REINVENTING VAT COLLECTION:
INDUSTRY VERTICAL ASSESSMENT, REVENUE INCREASE,
AND PUBLIC SECTOR RELIABILITY
Mbnica de Maria Santos Fornitani Pinhanez
Prior degree titles:
Bachelor of Laws, S5o Paulo Pontifical Catholic University, 1987
Master in Public Administration, GetOlio 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
M
/
Certified by
V
ica de Maria Sa tos Fornitani Pinhanez
epartment of rb n tudies and Planning
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
n licly paper and electronic copies of the thesis document in whole or in part.
OF TEGHNLOov"Y
MA 32AOTCH
MAR 0 3 2008
REINVENTING VAT COLLECTION:
INDUSTRY VERTICAL ASSESSMENT, REVENUE INCREASE,
AND PUBLIC SECTOR RELIABILITY
M6nica 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
19
Chapter 1 Introduction............................................................................................
19
1.1
What Makes a Good Administration .............................................................................................
1.2
A Novel Measure of Success: Quantitative Modeling of Tax Administration Reforms
1.3
A New Paradigm for Improving Tax Administration and Increasing Public Officials'
Professionalization: Industry Vertical Assessment....................................
26
S tructure of the Disse rtation ..............................................................................................................
30
Chapter 2 Research Design and Methodology.....................................................
33
2 .1
Re s e a rc h Q u e ry ................................................................................................................................
33
2 .2
Re s e a rc h De sign ...............................................................................................................................
34
2 .3
Q ue stio nn aire ....................................................................................................................................
43
1.4
....
.......... 25
Chapter 3 Human and Managerial Determinants of Tax Revenue Growth: A Novel
45
Statistical Model ...................................................................................
3.1
Intro d uc tion ........................................................................................................................................
45
3.2
Brief Literature Review on the Determinants of VAT Revenue ......................................................
46
3.3
New Empirical Evidence, Indicators, and Model ...........................................................................
51
3.4
Da ta ........................................
3.5
Co rre latio n s .......................................................................................................................................
56
3 .6
Re s ults ...............................................................................................................................................
60
3.7
Multiple Reform Variable Estimation Models ...................................
3.8
Other Alternative Estimation Models.............................................................................................
67
3.9
Co nclu sion .........................................................................................................................................
69
3 .10
DB Ind ic a to rs ....................................................................................................................................
70
.........................................................................................................
54
..... 66
Chapter 4 Vertical Economic Assessment of the Tax Basis: The Central Strategy
for Raising VAT Revenues..................................................................... 73
4 .1
Intro d uc tion ........................................................................................................................................
73
4 .2
Cha pte r S tru ctu re ..............................................................................................................................
75
4.3
Conversations Between the New Public Administration and the Administration Reform Literature:
Connections and Im plications ...........................................................................................................
76
4.4
Tax Administration Guidelines for the Brazilian States ................................................................
4.5
Important Tax Administration Traits in Brazil
....................................
81
.... 85
4.6
4.7
From Territories to Industry Segmentation: Production Chain in the Tax Sector and the New
Co g nitiv e P ro c e s s .............................................................................................................................
....................
From Learning by Doing to Learning by Rationally Processing
91
100
Chapter 5 How Technological Changes Reveal Production Chains, Increase Tax
Revenue, and Foster Alliances Across the Public-Private Divide...... 103
5 .1
Intro d uctio n ......................................................................................................................................
10 3
5.2
T he T echnological A rchetype ..........................................................................................................
104
5 .3
T he Enactm ent Fra me.....................................................................................................................
109
5.4
e-T a xin g ...........................................................................................................................................
112
5.5
Computers at Work: The Information Age in the STABs
5.6
Co mputerized S e rvices ...................................................................................................................
12 1
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
......
114
.....................
................
.....................
130
....
133
5.10 Conclusion: Information Systems in Developing Country Bureaucracies
...............
Chapter 6 The Dilemmas of Tax Collectors (TC) and How They Learn a New Way
135
to Increase Revenue Collection ............................................................
6 .1
Intro d u ctio n ......................................................................................................................................
13 5
6 .2
T he J o b ............................................................................................................................................
136
6 .3
T he Ind iv idual ..................................................................................................................................
14 1
6 .4
T h e Pro fe ssio n al .............................................................................................................................
144
6.5
New Processes and New Professionals
6.6
Everything is Illuminated: The Rationale for a Rational Change....................................................
6.7
A ccountability and A lliances............................................................................................................152
6.8
Conclusions: The New Professional and the New Identity .......................................................
..........................................
146
149
157
Chapter 7 Conclusion ..............................................................................................
161
7 .1
W ha t is th e Puzzle .........................................................................................................................
16 1
7 .2
Ax io ma tic Mo d el ..............................................................................................................................
16 2
7.3
The Lasting Power of Vintage Ideas
7.4
Empirical Results: Tax Administration, Reforms and Recommendations.......................................
164
7.5
Implications for Theory and Methodology .......................................................................................
167
7.6
Implications for Technology Implem entation...................................................................................
167
7.7
Implications and Contributions for Public Sector Management
..........................
168
7 .8
Futu re R es e a rc h ..............................................................................................................................
............................................
163
16 9
Appendix A Brazilian Tax Structure: Facts and Figures.................. 171
A .1
Brazilia n R egional D ifferences ........................................................................................................
17 1
A .2
Brazilia n T ax R eve nue ....................................................................................................................
173
A .3
VA T in B razil by S tate .....................................................................................................................
176
A .4
O th er Info rm a tio n ............................................................................................................................
184
Appendix B List of Interviews ................................................................................
185
Appendix C Indicators ............................................................................................
193
Appendix D State Econom ic Development Levels............................................... 199
Appendix E Time-Lagged Predictors.....................................................................
203
Acronyms .................................................................................................................
205
References.................................................................................................................
207
List of Figures
Figure 1. Model Diagram: Technological and Organizational Change => Industry Segmentation
29
=> New Cognition Process => New Institution Image and Identity. ......................
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. Sao Paulo STABS' On-Line Connections with Internal and Outside Agencies. ........ 126
Figure 10 . Integ rated Diagram . .................................................................................................
148
Figure 11. Cognition Axis Assessment Model: New Cognition Embedded in Technological and
Organizational Change =>New (Tax Administration Process)=> New Ties and
162
Ide ntity .....................................................................................................................
List of Graphs
Graph 1. Six Brazilian States' VAT Growth in Percentage from 1997 to 2004 .......................
Graph 2. Brazilian States' GDP Growth in Percentage from 1997 to 2004. .......
21
.............. 22
47
Graph 3. Brazilian Tax Revenue -- Business Related Taxes.................................................
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
65
Mo del .........................................................................................................................
74
Graph 6. VAT and Total State Revenue Growth from 1995 to 2005. ....................................
G raph 7. Tax R eturns O N-LIN E................................................................................................
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
12 3
S ta te s . .....................................................................................................................
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. ....................................................
Table 2. VAT Fiscal Effort, Potential Taxation, and Effective Taxation..................
20
............... 48
Table 3. Fiscal Effort and Trend. All states, 1987 - 1999............................
50
T able 4 . S umma ry S tatistics. ......................................................................................................
57
Table 5 . Co rrelatio n T able . .........................................................................................................
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
T a ble 9. La rg e T ax paye rs . .........................................................................................................
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
122
Table 13. Computerization Processes. .....................................................................................
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
173
Table 18. Brazilian Basic Statistics by State (various years). ...................................................
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
18 1
V alu es . ....................................................................................................................
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.........
...............................
T able 27 . V A T R ates pe r State . ................................................................................................
182
183
Table 28. Selected States Features (Capital City, Territorial Area, Number of Municipalities,
18 4
Po pulatio n). .............................................................................................................
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 destinationgroup, 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,
Marilia, 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 notfanciful, 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?
needs with no excess tax burden on citizens?
Enough revenues to fund social and developmental
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.
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).
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 subnational 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
FEDERAL
STATES
1992 1993
1994
1995 1996 1997
16.72
17.5
18.47
20.46
7.31
7.37
6.48
8
VAT (ICMS)
MUNICIPIOS
TOTAL
1998 1999
2000 2001 2002 2003 2004 2005
2006
21.51
22.24
23.25
23.75
8.4
8.6
8.74
9.14
7.4% 7.4% 7.0% 6.6% 6.7% 7.2% 7.7% 7.6% 7.5% 7.7%
8%
n.a.
20.01
19.35
19.8
20.73
22.47
22.97
23.37
22.08
8.34
8.21
7.95
7.89
8.17
8.69
9.02
8.4
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
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
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.
2
Chapter1. Introduction
rationalization and industry segmentation of tax collection, implementation of computerization and online 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.
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%
4*14I1ahia
00t
wCear6
60%0
Paraiba
0
Pernambuco
-4-Rio Grande do Norte
40%-
0*S~o Paulo
20%
0%A
-20%-
c
p
1b
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 (1DB) jointly coordinated the 1990s tax
administration reforms in Brazil. The 1DB funded the implementation of the reforms and established the criteria for
their evaluation (Febres, Cruz et al. 1998; 1DB 1996).
6 For
the coefficient of correlation and VAT tax buoyancy, refer to Appendix A.
Chapter1. 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%
-*-Bahia
4 Cears
Paraiba
Pernambuco
-Rio Grande do Norte
-6-Sgo Paulo
80%
60%40%20%
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
politicalfortunes of governments, and on the level of economic welfare - may be equally
important.A poor-quality tax administrationmay collect large amounts from easy-to-tax
sectors such as wage earners, while being unable to enforce taxes on business enterprises
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
Casanegrade Janstcher1992).
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, Sdo Paulo, and Pernambuco were all successful in achieving whereas Paraiba and Ceari
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 Sao 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 Ceari
and Paraiba 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
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.' 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 municipiosmunicipalities). 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).
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.' 0
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)," but failed
to achieve the expected gains in efficiency and effectiveness.
2
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.
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:
10Interview
with Yoshiaki Nakano, former head of the Sso Paulo STAB.
" http://www.ciat.org.
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.
1 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.
12
Chapter 1. Introduction
Equations
(1)
VA T= Po + P31GDP + %2Reform + f3GDP*Reform + E,
(2)
6VAT/6GDP=P1 + NReform
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)
computerization, and training).
and
administrative
reform
(i.e.
administrative
efficiency,
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.
1.3
14
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.
5
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 Paraiba
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, Ceari's VAT growth in the same
period was 21%, whereas Bahia's and Pernambuco's were 23% and 35% respectively. In the case of Sdo
14 VAT efficiency is simply defined as a VAT/GDP ratio. I discuss tax efficiency in chapter 3.
is 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 Ceari'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 Sio 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.
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?
governments able to increase taxes without raising tax rates?
Under which conditions are state
What factors caused disparities among
states within the same region and similar economic backgrounds? For example, why did Paraiba 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 subnational 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-basedtax collection system to a vertically-, industry-assessed,production-chainorganized 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
interactionbetween 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
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 participationand 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.
Chapter 1. Introduction
Figure 1. Model Diagram: Technological and Organizational Change =>
Industry Segmentation => New Cognition Process => New Institution Image and Identity.
New
Innovation in the Cognitive Process of Taxation
Institutional
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 cognitiveprocess
that emergedfrom the administrativeprocesses in shaping new identities andphilosophies. 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
interaction between people, technologies, and change in government reform conditions.
the
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.
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. Paraiba and
Rio Grande do Norte, Bahia, Pernambuco, and Ceari, and Sao 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.
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
This chapter analyses the learning paths of public officials, going beyond the literature
training.
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
administration.
for restructuring
public
sector bureaucracies
and
reforming
public
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
strengthenpublic sector bureaucraciesand 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
Chapter2. 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 1DB, the World Bank, and the IMF, and the Ministry of Finance's leading
office, 18 among others.
" Interviews with Jaime Mano, Orlando Reos, and Julieta Verleun in Brazil and in Washington, 2004-2005.
18At the Ministry of Finance, the controlling unit was the UCP, Central Unit, which coordinated and gathered
information about all STABs.
Chapter2. Methodology
Figure 2. Multi Level Methodology.
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
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 understandwhy 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 - interpretationsproduced
by situatedactors" (Giddens 1984, cited in Kaplan 2004). My goal is to generate a theory of the middlerange 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,
Ceari, Paraiba, Rio Grande do Norte, and Sio 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
Chapter2. 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
Chapter2. Methodology
of the poorest regions in the world,' 9 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 Sio Paulo), to rural and agricultural, to developing states (as is the case of Rio
Grande do Norte and Paraiba), encompassing states with a very mixed urban and rural economic reality,
such as Bahia, Pernambuco, and Ceari (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, Sio Paulo can be
compared to the most developed nations, while Bahia, Ceari, and Pernambuco fall among the striving
developing countries, and Rio Grande do Norte and Paraiba would serve as a model for poorer nations.
Figure 3. Brazil - Political Division - Regions and States and States Chosen in this Research.
(Is
R9,P. e
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,
19UNDP indicators on quality of life in Appendix A.
Chapter2. 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, Sao Paulo, Ceari, Rio Grande do Norte, Paraiba, 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 Paraiba fall among the states with the
poorest institutional capacity and lowest economic development.
