional ideal and a consciousness of their public service mission. However, most of the training offered is technical, encompassing Tax Code regulations and procedures. Under the IDB orientation, STABs increased both the number of courses and the number of TCs being re-trained and re-qualified (Table 15). Also, under the modernization push, some STAs reviewed the recruiting process, established stricter public entrance examinations,54 and enforced the rule that only college-graduated candidates qualified for the positions. Once the candidates were selected, the STAs trained the TCs both according to their previous education system and to the demands of the new system.
Training indeed pays off, at least with respect to greater revenue collection to the states. The regression model in chapter 3 shows that for each one percent increase in training, there is a corresponding increase of 0.2% of state GDP value in the VAT (refer to chapter 3). Therefore, states such as São Paulo, whose VAT revenues scored approximately R$ 50 billion in 2005, could have achieved an optimal revenue increase of R$ 85 million (i.e. approximately US$ 43 million).55 This is the equivalent to 330,000 times the national minimum wage or 1% the budget for education in São Paulo, this increase could have represented roughly 150,000 primary school new seats.56 At optimal levels, the gain for the Brazilian public economy could possibly generate approximately US$ 2.5 billion.
What the statistical analysis does not show is the type of training provided at most of the STABs. Analyzing the questionnaires answered by the 27 states between 1997 and 1998, most of the courses offered by the STABs referred to learning technical issues such as, tax administration regulations and procedures, use of computers and software programs, and management skills. A summary of all the courses offered by the STABs yields the results shown in Table 15 which represents an average of 3.2 courses/year/tax collector.
53 The RN/STAB is the only one with an organizational structure solely specialized in tax administration, as explained in chapter 4. 54 Not all STABs had public recruiting examinations before the 1990s. However, all the states studied recruited through public examination. 55 The descriptive statistics show that the range of variation for training goes from 0 to 5.9, i.e., the maximum to be achieved by an optimal utilization of training is 5.9. In the states where there is less training the increment in VAT revenue is consequently smaller than that the in the states that offer several courses. 56 According to UNESCO the cost for one student in primary school in 1997 was US$331 (Wolff and Schiefelbein 2003).
145 Chapter 6. Learning
Professionalization however, is not only about the technical tools of the profession, how to fill in tax forms, to check taxpayers’ information, and to learn Tax Code regulations. Conversely to the idea of “professional fix” which argues that the problem is narrowing the gap between the professional’s service orientation theory and practice, the true solution is creating a sense of vocational ideal (Lipsky 1980). However, intensive on-the-job training, good credentials, and good compensation have not necessarily given TCs a sense of public duty, nor prevented them from taking an extra cut from the lucrative state activities for themselves.
Table 15. Courses Offered at the STABs (between 1997 and 1998). Type of Course Number Staff Members Technical: Tax and Financial Administration 40,000 Human Resources 10,894 Management 8,136 Computer Training 29,582 Source: Questionnaire provided by Ministry of Finance and PNAFE (2004). Author’s compilation.
6.5 New Processes and New Professionals
In the Brazilian STABs, emphasis on technical and training capabilities dominates. However, a few STABs have taken advantage of the diversity, background and specialization among the TCs.
São Paulo, Pernambuco, and Bahia are a few of the STABs that were able to take advantage of the TC previous professional background. By introducing the economic segmentation collection system, a domino effect took place: first, STABs taught their TCs general economic processes and instructed them in the procedures of revenue collection.
Second, once technical skills were acquired, emphasis was placed on understanding the production chain process in a few specific economic industry segments.
Third, they identified individual’s knowledge of specific sectors or industries. For example, a chemist who had passed the examination was sent to evaluate the pharmaceutical segment in São Paulo. In Bahia, an agricultural engineer was placed in the ‘Wholesale Food Segments.’ In Pernambuco, production engineers were visiting factories. Specialized skills and a practical knowledge of industry operations were fundamental to setting up teams that could cooperate to identify under-performing industries.
146 Chapter 6. Learning
“Once we were auditing a Pharmaceutical company and the manager turned to us and said that the supplies that he was using were not taxable, therefore he should be exempt from paying tax and had not included his manufacturing yields in the tax return. We called our auditor who had a degree in Chemical Engineering and asked him to have a look at the projects. Once he checked the firms’ documents and receipts, he confirmed that a few ingredients of the product were exempt, but the combined substance resulting from the formula was not and that the taxpayer was supposed to collect taxes. He also mentioned that it was clear in the regulation and that the taxpayer did not want to pay for parts of the value-added in this manufacture. However, if that TC had not been there, we would not have argued with the owner for lack of knowledge of the exact process and products derived. When the second TC, the specialized one, audited the firm, the owner had to admit that we were right.” Pernambuco, Glauco.
Fourth, and most importantly, the knowledge generated in such practices led the TCs to feel recognized and valued as distinctive professionals, equivalent to consultants in private sector firms, such as “Mckinsey or Andersen Consulting (now Accenture).” (Interview).
Fifth, they started working in teams. Flexible specialization actually led to teamwork and organized the “smarts” as Cohen and Zysman suggest (Cohen and Zysman 1988; Zysman 1994). For that and for problem solving, the dependence on employees’ skills was critical. This change in the perception of the TC only came about because of the changes in the organizational processes explained in chapter 4. However, further change would not have been possible without the specialized training of these professionals in each industry segment. Concurrently, computerization served as the enabling and enacting tool (as explained in chapter 5); it added the needed element of skill and cooperation to the diagrams proposed in chapters 4 and 5 (see Figure 10).
Vertical segmentation translated into specialization (i.e., economic segmentation) and computerization, which in turn demanded better professionals. The “Escolas Fazendárias” (Tax Administration Schools) multiplied their courses and the number of TCs being trained and re-qualified. As Zuboff (1988) has shown, the impact of technology on skill demand is more productive if employees are given responsibility for using the technology intelligently. Vertical segmentation combined with computerization allowed TCs to make just such an impact.
Specialization and technological innovation demanded training and efforts to increase capacity. The more professionalized, valued, and better skilled the tax official, the stronger the commitment to the
147 Chapter 6. Learning
reforms and the more efficient the tax collection. Although professionalization and employee empowerment are parameters of the New Public Administration model, it seems that these parameters are not always equally developed in the restructuring and reform equation. State government leaders, as the general public, also perceive most tax officials as greedy, rent-seeking public servants, if not downright corrupt. Relevant training and better career opportunities are important factors in making public officials feel committed to the new system and in obtaining better results in terms of tax collection and compliance. This in turn seems to increase the agency’s credibility.
Figure 10. Integrated Diagram.
new rationale
teamwork
STAB
The change in the STABs suggests a knowledge transformation well beyond the technical training provided by the courses at the TASs. The functional specialization brought about by the vertical industry assessment changes described in chapter 4 led to a different understanding of the tax collection process. It changed the way TCs processed information and delivered services; it also altered the cognitive process of tax collection. The way this happened is explained in the following section.
148 Chapter 6. Learning
6.6 Everything is Illuminated: The Rationale for a Rational Change
Clearly, the economic progress is linked to knowledge and specialization. Adam Smith, in his work about the wealth of nations discusses the division of labor among workers.
“The greatest improvements in the productive powers of labor, and the greater part of the skill, dexterity, and judgment with which it is anywhere directed or applied, seem to have been the effects of the division of labor.” (Smith 1776)
A classic definition for functional specialization, from Piore and Sabel (1984), is that it is a key strategy which emerged in the second industrial divide, where industries became more dependent on multi-skilled workers as a precondition for a more agile production and greater effectiveness. Central to the idea of specialized productive clusters is the understanding that collaborative relations between workers and owners emerge in networks of small and medium-sized firms, that cooperation rather then competition between firms is based upon flexible, computerized machinery, and that this adds up to skill- intensive work (Stern 1992). The dependence of specialization on knowledge ties the division of labor to economic progress, since progress depends on growth in human capital and technologies. Becker (Becker 1976; Becker 1964) notes that the productivity of specialists at particular tasks depend on how much knowledge they have.
Given these assumptions, we draw on another set of discussions to illuminate the understanding of the ways TCs learned and absorbed their role in the context of revenue collection. The literature on cognitive sense-making and adult education helps shed light on the segmentation rationale of some of the STABs visited. This section tries to explain how specialization (industry segmentation) led to knowledge formation and transformation. Additionally, it explores how this specialization has implications for workers’ learning, identity formation, and team organization in public sector bureaucracies.
In cognitive analysis, scientists such as Lave and Wenger (Lave 1988; Wenger 1998) explore how learning depend on the “situated learning” condition and suggest that learning through the work process itself may, in general, be the best way to acquire work-related knowledge and skill. Scrbiner (Wagner and Sternberg 1984) says
“Skilled practical thinking incorporates features of the task environment (people, things, information) into the problem solving system. It is as valid to describe the environment as part of the problem-solving system as it is to observe that problem solving occurs ‘in’ the environment… [It] emphasizes the inextricability of task from environment, and the
149 Chapter 6. Learning
continual interplay between internal representations and operations and external reality throughout the course of the problem solving activity.”
In support of this position, Piore argues that it is the institutional arrangement within which each agent operates that imposes the behavioral pattern (Piore 1995). Piore also points out that although technology is useful, it is the causal understanding which enables us to convert resources into goods and services. Similarly, we could say that although the technological advances in the STABs were necessary and enabling, it is the new rationale in the environment and reality surrounding the realm of tax collection that determines the way TCs perform and perceive their work and themselves. As Piore says, “it does not for that matter, explain how we identify the resources that serve as a means, or the goods and services that are the ends, or how to distinguish means and ends.” (Piore 1995). But it does explain how they implemented the change from territorial to economic segmentation.
Thus, how can we promote knowledge57 transformation and consolidation? The TCs in the Rio Grande do Norte, Bahia, Pernambuco, and São Paulo STABs involved in the new tax collection process seemed to have gone through a learning process, that recalls what Freire argued for as a way to teach illiterates adults (Freire and Macedo 1998; Freire 1993; 1997; 2007).
Freire constructed a methodology “to acquire literacy.” He explained that the learning process
“is more than to psychologically and mechanically dominate by reading and writing techniques. It is to dominate these techniques in terms of consciousness; to understand what one reads and to write what one understands; it is to communicate graphically. [It is] rather an attitude of creation and re-creation, a self-transformation producing a stance of intervention in one’s context.”
There has been some criticism that the Freirean method should not be separated from its ideology of empowerment of oppressed classes. To apply Freire’s method to the empowerment of tax collectors, an already powerful public sector category, would seem a violation. However, the application of the Freirean method allows the recognition of a learning process that works across class divisions and
57 Throughout this dissertation I refer to the idea of knowledge. I refer to knowledge as to what is explicitly verbalized, written, or articulated in any way. Polanyi (1966) describes the importance of the tacit dimension of knowledge which defines and gives meaning to its complementary explicit dimension. Given that my work describes the known structural knowledge of tax collection, I am not addressing the cognitive process prior to the finding of my explicit knowledge: the vertical assessment of industries. The knowledge I describe here fits into the definition proposed by (Hedlung 1994) and (Weiss 1998): cognitive rationalized knowledge as a mental construct.
150 Chapter 6. Learning
intellectual dominions. Also, its use here remains applied in adult education.58 Notice the similar features of TC education reform to the manner in which Freire worked with illiterate adults in Brazil:
1. research of elements with which the group is working; 2. selection of the processes in which these elements are immersed; 3. codification or pattern establishment; and, 4. recognition of the process’ rationale. The new paradigm for the tax collection process, which I observed at the STABs, was indeed a process of self-recreation. The testimonies all emphasized empowerment and the knowledge that emerged from the functional specialization pattern and the input-output methodologies in the economic process once graphically grasped and visualized.
When I asked my interviewees in São Paulo and Pernambuco why they had not used the segmentation method before, considering that it was so particular to the economic process, they were very emphatic in saying:
“We could not visualize it. We did not have the tools for applying and understanding the system properly.” Nelson Fagundes, São Paulo STAB and similarly, Neil, Pernambuco STAB.
How did it work? Starting from the strong, basic rationale designed by Hirschman, addressed before (chapter 4), notice how the economic structures of backward and forward linkages pinned down the reality of industry segmentation in the STABs. It allowed for a schematic and structural representation of the world surrounding the TC. The TC’s perception of their roles changed and, subsequently, so did their reality. Workers came to understand their job and their internal connection with the exterior economy. In this sense, the TCs learned their specialized roles and that these new roles could be changed as the external circumstances required. Hirschman’s backward and forward linkages as well as the production chain rationale provided the fundamental epistemological connection for the TCs.
Learning involves developing new ways of understanding. It involves the acquisition and interpretation of knowledge (Lindsay and Norman 1977). Even if the rationale for the tax collection process was not predictable, once the TCs visualized it with the new IT tools available (addressed in
58 Adult learning theory is usually studied as a continuing professional development. To understand adult education theory, one should go back to John Dewey, who put into context the relationship among experience, interaction, and reflection, as elements of learning. Later on, Houles (1980) defines education as a direct participation in the events of life.
151 Chapter 6. Learning
chapter 5), they learned and understood the process. This changed the TCs’ behavior. Assuming that learning is the process of modifying one’s “cognitive maps or understanding” (Friedlander, 1983, p. 194) and patterns of individual and organizational learning, thereby changing the range of one’s potential behaviors, the vertical industry assessment paved the way for modifying TCs’ cognitive maps and understanding of their work (Huber 1991; Huber and Glick 1993).
Finally, a new self-awareness, discovered and acted upon increased, individual self-esteem. As Argyris explains:
“Manifestations of self-esteem are the predispositions to enlarge the awareness of one’s self and others and to enlarge the acceptance of self and others (Argyris).”
This new understanding and increased self-esteem led to a change in identity, an identity shaped by the social interplay of individuals and their realities, as proposed by (Berger and Luckmann 1966). Berger and Luckmann note that, “theories of identity are always embedded in a more general interpretation of reality” (1966). This reality-oriented identity reinterpretation can be recognized in the case of the TCs in Brazil. The reality of the new tax collection method allowed them to be compared to private sector consultants. The well-structured nature of the new vertical tax collection process, contributing to the change in understanding and knowledge of TCs, is predicted by Giddens, in his work about structuration. According to Giddens (1984), “structure is not ‘external’ to individuals,” rather, it is embedded in the recursive practices and routines of human action. The rationalization or intentionality of these practices involves a constant and reflexive monitoring, which later is reciprocated by other actors and systems, in this case, the Brazilian taxpayers, through social and systemic interaction. Reciprocation, through systemic interaction, entails the construction and solidification of the new identity.
6.7 Accountability and Alliances
6.7.1 Public Sector Credibility and Alliances with the Private Sector To change an identity, learning needs to be reciprocated in both interacting actors so that there is a recognition of such a change by the parties. In the case of tax administration in the Brazilian states, the realms of interaction were the private and the public sector. Reciprocation came from taxpayers and business’ associations.
Alliances with these business associations were quite important, although less prominent. Credibility and support from business associations resulted because of the new perception that the
152 Chapter 6. Learning
government had accurate information about their businesses and knowledge about the economy and individuals performance. From there, new ties were created and the business sector started discussing and “collaborating” with the government on the implementation of new procedures. From my interviews, a quote caught my attention:
“Before we could not visualize these events. We knew of some firms underreporting, cheating, or not filing, but we could not see the whole picture. Now we can visualize (with the computers and programs that we created) what is going on in the economy, in each sector, and we can control [the process]. It allows us to be selective, by targeting each firm without losing touch with the whole economy. For that, each TC specializes in one sector: TCs are almost as specialized and professionalized as private consultants. When business entrepreneurs come to us now, we show them what we know and that we know (our tables and our information). [Now] they know that we know [stuff]. We are right, they respect us.” (Interview with Alexandre, TC, Rio Grande do Norte’s STAB).
The desirable result of these alliances would be that the private and public sector would eventually walk hand in hand in the search for an optimal development policy, both by balancing public finances and by exchanging information on the market economy. However, in terms of taxation, the current relationship between the public and private sectors is more of a suspicious cooperation.
Some of the business’ associations that approached the government and questioned their policies with respect to taxation were representatives of the big firms. When the Pernambuco STAB started taking action in order to assess the restaurant production chain, the president of the Restaurant Owners’ Association decided that they had to act on it. He stated:
“If we were to be audited and assessed on a daily basis like that and we had no chance to escape, we decided that we would work ‘together’ with the government. I talked to the coordinator of my segment and told him that we needed to do something to tax the informal sector as well. Otherwise, they would be taxing, but not doing justice. … We are developing a policy to account not only for the big restaurants, but also for the small, invisible ones and even the informal sector.” President of Restaurants’ Owners’ Association, Pernambuco.
In another interview, the owner of a small retail business complained frustrated:
153 Chapter 6. Learning
“There is nothing new about their policy. They are just coming after us all and somehow they are much rougher now than before and they do not want to hear our complaints.”
While there was no unanimity about the quality and the outcomes of such alliances, it was clear that there was an increased respect for the Tax Collector. Another business sector entrepreneur told me:
“It is not that we are really collaborating with each other. But we are talking now.”
However, there was also cynicism on the side of tax collectors:
“It is not that we are helping out or consulting for the private sector. We are showing them that we know our business. They have to abide by the law. We have the data on them and that is it.”
Despite skepticism from some interviewees, the volunteer transfer of information among sectors became more fluid and business associations spearheaded courses about the new structure of the tax collection, organizing events where the STAB representatives were featured speakers.
Alliances and reciprocation between the public and private sectors have long been explored in the literature on economic development, the state, and the market. Amsden, for example, demonstrates how Korean economic development is highly related to the level of reciprocity between the public and private sector (Amsden 1989), when she explains how standard performance, monitoring, and government sanctions actually bridge the public-private divide. Schneider suggests that one should distinguish among performance standards, monitoring, and sanctions (Schneider 1998). In the case of the tax administration bureaus, it was clear the setting of standards, as the rules and regulations established by the tax codes, as well as the sanctions, were plainly established in the Tax Code. The monitoring part, as describe in the chapter 4, was less evident, but once unveiled it shows the way to understanding the industry performance and to connect with the private sector.
In monitoring, explains (Schneider 1998), “bureaucrats have to have access to information to allow them to determine if firms are complying.” If firms manipulate the information, the government will not be able to make assessments about the firms’ economic activity. What helped the STAB in collecting information about the firms was the advantage gained in the reform. The STAB was able to collect accurate data about the firms with the new IT systems and evaluate its economic transactions. Both the information gathering and evaluation followed the rationale of the production chain and backward and forward linkages presented earlier (chapter 4). To share this knowledge both within the
154 Chapter 6. Learning
STAB and outside did not demand require much more effort, once the right machine was in place (chapter 5).
As in the study by Amsden (1989), discipline was a key factor for empowering the STAB. This discipline came from a top-down determination by the head of the STAB and then via the police force of the state. Reciprocity derived out of a desire by the private sector not to affront the Tax Collectors, partly out of fear of the new force in the tax administration. As in Amsden’s account, the tension between discipline and reciprocity was present in the relationship between the state and the private sector.
The study of reciprocity is closely related to the study of state embeddedness (Evans 1995; 1996). Embeddedness allows officials to “count on the private sector for effective implementation.” In the past, TCs counting on private sector brought more than skepticism, it brought scorn. Yet due to the changes in the way Tax Collectors operated “the business sector could see the change” and “know what we know. So they respect us.” Therefore, the vision of a common ground—the economic segmentation and industry functioning understanding—brought together incompatible “allies” that in a few cases led to joint projects, as in the case of the fiscal education to businesses’ owners, and in others, to some sort of consensus building, as in the restaurant sector in Pernambuco.
A final comment about reciprocity: given the nature of the change put in place by the reform—the vertical assessment of firms—and the need to relate with the exterior, by way of capturing and understanding the economic relations, the STABs had to open up, contrary to the idea that professional communities tend to self-sealing groupings (Ferlie, Fitzgerald et al. 2005). Typically, when professional groupings develop a distinctive knowledge base, the information becomes less fluid and permeable and does not allow for fluid participation (Wenger 1998). However, in this case, the common structure and rationale was shared, i.e., common cognitive and the instrumental boundaries, the interaction between groups was facilitated.
6.7.2 Resistance and Cooperation with Workers’ Unions and Associations Conversely to the assumed resistance to change touted in the literature on unions and associations and industrial relations, the TCs, through promoting fiscal education, were able to pave the way to reach taxpayers and the broader population.
Contrary to the belief that public sector unions are disruptive to the reform process, in many states the Tax Collectors’ Unions joined the training effort by creating programs to disseminate tax
155 Chapter 6. Learning
education.59 Equally important was the role that the Accountants’ Unions played. Since accountants are responsible for filing the tax return for all firms, these unions had active participation in the reform. In fact, Accountants’ Unions’ representatives claimed that the computerization of tax procedures purged the system of undesirable intermediaries (e.g., non-certified accountants and despachantes, i.e., accountants with no license, intermediaries in the process of tax return, usually providing cheaper services and preferred by smaller firms) and facilitated their work. Those unions played a key role in mediating the implementation and dissemination of online procedures, such as the online VAT return. A true symbiosis occurred between the Accountants’ Unions and the state government, improving their relationship and establishing new roles for unions in the public service dissemination and delivery60
“We have always collaborated with the STAB. Without us they would not be able to reach the firms. After all, we are the ones who prepare the tax return, co-sign it, and answer questions about it. Some of us have even worked for the STAB in the past, so that we know the subject and the threads to do the work,” explained an accountant at the São Paulo Accountants’ Unions.