Ceari can be compared to Bahia and
Pernambuco, given its location, institutional development, and size. Sao Paulo, however, stands on its
own, due to its economic and advanced institutional development (for general information on the statese.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, Ceari, Paraiba, Pernambuco, Rio Grande do Norte, and Sao
Paulo, considered to be relevant cases in the reform process.
Other characteristics make these states an interesting comparative study. Paraiba 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
respectively.
economic and regional conditions apply to Bahia, Ceari, and Pernambuco,
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.
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.
2
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,
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).
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)
21
Chapter2. 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;
e
Archival investigation, based on internal documents and from each STAB and on internal
documents from some of the agencies sponsoring the projects;
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.
2.
3.
4.
5.
6.
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?
What were the first steps taken in the last reform? How did it differ from the previous reforms?
When does the reform really started in this state (e.g. in Bahia)? What was the launch event?
Who was the Secretary (Head of the Tax Administration Bureau) at the time those reforms were
implemented? What was his management style?
What is the functional structure of the STAB? Who does what? What are the specific functions for each
public official?
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:
Since when have entrance examinations been established in this State?
7.
How many years have you worked as a tax collector?
8.
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?
Do you have a choice about place? Have you chosen? Why?
11.
Why have you chosen to be a tax collector?
12.
What is the best thing about being a tax collector?
13.
Please comment on the following professional aspects (after the answer is given):
14.
a. Prestige
b. Wage
c. Working conditions
d. Freedom
15.
Which strategies have been more effective during the STAB modernization process?
Please comment on the following modernization features:
16.
a. Computerization
b. Reorganization
c. Training
Chapter2. Methodology
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
d. Service to taxpayers
Do you disagree with any reform policy strategy? Why?
Which steps in the modernization process do you think are mostly relevant/irrelevant? Could you elaborate
on your answer?
What would you have done differently?
Do you think the current policies are helpful to your work?
Which ones are helpful and which ones are not as helpful?
Which previous policies were more helpful than the new ones?
What was your education level before being a tax collector?
What is your background?
Do you think that your previous background helps/helped you with your career as a tax collector? How?
Do you think the Tax Administration Bureau is doing something to take advantage of it?
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:
What is your business?
39.
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?
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 Sdo 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
Chapter 3. StatisticalModels
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 revenuemaximizing rate" (Keen and Lockwood 2006). Keen and Lockwood model the revenue ratio r as:
ri,= aVi, +
',
+ /',Vi, X
+ p,,+
At+ ui,
where Vi is a dummy variable if a country has VAT, X, captures other variables affecting the tax ratio,
the term pi is a country-specific effect, A, 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).
Chapter3. StatisticalModels
Graph 3. Brazilian Tax Revenue
-
Business Related Taxes.
700000-
600000
500000
E Federal VAT
M BUSINESS INCOME TAX
400000
M Industrialization Tax
Total Import Tax
----300000
-VAT
E3 Total Federal Tax Revenue
200000
0-
--
- --
- -- -
-- - - - -
--
-
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 Pagin 2000) use a roll of variables including VAT rates,
manufacturing production, government spending; and exemptions, number of VAT rates, and the rate of
inflation.
Chapter3. StatisticalModels
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:
InqTCM )=A +A In(GD )+Al(,
InPg},) +A ln(X,)+A In(M,)+
In(rc,) +,86n(D)+v,,-ui,
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; rc 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.
MA
Pl
CE
RN
PB
PE
AL
SE
BA
CE
3,49
4,38
5,37
3,77
4,83
6,59
4,68
5,85
5,95
1991
EF
CP
68,11 5,13
84,40 5,18
83,72 6,42
65,52 5,75
77,64 6,22
99,99 6,59
96,29 4,86
98,99 5,91
88,67 6,71
1992
E F
3,85
76,39
4,69
98,70
4,94
76,97
3,41
59,88
4,40 76,14
6,17 93,57
4,39 88,25
5,16
96,61
6,30
96,26
CP
5,C4
4,75
6,42
5,69
5,77
6,60
4,98
5,34
6,54
CE
3,31
4,36
4,80
3,24
4,05
5,80
4,07
5,41
6,03
1993
EF
66,0B
91,12
74,89
59,19
71,13
90,27
89,55
97,18
92,70
CP
5,02
4,78
6,41
5,47
5,69
6,43
4,55
5,56
6,50
CE
33
4,06
5,41
3,73
4,19
6,27
4,42
6,75
6,41
1994
EF
61,36
65,55
68,55
51,93
59,49
79,74
72,32
9B,39
79,64
CP
6,07
6,20
7,89
7,18
7,05
7,87
6,11
6,86
8,05
CE
3,84
5,12
5,90
4,69
5,29
6,93
5,18
7,37
6,79
1995
EF
55,51
74,27
73,75
56,31
63,17
78,74
76,51
92,13
74,32
CP
6,92
6,89
8,00
8,15
8,37
8,80
6,78
8,00
9,14
CE
4,18
5,63
6,18
4,91
5,66
7,47
5,38
7,31
7,16
1996
EF
54,82
70,97
61,58
54,43
63,78
73,48
73,64
90,53
72,03
CP
7,63
7,94
10,03
9,03
8,87
10,16
7,30
8,08
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
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 colunm indicate the trend in the state to
increase or decrease the fiscal effort. Interestingly, Bahia, Ceari, Paraiba, and Rio Grande do Norte all
present positive trends.
Sao 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.
Chapter 3. StatisticalModels
Table 3. Fiscal Effort and Trend. All states, 1987 - 1999.
State
Effective Tax
Burden
Potential Tax
Burden
Fiscal Effort
Trend between
1987 and 1999
(%)
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
SC
SE
SP
TO
Source
5.221
9.251
16.367
5.976
12.918
14.363
8.582
18.738
17.744
9.325
13.269
15.524
20.668
7.502
12.506
13.819
12.802
10.070
12.029
9.743
10.743
8.334
14.494
14.704
11.471
14.621
20.795
5.363
5.363
16.13
4.256
12.702
14.272
10.237
17.513
16.626
9.101
12.660
13.359
19.765
7.420
12.325
13.681
13.161
8.610
12.300
9.579
11.640
8.074
13.724
14.035
11.530
14.225
22.170
0.971
1.067
1.016
1.417
1.013
0.999
0.996
1.046
1.033
1.012
1.024
1.157
1.044
1.002
1.013
1.016
0.971
1.175
1.021
1.005
0.962
1.034
1.025
1.027
0.993
1.023
0.967
1.13
-3.13
0.08
3.65
0.15
0.65
7.63
0.54
-132
1.33
-1.25
-9.68
-3.82
-2.10
0.52
-1.11
-0.42
-9.51
4.34
0.94
2.45
-4.77
-1.42
-1.67
-0.65
-1.61
5,41
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.
Chapter3. StatisticalModels
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.
Although the role of indicators in detennining tax revenue has been well explored in recent
23
studies , 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.
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).
22
Chapter3. StatisticalModels
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
VA Tit=
PO + J3GDP + J 2Reform + P3GDP*Reform + 8i
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
6VAT/GDP = P1
+
P3Reform
So, P1 can be interpreted as the average tax effort or effective VAT rate, and
P3
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
%3is 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
Chapter 3. StatisticalModels
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.
Chapter 3. StatisticalModels
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%.
Chapter 3. StatisticalModels
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
Chapter 3. StatisticalModels
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
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
Mean
Obs
Std. Dev.
Min
Max
Measures of Administrative Efficiency:
AdminCost
TaxCompliance
DebtNotifications
DebtCollected
DebtInstallments
DebtOutstanding
DebtDisputed
DebtJudged
LargeTaxpayers
191
197
161
163
180
177
171
165
198
3.52
67.40
1.11
0.14
0.07
0.57
0.25
0.59
0.08
2.59
25.30
2.53
0.25
0.09
0.72
0.30
0.60
0.15
0.00
4.25
0.00
0.00
0.00
0.00
0.00
0.00
0.00
20.0
100.00
22.60
1.73
0.75
4.42
1.75
3.62
1.00
Measures of Computerization:
OnlineBudgeting
OnlineFiling
OnlineValidation
AdminConnectivity
187
197
199
207
0.86
0.54
0.48
0.71
0.35
0.50
0.50
0.46
0.00
0.00
0.00
0.00
1.00
1.00
1.00
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. DebtOutstandingpositive correlation with DebtDisputedand
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 DebtOutstandingand 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 Online Validation 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
Chapter 3. StatisticalModels
correlation between DebtOutstandingand 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 DebtCollectedand 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, DebtJudgedis 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.
Table 5. Correlation Table.
AdminCost TaxComplianceDebtNotificationsDebtCollectedDebtInstallmentsDebtOutstanding DebtDisputed
1
AdminCost
1
TaxCompliance
-0.044
DebtNotifications
-0.136
0.220*
1
DebtCollected
-0.069
0.058
0.157
1
DebtInstallments
-0.138
-0.057
0.031
-0.072
1
1
DebtOutstanding
-0.053
-0.138
-0.051
-0.292*
0.353*
DebtDisputed
0.183*
0.112
-0.115
-0.157
-0.127
DebtJudged
-0.091
0.013
0.067
0.042
0.344*
0.600*
-0.096
-0.045
0.171*
LargeTaxpayers
0.099
-0.050
-0.102
-0.018
OnlineBudgeting
-0.013
0.025
0.113
0.042
0.128
-0.15*
1
-0.090
0.015
OnlineFiling
0.196*
0.059
-0.122
-0.018
0.178*
0.161*
-0.067
0.149
-0.001
0.134
0.048
0.115
OnlineValidation
-0.004
0.129
0.145
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
1
DebtJudged
1
LargeTaxpayers
-0.157
OnlineBudgeting
0.270*
-0.088
OnlineFiling
0.299*
-0.201*
0.141
Online Validation
0.294*
-0.179*
0.184*
0.410*
AdminConnectivity
0.175*
-0.155*
0.238*
0.201*
Training
0.050
Training
-0.019
1
0.058
1
0.049
1
0.309*
0.046
1
-0.013
1
Chapter3. StatisticalModel
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.
GDP explains approximately 92% of the variation of VAT collection, as the
R2
is 0.92. The value of pi is
0.079, or 7.9%, which is interpreted at the average effectual VAT rate. The standard error of Pi 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.
Chapter 3. StatisticalModel
Graph 4. Best Fitted Line (VAT by GDP). Simple Regression Model.
C0
0
0
0
0
0
0
0
0
400,000
200,000
GDP (1000's nominal Reais)
* VAT (1000's nominal Reais)
--
600,000
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.
Table 6. Results of the Base Model Estimation and Reform Indicators.
Model
Base Model
AdminCost
Constant
-283,572
(189,845)
-280,170
(173,493)
-106,001
(142,198)
103,002
(103,912)
-129,090
(121,976)
-105,483
(133,261)
-224,917*
(135,510)
-176,719**
(88,671)
p-GDP
0.0793***
(0.0047)
0.0761***
(0.0037)
0.0711***
(0.0042)
0.0687***
(0.0031)
0.0742***
(0.0034)
0.0745***
(0.0034)
0.0789***
(0.0038)
0.0722***
(0.0025)
-38,200**
(15,608)
-1,557
(1,193)
-59,751
(60,653)
-144,678
(280,605)
343,167
(485,564)
88,118*
(46,645)
-164,088
(141,551)
0.0025**
(0.0010)
0.000084
(0.000054)
0.000752
(0.000995)
0.012
(0.011)
-0.0019
(0.0206)
-0.0037**
(0.0019)
0.024***
(0.008)
191
0.929
197
0.952
161
0.954
163
0.970
180
0.945
177
0.955
171
0.975
P -Reform
P -Reform*GDP
N
R2
216
0.922
TaxCompliance
DebtNotification
DebtCollected DebtInstallments DebtOutstanding DebtDisputed
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.
AdminConnectivity
-302,426*
(167,697)
35,506
(52,308)
-230,344
(187,710)
-33,898
(76,141)
0.0755***
(0.0045)
0.0783***
(0.0047)
0.0665***
(0.0028)
0.0764***
(0.0049)
0.0715***
(0.0019)
303,676
(283,438)
-62,976
(51,572)
30,196
(54,840)
-220,409**
(95,785)
-60,623
(45,792)
-62,606**
(25,833)
0.0073***
(0.0028)
-0.021
(0.022)
0.0038***
(0.0011)
0.00096
(0.00217)
0.0118***
(0.0034)
0.0034***
(0.0008)
0.0017***
(0.0004)
165
0.987
198
0.915
187
0.929
197
0.922
199
0.987
207
0.913
203
0.982
LargeTaxpayers
OnlineBudgeting
Constant
4,966
(89,739)
-292,548*
(177,268)
-213,471
(176,418)
p-GDP
0.0680***
(0.0040)
0.0794***
(0.0048)
-146,713**
(69,773)
B - Reform
B - Reform*GDP
N
R2
Training
Online Validationt
DebtJudged
Model
OnlineFiling
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.
t: Online Validation: 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.
Chapter3. StatisticalModel
For each point-percentage increase in administrative cost, AdminCost, there is 0.25 percentagepoint 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 DebtJudgedleads 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.
Online Validation, 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
Chapter3. 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 Sio 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.
200,000
0
400,000
GDP (1000's nominal Reais)
600,000
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.
Chapter3. StatisticalModel
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
performance,
most
notably
OnlineFiling, OnlineBudgeting,
OnlineFiling
of administrative
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, Admin Connectivity, 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, DebtOutstandingand DebtDisputed'slevel 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.