Confirming the previous statement and the literature recommendation, the most significant new alliance was with the accountants’ associations (Bird 2003). Since VAT is collected from firms and firms usually work with an accountant to file their tax returns, reaching out to educate the taxpayers and pass the new technology and ideas happened mostly through the Accountants’ Associations. The STAB trained the accountants so that they would work out the change with the firms.
“Lately, we have been involved with training. We had a joint seminar about the new online tax return form and what the new instructions are. The seminar was for accountants. But afterwards we also open it to the society in general.” Accountants’ Unions, interview, São Paulo.
Both in São Paulo, and Rio Grande do Norte, as well as at the national level (i.e., at the FENAFISCO, the National Federation of State Tax Collectors), TCs’ Associations were quite involved in the economic process and very aware of the benefits and predicaments of reforms.
59 In São Paulo, the Union maintains a TV channel with tax education programs of diverse sort (Interview with Union president). 60 Bird (2003) comments on the necessity and practicality of involving private accountants in the work of the tax administration agencies for educational purposes.
156 Chapter 6. Learning
The TCs’ association in São Paulo not only had publications and courses for the general public and the Tax Collectors alike, but also created a TV program for fiscal education.
In Ceará, Paraíba, and Bahia, the TCs’ Associations were not as involved in the reform process. In Ceará and Pernambuco, TCs’ Associations were pushing for work stoppage and better salaries. The strongest antagonism between the union and the STAB was in Ceará (interviews and field trip). The Ceará TCs’ Association organized a strike and there was a lot of discontentment with STAB management and reforms, even from high ranks in the government.
“Despite our initial cooperation with the reform, we are not appreciated. The governor and the head of the bureau refuse to receive us and even mock us. They are very intransigent and they do not see that we could come to agreement. They talk about reform and how we must cooperate and cut cost. They are cutting costs; they have not hired new TCs since the 1980s. We need more resources and more people and we have not received a raise since the reform started in our state, which is to say since 1987. We are some of the lowest paid TCs in Brazil.” Interview with TC who asked to not be named.
There were also strikes in Pernambuco. However, there was less antagonism, despite being displeased with the overall situation and fighting for better wages. Some of these strikes seemed to concern lower administrative ranks in the STABs rather than the TCs.
The set of institutions supporting the state government implementing reform affects the government tax administration efficiency. Successful reform is dependent on whether the government is capable of establishing a network of support and institutional arrangements within diverse sectors of society such as labor institutions and professional organizations, political institutions, and firms and business’ associations. This also helps change the public image of a government agency.
6.8 Conclusions: The New Professional and the New Identity
Public sector workers, like all other workers, will perform better if they understand why things happen. Horizontal ties will develop, teamwork, cooperation and identity will improve, and connections among the workers will form when they see the logic vertical linkages of their work.
The conclusions in this chapter are two-pronged. First, we learned that using a similar to adult education method proposed by Freire, a public servant’s learning can be replicated and taught. The
157 Chapter 6. Learning
second part of this chapter emphasize that the kind of knowledge to be ingrained in people’s minds and behaviors, i.e., a structural kind of knowledge, the backward and forward linkages, with beginnings and ends, with structure, and causations. This latter conclusion corroborates the findings in chapter 4.
The turning point seems to reside in understanding the way public sector employees learn technical capabilities in order to maneuver their knowledge into effective action; and to meet the challenge of the modernization. Thus, functional specialization of TCs activities enabled by computerization, led to industry segmentation and a better understanding of the economy and the TCs role in society. Rethinking their internal organization and specialization, allowed them the awareness of the entire process. To have the best wages in the country, to have power, and even some local recognition was not enough to change their attitude. What was more effective was a broader understanding of their function as viewed through an economic rationale.
Summing up, the case of Brazilian STABs modernization reforms suggests that the predicaments and constraints that surround this public agency’s decisions are influenced by how TCs have managed technological and organizational aspects of the process in the past. It also suggests that a different approach ought to be in place: one that integrates technology, organizational aspects, institutional partners, and integrates the new knowledge about their work environment and societal roles.
Of course, high pay, better working conditions, recognition, and training play an enormously significant role in the success of the modernization of the STABs. Also, the circumstances that each STAB faces, such as local resistance and low performance on the one hand, as well as pressures for lower taxes from business associations on the other hand, may have hindered or reinforced the STAB’s management of reform procedures. As Taliércio states “non-institutional reforms, such as merit-based recruitment drive could be easily subject to political reversals and would therefore not be credible. Such reforms, though observable by taxpayers, do not establish the institutional context necessary to convince taxpayers to change their behavior.” However, according to my case studies, without the change in the sector rationale the dissemination of such knowledge and the reciprocity from the private sector, STABs would not be able to establish the necessary context to convince Tax Collectors and taxpayers alike of the new image and identity for the TC, and consequently for the STAB.
As Taliércio comments, basic reforms aim to create a realm of credibility for the government. This was true in most of the STABs, mainly in RN, SP, and BA, where fiscal education were implemented with full help and cooperation of the workers’ associations.
158 Chapter 6. Learning
In this chapter I made the argument that to accomplish more effective, competent, and fair tax administration reforms, STABs needed to employ reliable Tax Collectors and implement a system that works well. A well-oiled, functioning state apparatuses could happen if Tax Collectors were empowered in their own technicalities and capabilities. However, an education process was crucial to gain the taxpayers’ commitment. Taxpayers partaking in the reform, a usually opposing force against tax collection, closed the circle for attaining a credible commitment.
159
Chapter 7
Conclusion
“There are very few human beings who receive the truth, complete and staggering, by instant illumination. Most of them acquire it fragment by fragment, on a small scale, by successive developments, cellularly, like a laborious mosaic.”
Anais Nin
7.1 What is the Puzzle?
What is the core puzzle in this dissertation? My initial motivation to take on this research was to understand how to improve government bureaucracies so that they can improve the efficiency and effectiveness of their services. My interest in the public sector’s institutional capacity was heightened by my annoyance as a taxpayer and by my concern with the financial difficulties that government faces to deliver services. In my reflections, I found that I continued to return to the same questions: what would a good tax administration structure and process, capable of raising revenues and paying for better services without a significantly increasing my taxes, look like? Given the ongoing misdeeds of the public sector, my hope to find a workable answer was remote. Furthermore, since the nature of this investigation is interdisciplinary, possible solutions were blurred by the fuzziness of the environment in which this question is embedded: that of multiple and iterative processes of inquiry. Does the answer lie in organizational restructuring, in technological change, or in learning processes? Is it about imposing the
161 Chapter 7. Conclusion forceful hand of the state or about creating compliance consciousness in taxpayers and motivation in public officials (tax collectors)? How does one connect the dots?
7.2 Axiomatic Model
This dissertation constructs an argument for bridging together the above issues with a common thread, without losing the quality of the parts. I believe argument is that there is an axis around which issues concerning public sector reform, individual learning and transformation, and technological enactment in the public sector all revolve. This axis is the cognitive content of organizational restructuring, illustrated by the experience of the Brazilian State Tax Administration Bureaus, i.e., the change from a territorial to a vertical production chain assessment rationale. This axis “represents the teleological explanation for the change in process within an entity.” (Van de Ven 2004). In this sense, the Barnett and Carroll’s distinction between the content and the process of change made by is very useful: “Content refers to what actually change in the organizational entity, while process examines how the change occurs” (Barnett and Carroll 1995) This idea is depicted in Figure 11, below.
Figure 11. Cognition Axis Assessment Model: New Cognition Embedded in Technological and Organizational Change =>New (Tax Administration Process)=> New Ties and Identity.
New Institutional Image, New ties & New Cognition: Vertical Industry Assesment Identities
162 Chapter 7. Conclusion
Since organizations and people engage around this axis, this approach is not just about process. It is about content, too. In the case of the STAB’s VAT collection, the axis was the vertical industry assessment. This is what I call a “structured knowledge axis” for action.
My intention is to offer an alternative frame, not new per se, but novel in terms of application; this approach may be applied across different sectors and theories in order to explain organizational change, especially, in the public sector. I believe that content of change matters as much as its enabling circumstances (both organization and technological). Although these ideas need the organizational apparatuses and the technological engine to have a transformative impact, it is the content that is likely to engage the individual. This content is a logical one, one that fits Hirschman’s formulation of economic linkages.
7.3 The Lasting Power of Vintage Ideas
By using Hirschman’s theory of backward and forward linkages as a cognitive axis for the situated learning method, I try to give corpus to the idea of “change content”. In the past, the linkage theory has been particularly relevant in interpreting development experiences and it is intrinsically related to input-output economic modeling and to the industrialization process. Linkages refers to “interdependence… to interrelation in a general equilibrium system, where everything depends on everything…(Syrquin 1992),” To Hirschman it is about “the inducement mechanisms at work within sectors of directly productive activities.” (Hirschman 1958; 1963). The success of this framework is that it “seems more operational, less fuzzy [than the alternative concepts]” (Hirschman 1981).
Hirschman also tried to connect this rationale beyond pure economic activities to social and political components that shape economic development. In discussing the theory of linkages, Hirschman explains that it would be possible to underline the feasibility of a structural approach, because linkages provide an analytical tool—an operational device (Hirschman 1984).
The structural knowledge axis is not static; it is flexible and interacts with the environment, organizations, and technology. This process is iterative, rich, and alive. Individuals propel it and it is constantly renovated, and as a process, it resonated. The structural knowledge axis thus resonates with Giddens’ theories of structuration (1984).
163 Chapter 7. Conclusion
7.4 Empirical Results: Tax Administration, Reforms and Recommendations
Using the most straightforward measure of success the maximization of tax revenue and collection, the tax administration reform in the Brazilian STABs was a profound success—even in states with small percentage increases (Table 16). But this should not be the only criteria for success in any reform (Bird 1991; 2003; Bird and Casanegra de Janstcher 1992). I found that successful STAB reform was also dependent on whether the government was capable of establishing a network of support and institutional arrangements with diverse sectors in society, such as other state agencies, labor institutions and professional organizations, individual firms, and business’ associations. The set of institutions supporting the state government implementing reform affects government tax administration efficiency (World Bank 1994). This study shows that public sector bureaucracies that improve their relationship with the private sector also have a more user-oriented posture and offer better working conditions for their employees. This in turn allows STABs to obtain the cooperation of the private sector in implementing reforms and improves tax collectors’ self-esteem, identity, and commitment to the STABs. I propose that if this is possible in an “unfriendly” and antagonized sector, such as taxation, it should definitely possible in more service-oriented sectors, such as education and health.
The restructuring of tax administration around the lines of linkages and the production chain was invaluable to building the capacity of tax collectors, leading to changes in their professional identity. It allowed the STABs to match collectors’ previously acquired skills and better how understand the economies of specific industries. The most valued tax officials, highly professionalized and educated, could have resisted this new organizational paradigm. However, once the tax collectors were able to visualize the new collection model, they demonstrated a high commitment to participating in the reforms, leading to more efficient tax collection. Also, relevant training was an important factor in gaining tax collectors’ commitment to their STABs and to improving tax collection and compliance. As tax collectors began processing information and delivering services using the industry segmentation model, STAB credibility appeared to increase, as well. Vertical industry assessment was, thus, a catalyst for reforms’ success.
My findings in Ceará, Bahia, Pernambuco, São Paulo, Paraíba, and Rio Grande do Norte revealed that the reforms’ results were a function of a process: changing the cognitive understanding and practice of tax collection and changing institutional practices between the private and public sector. Both technological and organizational changes together enabled vertical industry assessment (specialization
164 Chapter 7. Conclusion
and change in the rationale of the tax collection), which, in turn changed the relationship between the public and private sectors, leading to a better tax administration and increased tax collection.
Table 16. Features and Changes in Brazilian STABs: Selected States.
turing c n nt ith io Growth t rme e 5) 0 w dination w T Revenue97-20 Empo omputeriza A oor rivate Sector State Vertical Restru C Training V (19 TC Idenity C P
State STAB Features Outcomes
BA Yes 100% Yes 23% HighElitist CE No 100% Yes 21% Low Weak - Elitist PB No 100% Yes 39% Low Weak PE Yes 100% Yes 35% High/Medium Good RN Yes 100% Yes 102% High Good SP Yes 100% Yes 1% High Medium Source: Author’s elaboration. VAT data from IPEA and computerization from PNAFE – Ministry of Finance.
This analysis suggests that differences in the growth of tax revenue across the states stem not only from local economic development, but also from the way governments adjusted their administrative apparatuses and used new technology. Governments that established a well-defined and appropriate strategy to modify the organizational structures and integrate technology experienced the best performance. For example, Rio Grande do Norte experienced significantly higher and better results than Paraíba, in spite of their being neighboring states with similar economic and territorial features. The RN/STAB, an earlier adopter of the segmentation rationale with a strong administration that aimed the entire universe of registered firms in the state, was by far the best performing state among the six studied. The Pernambuco and Bahia STABs were also early adopters of the vertical assessment model (Bahia was one of its pioneers) and also experienced good results. In contrast, Ceará STAB did not adopt vertical assessment until much later, showed weak efforts to reach out to its employees, and did not develop ties with the business sector beyond the business elites already in power.
Both the organizational management of tax collection and the technology used to facilitate it rotated around this new cognition axis: organizational restructuring, technological change, and the learning experiences constructed to support it. Successful modernization is thus a function of adjustments
165 Chapter 7. Conclusion in the administrative apparatus enabled by the technology (technological continuum). This had a profound effect on both tax collectors and taxpayers. Both of them accepted change and appropriated the new model, which helped to bridge the gap across the public-private divide. After all, as Montgomery states, most of “the information and changes came from the employees who were to be most affected by [the reforms], and who were to be its principal sources of implementation.” (Montgomery 1996)
Before modernization, STABs organized revenue collection horizontally—different taxpayers were aligned together with no connection, and tax collectors worked in a vertical hierarchical structure. Tax collectors work vertically—pieces of the process are connected by a common rationale—and are structured horizontally as a cooperative, although centrally controlled, team.
A true symbiotic relationship emerged between unions and private and public associations, particularly between the tax collectors’ associations, accountants’ unions, governmental agencies, and the private sector—that brought them close together, established new roles for the organizations and for public officials.
The literature on tax administration reform seldom explores whether or how alliances between “tax collectors—taxpayers” can be forged to facilitate the implementation of reforms. Such alliances were quite important in the dissemination of reform efforts in some Brazilian STABs, while these groups remained quite antagonistic in others. In São Paulo and in Rio Grande do Norte, for example, there has been a real synergy developed between business sectors and STABs with respect to fiscal education, dissemination of new practices, and extension of the new rationale to other sectors.
While tax collectors’ interests and taxpayers’ willingness to pay taxes are usually assumed to be intractable, I observed that public sector unions and business associations were strategic interlocutors in implementation of the reform in some STBAs. The comparison of performance among six STABs suggests that a beneficial tax-payer/tax collector relationship can exist that is conducive to tax compliance. These relationships emerged from distinct institutional re-arrangements, such as the new professionalism of tax collectors. Of course, there were fluctuations between periods of closer interaction and periods of antagonism. Yet, the new framework enabled a credible relationship with the taxpayer.
What did it take to legitimize this reform? Why did the tax administration reform follow successful trajectory? How has the process of new institutional and professional behaviors become enacted? Recognition, accreditation, and enactment occurred concomitantly, due to the clarity of the new rationale. Then, it made easy to involve business sector actors.
166 Chapter 7. Conclusion
How did having a new rationale for the sector contribute to these new alliances and modes of cooperation? First, the STABs knew exactly what kind of information to demand and how to organize it. Second, this information provided the necessary foundation for their work with firms. Third, tailored training and situated learning were put in place (Lave 1988; Orlikowski; Wenger 1998). Fourth, the new rationale facilitated teamwork. Instead of sending a couple of tax collectors to audit and collect taxes from a group of unrelated firms in a given territory, teams were created around specific industries.
7.5 Implications for Theory and Methodology
Organizational studies, by and large, deal with structural analysis, examining the centralization, formalization, and communication channels of organizations. In studying bureaucracies, structural analysis tries to sort out the characteristics of design, implementation, the use of technology, institutions, and personnel management. Studies of institutional economics examine arrangements such as laws, regulations, and other cognitive, cultural, or socio-structural constraints. Such studies also widely discuss technology adoption, as mechanical, as a deterministic enabler, or as an enacted, structured device that interacts with both organizations and their environments. One way of looking at this complex set of interactions is to examine at how these different factors and approaches interact, the recurrences of these interactions, the impact across them, and the content of the change per se. Investigating how learning occurs may provide clues to understanding why some changes are persistent and long-lasting and other are ethereal. I suggest that finding the sector linkages is a key to understanding the learning process.
On the other hand, using alternate methodologies to measure and capture different aspects of changes allows us to fill in the gaps about what we currently know. For instance, although the statistical method used here proved to be useful in tracing important features of the tax administration reform, it was not enough to exemplify the minutiae in the process. Although this does not invalidate the methodology, it shows how using different techniques expands the universe of perceptions and understandings. The methods complement each other.
7.6 Implications for Technology Implementation
It is widely accepted that information technology is a means to deliver reforms in a more efficient and accountable way. Information technology not only increases the capacity for control of the organizational elements, but also facilitates the decentralization of functions, resource management, and privatization (Heeks 1998; 1999). In the case of the Brazilian STABs, the technological change proved to
167 Chapter 7. Conclusion be very critical. However, computerization proved to be more than simply regeneration tool (Southern 2001). It was not enough to simply introduce computers and purchase more equipment. The timing of implementation was also crucial: in order for the new technology to be accepted, it was essential the people using it to b3e professionalized and adequately prepared.
The government must have a strategy to reallocate resources away from traditional tasks and toward ones that are more productive. In the tax collection sector, technological innovation and organizational change enabled economic specialization and opened a window for tax collectors to have a fuller view of tax collection and its relationship to the local economy. This is a continuous, circular process, that is only possible with proper technology, management, and workers’ skills. This process requires the participation of those directly affected and responsible for implementing the reform.
Finally, computerization tools for the taxation process were necessary and inevitable. Information technology enabled the transformation of taxation. Government agencies, in turn, also influenced the technology features that were selected, implemented, and used, as explained in the framework of technological enactment.
7.7 Implications and Contributions for Public Sector Management
Although some of my findings in terms of administrative reform, computerization, and the treatment of public employees are analogous to those previously proposed in the literature on New Public Management and developing countries reform (Barzelay 2000; Barzelay and Armajani 1992; Beetham 1996; Bhatnagar 2000; Grindle 1997; Kelly 2007; Kelman 1987; Kettl 1994; 1996; 1998; Kettl and DiIulio 1995; Kettl, Ingraham et al. 1996; Kettl and Milward 1996; Kickert 1997; Tendler 1997; 2001; Tendler and Freedheim 1994), my research differs by proposing that amidst accountability, client orientation, personnel valorization, and so on, it is necessary to recognize the ways individuals understand the functioning of their world. It is not just about mission, charisma, or empowerment. New theoretical categories and dimensions are required in order to fully understand what made the Brazilian STABs’ reforms succeed. In this case, the process of learning—specifically situated learning—created a paradigm shift in service delivery and professional identity.
This study suggests that researchers and policy makers should examine not only institutional, organizational, and technological interactions, but also the structural content and rationale for each sector. It is important to learning the reasoning behind and substance of a project, an organization, or a sector before examining how and why it works well. For example, if one is studying the healthcare sector, what
168 Chapter 7. Conclusion
are the connections and motivations that can be used to intervene? Focusing on these features should help researchers and policy makers to implement successful reforms.
7.8 Future Research
The achievements discussed in this study shows that there is not a monolithic approach to reforms. Neither is the empirical record of success. For example, both IDB evaluators and Brazilian tax administration reform experts have acknowledged that the São Paulo STAB is the most advanced tax administrations in the country. The São Paulo STAB was leader in moving to the vertical industry segmentation paradigm. However, Table 16, above, shows that VAT revenue grew the least in São Paulo. This suggests that other aspects must be evaluated, such as STAB’s ability of the to continue adopting new ideas and trends, their adherence to modernization guidelines, and their ability to maintain their commitment to the public in general and to their employees in particular. In Ceará, results were not good in spite of the excellent STAB initiatives because government leaders were unable to keep up with the employees’ needs and the empowerment of their employee associations. The CE STAB’s entrepreneurial approach (Tendler 1997) and exceptional leadership made it an early model successful reform (Bonfim 1999), but this reform was not sustained throughout all government sectors and the image of tax collectors did not improve. Finally, the Bahia and Pernambuco STABs faced different realities in terms of political environment and momentum. Yet both were early adopters of the reform process, both adopted the new tax collection rationale, and both made similar efforts towards involving their workforce. Their results in the reform process were similar, but not equal, which suggests that further research is needed into aspects such continuity, implementation, and subsequent results of tax administration reform.
Another area for further study is the role of judicial and tax court reforms that appears to have greatly contributed to the success of tax administration reform. These contributions are indicated in the statistical findings but have yet to be analyzed.