Chapter 3. StatisticalModel
Table 7. Multiple Variables Models.
DebtOutstanding
DebtDisputed
DebtJudged
OnlineBudgeting
AdminConnectivitv
Training
Constant
-212,148
(146,013)
-176,242*
(101,986)
17,381
(91,795)
-116,445
(172,584)
-174,686
(161,512)
-26,794
(86,595)
GDP
0.0751***
(0.0032)
0.0708***
(0.0026)
0.0673***
(0.0038)
0.0693***
(0.0031)
0.0713***
(0.0034)
0.0726***
(0.0017)
AdninCost
-40,229**
(16,198)
940
(19,110)
1,927
(13,280)
-47,014***
(15,029)
-40,516***
(13,510)
16,479
(16,720)
GDP *AdminCost
0.0026***
(0.0010)
0.0008
(0.0012)
0.0003
(0.0010)
0.0032***
(0.0007)
0.0030***
(0.0008)
-0.0010
(0.0011)
DebtOutstanding
102,789**
(46,197)
GDP * DebtOutstanding
-0.0038**
(0.00 16)
DetDisputed
-102,998
(220,470)
GDP * DebtDisputed
0.0220**
(0.0108)
Debtdudged
-137.695**
(67,921)
GDP * Debtiudged
0.0074***
(0.0027)
OnlineBudgeting
-143,546**
(71,123)
GDP * OnlineBudgeting
0.0057***
(0.0010)
AdminConnectivity
- 18,26 1
(54,679)
GDP *AdminConnectivity
0.0049***
(0.0010)
Training
-73,761**
(29,326)
0.002 1***
GDP * Training
(0.0005)
161
0.957
160
0.983
155
0.990
170
0.949
187
0.936
182
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
Chapter3. StatisticalModel
modified to include two dummies - for the least and second least developed states (5t and
62)
-
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-I) and t (,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 ofAdministrative 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.
Chapter3. StatisticalModel
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, A dminConnectivity, 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 DebtOutstandingas 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 AdrninCost 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.
Chapter 3. StatisticalModel
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 1DB 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
(Admin Connectivity
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 (Online Validation)
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
Chapter 3. StatisticalModel
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%)
Total installments
10. Revenue negotiated (DebtInstallments)
Total revenue
Total amount in Reals to be paid in installments (except those under dispute) divided by total STAB
revenue. Goal: 0. 07 (0. 7%).
Value under dispute
Total revenue
11. Revenue under dispute (DebtDisputed)
Total amount in Reais under dispute in 1s' and
2 "dAppeal
Court divided by total revenue. Goal: 0. 2
(20%)
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%).
Trained Tax Collectors
13. Training and Professionalization (Training)
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%).
Tax payers paying more than 50%
14. Tax concentration by large taxpayers (LargeTaxpayer)
Total # taxpayers registered
12. Revenue to be received judicially (DebtJudged)
Total number of taxpayers who are responsible for 50% of tax collection in each state until December 3 1
of each given year divided by total number of tax collectors registered by the STAB. Goal: 0. 0006 (0.
06%)
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.
designed by the Brazilian
Ministry of finance, international
donors, such
Following guidelines
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).
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 Sao Paulo, 90% in BA, 88% in CE, 85% in
PB, 87% in PE, and 85% in RN.
24
Chapter4. Vertical Assessment
Graph 6. VAT and Total State Revenue Growth from 1995 to 2005.
200000000
180000000
160000000
140000000
1
120000000
100000000
Total State Tax Revenue
-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 ofproperty investment-are connected to certain output variablese.g. expenditure on low-cost housing, growth in the urban economy-but it could tell us
little about why these outputs occurred. The typical quantitativepaper that links policies
Chapter 4. Vertical Assessment
to outcomes usually has a section at the end speculating on reasonsfor the relationship,
but unless it is accompaniedby 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 rationalfirm 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 decisionalprocesses of the firm. " (Fainstein2007)
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 revenuebut that is not captured by statistical analysis: the vertical industry segmentation of tax basis.2 6 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
For this dissertation I use vertical industry segmentation, economic segmentation, and industry specialization
interchangeably.
26
Chapter 4. VerticalAssessment
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: Sdo Paulo (SP), Rio Grande
do Norte (RN), Paraiba (PB), Pernambuco (PE), Ceari (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 supplychain 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 Dilulio 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 wellstructured 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.
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.
27
Chapter4. 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, standardizationof procedures, the improvement of techniques and the adjustment of
the public workers to theirjobs (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 managerialprescriptions.
In this sense, efficiency and accountability should be measured according to market standards. AngloAmerican 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
28Osborne
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).
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 needfor 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.
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 govermnent (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).3o 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 noncompliance, 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
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.
30
Chapter4. 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
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. 3 '
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
administrationagencies in Brazil;
strengthen and integratethe financial administrationand consolidateaudit and internal control;
See IDB (1996) and Cruz et al. (1999) for the 1DB indicators. See also Chapter 3 and Appendixes C.
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 actionsfor collection of delinquent tax debts and strengthen integration between
the tax administrationand the agencies responsiblefor 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 integratedtax administrationsystem. " (IDB 1999)
The financial subcomponents included:
"studies and analyses of public spending; integratedfinancial administrationsystem;
budget, cash management; and governmental accounting; control and management of
public spending; and institutionaland regulatoryframework 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.
Chapter4. 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 Ceari, 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 responsiblefor 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.
Chapter 4. VerticalAssessment
Table 9. Large Taxpayers.
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
1997
nd
0.15%
0.04%
nd
0.02%
0.05%
0.04%
nd
0.01%
0.02%
0.02%
0.08%
0.04%
0.10%
0.15%
0.04%
0.01%
0.01%
0.01%
nd
nd
0.14%
0.00%
Large Taxpayers
Taxpayers paying > 50% VAT Revenue
Total Taxpayer-s
2000
2001 | 2002 I
1998 1 1999
nd
0.03%
0.11%
0.09%
0.09%
0.11% 0.10%
0.05%
0.06%
0.05%
0.04%
0.04%
0.06% 0.05%
0.05%
0.08%
nd
nd
0.14%
nd
0.02% 0.01%
0.00%
0.00%
0.01%
0.02% 0.02%
0.03%
0.02%
0.02%
nd
nd
0.10%
0.05% 0.01%
0.04%
0.04%
0.02% 0.03%
0.03%
0.01%
0.02%
0.03% 0.02%
0.08%
0.03%
0.02%
nd
nd
0.02%
nd
0.00%
0.00%
0.01%
0.00%
0.06%
nd
0.00%
0.13%
0.67%
0.02% 0.01%
0.01%
0.01%
0.02%
0.04% 0.05%
0.33%
0.22%
0.63%
0.15% 0.20%
0.30%
0.10%
0.10%
0.06% 0.05%
0.05%
0.02%
0.02%
0.06% 0.08%
0.04%
0.04%
0.02%
0.01%
0.03% 0.01%
0.01%
0.01%
0.01% 0.00%
0.01%
0.00%
0.01%
0.03% 0.06%
0.05%
0.08%
0.08%
0.04% 0.08%
0.10%
0.25%
nd
nd
nd
0.09%
0.09%
0.30%
0.00%
0.00%
0.00%
0.00% 0.00%
SC
SE
0.04%
0.50%
0.03%
0.50%
SP
-:
0.02%
0.50%
0.01%
0.48%
0.02% 0.02% 0.01%
0.01%
.
0 0TW0% 0
. o~T
O . 00.oT-TF/77Tm
2003
0.20%
0.03%
0.03%
0.26%
0.00%
0.01%
0.65%
0.03%
0.00%
0.04%
0.00%
0.01%
0.02%
0.02%
0.06%
0.02%
0.04%
0.00%
0.00%
0.10%
0.05%
0.21%
0.00%
2004
0.17%
0.03%
0.03%
0.26%
0.00%
0.01%
0.30%
0.03%
0.00%
0.03%
0.01%
0.01%
0.02%
0.05%
0.06%
0.03%
0.04%
0.00%
0.00%
0.03%
0.00%
0.21%
0.00%
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
P1
PR
RJ
RN
RO
RR
RS
0.01%
0.97%
0.01%
0.07%
0.00%
0.06%
0.00%
0.05%
SC
SE
0.01%
.
0.01%
0
0.01%
O
0.00%
.
SP
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
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).
3
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 accompaniedby specialprograms
for controlling medium-size and small taxpayers, who represent the largestproportionof
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:
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, 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 transactionsare 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 intrastatetransactionscomplicates administrationandprovides
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."
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: Sio 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 Rond6nia 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,
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).
"Tax 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.
3
Chapter 4. Vertical Assessnent
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.
Tax Administration Reform
April 1997
January 1997
January 2000
March 1997
February 1997
January 1996
November 1997
January 1997
January 1997
December 1996
January 1997
March 1996
February 1996
August 1996
January 1996
July 1997
January 1997
January 1997
January 1997
March 1996
October 1997
July 1997
January 1997
January 1997
January 1997
June 1998
SP
April 1997
TO
Source: PNAFE 2004 (PNAFE).
State
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
P1
PR
RJ
RN
RO
RR
RS
SC
SE
Budgeting Administration Reform
January 1997
January 2000
April 1997
December 1999
July 1996
May 1997
January 1997
March 1996
October 1998
December 1996
January 1998
July 1997
January 1998
April 1997
January 1997
July 1997
January 1997
October 1997
March 1994
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:
Chapter4. 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 administrationreforms." Ednilton Soarez, 2001, priorhead of
the STAB/CE.
"Sdo Paulo is and has always been the locomotive of the Brazilian economy. Of course,
it was ahead in the reform process." Yoshiaro Nakano, 2001, priorhead of the STAB/SP.
"Bahia has been at the head of process since the project Baia Azul in the early 1990s.
Although the Baia Azul project focused on the cleanliness of the beach in the state's
capital city, we needed to resolve ourfinances because we were running out offunding
and internationalorganizationsrefused to lend more money, given the state'sfiscalcrisis.
Before continuing with the Baia Azul project, we had to "clean" ourfinances. "Dimas de
Melo, 2001, STAB/BA.
"Sdo Paulo, Bahia, and Ceari were way aheadin 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 administrationreform 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 Federalgovernment and by the Senate), the first
step toward the implementation of the reforms, the representatives of the other states
barredthem. All the states wanted to participateon the program. So we had to waitfor
all the states to get ready and put together projects so that we could present one whole
project. Only then did we get approvalfor the loan to find 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, Amapi, Roraima, Rond6nia, Tocantins, Sergipe, the Federal District, Piaui,
Alagoas, Mato Grosso do Sul, Rio Grande do Norte, Paraiba, and Mato Grosso; which together accounted
Chapter4. 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, Maranhio, Para, Espirito Santo,
Goiis, Ceari, 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 Sdo Paulo, Parani,
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 apartfrom the system and we were a part of the system. We worked with a
parallel structure that allowed us to work separatedfrom the original STAB structure
and that createdfor us a protected zone within which we could examine the Bureau. We
could work and rethink the traditionalstructure, 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 organizationfrom inside. But at the same time, we kept our
positions in the originalSTAB and continued working day-by-day. So we were somewhat
within and somewhat outside. We were insulatedfrom 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/Sao
Paulo, 2001.
An internal unit studied reform in each STAB.
(Programa de Modemizaeio Tributiria).
PROMOCAT, and PROMOSEFAZ.
In Sio Paulo, this group was called PROMOFAZ
In other states, it had similar names, such as PROMOAT,
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
"proudof 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 privilegedor a traitor,and to have a down-to-earth perspective on
the reform. Nelson Fagundes, Tax collector,Sdo 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
Chapter4. 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
Previously, tax collection had been performed according to the territorial allocation of tax
Brazil.
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 happenedin 1996 is still irregularnow. Then, we
would notify the taxpayer and leave. It did not guaranteethat 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
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
j.IiW
,,;i
""i""
"'Backward
~
Rationale:
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 Jinior 2000).
Chapter4. 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 Sao 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
accountabilityof these firms my raise questions and difficultiesfor 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
understandingeach aspect of the tax phenomenon, departingfrom a study ofjust afew of
the multiple levels of relationship and complexity. Taking into account the environment
Chapter 4. Vertical Assessment
with its diverse forms, interactions, and objectives, it allows an understanding of how
taxpayers work to survive and compete.. (Nascimento Junior 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."
(Hirschman1987)
"The trade off between "operationality"and "sweepingness,"the switch from strategy to
interpretation,the originalemphasis was on tradables and specifically industrialsectors
whose production was to be stimulated in a strategy of industrializations." (Syrquin
1992)
Chapter4. VerticalAssessment
Figure 5. Vertical Industry Assessment.
Axiomatic Rationale:
backward and
forward linkages
Information
.
-SS
6
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 Sao 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 Sdo Paulo has one million. Our goal is not to
reachjust sectors, but the whole economy." (Ludemilson Araujo Lopes, TC and chief-ofstaff STAB/RN).
In Pernambuco, the PROMOFAZ's tax administration procedures and modernization project was
called REMAF (Revisio do Modelo de Agao 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 infragdo). 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.
Chapter4. Vertical Assessment
The Bahia and Goiis' STABs were considered the role models for STABs in the group of poor
and developing states. Goiis 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 broughttogether 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 responsiblefor the more than 95% of the tax collection."
According to Ferreira, segmentation allowed for the implementation of another tool in taxation,
the 'substituigdo tributdria', 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 PernambucoSTAB.
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 evasionfor each firm, and
then ascribe appropriateactionsfor 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 operationalfactor. 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 similarfirms in the same economic
segment." Tax collector in Pernambuco, 2001.
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 crossedseveral 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 theirfiscal 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 itfrom 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 Piauistate, what is happening with a particulartaxpayer, even
ifI have never met them." Tax collector in Bahia.
The following describes the minutiae carried out to map the restaurant segment.
"The restaurantsegment is quite large and is known for evading taxes. Just think how
many times you go to a restaurantand leave without getting a receiptfor the total check,
much less the tax receipt. Of course, the restaurantowners deny any claim of evasion
and it was very difficult to get hold of their suppliers and inventories. First of all,
Chapter4. VerticalAssessment
inventories are not well kept and documented, and second, many suppliers are from the
informal sector. Thus, the restaurantindustry 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 restaurantin the mall. Usually, mall restaurantsare quite busy
and have a high turnover. The students had to check every single customer who left the
restaurantand 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 (fiscalizagdo de estabelecimentos); (3) transport and commodities circulation policies and audits
(mercadoriasem tr6nsito) 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 Bureaufor 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 approvedby the Director."
Figure 6 shows the results for the wholesale food segment when reporting to his superiors and
colleagues in the Pernambuco STAB.
Chapter4. 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
Month
1999
1999
2000
% Growth
Rev
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%
10291933
26611833
159%
39048236
41109605
3%
Total
.
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.
,7
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
understandwhole economy." Segment manager in PernambucoSTAB.
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 collectorsfor 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.
Chapter 5
How Technological Changes Reveal Production
Chains, Increase Tax Revenue, and Foster Alliances
Across the Public-Private Divide
"I have to conftss 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.3
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.
Chapter5. 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 nongovernmental agencies;
e
the elimination of intermediaries (despachantes) in the tax collection process;
e
dissemination of information both within and outside the STAB, which has fostered transparency
and credibility for the STAB; and,
e
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
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.3' This means that the computer
38Heeks
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 toolfor administrative reform andfor change in any sector.
Brazil is actively participatingin 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 intofour categories: eservices, 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-Garciaand 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
Chapter5. 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 infbrmation 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. Priorto carryingpersonal laptop computers to the audits, we
could not access this kind of information. The information about the taxpayer, even
whether the firm was legally registeredor 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 informationfrom 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:
"shareddatabases are not possible without standardizeddata, and once developed, they
create a platformfor 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 responsiblefor that particularinformation was
not available,we had to waitfor 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, informationflow, and
performance [providing] the technologicalfix. " (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
Chapter5. 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 institutionalizedsocio-structuralmechanisms even when
such actors tend to enact technology to preserve ongoing social or network relationships
and to maintainperformanceprograms." (Fountain2001).
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
institutionalstructures influences the design, perceptions, and uses of the Internet and
relatedIT." (Fountain2001)
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.
Chapter5. New Technology
Figure 7. Technological enactment framework (Fountain 2001).
IT
Features
Enacted
Organizational
IT
4
Outcomes
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-Garcia 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.
(Gil-Garcia and Cid 2003) make a similar argument with respect to e-budgeting in Mexico.
Chapter5. 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.
technology produces certain organizational outcomes.
The resulting enacted
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 allinclusive.
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
Chapter5. New Technology
Figure 8. Technological Enactment Framework for
e-Taxing based on Fountain (2001).
Oil
Organization:
New Rationale;
Vertical Industry
Assessment
STAB
(I)M
/
Institutional Settings
113
Chapter5. 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.4
"At that time, we had mammoth machinesfor 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, Brasilia, 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: Sdo Paulo, Bahia, Rio de Janeiro, Parani, Pernambuco, Rio Grande do Norte,
Paraiba, Maranhao, and Ceari (Interview with Dimas Melo in BA, STAB).
The joint effort reflected
needs and demands in each of these states. Some, such as Sdo Paulo and Bahia, had already equipment to
facilitate this reform. However, even wealthy states had to improve their technological infrastructure in
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 Luis 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.).
40
Chapter5. New Technology
order to develop their collection capabilities. Of the 27 STABs, only Sao Paulo, Parani, 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%
ECE
MRN
70%
60%
50%
Percentage of
Online Filing
40%
OPE
E BA
MPB
EsP
30%
1997
-20%
1999
Year
10%
2001
0%
PE
CE
RN
BA PB
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.
"4
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
Chapter5. 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 accumulatingdust. "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.
rapidly.
In Brazil, government expenditures on IT have grown
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?
Chapter5. New Technology
Graph 8. STABs' Administrative Costs as a Percentage of Tax Revenue, Selected States.
Administration Cost
*""BA
"'O'-CE
PB
PE
-4- RN
-'$SP
0
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, Sao Paulo, Rio Grande do Norte, and Ceari, 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 "marketoriented 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.
State
Poitca Prt
Political Party
HadSTB
Head STAB
BA
Paulo G. Souto - Liberal Party
Mascarenhas
CE
Ciro Gomes & Tasso Jereissati
PSDB
Ednilton
Soarez
PB
Jose Taguino Maranhao PSB
PE
Jorge Jatoba PSB
RN
Garibaldo Alves Fo. PMDB
Ubirajara
SP
Mario Covas PSDB
Y. Nakano
Internal
Opposition
Business
Class
Oscillating
Strong
Very strong
+ elitist
Inactive
n.d.
4
Strong
Strong
8
Conciliatory
Strong
8
Conciliatory
Strong
# of
Years
7
+ 10
-
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 Sao 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 Sao Paulo
STAB 2001.
When asked about the computerization of the offices, a TC said:
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 Sao Paulo, STAB, commented:
"On the one hand, we establishedthat in thefirst 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 Sio 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 Sao 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 notfighting 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 retiredtax collectors. But afew will of course
leave. Not because of the computers, but because they have been aroundfor too long."
The availability of new technologies in information processing and the possibility of crosschecking 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.
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 Sdo
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,
transaction history, and criminal records.
Automated Assessment
Using software to automate assessment of tax returns.
Taxpayer Accounting Record
Collection
Electronic collection of firms' accounting information.
Data Mining for Auditing and
Investigation
Creating automated links with other sources to obtain information
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
Taxpayers
Systems to identify new taxpayers using data from other firms.
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
previous
tax
returns,
Chapter 5. New Technology
Graph 9. On-Line Budgeting in Selected States.
On-Line Budgeting
120%
100%
0
.9
60%
40%
0
20%
0%
K
-
Source: IDB.
Graph 10. STAB On-line Connectedness (for data sharing)
with Other Agencies in Selected States.
Source: IDB.
123
Chapter5. New Technology
Graph 11. On-Line Validation in Selected States.
On-Line Validation
120%
100%
-
"a
> 80%
lliBA
""C
EC
=4PB
CWr 60%""
meP
*RN
0
40%
enp
20%
0%
-%
Nq
rr,
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
Chapter5. 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 Sao 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 Jinnior 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
establisheda new paradigmfor information transparencyand the management of public
expenditures [and tax collection] in Sdo 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 PublicExpenditure Data ManagementSystem (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 floorplan of the STAB. We knocked down walls, eliminatedpaper
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
STA B/SP)
Chapter 5. New Technology
Figure 9. Sio Paulo STABS' On-Line Connections with Internal and Outside Agencies.
L
Source: Harumi presentation.
Figure 9 shows the connectivity between Sao 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
Chapter5. New Technology
and PE was on-line, tax collectors would share diskettes with state infonnation 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 organizationaland 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 firns 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 Sio 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 territorialflexibility." (Glauco, tax
collector, interview presentationSTAB/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 appropriatereplies 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:
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." (Interviewfrom 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 Sao 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.
Chapter5. 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 Sdo 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
Chapter5. 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.
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 computerizationwere 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 Sao 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 Sao
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 outsiderprovider. We tried to have a vertical control of all phases
of the tax collection process. 46 (Guaragna2000).
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
For example, STABs in Sao 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 Gestao da Informa95o na Administragao Fiscal,
MF-UCP-PNAFE (PNAFE N.D.).
4
Chapter5. 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 presentationto 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 institutionalizationwere 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
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.
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 ftar." (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 Sao 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 Sdo
Paulo STAB, 26,708 candidates tested for 350 entry-level positions.4
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 Ceari, fifteen times in Rio Grande do Norte, twenty-one times in Paraiba, and thirty times in
Pernambuco and Sdo 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
47
http://www.centraldeconcursos.com.br/docs/Diversos/ICMS-estatistica.pdf and
http://www.centraldeconcursos.com.br/docs/edital/Edital-deAbertura-deInscricoes
FINALAPROVADO.pdf
137
Chapter6. 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%
*)
r M
0 a
w (
9.00%
8.00%
7.00%
6.00%
+
5.00%
y =-1 E-07*+ 0.0741
4.00%
3.00%
>
R2
2.00% -=~
1.00%
0.00%0
1
0
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.
48
Source: http://www.portalbrasil.net/salariorninimo.htm and,
http://www.data360.org/dsg.aspx?DataSet_GroupId=59 by Board of Governors of the Federal Reserve. System.
Chapter 6. Learning
Table 14.STAB Tax Collectors' Salaries by Brazilian State.
States
Acre
Alagoas
Amapi
Amazonas
~
Bia
Income
Per Capita
(a)
R$ 2,826.57
R$ 2,283.01
R$ 3,564.35
R$ 5,990.04
R$ilil0iiL1
Annual Tax
Collectors'
Nominal Wage
(c)
(nd)
R$ 73,200.00
R$ 30,600.00
R$ 78,000.00
Annual Tax Collector/
Minimum wage
(R$ 1,560)
(nd)
46.9
19.6
50.0
Tax Collectors'
Wage/IPC (b/a)
(nd)
32.1
8.6
13.0
oSe44,2.o n2
,.8
14.'9
R$ 13,052.89
R$ 65,400.00
41.9
5.0
Espirito Santo
Goias
Maranhao
Mato Grosso
Mato Grosso do Sul
Minas Gerais
Pari
R$
R$
R$
R$
R$
R$
R$
5,996.10
3,671.66
1,347.81
4,958.46
4,307.66
5,23 .31
2,697.94
R$ 40,476.00
R$ 36,632.16
R$ 20,376.00
R$ 57,000.00
R$ 42,955.32
R$ 40,560.00
R$ 38,400.00
25.9
23.5
13.1
36.5
27.5
6.8
10.0
15.1
11.5
10.0
26.0.
24.6
7.8
14.2
Parana
R$ 6,131.02
R$ 42,360.00
27.2
6.9
Piaui
Rio de Janeiro
R$ 1,624.31
R$ 7,354.21
R$ 60,063.84
R$ 69,030.00
38.5
44.3
37.0
9.4
Rio Grande do Sul
Rond6nia
R$ 7,145.08
R$ 3,610.80
R$ 51,720.00
R$ 48,000.00
33.2
30.8
7.2
13.3
Roraima
R$ 2,861.47
(nd)-
(nd)
(nd
Santa Catarina
R$ 6,446.26
R$ 52,941.60
33.9
8.2
Sergipe
R$ 2,984.06
R$ 44,616.00___
R$ 26,400.00
28.6
16.9
15.0
15.1
Distrito Federal
R$ 1,751.21
Tocantins
R$ 5,647.64
BRAZIL
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,
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-tostate, 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 hardto 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 Paraibaborderfiscal 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 Paraiba.
In these two states, not only were the
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 borderSTABs.
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
49
women joined the rank-and-file of the STABs and their contingent in the workforce grew significantly.
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 yearsforIBM, and having been "outsourced" to work as a contractor,
I found myself making one third of my originalprofessional wage. Applying to a stable
position was the best, if not my only option."
Newspaper interviews.
In the last entrance examination for TCs in Sio 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
4
I do not have accurate number on female TCs.
Chapter 6. Learning
from mathematics and engineering, were dentists, physicians, historians, accountants, and economists. 0
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 Ifirst 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 partnerand 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 praisedme said: 'Fromnow on, my son, you will be rich. You will
get the dough.' An accolade notfor the merit that I had in passing the entry examination
so Interviews with VAT TC's applicants in locus at the examination site. Sao Paulo, April 17-18 2006.
http://www.cursoaprovacao.com.br/cms/entrevista.php?cod=1819
http://www.editoraferreira.com.br/publique/cgi/cgilua.exe/sys/start.htm?infoid=28 19&sid=7
http://clipping.planejamento.gov.br/Noticias.asp?NOTCod=235374
142
Also:
Chapter6. 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 careerpressure, 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, Sdo 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
professionalstaff is the organizationalresponse 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 wouldfind irresistible:a world
ofpractically limitless power, nectar of the goddess to the ineffectual dreamer,for whom
lfe was not a pursuit of happiness, but a struggle for recognition and control. (Yancey
2004)"
"I'm sure it's occurred to you, Rick, thatyour 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,
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 personaltendency." (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." 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 Fazenddrias"(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 Sao Paulo, Bahia, and Rio Grande do Sul have their own TAS.
Frequently, they are responsible for training TCs from other states.
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
52
Sio Paulo STAB used to pay TCs by quotas of work, measured by number of audits carried out in a month.
The sending STABs cover the cost of training tax collectors in other states.