Although not discussed in this dissertation, many tax administration projects in other states were also successful. They achieved their goals and improved tax collection primarily by reforming the (1) the collection of information, (2) the organization of tax basis, and (3) the STAB accountability. Since many of these other states were late adopters of policies and strategies, research into their reforms may highlight alternative processes and shortcuts for the implementation of reform goals.
Of course, local context—the institutional arrangements, the complex and abundant legal frameworks, and the organizational politics, among other contextual factors—also affected the STABs’
169 Chapter 7. Conclusion performance. Therefore, this research invites other to study the effects of uneven state capacities at the micro level, unexpected institutional arrangements, and others features of the various state and local governments, to uncover alternative paths, and innovative ways to catalyze public sector reforms.
This dissertation focused on a rationale focused to the features of VAT administration. As a result, future research is needed to examine other types of government taxation. For example, I would like to test the rationale for other taxes at the federal, state, and local levels, particularly since the administrative reforms were implemented across these jurisdictions as well. I also would like to try to map the change content in other sectors of public administration. Studies on isomorphism using the STABs as role models would also be interesting. Finally, further development of the statistical methodology suggested in this dissertation may expose a series of unexplored interactions, correlations, and causations suggesting other incidental or idiosyncratic aspects of the reform process.
In terms of tax collection per se, the new methodology discussed here the door to myriad use of indicators and tests. The advent of computerization makes information current and accessible. This dissertation offers a model of organizational, technological, and cognitive change as a means for understanding successful tax administration reform in Brazil. It is the processes of linkages in this model that may allow policymakers to replicate success in other governmental sectors, and offer researchers new tools.
170
Appendix A
Brazilian Tax Structure: Facts and Figures
This appendix presents generic data on Brazil’s regional and state economies, such as population, GINI coefficient, GDP per capita, etc, tax collection growth and evolution, VAT participation in federal and state GDP, and tax elasticity.
A.1 Brazilian Regional Differences
Graph 13 shows the partition of wealth, in terms of GDP, population, and tax collection per sub- national region across Brazil. It depicts the supremacy of southern states over northern ones, denoting a hugely divided economy between South and North regions.
Comparatively, the participation of VAT on state GDP has also grown on the states analyzed. The sub-national governments of the three less developed macro-regions hold 32% of total disposable tax revenues of those spheres of government or far more than their participation in the Brazilian economy.
171 Appendix A. Brazilian Tax Structure
Graph 13. Regional Population, GDP, and Tax Revenue.
Population GDP Tax Revenue
Centralwest, North, 3.70% Centralwest, 6.80% North, 7.00% Centralwest, North, 4.80% NorthEast, NorthEast, 6.80% 7.20% 12.30% 12.90%
NorthEast, 28.40% South, South, 14.70% Southeast, 16.40% 42.40%
Southeast, 58.80% Southeast, South, 62.60% 14.80% Source: IBGE (J.R.R. Afonso, April 2001 (IBGE 2001)) Brazil Regional Disparities in Social Indicators Source: J. R. Afonso/IBGE. (Regional Partition of Sub-national Revenue JRR Afonso 2001 Brazilian Experience: Consumption and Production Taxes, SF/BNDES).
Table 17 shows the very large disparities in social indicators according to the sub-national regions.
Table 17. Brazil: Regional Disparities in Social Indicators.
Social Indicator Regions North- Center- South- National North South east west east average
Infant mortality 55.0 98.0 41.0 40.0 38.0 59.0 (per thousand; 1989) Years of life 64.5 55.7 67.7 65.3 65.7 62.6 expectancy (1985) Child malnutrition 42.3 46.1 25.7 21.7 17.8 31.0 rate (1989) Rate of elementary 88.7¹ 77.9 84.6 87.1 82.7 83.2 school attendance Rate of first grade 31.3 14.4² 8.3³ 19.8 repetition Illiteracy
Illiteracy rate 19.8¹ 43.3 22.5 16.4 16.2 24.6
Percentage of 80.0¹ 48.4 62.0 84.4 68.4 70.9 dwellings with Source: Affonso (1995). 1/ Urban areas only. 2/ State of São Paulo only. 3/ State of Santa Catarina only.
172 Appendix A. Brazilian Tax Structure
Table 18 provides basic statistics about population and labor by state in Brazil.
Table 18. Brazilian Basic Statistics by State (various years). Income IDB Loan for Workers Total Urban Rural Per Gini Reform Formal in Formal States Population Population Population GDP Capita (1991) HDI (thousands) Firms Firms
Acre 557 526 370 267 187 259 3,241,847.44 5,815 0.61 0,697 USS 4,557 10 903 82 942 Alagoas 2 822 621 1 919 739 902 882 11,556,231.65 4,094 0.6316 0,649 USS 13,662 41 128 373 935 Amapá 477 032 424 683 52 349 35,888,581.05 75,233 0.5677 0,713 USS 4,450 8 232 78 601 Amazonas 2 812 557 2 107 222 705 335 3,720,358.98 1,323 0.6208 0,753 USS 12,180 37 751 419 649 Bahia 13 070 250 8 772 348 4 297 902 86,882,057.01 6,647 0.6533 0,688 USS 15,000 271 428 1 715 595 Ceará 7 430 661 5 315 318 2 115 343 33,260,671.70 4,476 0.6474 0,700 USS 14,784 169 513 1 051 641 Distrito Federal 2 051 146 1 961 499 89 647 43,521,629.02 21,218 0.6089 0,844 USS 10,193 85 784 957 933 Espírito Santo 3 097 232 2 463 049 634 183 34,487,904.52 11,135 0.6115 0,765 USS 156,000 109 341 713 621 Goiás 5 003 228 4 396 645 606 583 41,316,490.56 8,258 0.6197 0,776 USS 16,320 167 232 1 019 131 Maranhão 5 651 475 3 364 070 2 287 405 16,547,449.02 2,928 0.6012 0,636 USS 14,438 65 361 446 673 Mato Grosso 2 504 353 1 987 726 516 627 166,586,326.69 66,519 0.6005 0,773 USS 20,813 91 093 513 969 Mato Grosso do Sul 2 078 001 1 747 106 330 895 19,953,528.62 9,602 0.6136 0,778 USS 21,713 67 151 417 368 Minas Gerais 17 891 494 14 671 828 3 219 666 27,935,498.73 1,561 0.6347 0,773 USS 25,000 661 461 3 934 832 Pará 6 192 307 4 120 693 2 071 614 34,195,676.42 5,522 0.6015 0,723 USS 15,270 71 747 699 151 Paraíba 3 443 825 2 447 212 996 613 14,863,056.88 4,316 0.6381 0,661 USS 14,985 62 325 455 189 Paraná 9 563 458 7 786 084 1 777 374 47,697,442.23 4,987 0.6098 0,705 USS 165,000 449 962 2 535 517 Pernambuco 7 918 344 6 058 249 1 860 095 8,611,415.34 1,088 0.6597 0,656 USS 24,848 142 542 1 152 090 Piauí 2 843 278 1 788 590 1 054 688 108,698,901.36 38,230 0.6298 0,787 USS 10,910 48 769 311 137 Rio de Janeiro 14 391 282 13 821 466 569 816 222,563,502.61 15,465 0.6238 0,807 USS 24,440 416 184 3 593 173 Rio Grande do Norte 2 776 782 2 036 673 740 109 15,906,123.89 5,728 0.6459 0,705 USS 18,870 58 905 475 087 Rio Grande do Sul 10 187 798 8 317 984 1 869 814 9,744,450.65 956 0.5981 0,735 USS 22,977 598 252 2 845 998 Rondônia 1 379 787 884 523 495 264 1,864,150.71 1,351 0,746 USS 6,380 34 325 243 637 Roraima 324 397 247 016 77 381 142,874,226.33 440,430 0.617 0,814 USS 4,280 9 340 40 363 Santa Catarina 5 356 360 4 217 931 1 138 429 70,207,923.77 13,107 0.565 0,822 USS 17,000 302 973 1 732 936 São Paulo 37 032 403 34 592 851 2 439 552 13,120,855.33 354 0.5797 0,682 USS 68,718 1 728 708 11 301 850 Sergipe 1 784 475 1 273 226 511 249 546,606,818.53 306,312 0.6321 0,820 USS 10,140 28 696 283 444 Tocantins 1 157 098 859 961 297 137 4,767,935.52 4,121 0.6283 0,710 USS 11,100 26 821 182 260 BRAZIL 169 799 170 161790311 913,734,000.00 5,381 0.6366 Source: Ministry of Finance. For GDP and Population: IBGE.
A.2 Brazilian Tax Revenue
Brazilian fiscal revenue is responsible for 34.23% of gross domestic product (GDP). This in turn means that approximately 54% of the Brazilian national tax revenue, including federal, state, local taxes, is the result of fiscal efforts (see Graph 14).
173 Appendix A. Brazilian Tax Structure
Graph 14. Participation of Social Contributions and Taxes to total Brazilian GDP. Tax % Participation in total Government Revenue
Federal Taxes 23% Others 7% State Taxes 26% 54% Fiscal Effort
Social Security Contributions Municipal Taxes 40% 4%
Source: IBGE.
Table 19 describes the assignments of the different taxes to the levels of government: Federal, State, and Local.
174 Appendix A. Brazilian Tax Structure
Table 19. Brazil Tax Assignments for Federal, State, and Local levels.
______Responsibility for______Tax Category Definition of base Setting of rate Administration
Foreign trade F F F Corporate income F F F Capital gains F F F Personal income F F F Transfers of property S, L F, S, L S, L Vehicles S S S Property L L L Rural property F F F Payroll F F F Civil servants’ payroll F, S, L F, S, L F, S, L Sales taxes IPI F F F ICMS F, S F, S S ISS F, L F, L L Fees, royalties F, S, L F, S, L F, S, L
Source: Affonso and Raimundo (1995). 1/ F = Federal; S = State; and L = Local. Regional Revenues States Tax Revenue.
Note : VAT:Value Added Tax (by states); IPI: Tax on Industrialized Products (Federal VAT); II: Import Tax; INSS: National Social Security; FGTS: Severance Pay Indemnity Fund; COFINS: Social Security Financing Contribution; CPMF: Provisional Tax (because such tax is to be basically collected on debits to bank accounts as a compulsory contribution); PIS/PASEP: Turnover Tax; IR: Personal Income Tax; CSLL: Social Contribution on Net Profit Tax.
Graph 15 details the distribution of tax revenues and their participation in the GDP, in 2000. It also shows that State-collected taxes in 2000 accounted for 27% of the total tax collection, of which the majority was VAT.
175 Appendix A. Brazilian Tax Structure
Graph 15. State Taxes as a % of Total Tax Revenue and as % GDP (2000).
PIS/PASEP 3% (0.87% GDP) CPMF 5% (1.32% GDP) FGTS 6% (1.72% GDP) ISS 2% (0.58% GDP) COFINS 12% (3.53% GDP) IPTU 2% (0.47% GDP)
VAT 25% (7.55 GDP) State 27% SS 18% (5.11% GDP) OIL TAX 2% (0.49% GDP)
IPI 6% (1.72% GDP) TR 19% (5.48% GDP)
Source: IBGE (Instituto Brasileiro de Geografia e Estatísticas) and Afonso 2001.
A.3 VAT in Brazil by State
Table 20 depicts the VAT growth from 1996 to 2003, by State. The States which have been more detailed studies are highlighted in bold.
176 Appendix A. Brazilian Tax Structure
Table 20. VAT Growth from 1996 to 2003 per Year.61
1995/6 1996/7 1997/8 1998/9 1999/0 2000/1 2001/2 2002/3 2003/4 2004/5 Acre 12.2% 7.5% 47.6% -17.9% 32.4% 10.8% -0.8% 14.2% 9.8% 27.1% Alagoas 7.8% 7.7% 8.7% -16.6% 11.4% 2.1% -9.6% 9.9% 7.4% 14.0% Amazonas 60.1% -28.9% -17.1% -11.6% 16.5% 6.7% -6.8% 4.8% 5.1% 14.4% Amapá 8.2% -1.1% 14.4% -12.0% 32.7% 7.6% -6.5% -0.5% 10.9% 29.1% Bahia 9.5% -5.8% 3.2% -7.1% 12.7% 3.1% -6.2% 7.2% 3.2% 2.8% Ceará 13.7% -2.4% 6.4% -5.4% 9.9% 2.7% -10.0% 1.4% 0.9% 4.7% Distrito Federal 16.0% 5.2% 5.5% -4.5% 15.3% 2.2% -9.4% 13.6% 5.5% 10.9% Espírito Santo -0.6% 9.1% -8.1% -15.0% 17.2% 9.4% -22.2% 13.9% 12.7% 22.4% Goiás 11.9% 1.0% -1.1% -3.7% 11.1% 8.1% -9.7% 15.1% -5.5% 5.4% Maranhão 20.2% -16.4% 5.7% -11.9% 26.4% 15.4% -10.8% -0.011% 9.1% 21.7% Minas Gerais 5.1% -4.0% -2.4% -4.9% 9.0% 8.5% -16.5% 7.1% 6.2% 16.3% Mato Grosso do Sul -1.8% 1.8% -4.5% 12.5% 7.3% 7.2% -12.8% 16.6% 11.9% 13.5% Mato Grosso 2.5% 9.2% -13.3% 13.3% 14.8% -13.3% 8.4% 19.7% 10.1% 4.1% Pará 6.9% -14.3% 14.7% -16.9% 34.8% -1.6% 5.5% 11.0% 3.7% 17.1% Paraíba 16.1% 2.3% 8.7% -8.0% 9.4% 12.7% -19.3% 2.3% 0.7% 14.5% Pernambuco 12.1% 2.0% 5.1% -13.1% 8.7% 1.0% -5.2% 3.1% 3.0% 16.8% Piauí 17.3% 1.0% 3.2% -11.5% 14.3% -0.9% -8.3% -21.3% 47.7% 17.4% Paraná -23.1% -7.6% 0.6% -1.4% 14.6% 39.3% -8.7% 11.2% 3.4% 11.1% Rio de Janeiro 5.7% -7.3% 19.3% -5.3% 2.5% 3.4% -11.8% 8.9% 5.4% 1.7% Rio Grande do Norte 17.2% 11.9% 9.7% -0.4% 17.3% 4.2% -11.3% 8.5% 4.7% 14.7% Rondônia 5.1% 45.5% -12.6% -6.6% 30.3% -4.0% -7.6% 26.0% 7.7% 16.2% Roraima 9.2% 12.2% 19.0% -11.1% 47.3% -1.5% -19.6% 6.9% -3.1% 23.9% Rio Grande do Sul 1.0% -7.6% 4.6% -8.2% 10.8% 7.3% -12.1% 10.7% -3.1% 15.3% Santa Catarina 2.3% -1.7% -2.6% -5.7% 10.5% 8.6% -6.8% 0.1% 12.5% 10.6% Sergipe 4.4% 10.6% -3.3% -7.6% 13.3% 4.5% -6.4% 6.5% 2.8% 15.0% São Paulo 4.7% -0.7% -3.4% -8.2% 9.6% 1.0% -12.8% 0.4% 2.1% 9.5% Tocantins 17.6% 4.0% 14.8% -2.6% 18.8% 11.1% -3.8% 22.9% -1.4% 9.6% Source: IBGE (Instituto Brasileiro de Geografia e Estatísticas). STABs, Ministry of Finance, National Council on Fiscal Policy (CONFAZ), and VAT Permanent Commission (CONFAZ), IPEA. I highlighted the states that I have visited.
Table 21 displays the cumulative VAT growth between 1997 and 2005 per state.
61 I have used nominal values deflated by the GPI-DI, the Brazilian General Consumer Price Index. The nominal values are also provided. All values are deflated according to year base 2006. Growth was calculated as a ratio between the current year tax collection and the previous year tax collection.
177 Appendix A. Brazilian Tax Structure
Table 21. Cumulative VAT Growth between 1997 and 2005.
1995/6 1995/7 1995/8 1995/9 95/00 95/01 95/02 95/03 95/04 95/05 AC Acre 12% 21% 78% 46% 93% 114% 113% 143% 166% 239% AL Alagoas 8% 16% 26% 5% 17% 20% 8% 19% 28% 45% AM Amazonas 60% 14% -6% -17% -3% 4% -3% 1% 6% 22% AP Amapá 8% 7% 22% 8% 43% 54% 44% 43% 59% 105% BA Bahia 10% 3% 6% -1% 11% 15% 8% 16% 19% 23% CE Ceará 14% 11% 18% 12% 23% 26% 13% 15% 16% 21% DF Distrito Federal 16% 22% 29% 23% 42% 45% 31% 49% 57% 74% ES Espírito Santo -1% 8% 0% -15% -1% 9% -15% -4% 9% 33% GO Goiás 12% 13% 12% 8% 20% 29% 17% 34% 27% 34% MA Maranhão 20% 0% 6% -6% 18% 36% 22% 22% 33% 62% MG Minas Gerais 5% 1% -1% -6% 2% 11% -7% -1% 5% 23% MS Mato Grosso do Sul -2% 0% -4% 7% 15% 24% 8% 26% 41% 60% MT Mato Grosso 2% 12% -3% 10% 26% 9% 19% 42% 56% 63% PA Pará 7% -8% 5% -13% 18% 16% 22% 36% 41% 65% PB Paraíba 16% 19% 29% 19% 30% 46% 18% 21% 22% 39% PE Pernambuco 12% 14% 20% 5% 14% 15% 9% 12% 16% 35% PI Piauí 17% 18% 22% 8% 24% 23% 12% -12% 31% 53% PR Paraná -23% -29% -28% -30% -19% 13% 3% 14% 18% 31% RJ Rio de Janeiro 6% -2% 17% 11% 14% 17% 4% 13% 19% 21% RN Rio Grande do Norte 17% 31% 44% 43% 68% 75% 55% 68% 76% 102% RO Rondônia 5% 53% 34% 25% 63% 56% 44% 82% 96% 128% RR Roraima 9% 23% 46% 30% 91% 88% 51% 62% 57% 94% RS Rio Grande do Sul 1% -7% -2% -10% -1% 6% -6% 4% 0% 16% SC Santa Catarina 2% 1% -2% -8% 2% 11% 3% 3% 16% 29% SE Sergipe 4% 16% 12% 3% 17% 22% 14% 22% 25% 44% SP São Paulo 5% 4% 0% -8% 1% 2% -11% -11% -9% 0% TO Tocantins 18% 22% 40% 37% 62% 81% 74% 113% 110% 131% Source: IBGE/IPEA, values deflated by the IGP-DI
Notice that according to Table 20, tax administration agencies have improved, in general, their VAT collection. States such as Amapá and Acre have had an outstanding performance. Their performance, however, reflects the existing underdeveloped situation in both states, which until recently had not even the status of State government. By the same token, São Paulo’s long running and well- developed economy indicates that its apparent low performance is linked to the initial stage from which state departs in the efforts to reform.
To compare the VAT growth with the economic growth, refer to Table 22 which depicts the GDP growth per State from 1994 to 2002.
178 Appendix A. Brazilian Tax Structure
Table 22. GDP Growth per State (1994-2002).62
1995/6 1996/7 1997/8 1998/9 1999/0 2000/1 2001/2 2002/3 2003/4 Acre 5.4% 6.4% 8.8% -10.8% -0.2% 2.0% -6.6% 11.7% 6.2% Alagoas 14.8% 5.5% 4.9% -12.8% -0.3% -2.6% -8.1% 9.4% -0.4% Amazonas 17.1% -4.9% 3.1% -14.3% 10.7% -0.7% -4.2% 4.2% 13.8% Amapá -0.9% 5.9% -3.3% -12.1% 13.4% 3.5% -6.6% 8.0% 7.4% Bahia 12.6% 4.3% 2.9% -9.7% 4.6% -2.0% -5.7% 9.5% 5.7% Ceará 14.4% 4.5% 5.3% -13.8% -2.7% -6.2% -11.0% 9.1% 4.1% Distrito Federal 14.2% 11.9% 24.0% -26.2% 21.3% 1.0% -14.3% -1.7% 2.6% Espírito Santo 6.0% 1.0% 5.4% -10.0% 4.6% -5.4% -12.9% 8.9% 5.9% Goiás 12.3% 2.1% 6.9% -14.4% 10.3% 4.5% -0.8% 9.3% -0.2% Maranhão 24.1% 0.2% -4.2% -8.7% 6.1% 1.1% -11.9% 13.8% 5.3% Minas Gerais 14.1% 3.1% 0.9% -12.8% 3.3% -3.3% -12.3% 7.1% 2.6% Mato Grosso do Sul 8.7% 3.8% 6.3% -9.7% -0.7% 4.7% -11.3% 14.9% -6.4% Mato Grosso 11.6% 7.1% 6.3% -1.6% 4.7% -2.7% -1.8% 17.5% 9.9% Pará 4.8% -1.3% 4.0% -10.8% 3.5% 3.9% -6.8% 6.3% 4.2% Paraíba 12.4% -0.8% 2.1% -9.0% 6.2% 0.5% -10.1% 9.5% -3.5% Pernambuco 12.0% 1.8% 4.1% -12.7% 2.1% -1.5% -8.7% 7.5% 0.4% Piauí 12.2% 0.6% 2.7% -10.7% 2.7% -5.4% -12.2% 10.4% 4.6% Paraná 13.7% 2.9% 5.7% -9.5% -2.5% -0.3% -11.2% 12.9% -2.3% Rio de Janeiro 6.6% 4.6% 1.3% -5.3% 9.9% -2.9% -8.8% 4.0% 4.0% Rio Grande do Norte 13.6% 5.5% 0.9% -7.0% 10.9% -4.3% -6.1% 9.4% 3.3% Rondônia 12.4% 7.2% 8.0% -9.3% 2.2% -2.2% -5.0% 8.3% 2.1% Roraima 6.5% 5.6% 18.1% -8.8% 24.7% -1.3% -3.1% 4.7% -1.1% Rio Grande do Sul 7.8% 1.7% 0.2% -10.9% 3.0% -0.1% -11.9% 13.9% -0.7% Santa Catarina 14.2% 0.6% 0.0% -8.4% 8.5% -0.9% -11.6% 11.5% 0.4% Sergipe 11.3% 4.6% 2.1% -10.1% -0.6% 25.3% -8.1% 14.5% -0.2% São Paulo 8.5% 5.5% 3.2% -12.6% -0.6% -2.3% -13.2% 4.9% -1.7% Tocantins 14.3% 4.3% 11.0% -9.7% 6.2% 13.1% -8.2% 9.8% 1.3% Source: IBGE (Instituto Brasileiro de Geografia e Estatísticas) and IPEA.63
Table 23 depicts the ratio of VAT per GDP by State in Brazil, from 1995 to 2004.