Chapter 6. Learning
in terms of tax revenue, professional satisfaction, and good connections with the TCs' association and the
business sector.
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 Sdo 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).
This is the
equivalent to 330,000 times the national minimum wage or 1% the budget for education in Sdo 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.
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.
5 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
5
2003).
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.
Sao 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 Sao Paulo.
In Bahia, an agricultural engineer was placed in the 'Wholesale Food Segments.'
production engineers were visiting factories.
In Pernambuco,
Specialized skills and a practical knowledge of industry
operations were fundamental to setting up teams that could cooperate to identify under-performing
industries.
146
Chapter6. Learning
"Once we were auditing a Pharmaceuticalcompany 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 manufacturingyields 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 afew 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 regulationand that the taxpayer did not want to pavjbr
parts of the value-addedin 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
(i.e.,
into specialization
computerization, which in turn demanded better professionals.
economic
segmentation)
and
The "Escolas Fazenddrias" (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
Chapter6. 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
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.
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 greaterpart 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." (Sinith 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 skillintensive 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.
Serbiner
(Wagner and Sternberg 1984) says
"Skilled practicalthinking incorporatesfeatures 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 representationsand 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 knowledge5 7 transformation and consolidation? The TCs in the Rio
Grande do Norte, Bahia, Pernambuco, and Sao 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
Chapter6. Learning
intellectual dominions. Also, its use here remains applied in adult education.5 8 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 Sdo 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 understandingthe
system properly." Nelson Fagundes, Sdo 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.
representation of the world surrounding the TC.
It allowed for a schematic and structural
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.
interpretation of knowledge (Lindsay and Norman 1977).
It involves the acquisition and
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 leaming. Later on, Houles (1980) defines education as a direct participation in the events
of life.
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 predispositionsto 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
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 entrepreneurscome to us now, we show them what we know and that we
know (our tables and our information). [Now] they know that we know [stufJ]. 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:
"Ifwe were to be auditedand assessedon 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 alsofor 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:
Chapter 6. Learning
"There is nothing new about theirpolicy. They arejust 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 collaboratingwith 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 consultingfor 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), "bureaucratshave to have access to information to
allow them to determine iffirms 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
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
Chapter 6. Learning
education.5 9 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' representativ.es 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 collaboratedwith 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 workedfor the STAB in the past, so that
we know the subject and the threads to do the work," explainedan accountantat the Sao
PauloAccountants' 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, Sao Paulo.
Both in Sao 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.
* In Sso 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.
Chapter6. Learning
The TCs' association in Sdo 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 Ceari, Paraiba, and Bahia, the TCs' Associations were not as involved in the reform process.
In Ceari and Pernambuco, TCs' Associations were pushing for work stoppage and better salaries. The
strongest antagonism between the union and the STAB was in Ceari (interviews and field trip). The
Ceari 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
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 Taliercio states "non-institutionalreforms, 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 Talidrcio 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.
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.
taxpayers' commitment.
However, an education process was crucial to gain the
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 itfragment by fragment,
on a small scale, by successive developments, cellularly,
like a laboriousmosaic."
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
Chapter7. 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 explanationfor 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 organizationalentity, 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.
Cognition: Vertical Indus1
Identities
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." (Hirschman1958; 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).
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 Ceari, Bahia, Pernambuco, Sio Paulo, Paraiba, 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
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.
State
41
0P
Y,
Outcomes
STAB Features
State
BA
CE
PB
Yes
No
No
100%
100%
100%
PE
Yes
100%
RN
SP
Yes
Yes
100%
100%
U11
Yes
Yes
Yes
Yes
Yes
Yes
23%_
_Elitist
21%
Low
Weak - Elitist
39%
35%
102%
1%
Low
High/Medium
High
High
Weak
Good
Good
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
Paraiba, 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, Ceari 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
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 caine from the employees who were to be most affected by [the
reforms], and who were to be its principalsources 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 Sdo 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?
successful trajectory?
Why did the tax administration reform follow
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
Dilulio 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 Sio Paulo STAB is the most advanced tax
administrations in the country.
The Sdo Paulo STAB was leader in moving to the vertical industry
segmentation paradigm. However, Table 16, above, shows that VAT revenue grew the least in Sio 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 Ceari, 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.
STABs as role models would also be interesting.
Studies on isomorphism using the
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.
Appendix A. Brazilian Tax Structure
Graph 13. Regional Population, GDP, and Tax Revenue.
Centralwst,
6.80%
Tax Revenue
GDP
Population
North, 3.70%
Centralwest
North, 7.00%
....
d:.
. - - - -.
. .- . - .
Centralwest,
7.20%
North 4 80%
.
-
6.80%
NorthEast,
12.90%
-
NorthEast,
28.40%
NorthEast,
N12.30%
South,
.
South,
16.40%
:..
Southeast,
42.40%
.
Southeast,
Southeast,
South,
14.80%
62.60%
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.
Regions
Social Indicator
North
Infant mortality
Northeast
Centerwest
Southeast
South
National
average
55.0
98.0
41.0
40.0
38.0
59.0
64.5
55.7
67.7
65.3
65.7
62.6
42.3
46.1
25.7
21.7
17.8
31.0
88.71
77.9
84.6
87.1
82.7
83.2
14.42
8.33
19.8
(per thousand; 1989)
Years of life
expectancy (1985)
Child malnutrition
rate
(1989)
Rate of elementary
school attendance
Rate of first grade
31.3
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 Sio Paulo only.
3/ State of Santa Catarina only.
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).
States
Acre
Amazonas
Bahia
Ceare
Distrito Federal
Espirito Santo
Goies
MaranhAo
Mato Grosso
Mato Grosso do Sul
Minas Gerais
Pare
Paraiba
Parana
Pernambuco
Piauli
Rio de Janeiro
Rio Grande do Norte
Rio Grande do Sul
Rondania
Roraima
Santa Catarina
Sao Paulo
Sergipe
Tocantins
BRAZIL
Total
Population
557526
2822621
477032
2 812 557
13 070 250
7430 661
2 051 146
3 097 232
5 003 228
5 651 475
2 504 353
2 078 001
17891 494
6192307
3 443 825
9563458,
7.918 344
2843278
14 391 282
2776782
10187 798
1 379787
324397
5 356 360
37032403
1 784 475
1 157098
169 799 170
Urban
Population
370267
1919739
424 683__
2 107222
8 772 348_
5315318
1 961 499
2 463 049
4 396 645
3 364 070
1 987 726
1 747 106
14671 828
4120693
2447212
7 786 084
6 058 249
1 788 590
13 821 466
2 036 673
8 317 984
884 523
247 016
4
.217
931
34 592 851
1 273 226
859 961
Rural
Population
GDP
Per
Capita
Gini
(1991)
HDI
Reform
(thousands)
Formal
Firms
187 259
3,241,847.44
5,815
0.61 0,697
USS 4,557
10903
11,556,231.65
4,094 0.6316 0,649 USS 13,662
902 882
41 128
USS 4,450
8232
52 349_ 35,888,581.05 75,233 0.5677 0,713
3,720,358.98
1,323 0.6208 0,753 USS 12,180
37 751
705 335
86,882,057.01
6,647 0.6533 0,688 USS 15,000
4 297 902
271 428
2115343
33,260,671.70
4,476 0.6474 0,700 USS 14,784
169 513
43,521,629.02 21,218 0.6089 0,844 USS 10,193
89 647
85784
34,487,904.52 11,135 0.6115 0,765 USS 156,000
634 183
109341
606 583
41,316,490.56
8,258 0.6197 0,776 USS 16,320
167 232
2,928 0.6012 0,636
USS 14,438
65 361
2
405
.287 16,547,449.02
516 627 166,586,326.69 66,519 0.6005 0,773 USS 20,813
91 093
330 895
19,953,528.62
9,602 0.6136 0,778 USS 21,713
67 151
27,935,498.73
1,561 0.6347 0,773 USS 25,000
3219 666
661 461
34,195,67642
5,522 0.6015 0,723 USS 15,270
71 747
2071 614
14,863,056.88
4,316 0.6381 0,661 USS 14,985
996 613
62 325
47,697,442.23
4,987 0.6098 0,705 USS 165,000
1 777 374
449 962
1 860 095
8,611,415.34
1,088 0.6597 0,656 USS 24,848
142542
1 054 688 108,698,901.36 38,230 0.6298 0,787 USS 10,910
48769
USS 24,440
416 184
569 816 222,563,502.61 15,465 0.6238 0,807
740 109
15,906,123.89
5,728 0.6459 0,705 USS 18,870
58905
1 869 814
9,744,450.65
956 0.5981 0,735 USS 22,977
598 252
495 264
1,864,150.71
1,351
0,746
USS 6,380
34325
USS 4,280
77381 142,874,226.33 440,430 0.617 0,814
9340
70,207,923.77 13,107 0.565 0,822 USS 17,000
302 973
1.138 429
2 439 552
13,120,855.33
354 0.5797 0,682
USS 68,718 1 728 708
511 249 546,606,818.53 306,312 0.6321 0,820 USS 10,140
28 696
4,767,935.52
4,121 0.6283 0,710 USS 11,100
297 137
26821
5,381 0.6366
161790311 913,734,000.00
in Formal
Firms
82942
373935
78601
419 649
1 715 595
1 051 641
957933
713621
1019 131
446 673
513969
417 368
3934832
699 151
455 189
2 535517
1 152090
311 137
3593 173
475 087
2 845 998
243 637
40363
1732936
11 301 850
283444
182260
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).
Appendix A. Brazilian Tax Structure
Graph 14. Participation of Social Contributions and Taxes to total Brazilian GDP.
Source: IBGE.
Table 19 describes the assignments of the different taxes to the levels of government: Federal,
State, and Local.
Appendix A. Brazilian Tax Structure
Table 19. Brazil Tax Assignments for Federal, State, and Local levels.
Tax Category
Foreign trade
Corporate income
Capital gains
Personal income
Transfers of property
Vehicles
Property
Rural property
Payroll
Civil servants' payroll
Sales taxes
IPI
ICM4S
ISS
Fees, royalties
Source: Affonso and Raimundo (1995).
1/ F = Federal; S = State; and L = Local.
Regional Revenues
States Tax Revenue.
Responsibility for
Setting of rate
Definition of base
Administration
F
F
F
F
S, L
S
F
F
F
F
F, S, L
S
F
F
F
F
S, L
S
L
L
L
F
F
F, S, L
F
F
F, S, L
F
F
F, S, L
F
F, S
F, L
F, S, L
F
F, S
F, L
F, S, L
F
S
L
F, S, L
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.
Appendix A. Brazilian Tax Structure
Graph 15. State Taxes as a % of Total Tax Revenue and as % GDP (2000).
PIS/PASEP 3% (0.87%
* CPMF 5% (1.32% GDP)
GDP)
n FGTS 6% (1.72% GDP)
U ISS 2% (0.58% GDP)
*IPTU 2% (0.47% GDP)
UCOFINS 12% (353% GDP)
m VAT 25% (7.55 GDP)
1
! SS 18% (5.11% GDP)
State 27%
SOIL
TAX 2% (0.49% GDP)
E IPI 6% (1.72% GDP)
*TR 19% (5.48% GDP)
Source: IBGE (Instituto Brasileiro de Geografia e Estatisticas) 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
2002/3 2003/4 2004/5
1995/6 1996/7 1997/8 1998/9 1999/0 2000/1 2001/2
14.2%
9.8% 27.1%
12.2%
7.5% 47.6% -17.9% 32.4% 10.8% -0.8%
Acre
2.1% -9.6%
9.9% 7.4% 14.0%
7.8% 7.7% 8.7% -16.6% 11.4%
Alagoas
4.8% 5.1% 14.4%
60.1% -28.9% -17.1% -11.6% 16.5% 6.7% -6.8%
Amazonas
-0.5% 10.9% 29.1%
8.2% -1.1% 14.4% -12.0% 32.7% 7.6% -6.5%
AmapA
7.2% 3.2% 2.8%
9.5% -5.8% 3.2% -7.1% 12.7% 3.1% -6.2%
Bahia
6.4% -5.4%
9.9%
2.7% -10.0%
1.4% 0.9% 4.7%
13.7% -2.4%
Ceara
13.6% 5.5% 10.9%
5.5% -4.5% 15.3% 2.2% -9.4%
16.0% 5.2%
Distrito Federal
13.9% 12.7% 22.4%
-0.6% 9.1% -8.1% -15.0% 17.2% 9.4% -22.2%
Espirito Santo
8.1% -9.7%
15.1% -5.5%
5.4%
1.0% -1.1% -3.7% 11.1%
11.9%
Goiis
5.7% -11.9% 26.4% 15.4% -10.8% -0.011% 9.1% 21.7%
20.2% -16.4%
Maranh~o
9.0% 8.5% -16.5%
7.1% 6.2% 16.3%
5.1%
-4.0% -2.4% -4.9%
Minas Gerais
16.6% 11.9% 13.5%
7.3% 7.2% -12.8%
-1.8%
1.8% -4.5% 12.5%
Mato Grosso do Sul
8.4%
19.7% 10.1% 4.1%
2.5%
9.2% -13.3% 13.3% 14.8% -13.3%
Mato Grosso
5.5%
11.0% 3.7% 17.1%
6.9% -14.3% 14.7% -16.9% 34.8% -1.6%
Par-i
2.3% 0.7% 14.5%
9.4% 12.7% -19.3%
16.1% 2.3% 8.7% -8.0%
Paraiba
5.1% -13.1%
8.7% 1.0% -5.2%
3.1% 3.0% 16.8%
2.0%
12.1%
Pernambuco
14.3% -0.9% -8.3% -21.3% 47.7% 17.4%
1.0% 3.2% -11.5%
17.3%
Piaui
11.2% 3.4% 11.1%
-23.1% -7.6% 0.6% -1.4% 14.6% 39.3% -8.7%
Parani
2.5%
3.4% -11.8%
8.9% 5.4% 1.7%
5.7% -7.3% 19.3% -5.3%
Rio de Janeiro
8.5% 4.7% 14.7%
9.7% -0.4% 17.3% 4.2% -11.3%
Rio Grande do Norte 17.2% 11.9%
26.0% 7.7% 16.2%
5.1% 45.5% -12.6% -6.6% 30.3% -4.0% -7.6%
Rond6nia
47.3% -1.5% -19.6%
6.9% -3.1% 23.9%
9.2% 12.2% 19.0% -11.1%
Roraima
10.7% -3.1% 15.3%
1.0% -7.6% 4.6% -8.2% 10.8% 7.3% -12.1%
Rio Grande do Sul
8.6% -6.8%
0.1% 12.5% 10.6%
2.3% -1.7% -2.6% -5.7% 10.5%
Santa Catarina
6.5% 2.8% 15.0%
4.4% 10.6% -3.3% -7.6% 13.3% 4.5% -6.4%
Sergipe
1.0% -12.8%
0.4% 2.1% 9.5%
9.6%
4.7% -0.7% -3.4% -8.2%
Sio Paulo
17.6% 4.0% 14.8% -2.6% 18.8% 11.1% -3.8% 22.9% -1.4% 9.6%
Tocantins
Source: IBGE (Instituto Brasileiro de Geografia e Estatisticas). 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.