62 To calculate State GDP growth, I deflated nominal values, using the IGP-DI (GPI) (General Consumer Price Index), published by the IBGE (Instituto Brasileiro de Geografia e Estatística). 63 The IBGE has finished compiling data for the State GDP until 2004.
179 Appendix A. Brazilian Tax Structure
Table 23. Ratio of VAT per GDP by State and Coefficients of Variation.
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 AC Acre 3.6% 3.9% 3.9% 5.3% 4.9% 6.5% 7% 8% 8% 8% AL Alagoas 6.94% 6.52% 6.65% 6.89% 6.59% 7.37% 7.72% 7.59% 7.62% 8.22% AM Amazonas 8.31% 11.36% 8.50% 6.83% 7.05% 7.42% 7.97% 7.75% 7.80% 7.20% AP Amapá 3.54% 3.87% 3.61% 4.27% 4.28% 5.01% 5.21% 5.21% 4.80% 4.96% BA Bahia 7.80% 7.59% 6.85% 6.87% 7.07% 7.62% 8.01% 7.97% 7.81% 7.63% CE Ceará 7.61% 7.56% 7.06% 7.14% 7.83% 8.84% 9.68% 9.78% 9.09% 8.81% DF Distrito Federal 4.76% 4.83% 4.54% 3.86% 5.00% 4.75% 4.81% 5.09% 5.88% 6.05% ES Espírito Santo 9.89% 9.27% 10.02% 8.74% 8.26% 9.25% 10.70% 9.56% 10.00% 10.64% GO Goiás 9.56% 9.53% 9.43% 8.72% 9.82% 9.89% 10.22% 9.31% 9.80% 9.27% MA Maranhão 6.63% 6.42% 5.36% 5.91% 5.70% 6.79% 7.76% 7.86% 6.90% 7.15% MG Minas Gerais 7.34% 6.76% 6.30% 6.09% 6.65% 7.01% 7.87% 7.49% 7.50% 7.76% MS Mato Grosso do Sul 8.10% 7.32% 7.17% 6.45% 8.03% 8.68% 8.89% 8.75% 8.87% 10.61% MT Mato Grosso 10.85% 9.97% 10.16% 8.29% 9.54% 10.47% 9.33% 10.30% 10.49% 10.51% PA Pará 5.25% 5.36% 4.65% 5.13% 4.78% 6.22% 5.89% 6.67% 6.96% 6.93% PB Paraíba 6.62% 6.83% 7.05% 7.50% 7.58% 7.81% 8.76% 7.86% 7.35% 7.67% PE Pernambuco 6.77% 6.78% 6.79% 6.86% 6.83% 7.27% 7.46% 7.74% 7.42% 7.61% PI Piauí 6.89% 7.21% 7.23% 7.27% 7.20% 8.01% 8.39% 8.77% 6.25% 8.83% PR Paraná 6.45% 4.36% 3.92% 3.73% 4.06% 4.78% 6.67% 6.85% 6.75% 7.14% RJ Rio de Janeiro 6.07% 6.02% 5.33% 6.28% 6.29% 5.86% 6.24% 6.04% 6.32% 6.41% RN Rio Grande do Norte 6.29% 6.49% 6.89% 7.49% 8.02% 8.48% 9.23% 8.72% 8.65% 8.76% RO Rondônia 6.81% 6.37% 8.65% 7.00% 7.21% 9.19% 9.02% 8.77% 10.21% 10.76% RR Roraima 7.66% 7.86% 8.35% 8.42% 8.21% 9.70% 9.68% 8.04% 8.21% 8.04% RS Rio Grande do Sul 6.54% 6.13% 5.57% 5.81% 5.99% 6.45% 6.92% 6.91% 6.71% 6.55% SC Santa Catarina 7.10% 6.36% 6.22% 6.05% 6.23% 6.35% 6.95% 7.33% 6.58% 7.37% SE Sergipe 7.48% 7.01% 7.42% 7.03% 7.22% 8.23% 6.87% 7.00% 6.50% 6.70% SP São Paulo 8.13% 7.84% 7.38% 6.91% 7.26% 8.00% 8.28% 8.32% 7.96% 8.27% TO Tocantins 8.68% 8.93% 8.90% 9.21% 9.94% 11.12% 10.93% 11.46% 12.82% 12.49%
Standard deviation 0.02 0.02 0.02 0.01 0.02 0.02 0.02 0.01 0.02 0.02 Mean 0.07 0.07 0.07 0.07 0.07 0.08 0.08 0.08 0.08 0.08 Coefficient Variation 0.29 0.29 0.29 0.14 0.29 0.25 0.25 0.13 0.25 0.25
The coefficients of variation shown in Table 23 portray how the VAT efficiency rates have varied over a period of ten years. According to my calculations, the variation is quite small, showing a quite small standard deviation among the states (between 0.01 and 0.02). As for the coefficient of variation, it shows that there is a constant and stable relationship in the VAT efficiency across all states.
Table 24 depicts the nominal values of VAT collected by the States, from 1995 to 2005, in thousands of reais (R$), in current value (corrected for inflation).
180 Appendix A. Brazilian Tax Structure
Table 24. VAT Collected by Brazilian States, from 1995 to 2005, in R$ (thousands), in Nominal Values. g g () 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 AC Acre 12 36279.56 44524.61 51485.41 77282.34 76224.01 110619.1 135615 169482.9 208225.9 256894.2 329757.037 AC AL Alagoas 27 280124 330483.2 382906.7 423403 423889.7 517600.7 584523.5 665566.9 787086.8 949553.5 1093364.669 AL AM Amazonas 13 913659 1600635 1224789 1032229 1096178 1399486 1651686 1939187 2187808 2584452 2985849.341 AM AP Amapá 16 43747 51812.78 55127.53 64162.48 67786.79 98611.62 117379.6 138255.5 148025.1 184405.6 240337.6446 AP BA Bahia 29 2088433 2503153 2537451 2663121 2970475 3670292 4187240 4950531 5712367 6625528 6877126.582 BA CE Ceará 23 950829 1182979 1242053 1344086 1526927 1838435 2089231 2368020 2585224 2930038 3097416.554 CE DF Distrito Fed53 629366 798877.6 903888.2 970418 1113493 1406623 1590928 1816261 2219996 2631363 2945232.93 DF ES Espírito Sa 32 1271968 1382798 1623200 1517876 1550384 1992134 2411699 2364264 2897949 3670195 4535688.795 ES GO Goiás 52 1135574 1390825 1511441 1520270 1759087 2142345 2560977 2914197 3608729 3831706 4078902.958 GO MA Maranhão 21 335698 441464 397030 426665.3 451485.6 625401.7 798496.7 897073 965388.8 1183771 1455053.558 MA MG Minas Gera31 4620148 5314785 5491781 5452013 6230332 7441456 8929614 9397793 10835788 12931306 15184191.14 MG MS Mato Gross50 566515 608766.5 666660.3 647882.1 875385.8 1029978 1221645 1341762 1683460 2117631 2426673.917 MS MT Mato Gross51 706470 792061.2 930555 820665.7 1116662 1405540 1348743 1841745 2372078 2934667 3085602.386 MT PA Pará 15 634416 742160.5 684615.3 798492.2 796606.4 1177133 1281217 1702946 2034436 2370520 2801569.094 PA PB Paraíba 25 352235 447390.1 492639 544646.4 601562.4 721597.1 899732.7 914663 1007196 1139713 1317435.888 PB PE Pernambuc26 1182161 1450380 1592327 1701954 1776714 2116784 2365983 2826721 3135394 3628096 4277778.27 PE PI Piauí 22 219046 281135.2 305511.6 320629.1 340619.6 426821.1 467972.6 540797.4 458021.3 760498.2 901060.8474 PI PR Paraná 41 2474033 2082779 2069974 2118421 2508332 3150977 4854435 5580693 6679675 7763904 8707026.435 PR RJ Rio de Jane33 4513979 5220055 5208807 6322927 7194223 8080750 9239976 10272526 12036014 14259396 14648234.6 RJ RN Rio Grande24 297431 381445.1 459312.3 512532.1 613099.9 788077.6 908011.2 1014478 1184697 1393621 1614254.689 RN RO Rondônia 11 201587 231829.3 363015.1 322848.1 362099.8 516946.5 548914.6 639159.8 866783.3 1048957 1231116.247 RO RR Roraima 14 35964 42979.82 51897.18 62826.7 67085.9 108309.8 118022.8 119578.4 137625.2 149909.1 187486.3116 RR RS Rio Grande43 3509483 3879186 3856958 4101730 4521426 5489652 6515175 7213311 8595599 9360614 10900617.79 RS SC Santa Cata42 1674549 1873491 1981274 1962928 2223841 2693476 3234880 3798879 4094442 5175747 5777081.563 SC SE Sergipe 28 264082 301780.9 359170.4 353445.9 392271.4 487336.7 563435.3 664264.8 761312 879533 1021126.519 SE SP São Paulo 35 18626437 21334002 22795013 22387602 24694373 29677851 33165923 36453327 39394704 45223079$ 49,988,622 SP TO Tocantins 17 106469 136942.7 153175.5 178854.4 209303.3 272553.9 335110.3 406195.6 537239.9 595288.6 658939.3299 TO
Table 25 depicts the nominal values of the GDP for each State, from 1995 to 2005, in thousands of reais (R$), in current value (corrected for inflation).
Table 25. GDP by Brazilian States, from 1995 to 2005, in R$ (thousands), in Nominal Values.
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 AC Acre 995065.4 1147584 1313635 1453505 1557195 1702621 1920769 2259000 2716123 3241847 AL Alagoas 4036622 5072157 5756057 6141038 6429095 7022923 7569188 8767000 10325908 11556232 AM Amazonas 10993593 14090343 14410659 15106983 15554778 18872885 20736037 25030000 28062624 35888581 AP Amapá 1235377 1339775 1526341 1501146 1584001 1968365 2253041 2652000 3083013 3720359 BA Bahia 26769072 32989503 37020911 38759097 42040109 48197174 52249320 62103000 73166488 86882057 CE Ceará 12494861 15641122 17589092 18835765 19510907 20799548 21581141 24204000 28425175 33260672 DF Distrito Federal 13231169 16539188 19916064 25119060 22256423 29587137 33051371 35672000 37752658 43521629 ES Espírito Santo 12857799 14908936 16198253 17369207 18772147 21530247 22538119 24723000 28979694 34487905 GO Goiás 11875395 14592142 16025469 17428114 17919856 21665356 25048231 31299000 36835111 41316491 MA Maranhão 5063120 6872683 7410423 7223687 7918384 9206845 10293103 11420000 13983802 16547449 MG Minas Gerais 62971008 78604474 87196507 89490142 93748370 1.06E+08 1.14E+08 1.25E+08 1.45E+08 1.67E+08 MS Mato Grosso do Sul 6994111 8317293 9292069 10049502 10901030 11861168 13736055 15343000 18969505 19953529 MT Mato Grosso 6510213 7945768 9155035 9901396 11701250 13428289 14452678 17888000 22615132 27935499 PA Pará 12081137 13855469 14716542 15571994 16673915 18913684 21747997 25530000 29215268 34195676 PB Paraíba 5324208 6549148 6989286 7261827 7936649 9237737 10271930 11634000 13710913 14863057 PE Pernambuco 17460791 21391139 23439349 24810247 26021483 29126796 31724962 36510000 42260926 47697442 PI Piauí 3179601 3901583 4224891 4412523 4733809 5329536 5574648 6166000 7325106 8611415 PR Paraná 38368590 47719783 52848951 56798250 61723959 65968713 72770350 81449000 98999740 1.09E+08 RJ Rio de Janeiro 74412360 86757651 97673810 1.01E+08 1.14E+08 1.38E+08 1.48E+08 1.7E+08 1.9E+08 2.23E+08 RN Rio Grande do Norte 4727278 5876000 6669225 6844467 7647781 9293319 9833650 11633000 13695517 15906124 RO Rondônia 2959448 3639348 4198007 4611191 5023344 5624964 6082841 7284000 8491977 9744451 RR Roraima 469440 546826.5 621369 746384.3 817192.3 1116581 1218984 1488000 1677318 1864151 RS Rio Grande do Sul 53652947 63262677 69221314 70541889 75450458 85137543 94084498 1.04E+08 1.28E+08 1.43E+08 SC Santa Catarina 23572606 29453733 31874672 32434064 35681851 42428004 46534519 51828000 62213541 70207924 SE Sergipe 3532532 4303611 4842561 5030502 5434375 5920725 8204018 9496000 11704013 13120855 SP São Paulo 2.29E+08 2.72E+08 3.09E+08 3.24E+08 3.4E+08 3.71E+08 4.01E+08 4.38E+08 4.95E+08 5.47E+08 TO Tocantins 1225936 1533584 1720259 1941479 2106171 2450498 3066502 3545000 4189864 4767936
181 Appendix A. Brazilian Tax Structure
Table 26 shows VAT elasticity in each State, from 1995 to 2005.
Table 26. VAT Elasticity by Brazilian States, from 1995 to 2005.
1995/6 1996/7 1997/8 1998/9 1999/0 2000/1 2001/2 2002/3 2003/4 1995/2004 1997/04 1995/8 AC Acre 1.06 1.01 1.36 0.92 1.33 1.09 1.06 1.02 1.03 2.17 2.02 1.46 AL Alagoas 0.94 1.02 1.04 0.96 1.12 1.05 0.98 1.00 1.08 1.18 1.24 0.99 AM Amazonas 1.37 0.75 0.80 1.03 1.05 1.07 0.97 1.01 0.92 0.87 0.85 0.82 AP Amapá 1.09 0.93 1.18 1.00 1.17 1.04 1.00 0.92 1.03 1.40 1.37 1.21 BABahia0.970.901.001.031.081.050.990.980.980.981.110.88 CECeará0.990.931.011.101.131.101.010.930.971.161.250.94 DF Distrito Fed 1.02 0.94 0.85 1.30 0.95 1.01 1.06 1.15 1.03 1.27 1.33 0.81 ES Espírito Sa 0.94 1.08 0.87 0.95 1.12 1.16 0.89 1.05 1.06 1.08 1.06 0.88 GOGoiás1.000.990.921.131.011.030.911.050.950.970.980.91 MA Maranhão 0.97 0.83 1.10 0.97 1.19 1.14 1.01 0.88 1.04 1.08 1.34 0.89 MG Minas Gera 0.92 0.93 0.97 1.09 1.05 1.12 0.95 1.00 1.04 1.06 1.23 0.83 MS Mato Gross 0.90 0.98 0.90 1.25 1.08 1.02 0.98 1.01 1.20 1.31 1.48 0.80 MT Mato Gross 0.92 1.02 0.82 1.15 1.10 0.89 1.10 1.02 1.00 0.97 1.03 0.76 PA Pará 1.02 0.87 1.10 0.93 1.30 0.95 1.13 1.04 1.00 1.32 1.49 0.98 PBParaíba1.031.031.061.011.031.120.900.931.041.161.091.13 PE Pernambuc 1.00 1.00 1.01 1.00 1.06 1.03 1.04 0.96 1.03 1.12 1.12 1.01 PI Piauí 1.05 1.00 1.00 0.99 1.11 1.05 1.04 0.71 1.41 1.28 1.22 1.05 PR Paraná 0.68 0.90 0.95 1.09 1.18 1.40 1.03 0.98 1.06 1.11 1.82 0.58 RJ Rio de Jane 0.99 0.89 1.18 1.00 0.93 1.07 0.97 1.05 1.01 1.06 1.20 1.04 RN Rio Grande 1.03 1.06 1.09 1.07 1.06 1.09 0.94 0.99 1.01 1.39 1.27 1.19 RORondônia0.941.360.811.031.270.980.971.161.051.581.241.03 RRRoraima1.031.061.010.981.181.000.831.020.981.050.961.10 RS Rio Grande 0.94 0.91 1.04 1.03 1.08 1.07 1.00 0.97 0.98 1.00 1.18 0.89 SC Santa Cata 0.90 0.98 0.97 1.03 1.02 1.10 1.05 0.90 1.12 1.04 1.19 0.85 SESergipe0.941.060.951.031.140.831.020.931.030.900.900.94 SPSão Paulo0.960.940.941.051.101.031.010.961.041.021.120.85 TOTocantins1.031.001.031.081.120.981.051.120.971.441.401.06 Brazil 0.95 0.94 0.97 1.05 1.07 1.07 1.00 0.98 1.03 1.05 1.17 0.87 Note : Values deflated by IGP.
I decided not to use the tax rates across the states because of the specificities of VAT rates. However, looking at the table below, it is noticeable the there are several common generic rates for the GDP across the States. Transactions among southern states are usually rated at 12%, otherwise, intra- states transactions are rated at 7%. Given the overlapping values and the difficulty in determining the accurate value for each state and for different products, I have not included rate predictors in my statistical analysis (DeltaTranslator N.D.; Nogueira 1999; Souza 2005).
Table 27 shows the different VAT rates used in different states.
182 Appendix A. Brazilian Tax Structure
Table 27. VAT Rates per State.
State VAT rates in 1998 and 1999 AC 98 12%, 13%, 17%, 25% AL 99 25%, 17%, 12% AM 99 4%, 12%, 17%, 25% AP 99 25%, 17%, 12%, 8%, 2% BA 99 7%, 17%, 25% CE n.a. DF 99 25%, 21%, 17%, 12% ES 99 12%, 13%, 17%, 25% GO 99 4%, 7%, 12%, 13%, 17%, 25% MA 99 4%, 12%, 17%, 25% MG 99 4%, 7%, 12%, 16%, 18%, 25%, 30% MS 99 12%, 20%, 25%, 25%, 25%, 17% MT 99 12%, 17%, 20%, 25%, 30% PA 99 25%, 12%, 7%, 17% PB 99 25%, 20%, 17%,12% PE 99 7%, 12%, 17%, 20%, 25% PI 99 4%, 12%, 17%, 20%, 25% PR 99 7%, 9%, 12%, 13%, 17%, 25% RJ 99 18%, 7%, 12%, 25%, 37% RN 99 25%, 17%, 13%, 12%, 4% RO 99 25%, 17%, 12%, 9% RR 99 12%, 17%, 25% RS 99 12%, 7%, 18%, 26% SC 99 25%, 17%, 12%, 7% SE 99 4%, 7%, 12%, 17%, 25% SP 99 4%, 7%, 12%, 18%, 25% TO 99 25%, 17%
183 Appendix A. Brazilian Tax Structure
A.4 Other Information
Table 28 contains some geographical information about the States which were object of this study.
Table 28. Selected States Features (Capital City, Territorial Area, Number of Municipalities, Population).
State Bahia Ceara Paraiba Pernambuco Rio Grande do Norte São Paulo Capital Salvador Fortaleza João Pessoa Recife Natal São Paulo Area (km²) 564.692,669 148.825,602 56.439,838 $ 98,311,616.00 $ 52,796,791.00 $ 248,209.43 # municípios 417.00 184.00 223.00 185.00 167.00 645.00 Population 13,815,334.00 8,097,276.00 3,595,886.00 8,413,593.00 3,003,087.00 40,442,795.00
184
Appendix B
List of Interviews
This appendix contains information about the interviews conducted in the study.
185 Appendix B. List of Interviews
RIO GRANDE DO NORTE and PARAÍBA
Name Position/Sector Date 1. Jaime Goudeiro Promosad – Projeto de Reestruturação dos 3/18/2001 Sistemas de Administração Tributária 2. Saulo Rocha Sub 3/18/2001 3. Jacaúna de Assunção Secretário da Tributação do RN 3/19/2001 4. João Flávio S. Medeiros Coordenadoria de Arrecadação Estatística e 3/20/2001 Controle 5. Edvani de Freitas Neto Setor de Pessoal 3/20/2001 6. Graça Queiroz Responsável por treinamento 7. Brenan Rembrant C. Coordenadoria de 3/20/2001 Vasconcelos Informática 3/23/2001 8. Neil Armstrong de Almeida Tax Collector 3/20/2001 9. Jorge Fiscal – Volante 10. Isnard Fiscal – Volante 11. Walter Fiscal – Volante 12. Olavo Posto Fiscal de Caraú 3/21/2001 13. Sayonara Pereira de Oliveira Fiscal SINDIFERN – Diretora de Formação 3/22/2001 Sindical FENAFISCO 14. Ludenilson Araújo Lopes Sub-Secretário RN 3/23/2001 15. Walter Correia de Aquino Neto Tributação e Assessoria Técnica (Cabinet) 3/26/2001 16. Alexandre Firmino Fiscal, Ex-Secretário Adjunto (1995-1996) 3/26/2001 17. Jane Carmen C. de Araújo Fiscala, Director 1a. URT (Unidade Regional) 3/27/2001 18. Manoel 3/27/2001 19. José Araújo da Silva Presidente do Sindicato (SINDIFERN) 3/27/2001 20. Lina Secretária da Sec. Da Tributação 3/28/2001 21. Bira Secretário da Sec. Da Tributação 1995 3/28/2001 FIERN 22. Maria Valéria Secretária do Dr. Nuto 3/30/2001 23. Francisco de Assis L. Souza Fo. Fiscal – Gerente de Organização e Gestão 4/3/2001 24. Dr. Nuto 4/3/2001 25. Primo do Guido
186 Appendix B. List of Interviews
SÃO PAULO Name Position/Sector Date 26. Yoshiaki Nakano SEFAZ – Secretaria Da Fazenda de São Paulo Secretário 27. Lilian de Toni Furquim SEFAZ – Secretaria Da Fazenda de São Paulo - Assessora Económica 02/07/2001 28. Harumi Fiscala 29. David Torres SINAFRESP – Sindicato ods Agentes 7/12/2000 Fiscais de Rendas Do Estado De São Paulo – Presidente 30. Dalvanira Paes Lima APECAT - 06/2000 several 31. Nelson Fagundes 06/2000 several 32. Evandro de Mello Xavier Fiscal 33. Pedro Paulo Fiscal 01/11/2001 01/16/2001 34. Luzia Fiscal 01/11/2001 35. Cata Preta Fiscal 36. Dr. Soto SINAFRESP 37. Dirceu Pereira Fiscal / Consultor 38. Nelson Machado FAZESP – Escola Fazendária Several interviews 39. Glauco Honório Fiscal / DEAT/COMEX 01/22/2001 40. João Carlos Csillag Fiscal - Diretoria de Informações 41. Jaider Pereira Paiva Consultor UCE – Unidade de Coordenação Estadual SAF – Subcoordenadoria Administrativa Financeira 42. Helena Kerr do Amaral Escola Fazendária Do Estado de São Paulo – Diretora 43. Nelson Fagundes APECAT - Fiscal 06/06/2000 44. J. R. Lobato Fiscal 45. Carlos Leony Fonseca da Promocat – Programa de Modernização da 01/03/2001 Cunha CAT Diretoria de Informações - Diretor 46. Marta Ma. Alvarenga Freire Fiscal – APECAT 01/14/2001 47. Sr. Élcio Sindicato dos Contabilistas de São Paulo 2/19/2001 48. Evelyn Levy Consultora 2/20/2001 49. Marcos Fernandes Gonçalves Professor EAESP 2/21/2001 da Silva 50. João Adolfo Ponchio Fiscal – CECI (Controladoria do Estado SP) 2/21/2001 51. Victor Fiscal 52. Walter Soboll Coordenador – Coordenadoria Estadual de Controle Interno 53. Henrique Fingermann Professor EAESP 2/21/2001 54. Clóvis Panzarini Coordenador de Administração Tributária 2/22/2001 55. Carlos Eugênio FGV 56. José Ernesto FGV 57. Isabel FGV
187 Appendix B. List of Interviews
58. Claudio Couceiro D’Amorim Fiscal – Diretor Geral da Receita/Diretor da 3/15/2001 Adm. Tributária 59. Ana Claúdia Salgado Ouvidora Fazendária – Ombudswoman 60. Sílvia Franco Comunicação Institucional 06/06/2000 07/12/2000 61. Francisco Moura GIA launching 62. Wagner 63. Two people Sindicato de Contabilistas 64. Carlos Alberto Agostini FENAFISCO – Federação Nacional Do Fisco Estadual – Presidente São Paulo – Brasília 65. Mauro Bogea Ministério da Fazenda – Brasília 07/20/2000 66. Benedicto Ludgero Fornitani Fiscal – Retired 67. Benedito Franco Da Silveira Fo. AFRESP – Associação Dos Agentes Fiscais 01/11/2001 de Rendas do Estado De São Paulo – 01/17/2001 Presidente do Conselho Deliberativo 68. Alexandre Felix SINAFRESP – Imprensa 69. Rui de Britto Álvares Affonso FUNDAP – Diretor de Economia do Setor Público 70. Fernando Collet UCE/SP – Unidade de Controle Estadual 71. Conceição Fraga Política Salarial 01/17/2001
188 Appendix B. List of Interviews
PERNAMBUCO Name Position/Sector Date 72. Vânia Pernambuco Divisão de Treinamento 73. Jorge Jatobá Palestra Secretário da Fazenda 74. Felipe Chaves 3/15/2001 75. Aldemir Souza 76. Sol Garson Braule Pinto BNDES – Secretaria Para Assuntos 3/16/2001 Fiscais 77. Eli UCE – Coordenador 3/5/2001 78. Maria Roseana Gabinete Secretária da FAzenda 3/5/2001 79. Edilberto 3/6/2001 80. Alexandre Ribeiro 3/6/2001 3/7/2001 81. Maria do Carmo 82. José Cândido B. de Miranda SINDIFISCO – Sindicato dos 3/9/2001 Funcionários Integrantes do Grupo Ocupacional Do Tesouro Estadual de Pernambuco - Presidente 83. Aldo Ramos 3/6/2001 84. Luzanita Monteiro 85. Airan Castilho 86. Carlos Rogério 87. Valter Jarocki Jr. ABRASEL – Associação Brasileira de 3/9/2001 Restaurantes e Empresas de Entretenimento 88. Carlos Egito 89. Eziel 3/12/2001 90. Nevton 3/9/2001 91. Francisco 3/13/2001 92. Roberto Abreu 3/13/201 93. Roberto Cavacanti Tavares Diretoria De Segmentos Econômicos 3/13/2001
189 Appendix B. List of Interviews
BAHIA Name Position/Sector Date 94. Dimas Fonseca Mello SEFAZ BA 8/1/2000 95. Emília Maria D. Gonçalves Diretora Operacional – Superintendência de Desenvolvimento do Serviço Público e Atendimento ao Cidadão – SESAC 96. Monique Badaró Campos Fundação Luís Eduardo Magalhães 97. Monica Simões Bandeira FISEPE – Empresa De Fomento da Informática do Estado de Pernambuco – Assessora 98. Moysés de Oliveira Andrade Jr. Diretor – Diretoria de Arrecadaçãqo, 8/3/2000 Crédito e Contrôle – DARC 99. Fátima Freire UCP – Unidade Contrôle 100. José Sérgio Guanabara Superintendência da Administração 08/02/2001 Tributária – Diretor 101. João Batista Aslan Ribeiro UCE/PROMOSEFAZ – Subcoordenador Técnico 102. Celso Tavares Ferreira Superintendência de Administração Tributária/Diretoria de Planejamento da Fiscalização – Diretor 103. Alexandre Cialdini Secretaria da Fazenda/Ceará – Fiscal 7/25/2000 104. 105. Geraldo Machado Fiscal
190 Appendix B. List of Interviews
OTHERS Others Name Position/Sector Date 106. Elia Yi Armstrong Public Administration Officer 06/2003 DPADM/DESA/UNDP 107. John-Mary Kauzya Chief of Governance and Public 06/2003 Administration Branch – DESA/UNDP 108. Esdras Borges Costa FUNDAP – Executive Director 109. Francisco Mejia IADB – Inter-American Development Bank 110. Anwar Sha World Bank/Evaluation Officer and Public Sector Reform Coordinator 111. Carlos Pimenta IADB – Inter-American Development Bank Modernization of State Specialist 112. Clotilde Fonseca Fundación Omar Dengo – Diretora Ejecutiva 113. Glauco Arbix USP – Universidade de São Paulo – 06/2000 Professor several 114. Edmilton Secretaria da Fazenda/Ceará – Fiscal 7/25/2000 115. Sindicato 116. Márcio Gomes da Cruz IADB 117. Ferbes 118. M. Dilip Professor Boston University 119. Ceara/sindicato 120. Dilna PNAFE 7/26/2000 121. Lúcia Callou 122. Alberto Teixeira IMESP 123. Rejane 124. Jocélio SINTAF – Sindicato 125. Peter Spink FGV 126. Peter Spink FGV 127. Dilma M. da Fonseca 128. Jaime Mano 129. Luiz Tacca 130. João Dias 131. Carlos Silvani IMF 132. Pai do Henrique 133. Julieta Verleun IADB 2005 134. Jaime Mano IADB 2005 135. Orlando 2003 136. 2003 137. Cristina Waldwoger 138. James Poterba
191
Appendix C
Indicators
NATIONAL FISCAL ADMINISTRATION PROGRAM FOR THE BRAZILIAN STATES PNAFE
Reform Indicators
Reform 1 AdminCost Reform 2 AdminAconnect STAB Expenditure # State Agencies on-line STAB Revenue Total Budgeting State Agencies 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd 0.240 0.027 0.096 0.163 0.150 0.169 nd AC AC 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 AC AL 0.020 0.040 nd 0.029 0.032 0.033 0.035 0.035 AL AL 0.760 0.960 1.000 1.000 1.000 1.000 0.887 0.887 AL AM 0.040 0.071 0.076 0.069 0.071 0.071 0.042 0.042 AM AM 1.000 0.952 0.955 0.962 0.980 0.980 1.000 1.000 AM AP 0.011 0.040 0.040 0.020 0.080 0.070 0.002 0.002 AP AP 0.910 0.920 0.930 0.930 0.930 0.910 0.938 0.938 AP BA 0.037 0.064 0.043 0.030 0.035 0.030 0.031 0.028 BA BA 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 BA CE 0.058 0.047 0.047 0.040 0.034 0.029 0.046 0.036 CE CE 0.840 0.830 0.867 0.880 0.877 0.909 1.000 1.000 CE DF nd nd 0.010 nd nd nd nd nd DF DF 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 DF ES nd nd 0.190 0.070 0.020 0.020 0.023 0.014 ES ES nd 0.910 1.000 1.000 1.000 1.000 1.000 1.000 ES GO nd 0.200 0.080 0.047 0.040 0.005 0.011 0.010 GO GO nd nd 1.000 1.000 1.000 1.000 1.000 1.000 GO MA 0.030 0.090 0.025 0.020 0.022 0.020 0.019 0.020 MA MA 1.000 1.000 1.000 0.980 0.990 1.000 1.000 1.000 MA MG 0.026 nd 0.020 0.020 0.020 0.020 0.026 0.018 MG MG 0.920 nd 0.970 0.910 0.970 0.930 0.933 0.858 MG MS 0.057 0.019 0.006 0.000 0.058 0.055 0.085 0.085 MS MS 0.770 0.900 0.760 0.960 0.897 0.867 0.972 0.972 MS MT 0.071 0.061 0.044 0.065 0.057 0.060 0.031 0.065 MT MT 1.000 1.000 0.920 1.000 1.000 1.000 1.000 1.000 MT PA 0.034 0.030 0.028 0.025 0.023 0.022 0.019 0.021 PA PA 0.920 0.920 0.940 0.989 1.000 1.000 1.000 1.000 PA PB 0.030 0.040 0.030 0.040 0.030 0.030 0.027 0.033 PB PB 0.700 0.700 0.910 0.920 0.930 0.930 1.000 1.000 PB PE 0.050 0.060 0.040 0.043 0.039 0.041 0.047 0.029 PE PE 0.430 0.550 0.700 0.950 1.000 1.000 1.000 1.000 PE PI 0.067 0.072 0.062 0.051 0.049 0.050 0.040 0.039 PI PI 1.000 1.000 1.000 1.000 1.000 1.000 0.991 1.000 PI PR 0.015 0.014 0.015 0.013 0.015 0.016 0.014 0.014 PR PR 0.910 1.000 1.000 0.990 0.980 0.990 0.985 0.985 PR RJ 0.031 0.028 0.030 0.030 0.030 0.020 0.007 0.007 RJ RJ 0.360 0.420 1.000 1.000 1.000 1.000 1.000 1.000 RJ RN nd 0.025 0.022 0.020 0.019 0.019 0.024 0.020 RN RN nd 0.000 0.717 0.717 0.597 0.597 1.000 1.000 RN RO 0.050 0.030 0.030 0.004 nd nd 0.029 0.029 RO RO 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 RO RR 0.060 nd 0.000 0.000 0.000 nd 0.021 0.021 RR RR 0.000 nd 0.000 nd nd 0.844 0.882 0.882 RR RS 0.027 0.020 0.020 0.020 0.020 0.020 0.011 0.011 RS RS 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 RS SC 0.026 0.030 0.023 0.019 0.014 0.014 0.017 0.011 SC SC 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 SC SE 0.110 0.060 0.040 0.040 0.060 0.045 0.043 0.043 SE SE 0.930 0.970 0.980 nd 0.980 1.000 1.000 0.984 SE SP 0.022 0.023 0.017 0.031 0.030 0.028 0.016 0.016 SP SP 0.870 0.880 0.981 1.000 1.000 1.000 1.000 1.000 SP TO 0.040 0.036 0.030 0.020 0.020 0.020 0.076 0.026 TO TO 0.950 0.960 1.000 1.000 1.000 1.000 1.000 1.000 TO
193 Appendix C. Indicators
Reform 3 TaxComplaince Reform 4 OnlineBudget Total VAT Returns Expenditure Budgeted on-line Total Taxpayers Registered Total Expenditure 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd nd nd nd 0.398 1.000 0.043 AC AC nd nd 1.000 1.000 1.000 1.000 1.000 0.001 AC AL 0.450 0.620 0.480 0.590 1.000 0.964 0.943 0.964 AL AL nd 1.000 nd nd 1.000 1.000 1.000 1.000 AL AM 0.510 nd 0.300 0.354 0.329 0.329 0.298 0.329 AM AM 1.000 0.980 0.984 0.980 0.985 0.985 1.000 1.000 AM AP 0.770 0.830 0.780 0.790 0.720 0.450 0.587 0.450 AP AP nd 0.240 1.000 1.000 1.000 1.000 1.000 1.000 AP BA 0.750 0.720 0.680 0.690 0.690 0.690 0.702 0.688 BA BA nd 0.980 1.000 1.000 1.000 1.000 1.000 1.000 BA CE 0.960 0.230 0.910 0.950 0.953 0.905 0.993 0.857 CE CE 1.000 1.000 0.926 1.000 1.000 1.000 1.000 1.000 CE DF nd 0.850 0.790 nd 0.350 0.350 0.403 0.403 DF DF 1.000 1.000 0.990 0.760 1.000 0.000 0.999 0.999 DF ES nd nd 0.980 1.000 0.960 0.960 1.000 1.000 ES ES nd 1.000 1.000 1.000 1.000 1.000 1.000 0.954 ES GO 0.530 0.350 0.420 0.421 0.412 0.324 0.421 0.394 GO GO nd nd 1.000 1.000 1.000 1.000 1.000 1.000 GO MA 0.560 nd 0.520 0.600 0.520 0.970 0.861 0.815 MA MA 1.000 nd 0.570 0.670 1.000 1.000 1.000 1.000 MA MG 0.950 nd 0.340 0.980 0.980 0.980 0.983 0.984 MG MG 0.980 nd nd 0.970 0.980 1.000 0.973 0.975 MG MS 0.920 0.970 0.980 1.000 1.000 1.000 0.707 1.000 MS MS 0.920 0.970 0.880 0.840 0.888 0.905 0.961 0.961 MS MT 0.530 1.000 1.000 0.812 0.844 0.745 0.745 1.025 MT MT 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 MT PA 0.060 0.150 0.150 0.149 0.853 0.670 0.558 0.922 PA PA 0.940 0.950 0.960 1.000 1.000 1.000 1.000 1.000 PA PB 0.500 0.500 0.960 0.520 0.510 0.480 0.508 0.488 PB PB 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PB PE 0.570 0.910 0.830 0.910 0.890 0.770 nd 0.813 PE PE 0.980 0.990 0.990 1.000 1.000 1.000 1.000 1.000 PE PI 0.340 0.439 0.319 0.313 0.380 0.362 0.323 0.378 PI PI 0.930 0.901 nd nd 0.965 1.000 1.000 0.965 PI PR 0.510 0.490 0.510 0.540 0.580 0.680 1.000 1.000 PR PR 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PR RJ 0.900 0.880 0.880 0.880 0.890 0.990 0.558 0.829 RJ RJ 0.870 0.980 0.980 0.980 0.980 0.870 1.000 1.000 RJ RN nd 0.370 0.458 0.467 0.677 0.677 0.680 0.586 RN RN nd 0.000 0.821 0.967 1.000 1.000 1.000 1.000 RN RO 0.280 0.610 0.700 0.880 0.910 0.580 0.287 0.580 RO RO 1.000 1.000 1.000 1.000 0.980 0.980 0.980 0.980 RO RR 1.000 nd 0.610 0.614 0.614 0.992 0.523 0.523 RR RR 0.000 nd nd nd nd 0.925 1.000 1.000 RR RS 0.360 0.370 0.390 0.360 0.720 nd 0.472 0.258 RS RS 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 RS SC 0.960 0.950 0.894 0.946 0.946 0.994 0.994 0.994 SC SC nd 0.810 1.000 1.000 1.000 1.000 1.000 1.000 SC SE 0.540 0.560 0.790 1.000 0.880 0.910 0.969 1.071 SE SE nd nd nd nd nd nd nd nd SE SP 0.370 0.370 nd nd 0.828 0.828 0.766 0.828 SP SP 0.900 0.940 1.000 1.000 1.000 1.000 1.000 1.000 SP TO 0.620 nd 0.740 0.280 0.490 0.490 0.686 0.740 TO TO 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 TO
194 Appendix C. Indicators