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.
61
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
78%
46%
93%
114%
5%
-17%
8%
-1%
12%
17%
-3%
43%
11%
23%
20%
4%
54%
15%
26%
143%
19%
1%
166%
28%
6%
239%
16%
26%
14%
-6%
7% 22%
3%
6%
11%
18%
113%
8%
-3%
44%
8%
13%
43%
16%
15%
59%
19%
16%
105%
23%
21%
16%
-1%
22%
8%
29%
0%
23%
-15%
42%
-1%
45%
9%
31%
-15%
49%
-4%
57%
9%
74%
33%
Goi6s
Maranhao
12%
20%
13%
0%
12%
6%
8%
-6%
20%
18%
29%
36%
17%
22%
34%
22%
27%
33%
34%
62%
MG
MS
Minas Gerais
Mato Grosso do Sul
5%
-2%
1%
0%
-1%
-4%
-6%
7%
2%
15%
11%
24%
-7%
8%
-1%
26%
5%
41%
23%
60%
MT
PA
PB
Mato Grosso
Pard
Paraiba
-3%
5%
29%
20%
10%
-13%
19%
5%
26%
18%
30%
14%
9%
16%
46%
15%
19%
22%
18%
9%
42%
36%
21%
12%
63%
65%
Pernambuco
Piaui
Paran6
12%
-8%
19%
14%
56%
41%
PE
PI
PR
2%
7%
16%
12%
22%
16%
39%
35%
22%
-28%
8%
-30%
24%
-19%
23%
13%
12%
3%
-12%
14%
31%
18%
53%
31%
RJ
RN
Rio de Janeiro
Rio Grande do Norte
-2%
31%
17%
44%
11%
43%
14%
68%
17%
75%
4%
55%
13%
68%
19%
76%
21%
102%
RO
RR
Rond6nia
Roraima
5%
9%
53%
23%
34%
46%
25%
30%
63%
91%
56%
88%
44%
51%
82%
62%
96%
57%
128%
94%
RS
SC
Rio Grande do Sul
Santa Catarina
1%
2%
-7%
1%
-2%
-2%
-10%
-8%
-1%
2%
6%
11%
-6%
3%
0%
16%
16%
29%
SE
SP
TO
Sergipe
Sio Paulo
Tocantins
16%
4%
22%
12%
0%
40%
3%
-8%
37%
17%
1%
62%
22%
2%
81%
AC
Acre
AL
AM
AP
Alagoas
Amazonas
Amape
BA
CE
21%
Bahia
Cear6
12%
8%
60%
8%
10%
14%
DF
ES
Distrito Federal
Espirito Santo
GO
MA
17%
-23%
6%
17%
4%
5%
18%
18%
-29%
14%
-11%
74%
4%
3%
22%
-11%
113%
25%
-9%
110%
45%
22%
44%
0%
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 Amapi 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, Sio Paulo's long running and welldeveloped 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.
Appendix A. Brazilian Tax Structure
Table 22. GDP Growth per State (1994-2002).
Acre
Alagoas
Amazonas
Amapi
Bahia
Ceari
Distrito Federal
Espirito Santo
Goias
Maranhao
Minas Gerais
Mato Grosso do Sul
Mato Grosso
Pafi
Paraiba
Pernambuco
Piaui
Parani
Rio de Janeiro
Rio Grande do Norte
Rond6nia
Roraima
Rio Grande do Sul
Santa Catarina
Sergipe
S1o Paulo
Tocantins
2001/2 2002/3 2003/4
1999/0 2000/1
1997/8 1998/9
1995/6 1996/7
6.2%
11.7%
-6.6%
2.0%
-0.2%
-10.8%
8.8%
6.4%
5.4%
-0.4%
9.4%
81%
26%
03%
12.8%
49/
5.5%
14.8%
13.8%
4.2%
-4.2%
-0.7%
10.7%
-14.3%
3.1%
-4.9%
17.1%
7,4%
8.0%
-6.6%
3.5%
13.4%
-3.3% -12.1%
5.9%
-0.9%
5.7%
9.5%
-5.7%
-2.0%
4.6%
-9.7%
2.9%
4.3%
12.6%
4.1%
9.1%
-6.2% -11.0%
-2.7%
5.3% -13.8%
4.5%
14.4%
2.6%
-1.7%
-14.3%
1.0%
21.3%
-26.2%
24.0%
11.9%
14.2%
5.9%
8.9%
-12.9%
-5.4%
4.6%
-10.0%
5.4%
1.0%
6.0%
-0.2%
9.3%
-0.8%
4.5%
10.3%
-14.4%
6.9%
2.1%
12.3%
5.3%
13.8%
-11.9%
1.10
6.1%
-8.7%
-4.2%
0.2%
24.1%
2.6%
7.1%
-3.3% -12.3%
3.3%
-12.8%
0.9%
3.1%
14.1%
-6.4%
14.9%
-11.3%
4.7%
-0.7%
-9.7%
6.3%
3.8%
8.7%
9.9%
17.5%
-1.8%
-2.7%
4.7%
-1.6%
6.3%
7.1%
11.6%
4.2%
6.3%
-6.8%
3.9%o
3.5%
4.0% -10.8%
-1.3%
4.8%
-3.5%
9.5%
0.5% -10.1%
6.2%
-9.0%
2.1%
-0.8%
12.4%
0.4%
7.5%
-8.7%
-1.5%
2.1%
4.1% -12.7%
1.8%
12.0%
4.6%
10.4%
-12.2%
-5.4%
2.7%
-10.7%
2.7%
0.6%
12.2%
-2.3%
12.9%
11.2%
95% 25% .3%
57%
2.9%
13.7%
4.0%
4.0%
-8.8%
-2.9%
9.9%
-5.3%
1.3%
4.6%
6.6%
3.3%
9.4%
-6.1%
-4.3%
-7.0% 10.9%
0.9%
5.5%
13.6%
2.1%
8.3%
-5.0%
-2.2%
2.2%
-9.3%
8.0%
7.2%
12.4%
-1.1%
4.7%
-3.1%
-1.3%
24.7%
-8.8%
5.6% 18.1%
6.5%
-0.7%
13.9%
3.0% -0.1% -11.9%
0.2% -10.9%
1.7%
7.8%
0,4%
11.5%
-0.9% -11.6%
8.5%
-8.4%
0.00
0.6%
14.2%
-0.2%
14.5%
-8.1%
25.3%
-0.6%
-10.1%
2.1%
4.6%
11.3%
4.9% -1.7%
-2.3% -13.2%
-0.6%
3.2% -12.6%
5.5%
8.5%
1.3%
9.8%
-8.2%
13.1%
6.2%
-9.7%
11.0%
4.3%
14.3%
IPEA. 6
Source: IBGE (Instituto Brasileiro de Geografia e Estatisticas) and
Table 23 depicts the ratio of VAT per GDP by State in Brazil, from 1995 to 2004.
To calculate State GDP growth, I deflated nominal values, using the LGP-DI (GPI) (General Consumer Price
Index), published by the IBGE (Instituto Brasileiro de Geografia e Estatistica).
63 The IBGE has finished compiling data for the State GDP until 2004.
62
179
Appendix A. Brazilian Tax Structure
Table 23. Ratio of VAT per GDP by State and Coefficients of Variation.
AC Acre
AL Alagoas
AM Amazonas
AP Amapa
BA Bahia
CE CearA
DF Distrito Federal
ES Espirito Santo
GO Goins
MA Maranh o
MG Minas Gerais
MS Mato Grosso do Sul
MT Mato Grosso
PA Pari
PB Paraiba
PE
PI
PR
RJ
Pernambuco
Piaui
Parana
Rio de Janeiro
RN Rio Grande do Norte
RO
RR
RS
SC
Rond6nia
Roraima
Rio Grande do Sul
Santa Catarina
SE Sergipe
SP Sio Paulo
TO Tocantins
Standard deviation
Mean
Coefficient Variation
1995
3.6%
6.94%
1996
3.9%
6.52%
8.31% 11.36%
3.54% 3.87%
7.80% 7.59%
7.61%
4.76%
9.89%
9.56%
6.63%
7.34%
8.10%
10.85%
5.25%
2001
2000
2002
1997 1998 1999
2003 2004
3.9% 5.3% 4.9%
6.5%
7%
8%
8%
8%
6.65% 6.89% 6.59% 7.37% 7.72% 7.59% 7.62% 8.22%
8.50% 6.83% 7.05% 7.42% 7.97% 7.75% 7.80% 7.20%
3.61% 4.27% 4.28% 5.01% 5.21% 5.21% 4.80% 4.96%
i.85% 6.87% 7.07% 7.62% 8.01% 7.97% 7.81% 7.63%
7.06% 7.14% 7.83% 8.84% 9.68% 9.78% 9.09% 8.81%
4.54% 3.86% 5.00% 4.75% 4.81%
5.09% 5.88% 6.05%
10.02% 8.74% 8.26% 9.25% 10.70% 9.56% 10.00% 10.64%
9.43% 8.72% 9.82% 9.89% 10.22% 9.31% 9.80% 9.27%
5.36% 5.91% 5.70% 6.79% 7.76% 7.86% 6.90% 7.15%
7.56%
4.83%
9.27%
9.53%
6.42%
6.76% 6.30%
7.32% 7.17%
9.97% 10.16%
5.36% 1.65%
6.09% 6.65%
6.45% 8.03%
7.87%
7.49%
7.50%
7.76%
8.68% 8.89% 8.75% 8.87%
8.29% 9.54% 10.47% 9.33% 10.30% 10.49%
5.13% 4.78% 6.22% 5.89% 6.67% 6.96%
6.62% 6.83% 7.05% 7.50% 7.58% 7.81% 8.76% 7.86% 7.35%
6.77% 6.78% 6.79% 6.86% 6.83% 7.27% 7.46% 7.74% 7.42%
6.89% 7.21% 7.23% 7.27% 7.20% 8.01% 8.39% 8.77% 6.25%
6.45% 4.36% 3.92% 3.73% 4.06% 4.78% 6.67% 6.85% 6.75%
6.07% 6.02% 5.33% 6.28% 6.29% 5.86% 6.24% 6.04% 6.32%
6.29% 6.49% 6.89% 7.49% 8.02% 8.48% 9.23% 8.72% 8.65%
6.81% 6.37% 8.65% 7.00% 7.21% 9.19% 9.02% 8.77% 10.21%
7.66% 7.86% 8.35% 8.42% 8.21% 9.70% 9.68% 8.04% 8.21%
6.54% 6.13% 5.57% 5.81% 5.99% 6.45% 6.92% 6.91% 6.71%
7.10% 6.36% 6.22% 6.05% 6.23% 6.35% 6.95% 7.33% 6.58%
7.48% 7.01% 7.42% 7.03% 7.22% 8.23% 6.87% 7.00% 6.50%
10.61%
8.13%
8.68%
0.02
0.07
0.29
7.01%
10.51%
6.93%
7.67%
7.61%
8.83%
7.14%
6.41%
8.76%
10.76%
8.04%
6.55%
7.37%
6.70%
7.38% 6.91% 7.26% 8.00% 8.28% 8.32% 7.96% 8.27%
8.93% 8.90% 9.21% 9.94% 11.12% 10.93% 11.46% 12.82% 12.49%
7.84%
0.02
0.07
0.29
0.02
0.07
0.29
0.01
0.07
0.14
0.02
0.07
0.29
0.02
0.08
0.25
0.02
0.08
0.25
0.01
0.02
0.02
0.08
0.08
0.08
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).