Reform 5 OnlineFiling Reform 6 OnlineValidation VAT Re turn On-line Rvenue Validated On-line Total VAT Returns Totlat STAB Revenue 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd nd nd 0.400 0.315 0.767 0.767 AC AC nd 1.000 nd 1.000 1.000 1.000 1.000 1.000 AC AL nd 1.000 0.910 1.000 nd 1.000 1.000 1.000 AL AL nd nd 1.000 1.000 0.950 1.000 1.000 1.000 AL AM 0.060 0.428 1.000 1.000 1.000 1.000 1.000 1.000 AM AM 0.130 0.357 0.504 0.974 0.987 0.987 1.000 1.000 AM AP 0.000 0.000 0.000 0.320 0.420 0.290 0.727 0.727 AP AP 0.000 0.000 0.000 0.000 0.410 0.850 1.000 1.000 AP BA nd 0.950 0.990 1.000 1.000 1.000 1.000 1.000 BA BA 0.000 1.000 0.980 0.990 0.990 0.910 0.996 0.997 BA CE 0.710 0.320 0.660 0.976 0.889 1.000 0.827 0.852 CE CE nd 0.000 0.795 0.860 0.979 1.000 1.000 1.000 CE DF nd 0.350 0.830 nd 0.820 0.820 0.521 1.000 DF DF 1.000 1.000 0.150 nd 0.950 0.950 1.024 1.024 DF ES nd nd 0.560 0.520 0.540 0.540 1.000 1.000 ES ES nd 1.000 1.000 1.000 0.940 0.940 1.000 1.000 ES GO 0.750 1.000 0.990 1.000 1.000 1.000 1.000 1.000 GO GO 0.250 nd 0.810 0.538 0.900 1.003 1.000 1.000 GO MA 0.000 0.000 1.000 1.000 1.000 1.000 1.000 1.000 MA MA 0.000 0.430 0.630 0.530 0.550 0.950 0.994 0.997 MA MG 0.000 nd 0.490 0.760 0.730 1.000 0.995 0.945 MG MG 0.780 nd 0.720 0.870 0.870 0.950 0.946 0.959 MG MS 0.950 0.980 0.800 1.000 0.972 1.000 1.000 1.000 MS MS 0.890 0.670 0.570 0.630 0.974 0.952 1.000 1.000 MS MT 0.410 0.870 0.615 0.727 0.809 1.000 1.000 1.000 MT MT 0.000 0.150 0.731 0.720 0.814 0.980 0.980 1.000 MT PA 0.280 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PA PA 0.690 0.820 0.700 0.714 0.842 0.942 0.966 0.925 PA PB 0.410 0.790 0.970 0.990 1.000 1.000 1.000 1.000 PB PB 0.250 0.330 0.600 0.630 0.750 0.930 1.000 1.000 PB PE 0.030 0.400 0.970 0.990 1.000 1.000 nd 1.000 PE PE 1.000 0.980 1.000 1.000 1.000 1.000 1.000 1.000 PE PI 0.710 0.552 0.836 0.785 0.357 0.540 0.647 1.000 PI PI 0.000 0.000 0.000 0.799 0.900 0.950 0.950 0.990 PI PR 0.100 0.760 0.940 0.940 1.000 1.000 0.986 1.000 PR PR 0.890 0.880 0.810 0.860 0.970 0.930 0.971 0.990 PR RJ 0.450 0.710 0.820 0.880 0.970 0.990 1.000 1.000 RJ RJ 0.000 0.150 0.950 0.930 1.000 1.000 1.000 1.685 RJ RN nd 0.850 0.893 0.981 1.000 1.000 1.000 1.000 RN RN nd 0.270 0.639 0.933 0.889 0.889 0.878 0.859 RN RO nd nd 0.760 1.000 1.000 1.000 1.000 1.000 RO RO 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 RO RR 0.000 nd 0.000 nd nd 0.489 0.900 0.900 RR RR 0.000 nd 0.036 0.189 0.190 0.993 0.862 0.862 RR RS 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 RS RS 0.070 0.240 0.790 0.960 0.960 0.990 1.000 1.000 RS SC 0.790 0.660 0.842 0.902 0.902 0.988 0.988 0.988 SC SC 0.060 0.720 1.000 1.000 1.000 1.000 1.000 1.000 SC SE 0.000 0.000 0.680 0.830 0.410 0.850 1.000 1.000 SE SE 0.000 0.000 0.000 nd nd nd nd nd SE SP 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 SP SP 1.000 1.000 nd 0.986 1.000 1.000 1.000 1.000 SP TO 0.000 0.000 0.360 0.802 0.860 0.860 1.000 1.000 TO TO 0.520 0.540 0.520 0.420 0.850 0.850 1.000 1.000 TO
Reform 7 DebtNotification Reform 8 DebtCollection Number Administrative Suits Revenue from Previous Debts Number Total Debts Total Debt from Taxpayers 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd nd nd 0.600 0.750 0.401 0.132 AC AC nd nd nd 0.097 0.250 0.068 0.000 0.281 AC AL nd nd nd nd nd nd nd nd AL AL nd nd nd nd nd nd nd nd AL AM 0.640 0.659 0.484 0.239 0.267 0.267 0.140 0.140 AM AM 0.092 0.072 0.056 0.046 0.052 0.052 0.017 nd AM AP nd 0.360 0.320 0.190 0.200 0.050 0.227 0.227 AP AP nd 0.040 0.060 0.070 0.190 0.030 0.000 nd AP BA 1.130 1.130 1.166 0.470 1.320 1.410 1.265 1.770 BA BA 0.020 0.030 0.037 0.050 0.065 0.150 0.166 0.201 BA CE 0.370 0.450 0.245 0.096 0.156 0.297 0.768 0.347 CE CE 0.030 0.030 0.023 0.015 0.007 0.047 0.003 0.011 CE DF nd nd 0.010 nd 0.470 0.470 0.616 0.616 DF DF nd 0.010 0.140 nd 0.030 0.030 0.066 0.004 DF ES nd 0.040 1.000 0.470 0.470 0.470 0.443 0.443 ES ES nd nd 0.100 0.090 nd nd 1.730 nd ES GO 0.070 0.000 nd 0.595 0.380 0.380 0.782 1.265 GO GO 0.070 nd 0.030 0.018 0.012 0.012 0.001 0.023 GO MA nd nd nd nd nd nd nd 0.716 MA MA nd nd nd nd nd nd nd 0.280 MA MG 0.870 nd 0.760 1.230 2.060 3.950 4.000 4.000 MG MG 0.030 nd 0.050 0.030 0.030 0.030 0.039 0.014 MG MS 0.550 0.400 0.210 0.200 nd nd 1.279 1.279 MS MS 0.140 0.300 0.180 0.100 0.119 0.095 0.042 0.000 MS MT nd nd nd 0.154 0.814 0.379 0.751 0.073 MT MT nd nd 0.505 0.008 0.065 0.160 0.207 0.112 MT PA nd 0.030 0.030 0.261 nd nd 0.194 0.837 PA PA nd 0.041 0.024 0.021 0.019 0.007 0.016 0.020 PA PB nd 0.630 0.870 1.980 2.720 1.710 0.813 0.794 PB PB nd 0.030 0.060 0.850 0.660 0.100 0.007 0.005 PB PE 0.230 2.080 3.120 1.640 0.960 0.920 1.346 0.425 PE PE 0.040 0.080 0.130 0.160 0.150 0.330 1.272 0.226 PE PI nd nd nd 0.155 0.003 0.054 0.021 0.062 PI PI nd nd nd nd 0.002 0.362 0.193 0.387 PI PR 1.090 1.880 1.280 1.270 0.990 1.180 0.265 0.145 PR PR 0.020 0.010 0.010 0.010 0.010 0.020 0.008 0.007 PR RJ nd 0.040 0.020 0.990 0.990 0.970 0.523 0.555 RJ RJ nd 0.010 0.003 0.005 0.040 0.050 0.075 0.077 RJ RN nd 0.460 0.740 0.779 2.257 2.257 0.755 0.543 RN RN nd nd 0.308 0.449 0.231 0.231 1.524 0.901 RN RO nd 1.000 1.000 nd nd nd 0.010 0.010 RO RO nd 0.010 0.200 nd 0.061 0.150 0.001 0.000 RO RR 0.000 nd 0.000 0.000 0.000 0.004 0.627 0.627 RR RR nd nd 0.596 0.022 0.022 0.017 0.279 0.279 RR RS 1.270 0.250 0.610 1.750 3.490 3.100 3.051 2.313 RS RS 0.100 0.090 0.090 0.040 0.040 0.160 0.000 0.034 RS SC nd 1.200 2.338 1.688 1.688 8.943 22.614 20.409 SC SC nd 0.040 0.125 0.095 0.095 0.078 0.078 0.945 SC SE nd 1.870 0.480 nd 0.530 0.170 0.106 0.710 SE SE nd 0.310 0.020 0.020 0.050 0.050 nd nd SE SP 0.000 0.000 nd nd nd nd nd nd SP SP nd nd 0.202 0.144 0.201 0.201 0.158 nd SP TO nd 0.010 0.190 0.530 0.060 0.060 0.891 1.207 TO TO nd 0.030 0.080 0.005 0.030 0.030 0.012 0.568 TO
195 Appendix C. Indicators
Reform 9 DebtOutstanding Reform 10+L185 DebtInstallments Outstanding Debt Total Debt Negotiated STAB Total Revenue STAB Total Revenue 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd nd 0.382 0.350 0.417 0.411 0.057 AC AC nd nd nd 0.261 0.292 0.244 0.206 nd AC AL nd nd 0.001 nd nd nd 2.020 0.000 AL AL nd nd nd nd nd nd 0.017 nd AL AM 0.500 0.733 0.997 0.925 1.010 0.010 1.980 0.000 AM AM 0.060 0.029 0.063 0.012 0.015 0.015 0.013 nd AM AP nd 0.360 0.410 0.300 1.470 0.600 0.773 0.000 AP AP 0.050 0.070 0.060 0.030 0.020 0.030 0.037 nd AP BA 1.670 1.060 1.067 0.390 0.420 0.290 0.222 0.141 BA BA 0.140 0.140 0.044 0.060 0.027 0.027 0.027 0.068 BA CE 0.240 0.330 0.370 0.373 0.478 0.658 1.801 0.000 CE CE 0.020 0.020 0.015 0.020 0.030 0.017 0.018 0.008 CE DF nd 0.220 0.070 0.370 0.330 0.327 0.385 0.000 DF DF 0.050 0.020 0.750 0.059 0.036 0.025 0.072 0.121 DF ES nd nd 0.150 0.190 nd nd 0.007 0.000 ES ES nd nd 0.010 0.040 0.030 0.030 0.093 0.105 ES GO 0.370 0.000 2.180 2.468 2.167 2.389 4.420 2.281 GO GO 0.080 0.000 0.170 0.038 0.087 0.080 0.143 0.156 GO MA nd nd nd nd nd nd nd 0.098 MA MA nd nd 0.030 0.080 nd nd nd 0.060 MA MG 1.060 nd 0.990 1.180 1.130 1.080 1.136 0.910 MG MG 0.040 nd 0.040 0.050 0.080 0.070 0.069 0.082 MG MS 0.260 0.090 0.090 0.140 0.175 0.218 0.318 0.000 MS MS 0.030 0.040 0.030 0.020 0.025 0.025 0.052 0.000 MS MT nd nd 0.060 0.060 0.080 0.034 0.178 0.000 MT MT nd 0.002 0.004 0.006 0.036 0.034 0.014 0.006 MT PA nd 0.150 0.310 0.326 0.323 0.654 0.837 0.455 PA PA nd 0.030 0.010 0.025 nd 0.059 0.019 0.016 PA PB nd 1.140 1.010 0.040 0.070 0.210 4.027 3.744 PB PB nd 0.010 0.010 0.010 0.010 0.020 0.450 0.466 PB PE 1.840 0.440 0.310 0.250 0.220 0.120 0.026 0.158 PE PE 0.240 0.080 0.130 0.060 0.080 0.050 0.037 0.158 PE PI nd nd nd 0.002 0.036 0.069 0.077 0.034 PI PI nd nd 0.043 0.044 0.024 0.024 0.014 0.004 PI PR 0.630 0.650 0.820 0.760 0.870 0.870 1.010 1.108 PR PR 0.160 0.250 0.100 0.130 0.090 0.070 0.053 0.034 PR RJ nd 0.450 0.470 0.500 0.550 0.470 0.631 0.312 RJ RJ nd 0.040 0.040 0.010 0.020 0.030 0.135 0.071 RJ RN nd nd 0.040 0.038 0.024 0.024 0.014 0.018 RN RN nd 0.030 0.066 0.099 0.042 0.042 0.026 0.022 RN RO nd 0.040 0.030 nd 1.150 1.170 1.316 0.000 RO RO nd 0.010 0.200 nd 0.110 0.130 0.173 0.000 RO RR nd nd 0.670 0.505 0.505 0.013 0.039 0.039 RR RR nd nd 0.035 0.020 0.020 0.087 0.040 0.040 RR RS 0.740 0.930 1.130 1.220 1.100 0.270 0.000 1.310 RS RS 0.130 0.070 0.060 0.130 0.090 0.060 0.000 0.147 RS SC nd 0.500 0.256 0.378 0.378 0.394 0.277 0.021 SC SC 0.060 0.080 0.083 0.146 0.146 0.143 0.040 0.108 SC SE nd 0.140 1.030 0.193 0.430 0.808 2.342 1.084 SE SE 0.140 0.140 0.200 nd 0.110 0.080 0.131 0.061 SE SP 0.110 0.110 0.042 0.029 0.034 0.034 0.001 0.000 SP SP 0.000 0.000 nd 0.094 0.038 0.038 0.018 nd SP TO nd 0.220 0.140 0.200 0.420 0.420 0.900 0.011 TO TO 0.020 0.010 0.010 0.008 0.210 0.210 0.214 0.119 TO
Reform 11 DebtDisputed Reform 12 DebtJudged Value under Dispute Estoque valor em execução STAB Total Revenue STAB Total Revenue 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC nd nd nd 0.092 0.063 0.453 nd nd AC AC 0.060 0.180 0.340 nd nd 0.387 nd 0.057 AC AL nd0.010ndndndndndndAL AL 0.200 0.360 nd nd nd nd 0.024 0.000 AL AM 0.260 0.313 0.430 0.381 0.490 0.490 1.034 nd AM AM 0.270 0.391 0.504 0.488 0.516 0.516 0.913 0.000 AM AP 0.210 0.020 0.540 0.390 0.460 0.630 0.432 nd AP AP 0.010 0.090 0.090 0.130 0.120 0.010 0.264 nd AP BA 0.120 0.030 0.236 0.470 0.200 0.100 0.036 0.070 BA BA 0.580 nd 0.565 1.090 0.620 0.620 0.563 0.708 BA CE 0.680 0.760 0.872 0.840 0.841 0.841 0.864 0.387 CE CE 0.050 0.160 0.220 0.261 0.220 0.260 0.821 0.497 CE DF nd nd 0.090 0.240 0.030 0.030 nd 0.108 DF DF nd nd 0.400 0.440 nd 0.390 nd 0.826 DF ES nd nd 1.130 1.750 1.550 1.550 nd nd ES ES nd nd nd nd nd nd nd nd ES GO 0.620 0.000 0.260 0.571 0.492 0.495 0.803 0.467 GO GO 0.230 0.000 0.710 0.393 1.616 1.894 3.617 1.228 GO MA nd nd nd nd 0.300 0.260 0.231 0.353 MA MA 0.020 nd 0.060 0.090 0.170 0.150 0.437 0.430 MA MG 0.120 nd 0.040 0.160 0.120 0.030 0.063 0.057 MG MG 0.690 nd 0.780 0.860 0.960 0.930 0.991 0.910 MG MS 0.220 0.020 0.020 0.010 0.289 0.169 0.064 nd MS MS 0.350 0.060 0.020 0.520 0.427 0.401 0.057 nd MS MT nd 0.060 0.035 0.138 0.051 0.180 0.362 0.104 MT MT 0.110 0.770 0.653 0.590 0.779 0.779 2.271 1.028 MT PA nd 0.440 0.510 0.395 0.304 0.465 0.496 0.234 PA PA nd 0.060 nd 0.126 nd 0.014 0.031 0.129 PA PB nd 0.020 0.030 0.020 0.050 0.070 0.231 0.096 PB PB 0.980 1.100 1.010 1.130 1.130 1.520 3.414 3.209 PB PE 0.670 0.420 0.320 0.190 0.210 0.050 0.057 0.036 PE PE 1.040 1.050 1.260 1.480 1.660 1.700 1.871 1.879 PE PI 0.060 0.059 0.043 0.074 0.052 0.061 0.081 0.176 PI PI 0.322 0.378 0.527 0.473 0.500 0.481 0.468 nd PI PR 0.002 0.002 0.002 nd nd nd 0.012 0.106 PR PR nd nd nd nd nd nd nd 1.018 PR RJ nd nd nd nd nd 0.200 0.256 0.128 RJ RJ nd nd 0.370 0.380 0.510 0.380 0.621 0.328 RJ RN nd 0.090 0.063 0.022 0.106 0.106 0.072 0.026 RN RN nd 0.110 0.248 0.461 0.485 0.485 0.515 0.444 RN RO 0.010 0.130 0.130 0.360 0.260 0.020 0.195 nd RO RO 0.020 0.400 0.300 0.101 nd nd nd 0.000 RO RR nd nd 0.970 0.052 0.052 0.668 0.028 0.028 RR RR 0.110 nd 0.590 0.030 0.030 0.082 0.073 0.073 RR RS 0.040 0.010 0.010 0.080 0.060 0.060 0.046 0.063 RS RS 0.570 0.770 0.840 0.970 0.890 1.010 0.982 1.123 RS SC nd 0.100 0.107 0.113 0.113 0.126 0.160 0.199 SC SC nd nd nd 0.446 0.446 0.515 0.571 0.548 SC SE 0.500 0.580 0.030 0.125 0.045 nd 0.250 0.220 SE SE 0.180 0.140 0.320 0.330 0.800 0.810 1.902 1.002 SE SP 0.140 0.140 0.028 0.269 0.271 0.271 0.169 nd SP SP 0.000 0.000 nd nd nd nd 1.506 nd SP TO 0.020 0.040 0.040 0.009 0.030 0.030 0.036 0.107 TO TO nd 0.120 0.250 0.160 0.320 0.320 0.900 0.011 TO
196 Appendix C. Indicators
Reform 13 Training Reform 14 LargeTaxpayers Number of Trainned Officials Taxpayers paying > 50% VAT Revenue Total Number of Officials Total Taxpayers 1997 1998 1999 2000 2001 2002 2003 2004 1997 1998 1999 2000 2001 2002 2003 2004 AC 0.000 0.290 1.870 2.209 0.768 1.017 1.021 0.397 AC AC nd nd 0.0003 0.0011 0.0009 0.0009 0.002 0.002 AC AL 0.660 1.330 0.260 1.460 1.820 0.500 3.150 nd AL AL 0.0015 0.0011 0.0010 0.0005 0.0006 0.0005 0.0003 0.0003 AL AM nd 0.749 0.192 0.470 0.964 0.964 0.514 nd AM AM 0.0004 0.0006 0.0005 0.0005 0.0004 0.0004 0.0003 0.0003 AM AP 0.000 3.130 3.330 0.950 0.930 0.960 1.000 nd AP AP nd 0.0008 nd nd 0.0014 nd 0.0026 0.0026 AP BA nd 3.210 2.849 0.170 2.140 4.150 2.100 1.452 BA BA 0.0002 0.0002 0.0001 0.0000 0.0000 0.0001 0.000 0.000 BA CE 2.230 2.470 0.911 0.900 0.859 0.876 0.507 0.974 CE CE 0.0005 0.0002 0.0002 0.0003 0.0002 0.0002 0.0001 0.0001 CE DF 0.420 0.480 0.800 1.030 0.400 1.630 0.432 0.217 DF DF 0.0004 0.0005 0.0001 nd nd 0.0010 0.006 0.003 DF ES nd 0.150 0.870 3.990 1.890 1.890 0.329 0.628 ES ES nd 0.0002 0.0003 0.0003 0.0004 0.0004 0.0003 0.0003 ES GO 0.270 0.510 0.940 0.666 0.626 0.849 0.170 0.384 GO GO 0.0001 0.0003 0.0002 0.0008 0.0001 0.0002 0.0000 0.0000 GO MA 0.370 1.430 1.004 0.390 0.450 0.620 0.455 0.268 MA MA 0.0002 0.0002 nd nd 0.0002 0.0003 0.0004 0.0003 MA MG 1.610 nd 0.260 1.190 1.010 1.960 2.951 2.624 MG MG 0.0002 nd 0.0000 0.0000 0.0001 0.0000 0.0000 0.0001 MG MS 0.730 0.630 1.780 0.490 1.685 2.603 0.628 nd MS MS 0.0008 0.0006 nd 0.0000 0.0013 0.0067 0.0001 0.0001 MS MT 0.480 0.740 1.730 0.440 0.700 0.724 0.122 0.755 MT MT 0.0004 0.0002 0.0001 0.0001 0.0001 0.0002 0.0002 0.0002 MT PA 0.460 0.520 0.480 0.540 0.550 0.267 0.675 0.758 PA PA 0.0010 0.0004 0.0005 0.0033 0.0022 0.0063 0.0002 0.0005 PA PB 0.640 1.010 1.130 1.450 1.750 1.070 0.406 0.476 PB PB 0.0015 0.0015 0.0020 0.0030 0.0010 0.0010 0.0006 0.0006 PB PE 0.490 0.510 1.960 2.330 1.600 1.860 1.001 0.636 PE PE 0.0004 0.0006 0.0005 0.0005 0.0002 0.0002 0.0002 0.0003 PE PI 0.340 1.719 0.420 0.263 0.261 0.439 1.156 1.015 PI PI 0.0001 0.0006 0.0008 0.0004 0.0004 0.0002 0.0004 0.0004 PI PR 0.350 0.820 0.570 0.670 0.240 0.330 0.166 0.463 PR PR 0.0001 0.0003 0.0001 0.0001 0.0001 0.0001 0.0000 0.0000 PR RJ 0.260 0.130 0.310 1.910 1.530 0.560 0.139 0.325 RJ RJ 0.0001 0.0001 0.0000 0.0001 0.0000 0.0001 0.0000 0.0000 RJ RN nd 0.140 0.443 0.642 0.406 0.406 0.679 0.616 RN RN nd 0.0003 0.0006 0.0005 0.0008 0.0008 0.0010 0.0003 RN RO 0.430 0.010 0.250 0.370 0.490 0.490 0.728 nd RO RO nd 0.0004 0.0008 0.0010 0.0025 nd 0.0005 0.0000 RO RR 0.000 nd 0.207 0.230 0.230 0.223 0.394 0.394 RR RR 0.0014 nd nd 0.0009 0.0009 0.0030 0.002 0.002 RR RS 0.790 2.100 0.940 2.690 2.140 0.610 0.279 0.325 RS RS 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 RS SC 0.790 2.190 0.738 1.057 1.057 0.513 0.199 0.746 SC SC 0.0004 0.0003 0.0002 0.0001 0.0001 0.0001 0.0000 0.0000 SC SE 0.220 1.040 1.880 1.430 2.410 1.070 0.723 1.610 SE SE 0.0050 0.0050 0.0050 0.0048 0.0097 0.0007 0.0006 0.0005 SE SP 0.030 0.040 nd 5.900 4.960 5.297 2.891 nd SP SP 0.0002 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0000 SP TO 0.010 0.340 1.370 1.970 0.360 0.360 0.053 0.610 TO TO 0.0100 0.0013 0.0010 0.0040 0.0010 0.0010 0.0002 0.0005 TO
197
Appendix D
State Economic Development Levels
Table 29 on next page displays the results of the base model estimation, reform indicators, and state economic development levels.