Appendix A. Brazilian Tax Structure
Table 24. VAT Collected by Brazilian States, from 1995 to 2005,
in R$ (thousands), in Nominal Values.
Acre
12
Alagoas
27
Amazonas 13
Amap6
16
Bahia
29
Cears
23
Distrito Fec53
Espirito Sa 32
Goias
52
Maranhao 21
Minas Gerc31
Mato Gros 50
Mato Gros 51
Para
15
Paraiba
25
Pernambuc 26
Piaui
22
Parana
41
Rio de Jan 33
Rio GrandE 24
Rond6nia 11
Roraima
14
Rio GrandE 43
Santa Catz 42
Sergipe
28
Sao Paulo 35
Tocantins 17
1995
36279.56
280124
913659
43747
2088433
950829
629366
1271968
1135574
335698
4620148
566515
706470
634416
352235
1182161
219046
2474033
4513979
297431
201587
35964
3509483
1674549
264082
18626437
106469
1996
44524.61
330483.2
1600635
51812.78
2503153
1182979
798877.6
1382798
1390825
441464
5314785
608766.5
792061.2
742160.5
447390.1
1450380
281135.2
2082779
5220055
381445.1
231829.3
42979.82
3879186
1873491
301780.9
21334002
136942.7
1997
1998
1999
51485.41 77282.34 76224.01
382906.7
423403 423889.7
1224789 1032229 1096178
55127.53 64162.48 67786.79
2537451 2663121 2970475
1242053 1344086 1526927
970418 1113493
903888.2
1623200 1517876 1550384
1511441 1520270 1759087
397030 426665.3 451485.6
5491781 5452013 6230332
666660.3 647882.1 875385.8
930555 820665.7 1116662
684615.3 798492.2 796606.4
492639 544646.4 601562.4
1592327 1701954 1776714
305511.6 320629.1 340619.6
2069974 2118421 2508332
5208807 6322927 7194223
459312.3 512532.1 613099.9
363015.1 322848.1 362099.8
67085.9
51897.18
62826.7
3856958 4101730 4521426
1981274 1962928 2223841
359170.4 353445.9 392271.4
22795013 22387602 24694373
153175.5 178854.4 209303.3
2000
110619.1
517600.7
1399486
98611.62
3670292
1838435
1406623
1992134
2142345
625401.7
7441456
1029978
1405540
1177133
721597.1
2116784
426821.1
3150977
8080750
788077.6
516946.5
108309.8
5489652
2693476
487336.7
29677851
272553.9
2001
2002
135615 169482.9
584523.5 665566.9
1651686 1939187
117379.6 138255.5
4187240 4950531
2089231 2368020
1590928 1816261
2411699 2364264
2560977 2914197
798496.7
897073
8929614 9397793
1221645 1341762
1348743 1841745
1281217 1702946
899732.7
914663
2365983 2826721
467972.6 540797.4
4854435 5580693
9239976 10272526
908011.2 1014478
548914.6 639159.8
118022.8 119578.4
6515175 7213311
3234880 3798879
563435.3 664264.8
33165923 36453327
335110.3 406195.6
2003
208225.9
787086.8
2187808
148025.1
5712367
2585224
2219996
2897949
3608729
965388.8
10835788
1683460
2372078
2034436
1007196
3135394
458021.3
6679675
12036014
1184697
866783.3
137625.2
8595599
4094442
761312
39394704
537239.9
2004
2005
329757.037 AC
256894.2
949553.5 1093364.669 AL
2584452 2985849.341 AM
184405.6 240337.6446 AP
6625528 6877126.582 BA
2930038 3097416.554 CE
2631363
2945232.93 DF
3670195 4535688.795 ES
3831706 4078902.958 GO
1183771 1455053.558 MA
12931306 15184191.14 MG
2117631 2426673.917 MS
2934667 3085602.386 MT
2370520 2801569.094 PA
1139713 1317435.888 PB
3628096
4277778.27 PE
760498.2 901060.8474 PI
7763904 8707026.435 PR
14259396
14648234.6 RJ
1393621 1614254.689 RN
1048957 1231116.247 RO
149909.1 187486.3116 RR
9360614 10900617.79 RS
5175747 5777081.563 SC
879533 1021126.519 SE
45223079 $49,988,622 SP
595288.6 658939.3299 TO
Table 25 depicts the nominal values of the GDP for each State, from 1995 to 2005, in thousands
of reais (R$), in current value (corrected for inflation).
Table 25. GDP by Brazilian States, from 1995 to 2005, in R$ (thousands), in Nominal Values.
Acre
Alagoas
Amazonas
Amapa
Bahia
Ceara
Distrito Federal
Espirito Santo
Goias
Maranhao
Minas Gerais
Mato Grosso do Sul
Mato Grosso
Para
Paraiba
Pernambuco
Piaui
Parana
Rio de Janeiro
Rio Grande do Norte
Rond6nia
Roraima
Rio Grande do Sul
Santa Catarina
Sergipe
58o Paulo
Tocantins
1995
995065.4
4036622
10993593
1235377
26769072
12494861
13231169
12857799
11875395
5063120
62971008
6994111
6510213
12081137
5324208
17460791
3179601
38368590
74412360
4727278
2959448
469440
53652947
23572606
3532532
2.29E+08
1225936
1996
1147584
5072157
14090343
1339775
32989503
15641122
16539188
14908936
14592142
6872683
78604474
8317293
7945768
13855469
6549148
21391139
3901583
47719783
86757651
5876000
3639348
546826.5
63262677
29453733
4303611
2.72E+08
1533584
1997
1313635
5756057
14410659
1526341
37020911
17589092
19916064
16198253
16025469
7410423
87196507
9292069
9155035
14716542
6989286
23439349
4224891
52848951
97673810
6669225
4198007
621369
69221314
31874672
4842561
3.09E+08
1720259
1998
1453505
6141038
15106983
1501146
38759097
18835765
25119060
17369207
17428114
7223687
89490142
10049502
9901396
15571994
7261827
24810247
4412523
56798250
1.01E+08
6844467
4611191
746384.3
70541889
32434064
5030502
3.24E+08
1941479
1999
1557195
6429095
15554778
1584001
42040109
19510907
22256423
18772147
17919856
7918384
93748370
10901030
11701250
16673915
7936649
26021483
4733809
61723959
1.14E+08
7647781
5023344
817192.3
75450458
35681851
5434375
3.4E+08
2106171
2000
1702621
7022923
18872885
1968365
48197174
20799548
29587137
21530247
21665356
9206845
1.06E+08
11861168
13428289
18913684
9237737
29126796
5329536
65968713
1.38E+08
9293319
5624964
1116581
85137543
42428004
5920725
3.71E+08
2450498
2001
1920769
7569188
20736037
2253041
52249320
21581141
33051371
22538119
25048231
10293103
1.14E+08
13736055
14452678
21747997
10271930
31724962
5574648
72770350
1.48E+08
9833650
6082841
1218984
94084498
46534519
8204018
4.01 E+08
3066502
2002
2259000
8767000
25030000
2652000
62103000
24204000
35672000
24723000
31299000
11420000
1.25E+08
15343000
17888000
25530000
11634000
36510000
6166000
81449000
1.7E+08
11633000
7284000
1488000
1.04E+08
51828000
9496000
4.38E+08
3545000
2003
2716123
10325908
28062624
3083013
73166488
28425175
37752658
28979694
36835111
13983802
1.45E+08
18969505
22615132
29215268
13710913
42260926
7325106
98999740
1.9E+08
13695517
8491977
1677318
1.28E+08
62213541
11704013
4.95E+08
4189864
2004
3241847
11556232
35888581
3720359
86882057
33260672
43521629
34487905
41316491
16547449
1.67E+08
19953529
27935499
34195676
14863057
47697442
8611415
1.09E+08
2.23E+08
15906124
9744451
1864151
1.43E+08
70207924
13120855
5.47E+08
4767936
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
1995/8
1995/2004 1997/04
2003/4
2002/3
2001/2
2000/1
1999/0
1998/9
1997/8
1996/7
1.46
2.02
2.17
1.03
1.02
1.06
1.09
1.33
0.92
1.36
1.01
0.99
1.24
1.18
1.08
1.00
0.98
1.05
1.12
0.96
1.04
1.02
0.82
0.85
0.87
0.92
1.01
0.97
1.07
1.05
1.03
0.80
0.75
1.21
1.37
1.40
1.03
0.92
1.00
1.04
1.17
1.00
1.18
0.93
AC
AL
AM
AP
Acre
Alagoas
Amazonas
Amape
1.06
0.94
1.37
1.09
BA
Bahia
0.97
1.03
110
1,01
0.98B
0,93
0,96
0.9,7
1.16
1.11
1.25
0.88
113
1-05
1A0
0.98
0.99
1.00
101
0.99
cearb4
0.90
0-93
1.08
CE
DF
ES
GO
MA
MG
MS
MT
PA
Distrito Fec
Espirito Sa
Goies
Maranhao
Minas Gen
Mato Gros!
Mato Gros!
Parn
1.02
0.94
1.00
0.97
0.92
0.90
0.92
1.02
0.94
1.08
0.99
0.83
0.93
0.98
1.02
0.87
0.85
0.87
0.92
1.10
0.97
0.90
0.82
1.10
1.30
0.95
1.13
0.97
1.09
1.25
1.15
0.93
0.95
1.12
1.01
1.19
1.05
1.08
1.10
1.30
1.01
1.16
1.03
1.14
1.12
1.02
0.89
0.95
1.06
0.89
0.91
1.01
0.95
0.98
1.10
1.13
1.15
1.05
1.05
0.88
1.00
1.01
1.02
1.04
1.03
1.06
0.95
1.04
1.04
1.20
1.00
1.00
1.27
1.08
0.97
1.08
1.06
1.31
0.97
1.32
1.33
1.06
0.98
1.34
1.23
1.48
1.03
1.49
0.81
0.88
0.91
0.89
0.83
0.80
0.76
0.98
PB
PE
Paralba
Pemamnbut
1-03
1.00
1.03
1 06
1.01
1.61
1.00
-F03
1.06
112
1.03
0.90
0.93
1.04
0.96
116
1,04
10113 112
i09
1.00
PI
PR
RJ
Piaul
Parana
Rio de Jan.
1.05
0.68
099
1.00
0.90
0.89
1.00
0.95
1.18
0.99
1.09
1.00
1.11
1.18
0.93
1.05
1.40
1.07
1.04
1.03
0.97
0.71
0.98
1.05
1.41
1.06
1.01
1.28
1.11
1.06
1.22
1.82
1.20
1.05
0.58
1.04
RN
Rio Grandk
1.03
1.06
09
1.07
l'06
1,09
0.94
0,99
1,01
1.39
1,27
1.19
RO
RR
RS
Sc
SE
Rond6nia
Roraima
Rio GrandE
Santa CatE
Sergipe
0.94
1.03
0.94
0.90
0.94
1.36
1.06
0.91
0.98
1.06
0.81
1.01
1.04
0.97
0.95
1.03
0.98
1.03
1.03
1.03
1.27
1.18
1.08
1.02
1.14
0.98
1.00
1.07
1.10
0.83
0.97
0.83
1.00
1.05
1.02
1.16
1.02
0.97
0.90
0.93
1.05
0.98
0.98
1.12
1.03
1.58
1.05
1.00
1.04
0.90
1.24
0.96
1.18
1.19
0.90
1.03
1.10
0.89
0.85
0.94
TO
Brazil
Tocantins
1.03
0.95
1.00
0.94
1.03
0.97
1.08
1.05
1.12
1.07
0.98
1.07
1.05
1.00
1.12
0.98
0.97
1.03
1.44
1.05
1.40
1.17
1.06
0.87
1.12'
0,94
1.13
1.01
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, intrastates 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.
Appendix A. Brazilian Tax Structure
Table 27. VAT Rates per State.
VAT rates in 1998 and 1999
State
12%, 13%, 17%, 25%
AC 98
25%,17%,12%
AL 99
4%, 12%, 17%, 25%
AM 99
25%, 17%, 12%, 8%, 2%
AP 99
7%,17%,25%
BA 99
n.a.
CE
25%, 21%, 17%, 12%
DF 99
12%, 13%, 17%, 25%
ES 99
4%, 7%, 12%, 13%, 17%, 25%
GO 99
4%, 12%, 17%,25%
MA 99
MG 99 4%, 7%, 12%, 16%, 18%, 25%, 30%
12%, 20%, 25%, 25%, 25%, 17%
MS 99
MT 99
12%, 17%, 20%, 25%, 30%
25%, 12%, 7%, 17%
PA 99
25%, 20%, 17%,12%
PB 99
7%, 12%, 17%, 20%, 25%
PE 99
4%, 12%, 17%, 20%, 25%
P 99
7%, 9%, 12%, 13%, 17%, 25%
PR 99
18%, 7%, 12%, 25%, 37%
RJ 99
25%, 17%, 13%, 12%, 4%
RN 99
25%,17%,12%, 9%
RO 99
12%,17%,25%
RR 99
12%, 7%, 18%, 26%
RS 99
25%, 17%, 12%, 7%
SC 99
SE 99
4%, 7%, 12%, 17%, 25%
SP 99
4%, 7%, 12%, 18%, 25%
25%,17%
TO 99
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).