199
ity g n ce o tiv tin c n e ost a ne pli Filing Model nC n udg e alidati o B V C e ase nlin B Admi line O n TaxCom Onlin O Admin
Constant -283571.7483 -2184991.9412 -1664912.2854 -1247636.8168 -1838056.9146 -1885372.3086 -1090066.6189 se 189845.2526 538629.8638 704525.9303 540323.7568 544202.4282 632604.1044 1222588.2694 t -1.4937 -4.0566 -2.3632 -2.3091 -3.3775 -2.9803 -0.8916 Beta - GDP 0.0793 0.0803 0.0805 0.0747 0.0803 0.0836 0.0765 se 0.0047 0.0038 0.0058 0.0049 0.0043 0.0053 0.0082 t 16.7170 21.2916 13.9735 15.2438 18.5033 15.7182 9.2919 Beta - Reform -4893.5929 -78541.2475 -3132.6143 63304.2342 57455.6024 -10072.9796 se 9217.5356 150324.3412 1628.4988 85151.6738 85467.4081 68480.4916 t -0.5309 -0.5225 -1.9236 0.7434 0.6723 -0.1471 Beta - Reform*GDP 0.0029 0.0032 0.0001 0.0034 -0.0001 0.0024 se 0.0008 0.0010 0.0001 0.0016 0.0017 0.0016 t 3.5885 3.1223 1.7444 2.1220 -0.0643 1.4951 Beta - dev level 1 216.0000 2148966.4232 1758606.9933 1465617.4424 1867008.1744 1928765.6100 1047963.5940 se 0.9217 543284.6380 711892.5620 549342.5442 541809.8628 639657.6293 1236171.5037 t 3.9555 2.4703 2.6679 3.4459 3.0153 0.8477 Beta - dev level 1 * GDP 0.0019 -0.0137 -0.0299 -0.0104 -0.0173 0.0013 se 0.0147 0.0152 0.0131 0.0090 0.0129 0.0145 t 0.1319 -0.8991 -2.2888 -1.1555 -1.3436 0.0876 Beta - dev level 1* Reform * GDP -0.0013 0.0040 0.0003 -0.0030 0.0025 0.0000 se 0.0014 0.0128 0.0002 0.0041 0.0067 0.0057 t -0.9123 0.3130 2.1744 -0.7425 0.3806 -0.0017 Beta - dev level 2 2247820.4010 1694602.9614 1377131.8706 1716782.4216 1817253.5173 1031585.8857 se 583601.3845 750540.5503 581476.3882 565667.4084 676631.8092 1260274.4330 t 3.8516 2.2578 2.3683 3.0350 2.6857 0.8185 Beta - dev level 2 * GDP -0.0011 -0.0063 -0.0047 0.0045 -0.0050 0.0048 se 0.0087 0.0131 0.0086 0.0101 0.0102 0.0138 t -0.1252 -0.4802 -0.5400 0.4487 -0.4946 0.3458 Beta - dev level 2* Reform * GDP -0.0037 0.0024 0.0000 -0.0091 0.0004 -0.0023 se 0.0010 0.0069 0.0001 0.0059 0.0036 0.0035 t -3.8585 0.3463 0.7209 -1.5355 0.1162 -0.6575 obs 191.0000 207.0000 197.0000 187.0000 197.0000 199.0000 R-squared 0.9577 0.9390 0.9681 0.9671 0.9518 0.8698
Table 29. Results of the Base Model Estimation, Reform Indicators and State Economic Development Levels.
ts tions ed ed a nding g ic a dged in del f u n stallmen tst J Mo u Disput Noti In O Trai bt Debt DebtCollect Debt Debt Debt De
Constant -532458.8948 -1245495.7968 81131.2277 -1702503.1570 -1362144.0261 -569963.0096 -1118278.1617 se 480086.9324 493702.5790 970554.2674 557499.0157 399189.4622 338412.5987 329530.8680 t -1.1091 -2.5228 0.0836 -3.0538 -3.4123 -1.6842 -3.3935 Beta - GDP 0.0712 0.0780 0.0714 0.0833 0.0769 0.0705 0.0757 se 0.0046 0.0059 0.0063 0.0048 0.0033 0.0048 0.0022 t 15.3626 13.1729 11.2665 17.3655 23.1096 14.6304 34.9738 Beta - Reform -49200.1742 -80657.3209 -506463.0052 90714.4187 -119266.3144 -87872.4033 -51627.0085 se 61427.0181 193514.3985 666380.4993 87800.7381 88271.0204 101077.7813 43675.7392 t -0.8010 -0.4168 -0.7600 1.0332 -1.3511 -0.8694 -1.1821 Beta - Reform*GDP 0.0010 0.0168 -0.0021 -0.0018 0.0225 0.0067 0.0016 se 0.0012 0.0249 0.0202 0.0019 0.0076 0.0029 0.0003 t 0.8546 0.6753 -0.1032 -0.9114 2.9608 2.2977 5.4914 Beta - dev level 1 601815.0195 1352403.5390 -109624.6887 1747257.1542 1454482.6142 677889.5339 1221171.9602 se 477962.5436 489560.9404 963129.5226 553363.1689 401466.9466 326466.3418 329269.3560 t 1.2591 2.7625 -0.1138 3.1575 3.6229 2.0764 3.7087
Beta - dev level 1 * GDP -0.0017 -0.0131 0.0108 -0.0101 -0.0118 -0.0087 -0.0089 se 0.0116 0.0104 0.0118 0.0100 0.0077 0.0093 0.0074 t -0.1465 -1.2519 0.9173 -1.0132 -1.5299 -0.9308 -1.2140
Beta - dev level 1* Reform * GDP 0.0043 -0.0027 0.0452 -0.0079 -0.0092 -0.0014 0.0011 se 0.0073 0.0243 0.0185 0.0065 0.0234 0.0081 0.0035 t 0.5950 -0.1095 2.4435 -1.2201 -0.3924 -0.1686 0.3064 Beta - dev level 2 387411.5079 1415827.0210 -217163.3886 1652919.9855 1327194.8645 537005.8222 1098841.3056 se 498836.7879 514210.8617 968203.0815 588633.4355 446850.0067 368607.5269 361792.3593 t 0.7766 2.7534 -0.2243 2.8081 2.9701 1.4568 3.0372
Beta - dev level 2 * GDP 0.0143 -0.0064 0.0069 -0.0038 -0.0026 0.0029 0.0026 se 0.0076 0.0088 0.0093 0.0093 0.0083 0.0092 0.0073 t 1.8786 -0.7282 0.7430 -0.4055 -0.3130 0.3147 0.3498
Beta - dev level 2* Reform * GDP -0.0008 -0.0113 0.0640 0.0009 -0.0080 -0.0015 0.0005 se 0.0010 0.0235 0.0220 0.0027 0.0081 0.0034 0.0025 t -0.7562 -0.4812 2.9037 0.3279 -0.9814 -0.4453 0.2025 obs 161.0000 163.0000 180.0000 177.0000 171.0000 165.0000 203.0000 R-squared 0.9570 0.9796 0.8343 0.9586 0.9741 0.9857 0.9847
Appendix E
Time-Lagged Predictors
Table 30 on next page displays the time-lagged predictors.
203
Table 30. Time Lagged Predictors.
s er ay ns _P nce io d ding ax n ted ia icat ed _T del pl if lecte pu g T o m tsta u is PC el inCost Not tCol Jud d m Co bt b btO btD bt ining ax e e a ifty_ Mo Base M Ad T De De DebtInstallments D D De Tr F Constant -283571.748 -330520.895 -481394.326 299332.121 136307.779 -49284.729 -282655.523 1021.091 118887.602 259327.462 -244057.413 se 189845.253 228543.451 2748246.818 94836.634 112950.291 108669.317 131539.779 115084.352 98357.924 104119.906 220519.507 t -1.494 -1.446 -0.175 3.156 1.207 -0.454 -2.149 0.009 1.209 2.491 -1.107 Beta - GDP 0.079 0.076 0.082 0.064 0.068 0.077 0.080 0.068 0.065 0.065 0.080 se 0.005 0.005 0.014 0.003 0.004 0.003 0.003 0.002 0.004 0.002 0.005 t 16.717 14.980 6.095 20.641 18.693 29.388 23.620 31.450 14.905 28.507 16.649 Beta - Reform -26661.238 -1180.267 -16199.255 -501461.877 399194.441 93495.216 -497891.731 -77991.892 -62914.645 844579.080 se 12014.308 1118.896 59083.799 282768.727 379638.142 72498.631 126843.317 69971.781 28867.235 686401.019 t -2.219 -1.055 -0.274 -1.773 1.052 1.290 -3.925 -1.115 -2.179 1.230 Beta - Reform*GDP 0.003 0.000 0.000 0.023 -0.023 -0.003 0.036 0.004 0.001 -0.066 se 0.001 0.000 0.001 0.010 0.016 0.003 0.003 0.003 0.000 0.051 t 5.214 1.050 0.217 2.319 -1.460 -1.238 11.467 1.456 3.020 -1.300 Beta - Refo 216.000 14068.012 -1751.459 -122990.802 -488893.714 745182.565 160759.778 -179782.631 -5659.683 -39227.725 742206.010 se 0.922 20234.617 907.394 58994.153 212920.907 423520.615 132069.906 86298.282 92127.697 33041.575 582208.751 t 0.695 -1.930 -2.085 -2.296 1.759 1.217 -2.083 -0.061 -1.187 1.275 Beta - Reform*GDP Lag -0.001 0.000 0.002 0.029 -0.051 -0.005 0.014 0.002 0.002 -0.067 se 0.002 0.000 0.001 0.011 0.024 0.007 0.002 0.004 0.001 0.047 t -0.441 2.319 2.109 2.681 -2.140 -0.767 6.208 0.552 2.946 -1.424 Beta - Reform Lag 2 6017.125 268.190 8927.401 -241755.217 -1828443.252 233221.930 -111030.336 166063.509 -35983.151 558158.882 se 18920.274 1318.421 73783.492 236985.271 751704.312 114650.143 95136.360 91187.640 25325.427 300278.076 t 0.318 0.203 0.121 -1.020 -2.432 2.034 -1.167 1.821 -1.421 1.859 Beta - Reform*GDP Lag2 0.000 0.000 0.000 0.016 0.063 -0.007 0.003 -0.002 0.001 -0.057 se 0.001 0.000 0.002 0.021 0.032 0.007 0.002 0.004 0.001 0.042 t -0.021 -0.712 0.145 0.773 2.002 -1.059 1.270 -0.449 0.909 -1.353 obs 132.000 104.000 97.000 105.000 117.000 118.000 112.000 100.000 145.000 132.000 R-squared 0.961 0.794 0.944 0.984 0.982 0.976 0.984 0.934 0.987 0.932 LR test
Acronyms
For the Brazilian States:
AC Acre AL Alagoas AM Amazonas AP Amapá BA Bahia CE Ceará DF Distrito Federal ES Espírito Santo GO Goiás MA Maranhão MG Minas Gerais MS Mato Grosso do Sul MT Mato Grosso PA Pará PB Paraíba PE Pernambuco PI Piauí PR Paraná RJ Rio de Janeiro RN Rio Grande do Norte RO Rondônia RR Roraima RS Rio Grande do Sul SC Santa Catarina SE Sergipe SP São Paulo TO Tocantins
BNDES Federal Bank for Social and Economic Development CIAT Inter-American Center of Tax-Administrations
205 Acronyms
COFINS Social Security Financing Contribution CONFAZ National Council of Tax Policy (Conselho Nacional de Política Fazendária) CPMF Provisional Tax CSLL Social Contribution on Net Profit Tax FENAFISCO Brazilian Federation of State Tax Collectors FGTS Severance Pay Indemnity Fund GDP Gross Domestic Product IBGE Brazilian Institute of Geography and Statistics ICMS Value Added Tax (by states); IDB Inter-American Development Bank II Import Tax IMF International Monetary Fund INSS National Social Security IPEA Institute for Applied Economic Research IPI Tax on Industrialized Products (Federal VAT) IR Personal Income Tax PIS/PASEP Turnover Tax PNAFE National Fiscal Administration Program for the Brazilian States SEFAZ Same as STAB, local denomination in Portuguese SIAFEM Integrated Financial Management System for the Public Sector SINTAF State Tax and Financial Workers’ Association STAB State Tax Administration Bureau (one per state, e.g. RN/STAB is located in Rio Grande do Norte) TA Tax Administration TAS Tax Administration School TC Tax Collector UCE State Coordination Unit for the PNAFE UCP National Coordination Unit for the PNAFE USAID United States Agency for International Development VAT Value-Added Tax WB World Bank
206
References
Ackroyd, Stephen, Ian Kirkpatrick, et al. (2007). "Public Management Reform in the UK and its Consequences for Professional Organization: a Comparative Analysis." Public Administration, 85 (1): 9-26.
Aidt, Toke S. (2003). "Economic analysis of corruption: a survey." The Economic Journal, 113 (491): 632-652.
Aizenman, Joshua, Kenneth Kletzer, et al. (2007). "Economic Growth with Constraints on Tax Revenues and Public Debt: Implications for Fiscal Policy and Cross-Country Differences." National Bureau of Economic Research.
Amsden, Alice H. (1989). Asia's Next Giant: South Korea and Late Industrialization. New York. Oxford University Press.
Azevedo, Clovis Bueno de (1998). "Os Trabalhadores (Des)Regulados: Quem são os Servidores Públicos Brasileiros?" São Paulo Universidade de São Paulo.
Bahl, Roy W. and Johannes F. Linn (1992). Urban Public Finance in Developing Countries. Oxford [England] ; New York, N.Y. Published for the World Bank [by] Oxford University Press.
Bangura, Yusuf (2000). "Public Sector Restructuring: the Institutional and Social Effects of Fiscal, Managerial, and Capacity Building Reforms." UNSRID (United Nations Research Institute for Social Development).
Barbone, Luca, Arindam Das-Gupta, et al. (N.D.). "Reforming Tax Systems: the World Bank Record in 1990s." Washington, D.C. World Bank.
207 References
Barnett, William P. and Glenn R. Carroll (1995). "Modeling Internal Organizational Change." Annual Review of Sociology, 21: 217-236.
Barzelay, Michael (2000). The New Public Management: Improving Research and Policy Dialogue. Berkeley, CA. University of California Press.
Barzelay, Michael and Babak J. Armajani (1992). Breaking Through Bureaucracy: a New Vision for Managing in Government. Berkeley, CA. University of California Press.
Becker, Gary (1976). The Economic Approach to Human Behavior. Chicago, Illinois. University of Chicago Press.
Becker, Gary S. (1964). Human Capital. A Theoretical and Empirical Analysis with Special Reference to Education. Chicago and London. The University of Chicago Press.
Beetham, David (1996). Bureaucracy. Minneapolis, Minnesota. University of Minnesota Press.
Bendix, Reinhard (1968). "Bureaucracy." In International Encyclopedia of the Social Sciences. New York, NY and Boston, MA. Crowell Collier and The Macmillan Company. p. 206-219.
Berger, Peter and Thomas Luckmann (1966). The Social Construction of Reality. A Treatise in the Sociology of Knowledge. New York, N.Y. Anchor books, Random House, Inc.
Bhatnagar, Subash (2000). "Modernizing Tax Administration in Singapore." Washington, D.C. World Bank.
Binz-Scharf, Maria Christina (2003). "Exploration and Exploitation: Toward a Theory of Knowledge Sharing in Digital Government Projects." Ph.D. Dissertation. Zurich Universistat St. Gallen.
Bird, Richard M. (1991). "Tax Administration and Tax Reform: Reflections on Experience." In Tax Policy in Developing Countries. J. Kalilzadeh-Shirazi and A. Shah (eds.). Washington, D.C. World Bank. p. 38-56.
Bird, Richard M. (2000). "Subnational Revenues: Realities and Prospects." In Proceedings of Decentralization and Accountability of the Public Sector, Annual World Bank Conference on DEvelopment in Latin AMerica and the Caribbean. Washington, D.C. World Bank. p. 319-36.
Bird, Richard M. (2001). "Subnational VATs: Experience and Prospects." In Proceedings of 93rd Annual Conference on Taxation. Washington, D.C. National Tax Association. p. 223-228.
208 References
Bird, Richard M. (2003). "Administrative Dimensions of Tax Reform." Joseph L. Rotman School of Management University of Toronto.
Bird, Richard M., Amaresh Bagchi, et al. (1995). "An Economic Approach to Tax Administration Reform." International Center for Tax Studies, Faculty of Management, University of Toronto. Discussion Paper no. 3.
Bird, Richard M. and Milka Casanegra de Janstcher eds. (1992). Improving Tax Administration in Developing Countries. Washington, D.C., International Monetary Fund.
Bird, Richard M. and Pierre-Pascal Gendron (2001). "VATs in Federal States: International Experience and Emerging Possibilities." University of Toronto International Tax Program.
Blundell, Brian and Alex Murdock (1997). Managing in the Public Sector. Oxford, England. Butterworth- Heinemann.
Bonfim, Washington (1999). "A Reforma Fiscal no Ceará 1987-1997."
Brinton, M. C. and V. Nee (1998). The New Institutionalism in Sociology. Stanford, Ca. Stanford University Press.
Brusoni, Stefano (2005). "The Limits of Specialization: Problem Solving and Coordination in 'Modular Networks'." Organization Studies, 26 (12): 1885-1907.
CIAT, CIAT General Assembly (1996). "Minimum Necessary Attritubes for a Sound and Effective Tax Administration." Santo Domingo, Dominican Repubic CIAT - Inter-American Center of Tax Administrations.
Cohen, Stephen S. and John Zysman (1988). "Manufacturing Innovation and American Industrial Competitiveness." Science, 239 (4844): 1110-1115.
Cornfield, Dan, Holly McCammon, et al. (1996). "In the Community or In the Union? The Impact of Community Involvement on Non-Union Worker Attitudes About Unionizing." In Proceedings of Presented at the AFL-CIO-Cornell University ILR School Union/University Research Conference.
Crandall, William and Jean-Paul Bodin (2005). "Revenue Administration Reform in Middle Eastern Countries, 1994-2004." International Monetary Fund. IMF Working Paper.
209 References
Das-Gupta, Arindam (2001). "Reforming State Level Tax Administration in India: Some Issues." In Proceedings of National Seminar on Public Finance in Indian States: The Emerging Scenario. Bangalore, India, July 11-12. p. 24.
Das-Gupta, Arindam and Dilip Mookherjee (1998). Incentives and Institutional Reform in Tax Enforcement. Oxford and New York. Oxford University.
DEAT (1999). "Restrucuturação DEAT." São Paulo Departamento de Administratação Tributária, SEFAZ (Secretaria da Fazenda do Estado de São Paulo).
DeltaTranslator (N.D.). "Brazilian Taxes." 2007.
DeSanctis, G. and M. S. Pool. (1994). "Capturing the Complexity in Advanced Technology Use: Adaptive Structuration Theory." Organization Science, 5 (2): 121-147.
Dewey, John (1933). How We Think: A Restatement of the relation of Reflective Rhinking to the Educative Process. Boston, New York, D.C. Heath and Company.
Dewey, John (1963). Experience and Education. New York. Collier Books.
Dunford, Richard, Tom Bramble, et al. (1998). "Gain and Pain: The Effects of Australian Public Sector Restructuring." Public Productivity and Management Review, 21 (4): 386-402.
Ebrill, Liam, Michael Keen, et al. (2000). "Value-Added Tax: Principles and Practice (Draft)." Washington, D.C. Fiscal Affairs Department, International Monetary Fund: 201.
Evans, Peter (1995). Embedded Autonomy: States and Industrial Transformation. Princeton, NJ. Princeton University Press.
Evans, Peter (1996). "Introduction: Development Strategies Across the Public-private Divide." World Development, 24: 1033-1037.
Evans, Peter, Claudio Frischtak, et al. (1992). High Technology and Third World Industrialization : Brazilian Computer Policy in Comparative Perspective. Berkeley, CA. Berkeley: International and Area Studies, University of California at Berkeley.
Fainstein, Susan S. (2007). "Using Case Studies to Answer Important Questions." In Proceedings of Association of Collegiate Schools of Planning, October 2007. Milwaukee, WI.
Febres, Jorge, Márcio Goes da Cruz, et al. (1998). "Indicadores de Géstion para El Área Tributária." Washington, D.C.
210 References
Febres, Jorge, Marcio Gomes da Cruz, et al. (1998). "Management Indicators for Tax Administration in Latin America and the Caribbean." Washington, D.C. Inter-American Development Bank.
Ferlie, Ewan, Louise Fitzgerald, et al. (2005). "The Nonspread of Innovatios: The Mediating Role of Professionals." Academy of Management Journal, 48 (1): 117-134.
Fernandes, Andréa Gomes and José Roberto Rodrigues Afonso (2001). "e-Governo no Brasil: Experiências e Perspectivas." In Revista do BNDES, 8: 15. p. 21-64.
Fountain, Jane (2001). Building the Virtual State: Informational Technology and Institutional Change. Washington, D.C. Brookings Institution Press.
Fountain, Jane E. (1994). "Comment: Disciplining Public Management Research." Journal of Policy Analysis and Management, 13 (2): 269-277.
Freeman, Richard (1996). "Through Public Sector Eyes: Employee Attitudes Toward Public Sector Labor Relations In The U.S." In Public sector employment in a time of transition. Dale Belman, Morley Gundersonet al (eds.). Madison, WI. Industrial Relations Research Association. p. iii, 307.
Freeman, Richard B. and James L. Medoff (1984). What Do Unions Do? New York, NY. Basic Books.
Freeman, Richard and Casey Ichniowski (1988). When Public Sector Workers Unionize. Chicago. University of Chicago Press.
Freire, Ana Maria Araújo and Donaldo Macedo (1998). The Paulo Freire Reader. New York. Continuum.
Freire, Paulo (1993). Pedagogy of the City. New York. Contitnuum.
Freire, Paulo (1997). Education for Critical Consciousness. New York and London. Continuum.
Freire, Paulo (2007). Pedagogy of the Oppressed. New York and London. Continuum.
Giddens, Anthony (1984). The Constitution of Society : Outline of the Theory of Structuration. Berkeley. University of California Press.
Gil-García, J. Ramón and Gabriel Purón Cid (2003). "Enacting E-Budgeting in Mexico: Bureaucracy, Institutions and the Budget Process." In Proceedings of International Conference on Public Participation and Information Technologies (ICPPIT-03). Cambridge, Massachusetts, November 10- 12.
211 References
Glenday, Graham (1996). "Perspectives on the Computerization of Tax Systems in Sub-Saharan Africa." In Information Technology and innovation in Tax Administration. Glenn Jenkins (ed.). The Hague, The Netherlands. Kluwer Law International.
Grindle, Merilee S. ed. (1997). Getting Good Government. Capacity Building in the Public Sector of Developing Countries. Cambridge, MA, Harvard University Press.
Grindle, Merilee S. (2000). Audacious Reforms: Institutional Invention and Democracy in Latin America. Baltimore. The Johns Hopkins University Press.
Guaragna, Paulo Ricardo Saldanha (2000). "Secretaria de Estado da Tributação: Fatores de Êxito na Nova Modelagem Institutional da Gestão Tributária do Rio Grande do Norte." In Proceedings of I Plenária Nacional da FENAFISCO. Rio de Janeiro, June 12-16.
Hadler, Sandra C. (2000). "Best Practices in Tax Administration in Sub-Saharan Africa." Cambridge, MA International Tax Program: 33.
Haldenwang, Christian von (2003). "E-Government - An Approach to State Reform in Developing Countries." (German Developement Institute) GDI.
Hedlung, G. (1994). "A Model of Knowledge Managementand the N-form Corporation." Strategic Management Journal, 15 (Summer): 73-90.
Heeks, Richard (1998). "Information Age Reform of the Public Sector: the Potential and Problems of IT for India"." The Univeristy of Manchester Instittue for Development Policy and Management.
Heeks, Richard ed. (1999). Reinventing Government in the Information Age : International Practice in IT- enabled Public Sector Reform. Routledge Research in Information Technology and Society. London ; New York, Routledge.
Hirschman, Albert O. (1958). The Strategy of Economic Development. New Haven. Yale University Press.
Hirschman, Albert O. (1963). Journeys Toward Progress: Studies of Economic Policy-Making in Latin America. New York, NY. The Twentieth Century Fund.
Hirschman, Albert O. (1981). Essays in Trespassing : Economics to Politics and Beyond. Cambridge [Eng.] ; New York. Cambridge University Press.
212 References
Hirschman, Albert O. (1987). "The Political Economy of Latin American Development: Seven Exercises in Retrospection." Latin American Research Review, 22 (3): 7-36.
Hirschman, Alberto O. (1984). "A Dissenter's Confession: "The Strategy of Economic Development" Revisited." In Pioneers in Development. Gerald M. Meier and Dudley Seers (eds.). Washington, D.C. World Bank.
Huber, George P. (1991). "Organization Learning: An Examination of the Contributing Processes and the Literatures, 2 88-115." Organization Science, 2: 88-115.
Huber, George P. and William H. Glick (1993). Organizational Change and Redesign: Ideas and Insights for Improving Performance. New York ;. Oxford University Press.
IBGE (2001). "Contas Regionais do Brasil 2001." IBGE (Instituto Brasileiro de Geografia e Estatística), Brazil. 2002.
IDB (1996). "Brazil: Proposal for a Loan for a National Fiscal Administration Program for Brazilian States." IDB (Inter-American Development Bank). Loan Proposal. BR-0171.
IDB (1999). "Evaluation of Three Bank-Financed Operations of Tax Administration Reforms." IDB (Inter-American Development Bank). Report of Evaluation. RE-237.
Jenkins, Glenn P. (1994). "Modernizing of Tax Administration: Revenue Boards as an Instrument for Change." Bulletin for International Fiscal Documentation, 48 (2).
Johnson, Burce Barner, Flavio Azevedo Marques de Saes, et al. (1996). Serviços Públicos no Brasil. Mudanças e Perspectivas. Concessão, Regulamentacão, Privatizacão e Melhoria da Gestão Pública. São Paulo. Editora Edgard Blucher LTDA.
Johnston, Paul (1994). Success While Others Fail: Social Movement Unionism and the Public Workplace. Ithaca. ILR Press.
Kalilzadeh-Shirazi, J. and A. Shah (eds.) (1991). Tax Policy in Developing Countries. Washington, D.C. World Bank.
Kaplan, Sarah (2004). "Framing the Future: Cognitive Frames, Strategic Choice, And Firm Response to the Fiber Optic Revolutions." Cambridge, MA M.I.T.
Keen, Michael (1991). "Tax Reform." Oxford Review of Economic Policy, 7 (3): 50-67.
Keen, Michael and Ben Lockwood (2006). "Is VAT a Money Machine?" National Tax Journal.
213 References
Keen, Michael and Ben Lockwood (2006). "The Value-Added Tax: Its Causes and Consequences." International Monetary Fund, IMF. IMF Working Paper. WP/07/183.
Kelly, Josi (2007). "Public Service Improvement: Policies, Progress and Prospects." Public Administration, 85 (1): 241-244.
Kelman, Steven (1981). Regulating America, Regulating Sweden: A Comparative Study of Occupational Safety and Health Policy. Cambridge, Mass. MIT Press.
Kelman, Steven (1987). Making Public Policy: A Hopeful View of American Government. New York. Basic Books.
Kelman, Steven (1990). Procurement and Public Management : The Fear of Discretion and the Quality of Government Performance. Washington, D.C.
Lanham, Md. AEI Press with University Press of America.
Kettl, Donald F. (1992). Deficit Politics: Public Budgeting in its Institutional and Historical Context. New York, Toronto. Macmillan; Maxwell Macmillan Canada.
Kettl, Donald F. (1994). Reinventing Government? Appraising the National Performance Review. Washington, D.C. Brookings Institution.
Kettl, Donald F. (1996). Civil service reform : building a government that works. Washington, D.C. Brookings Institution Press.
Kettl, Donald F. (1996). "Governing at the Millennium." In Handbook of Public Administration. James Perry (ed.). California, CA. Jossey-Bass. p. 5-18.
Kettl, Donald F. (1998). "Reinventing Government: A Fifth-Year Report Card." A Report of the Brookings Institution's Center for Public Management. The Brookings Institution.
Kettl, Donald F. and John J. DiIulio (1995). Inside the Reinvention Machine: Appraising Governmental Reform. Washington, D.C. Brookings Institution.
Kettl, Donald F. and John J. DiIulio (1995). "Learning form Public and Private Reform?" In Cutting Government. Washington, D.C. The Brookings Institution.
Kettl, Donald F., Patricia W. Ingraham, et al. (1996). Civil Service Reform: Building a Government that Works. Washington, D.C. Brookings Institution Press.
214 References
Kettl, Donald F. and H. Brinton Milward eds. (1996). The State of Public Management. Baltimore, Maryland, The John Hopkins University Press.
Kickert, Walter J. M. ed. (1997). Public Management and Administrative Reform in Western Europe. Cheltenham, UK and Northampton, MA, Edward Elgar.
Kling, Rob (2000). "Learning about Information Technologies and Social Change: The Contribution of Social Informatics." The Information Society, 16 (3): 217-232.
Kling, Rob (2001). "Social Informatics." In Encyclopedia of Library and Information Science. Kluwer Publishing.
Kling, Rob and Walt Scacchi (1982). "The Web of Computing: Computer Technology as Social Organization." Advances in Computers, 21 (1-90).
Lam, Jermain T. M. (1997). "Transformation from Public Administration to Management: Success and Challenges of Public Sector Reform in Hong Kong." Public Productivity and Management Review, 20 (4): 405-418.
Lasheras, Miguel A. and Isabel Menendez Ros (N.D.). "Tax Administration and Information Systems in Spain."
Lave, Jean (1988). Cognition in Practice. Cambridge and New York. Cambridge University Press.
Lipsky, Michael (1980). Street-Level Bureaucracy : Dilemmas of the Individual in Public Services. New York, New York. Russell Sage Foundation.
Machado, Sandra Maria Olimpio (1998). "Motivação e Satisfação do Servidor Público." Fortaleza, CE Univerdade de Fortaleza - Unifor.
Management, OECD Forum on Strategic (2001). "Tax Administration Aspects of Electronic Commerce: REsponding to the Challenges and Opportunities." OECD.
March, James and Johan Olsen (1989). Rediscovering Institutions: The Organizational Basis of Politics. London Macmillan.
Marinho, Emerson Luís Lemos and Aldemir Freire Moreira (1999). "Esforço Fiscal e Carga Tributária Potencial Dos Estados do Nordeste." CAEN, Universidade Federal do Ceará. Texto para Discussão No. 194/99.
215 References
Marini, Frank (1968). "Public Administration." In International Encyclopedia of the Public Policy and Administration. Jay Shafriz (ed.). New York, NY and Boston, MA. Crowell Collier and The Macmillan Company. p. 1782-1788.
Minogue, Martin, Charles Polidano, et al. (1998). Beyond the New Public Management: Changing Ideas and Practices in Governance. Cheltenham, UK and Northampton, MA. Edward Elgar.
Montgomery, John D. (1981). Bureaucrats and People: Grassroots Participation in Third World Development. Baltimore, Maryland. The John Hopkins University Press.
Nascimento Júnior, Wilson Silva do (2000). "O Auditor do Conhecimento e as Tecnologias de Auditoria Informática." In Proceedings of I Plenária Nacional da FENAFISCO. Rio de Janeiro, June 12-16.
Nogueira, Danilo A. (1999). "The Language of the ICMS Tax in Brasil." Business and Finance, 3 (3).
North, Douglass (1990). Institutions, Institutional Change, and Economic Performance. Cambridge, UK. Cambridge Univ. Press.
Nunberg, Barbara (1990). "Public Sector Management Issues in Structural Adjustment Lending." The World Bank. World Bank Discussion Papers. No. 99.
Nunberg, Barbara (1992). "Managing the Civil Service: What LDCs Can Learn from Developed Country Reforms." The World Bank. Policy Research Working Papers. WPS 945.
Nunberg, Barbara and John Nellis (1990). "Civil Service Reform and the World Bank." The World Bank. Policy Research Working Papers. WPS 422.
Orlikowski, W. J. and C. S. and Iacono (2001). "Research Commentary: Desperately Seeking the "IT" in IT Research--A Call to Theorizing the IT Artifact." Information Systems Research, 12 (2): 121-134.
Orlikowski, Wanda (2002). "Knowing in Practice: Enacting a Collective Capability in Distributed Organizing." Organization Science, 13 (3): 249-273.
Orlikowski, Wanda and Jack J. Baroudi (1991). "Studying Information Technology in Organizations: Research Approaches and Assumptions." information Systems Research, 2 (1).
Orlikowski, Wanda J. (1996). "Improvising Organizational Transformation Over Time: A Situated Change Perspective." Information Systems Research, 7 (1): 63-92.
Orlikowski, Wanda J. (2000). "Using Technology and Constituting Structures: A Practice Lens for Studying Technology in Organizations." Organization Science, 11 (4): 404-428.
216 References
Orlikowski, Wanda J. and JoAnne Yates (2002). "It's About Time: Temporal Structuring in Organizations." Organization Science, 13 (6): 684-700.
Osborne, David and Ted Gaebler (1993). Reinventing Government : How the Entrepreneurial Spirit is Transforming the Public Sector. New York, N.Y. Plume.
Osborne, David and Peter Plastrik (1991). Banishing Bureaucracy: The Five Strategies for Reinventing Government. Reading, MA. Addison-Wesley.
Peled, Alon (2000). "First-class Technology--third-rate Bureaucracy: the Case of Israel." Information Technology for Development, 9: 45-58.
Pereira, Francisco C. (2001). "Letter to Tax Collectors at the STAB/Pernambuco."
Peterson, Steven B. (1996). "Making it Work: Implementing Effective Financial Information Systems in Bureaucracies in Developing Countries." In Information Technology and Innovation in Tax Administration. Glenn Jeckins (ed.). The Hague, The Netherlands. Kluwer Law International.
Peterson, Steven B. (1997). "A Tale of Too Many Systems: The Role of Informatics in Reforming Tax Administration in Developing Countries." Harvard, MA Harvard Institute for International Development.
Piore, Michael J. (1995). Beyond Individualism. Cambridge, MA. Harvard University Press.
Piore, Michael and Charles F. Sabel (1984). The Second Industrial Divide: Possibilities for Prosperity. Basic Books.
PNAFE (2004). "Relatório do PNAFE." Programa Nacional de Apoio à Administração Fiscal para os Estados Brasileiros.
PNAFE (N.D.). "PNAFE é Realidade: A Nova Face da Gestão Fiscal." PNAFE, Programa Nacional de Apoio à Administração Fiscal para os Estados Brasileiros.
Powell, Walter W. and Paul J. DiMaggio eds. (1991). The New Institutionalism in Organizational Analysis. Chicago, IL and London, England, The University of Chicago Press.
Ramos, Alberto (2001). "Medium and Long-Term Fiscal Sustainability in Brazil: 2000-10." In Brazil - Selected Issues and Statistical Appendix. Washington, D.C. IMF.
Rose-Ackerman, Susan (1997). "The role of the World Bank in controlling corruption." Law and Policy in International Business, 27.
217 References
Ruediger, Marco Aurelio (n.d.). "e-Government and Its Impacts on the State and Civil Society: A Latin American Case." In Proceedings of (2003 International Conference on Public Management and Information Technologies) ICPPIT-03, 2003.
Rutherford, M. (1999). Institutions in Economics. The Old and the New Institutionalism. New York. Cambridge University Press.
Schneider, Ben Ross (1998). "Review: Elusive Synergy: Business-Government Relations and Development." Comparative Politics, 31 (1): 101-122.
Scott, W. Richard, Martin Ruef, et al. (2000). Institutional Change and Health Care Organizations: From Professional Dominance to Managed Care. Chicago, IL and London, England. The University of Chicago Press.
SEFAZ (Secretaria da Fazenda de São Paulo) and (Escola Fazendaria do Estado de São Paulo) FAZESP (1999). "Relatorio de Atividades."
Shepard, Jon M. (1970). "Functional Specialization, Alienation, and Job Satisfaction." Industrial and Labor Relations Review, 23 (2): 207-219.
Silvani, Carlos (1993). "The Outlook of Tax Administration." International Monetary Fund/ Fiscal Affairs Department.
Silvani, Carlos and Katherine Baer (1997). "Designing a Tax Administration Reform Strategy: Experience and Guidelines." International Monetary Fund. Working Paper WP/97/30.
Smith, Adam (1776). "The Wealth of Nations."
Smoke, Paul (2003). "Decentralization in Africa: Goals, Dimensions, Myths and Challenges , Vol. 23 (2003), No. 1." Public Administration and Development, 23 (1).
Soboll, Walter (1998). "Implantacao do SIAFEM no Estado de Sao Paulo." In Proceedings of SEminario Nacional O Controle Interno no Contexto da Modernizacao do Estado. Sao Paulo. SEFAZ/SP- FAZESP.
Southern, Alan (2001). "How, Where, Why and When information Technology Are Seen As Regeneration Policy." The International Journal of Public Sector Management, 14 (5): 423-38.
Souza, Celina (2005). "Brazil's Tax System: The Dilimmas of Policy Reform."
218 References
SRF (2002). "Análise da Arrecadação das Grandes Empresas em 1999." Secretaria da Receita Federal, Ministério da Fazenda, Brasil. Estatisticas Tributarias 09.
SRF (2007). "Carga Tributária no Brasil 2006." Secretaria da Receita Federal, Ministério da Fazenda, Brasil.
Stanback, Thomas M. and Jr. (1987). Computerization and the Transformation of Employment. Government, Hospitals, and Universities. Boulder and London. Westview Press.
Stern, David (1992). "Institutions and Incentives for Developing Work-Related knowledge and Skill." In Technology and the Future of Work. Paul Adler (ed.). New york-Oxford. Oxford University Press.
Surrey, S. (1973). "Tax Policy and Tax Administration." In Tax Policy and Tax Reform. W. Hellmuth and O. Oldman (eds.). CCH. p. 531-41.
Syrquin, Moses (1992). "Linkages and the Strategy of Development." In Towards a New Development Strategy for Latin America. Simon Teitel (ed.). Washington, D.C. IDB.
Tanzi, Vito and Anthony Pellechio (2000). "The Reform of Tax Administration." In Institutions and Economic Development. Growth and Governance in Less-Developed and Post-Socialist Countries. Christopher Clague (ed.).
Tendler, Judith (1993). "New Lessons from Old Projects: The Workings of Rural Development In Northeast Brazil." Operations Evaluation Department. World Bank Evaluation Study.
Tendler, Judith (1997). Good Government in the Tropics. Baltimore and London. The Johns Hopkins Univesity Press.
Tendler, Judith (2001). "Comments on "Governace and Upgrading." Paper for Workshop on "Local Upgrading in Global Chains."
Tendler, Judith and Sara Freedheim (1994). "Bringing Hirschmand Back In: A Case of Bad Government Turned Good." In Rethinking The Development Experience. Essays provoked by the work of Albert O. Hirschman. Lloyd Rodwin and Donald A. Schon (eds.).
Tendler, Judith and Sara Freedheim (1994). "Trust in a Rent-seeking World: Health and Government Transformed in Northeast Brazil." World Development, 22 (12): 1771-1791.
Tettey, Wisdom J. (2000). "Computerization, Institutional Maturation, and Qualitative Change: Analysis of a Ghana Public Corporation." Information Tecnolgoy for Development, 9: 59-76.
219 References
Tijerina-Guajardo, José A. and José A. Pagán (2000). "Value-Added Tax Revenues in Mexico: An Empirical Analysis." Public Finance Review, 28 (6): 561-575.
Tyler, Willian (1973). "Measuring Organizational Specialization. The Concept Role of Variety." Administrative Science Quarterly, 18 (3): 383-392.
Ueng, K.L. Glen and C.C. Yang (2000). "Taxation with Little Administration." Journal of Public Economics (75): 145-156.
UN, (Department for Development Support and Management Services United Nations) (1997). "Guidelines for Improving Tax Administration in Developing Countries." New York United Nations.
UNDP, (United Nations Development Programme) (2001). "Making New Technologies Work for Human Development."
Van de Ven, A. (2004). "Organizational Change." Blackwell Encyclopiedic Dictionary of Organizational Behavior. Nigel Nicholson, Pino Audia, and Madan Pillutla (eds.). London: Blackwell Publishers, forthcoming, 2004.
Varsano, Ricardo (1999). "Subnational Taxation and Treatment of Interstate Trade in Brazil: Problems and a Proposed Solution." Washington, D.C.
Vasconcelos, José Romeu de, Marcelo Piancastelli Siqueira, et al. (2005). "Esforço Fiscal dos Estados Brasileiros." In Finanças Públicas: IX Prêmio Tesouro Nacional - 2004. Brasília, Brazil. Editora Universidade de Brasília. p. 175-217.
Vehorn, Charles L. and John Brondolo (1999). "Organizational Options for Tax Administration." Bulletin for International Fiscal Documentation (November).
Wagner, Richard K. and Robert J. Sternberg (1984). "Alternative Conceptions of Intelligence and Their Implications for Education." Review of Educational Research, 54 (2): 179-223.
Waldo, Dwight (1968). "Public Administration." In International Encyclopedia of the Social Sciences. New York, NY and Boston, MA. Crowell Collier and The Macmillan Company. p. 145-156.
Weber, Max (1958). "Bureaucracy." In Max Weber.
Weiss, L. M. (1998). "Collection and Connection: Rationalized and Embedded Knowledge in Knowledge-Intensive Organizations." Cambridge, MA Harvard University.
220 References
Wenger, Etienne (1998). Communities of Practice: Learning, Meaning, and Identity. Cambridge and New York. Cambridge University Press.
White, Leonard (1930). "Public Administration." In Enciclopedia of the Social Sciences. E. Seligman and Alvin Johnson (eds.). New York, NY and Boston, MA. The Macmillan Company. p. 440-450.
Wilson, James Q. (1988). Bureaucracy: What Government Agencies Do and Why. New York, NY. Basic Books.
World Bank (1994). Governace: The World Bank's Experience. Washington, D.C. World Bank.
World Bank (2001). "Tax Administration Organization." World Bank. 2001.
Yancey, Richard (2004). Confessions of a Tax Collector. One Man's Tour of Duty Inside the IRS. New York, N.Y. HarperCollins Publishers.
Zysman, John (1994). "How Institutions Create historically Rooted Trajectories of Growth." Industrial and Corporate Change, 3 (1): 243-283.
221