Sso Paulo
Ceara
Paraiba
Pernambuco
Rio Grande do Norte
State
Bahia
Sso Paulo
Fortaleza
Joso Pessoa
Recife
Natal
Capital
Salvador
2
Area (km )
564.692,669
148.825,602
56.439,838
$ 98,311,616.00 $
52,796,791.00 $
248,209.43
# municipios
417.00
184.00
223.00
185.00
167.00
645.00
5930 -----. ...... 3,-003,087.0 0-4' 40,442,795.60-03,595,886.00.. 8.413,5.0.....
Population1381,33.0.....- 8,0--'97,276.0-0
Appendix B
List of Interviews
This appendix contains information about the interviews conducted in the study.
Appendix B. List of Interviews
RIO GRANDE DO NORTE and PARAIBA
.
Jaime Goudeiro
2.
3.
4.
Saulo Rocha
Jacanna de Assungao
Joao Flavio S. Medeiros
5.
6.
7.
8.
9.
10.
11.
12.
13.
Edvani de Freitas Neto
Graga Queiroz
Rembrant
Brenan
Vasconcelos
Neil Armstrong de Almeida
Jorge
Isnard
Walter
Olavo
Sayonara Pereira de Oliveira
14.
15.
16.
17.
18.
19.
20.
21.
Ludenilson Araijo Lopes
Walter Correia de Aquino Neto
Alexandre Firmino
Jane Carmen C. de Aradjo
Manoel
Jos6 Aranjo da Silva
Lina
Bira
22.
23.
24.
25.
Maria Val6ria
Francisco de Assis L. Souza Fo.
Dr. Nuto
Primo do Guido
C.
Promosad - Projeto de Reestruturagdo dos
Sistemas de Administragdo Tributiria
Sub
Secretario da Tributagao do RN
Coordenadoria de Arrecadag5o Estatistica e
Controle
Setor de Pessoal
Responsivel por treinamento
de
Coordenadoria
InformAtica
Tax Collector
Fiscal -Volante
Fiscal - Volante
Fiscal - Volante
Posto Fiscal de Cara6
Fiscal SINDIFERN - Diretora de Formagao
Sindical FENAFISCO
Sub-SecretArio RN
Tributagio e Assessoria T6cnica (Cabinet)
Fiscal, Ex-Secretario Adjunto (1995-1996)
Fiscala, Director Ia. URT (Unidade Regional)
Presidente do Sindicato (SINDIFERN)
Secretaria da Sec. Da TributagAo
Secretario da Sec. Da Tributag5o 1995
FIERN
Secretiria do Dr. Nuto
Fiscal - Gerente de Organizagao e Gestio
3/18/2001
3/18/2001
3/19/2001
3/20/2001
3/20/2001
3/20/2001
3/23/2001
3/20/2001
3/21/2001
3/22/2001
3/23/2001
3/26/2001
3/26/2001
3/27/2001
3/27/2001
3/27/2001
3/28/2001
3/28/2001
3/30/2001
4/3/2001
4/3/2001
Appendix B. List of Interviews
SAO PAULO
SEFAZ - Secretaria Da Fazenda de Sao
Paulo
Secretario
SEFAZ - Secretaria Da Fazenda de Sao
Paulo - Assessora Econ6mica
Fiscala
SINAFRESP - Sindicato ods Agentes
Fiscais de Rendas Do Estado De Sio Paulo
- Presidente
APECAT -
26.
Yoshiaki Nakano
27.
Lilian de Toni Furquim
28.
29.
Harumi
David Torres
30.
Dalvanira Paes Lima
31.
Nelson Fagundes
32.
33.
Evandro de Mello Xavier
Pedro Paulo
Fiscal
Fiscal
34.
35.
36.
37.
38.
39.
40.
41.
Luzia
Cata Preta
Dr. Soto
Dirceu Pereira
Nelson Machado
Glauco Hon6rio
Jodo Carlos Csillag
Jaider Pereira Paiva
42.
Helena Kerr do Amaral
Fiscal
Fiscal
SINAFRESP
Fiscal / Consultor
FAZESP - Escola Fazendaria
Fiscal / DEAT/COMEX
Fiscal - Diretoria de Informagoes
Consultor
UCE - Unidade de Coordenag5o Estadual
SAF - Subcoordenadoria Administrativa
Financeira
Escola Fazendaria Do Estado de Sio Paulo
43.
44.
45.
Nelson Fagundes
J. R. Lobato
Carlos Leony Fonseca
Cunha
46.
47.
48.
49.
50.
51.
52.
Marta Ma. Alvarenga Freire
Sr. tlcio
Evelyn Levy
Marcos Fernandes Gongalves
da Silva
Joao Adolfo Ponchio
Victor
Walter Soboll
53.
54.
55.
56.
57.
Henrigue Fingermann
Cl6vis Panzarini
Carlos Eugenio
Jos6 Ernesto
Isabel
02/07/2001
7/12/2000
06/2000
several
06/2000
several
01/11/2001
01/16/2001
01/11/2001
Several interviews
01/22/2001
- Diretora
da
APECAT - Fiscal
Fiscal
Promocat - Programa de Modemizagio da
CAT
Diretoria de Informagoes - Diretor
Fiscal - APECAT
Sindicato dos Contabilistas de Sio Paulo
Consultora
Professor EAESP
06/06/2000
Fiscal - CECI (Controladoria do Estado SP)
Fiscal
Coordenador - Coordenadoria Estadual de
Controle Interno
Professor EAESP
Coordenador de Administragao Tributaria
FGV
FGV
FGV
2/21/2001
01/03/2001
01/14/2001
2/19/2001
2/20/2001
2/21/2001
2/21/2001
2/22/2001
Appendix B. List of Interviews
58.
Claudio Couceiro D'Amorim
59.
60.
Ana Claidia Salgado
Silvia Franco
61.
62.
63.
64.
Francisco Moura
Wagner
Two people
Carlos Alberto Agostini
65.
66.
67.
Mauro Bogea
Benedicto Ludgero Fornitani
Benedito Franco Da Silveira Fo.
68.
69.
Alexandre Felix
Rui de Britto Alvares Affonso
70.
71.
Fernando Collet
Conceigdo Fraga
Fiscal - Diretor Geral da Receita/Diretor da
Adm. Tributaria
Ouvidora Fazendaria - Ombudswoman
Comunicagio Institucional
3/15/2001
06/06/2000
07/12/2000
GIA launching
Sindicato de Contabilistas
FENAFISCO - Federagdo
Fisco Estadual - Presidente
Nacional Do
Sio Paulo - Brasilia
188
Minist6rio da Fazenda - Brasilia
Fiscal - Retired
AFRESP - Associagao Dos Agentes Fiscais
de Rendas do Estado De Sio Paulo Presidente do Conselho Deliberativo
SINAFRESP - Imprensa
FUNDAP - Diretor de Economia do Setor
Pnblico
UCE/SP - Unidade de Controle Estadual
Politica Salarial
07/20/2000
01/11/2001
01/17/2001
01/17/2001
Appendix B. List of Interviews
PERNAMBUCO
72.
73.
74.
75.
76.
Vania Pernambuco
Jorge Jatobea
Felipe Chaves
Aldemir Souza
Sol Garson Braule Pinto
77.
78.
79.
80.
Eli
Maria Roseana
Edilberto
Alexandre Ribeiro
81.
82.
Maria do Carmo
Jos6 Candido B. de Miranda
83.
84.
85.
86.
87.
Aldo Ramos
Luzanita Monteiro
Airan Castilho
Carlos Rog6rio
Valter Jarocki Jr.
88.
89.
90.
91.
92.
93.
Carlos Egito
Eziel
Nevton
Francisco
Roberto Abreu
Roberto Cavacanti Tavares
Divisio de Treinamento
nSecretario
da Fazenda
3/15/2001
BNDES - Secretaria Para Assuntos
Fiscais
UCE - Coordenador
Gabinete Secretaria da FAzenda
3/16/2001
dos
SINDIFISCO
Sindicato
Funcionarios Integrantes
do Grupo
Ocupacional Do Tesouro Estadual de
3/9/2001
3/5/2001
3/5/2001
3/6/2001
3/6/2001
3/7/2001
Pernambuco - Presidente
3/6/2001
ABRASEL - Associa ao Brasileira de
Restaurantes
e
Empresas
de
Entretenimento
Diretoria De Segmentos Econ6micos
3/9/2001
3/12/2001
3/9/2001
3/13/2001
3/13/201
3/13/2001
Appendix B. List of Interviews
BAHIA
94.
95.
Dimas Fonseca Mello
Emilia Maria D. Gongalves
96.
97.
Monique Badar6 Campos
Monica Sim~es Bandeira
98.
Moysds de Oliveira Andrade Jr.
99.
100.
Fitima Freire
Jose Sergio Guanabara
SEFAZ BA
Diretora Operacional - Superi ntend6nc ia
de Desenvolvimento do Servigo Pdblico e
Atendimento ao Cidaddo - SESAC
Fundagio Luis Eduardo Magalhaes
FISEPE - Empresa De Fomento da
Informitica do Estado de Pernambuco Assessora
Diretor - Diretoria de Arrecadagiqo,
Crddito e Contr6le - DARC
UCP - Unidade Contr6le
Superintendencia
da
Administragio
8/1/2000
8/3/2000
08/02/2001
Tribut ria - Diretor
101.
Joio Batista Aslan Ribeiro
102.
Celso Tavares Ferreira
103.
104.
105.
Alexandre Cialdini
UCE/PROMOSEFAZ - Subcoordenador
T6cnico
Superintendencia
de
AdministragAo
Tributiria/Diretoria de Planejamento da
Fiscalizagio- Diretor
Secretaria da Fazenda/Ceari - Fiscal
Geraldo Machado
Fiscal
190
7/25/2000
Appendix B. List of Interviews
OTHERS
106.
Elia Yi Armstrong
107.
John-Mary Kauzya
108.
109.
110.
Esdras Borges Costa
Francisco Mejia
Anwar Sha
111.
Carlos Pimenta
112.
Clotilde Fonseca
113.
Glauco Arbix
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
Edmilton
Mircio Gomes da Cruz
Ferbes
M. Dilip
Ceara/sindicato
Dilna
Lncia Callou
Alberto Teixeira
Rejane
Joc6lio
Peter Spink
Peter Spink
Dilma M. da Fonseca
Jaime Mano
Luiz Tacca
Joio Dias
Carlos Silvani
Pal do Henrique
Julieta Verleun
Jaime Mano
Orlando
Cristina Waldwoger
James Poterba
Administration
Officer
Public
DPADM/DESA/UNDP
Public
Governance
and
Chief
of
Administration Branch - DESA/UNDP
FUNDAP - Executive Director
IADB - Inter-American Development Bank
World Bank/Evaluation Officer and Public
Sector Reform Coordinator
IADB - Inter-American Development Bank
Modernization of State Specialist
Fundaci6n Omar Dengo - Diretora
Ejecutiva
USP - Universidade de Sio Paulo Professor
Secretaria da Fazenda/Ceara - Fiscal
Sindicato
IADB
06/2003
06/2003
06/2000
several
7/25/2000
Professor Boston University
PNAFE
7/26/2000
IMESP
SINTAF - Sindicato
FGV
FGV
IMF
IADB
IADB
2005
2005
2003
2003
Appendix C
Indicators
NATIONAL FISCAL ADMINISTRATION PROGRAM FOR THE BRAZILIAN STATES
PNAFE
Reform Indicators
1997
1998
1999
2000
2001
2002
2003
2004
1997
1998
Reform 2 AdminA connect
# State Agencies on-line
Agenr~I~
ipJ
1999
2000
2001
2002
2003
2004
AC
AL
AM
AP
BA
CE
nd
0.020
0.040
0.011
0.037
0.058
0.240
0.040
0.071
0.040
0.064
0.047
0.027
nd
0.076
0.040
0.043
0.047
0.096
0.029
0.069
0.020
0.030
0.040
0.163
0.032
0.071
0.080
0.035
0.034
0.150
0.033
0.071
0.070
0.030
0.029
0.169
0.035
0.042
0.002
0.031
0,046
nd
AC AC
0.035 AL AL
0.042 AM AM
0.002 AP AP
0.028 BA BA
0.036 CE CE
1.000
0.760
1.000
0.910
1.000
0.840
1.000
0.960
0.952
0.920
1.000
0.830
1.000
1.000
0.955
0.930
1.000
0.867
1.000
1.000
0.962
0.930
1.000
0.880
1.000
1.000
0.980
0.930
1.000
0.877
1.000
1.000
0.980
0.910
1.000
0.909
1.000
0.887
1.000
0.938
1.000
1.000
1.000 AC
0.887 AL
1.000 AM
0.938 AP
1.000 BA
1.000 CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
nd
nd
nd
0.030
0.026
0.057
0.071
0.034
0.030
0.050
0.067
0.015
0.031
nd
0.050
0.060
0.027
nd
nd
0.200
0.090
nd
0.019
0.061
0.030
0.040
0.060
0.072
0.014
0.028
0.025
0.030
nd
0.020
0.010
0.190
0.080
0.025
0.020
0.006
0.044
0.028
0.030
0.040
0.062
0.015
0.030
0.022
0.030
0.000
0.020
nd
nd
0.070
0.020
0.047
0.040
0.020
0.022
0.020
0.020
0.000
0.058
0.065
0.057
0.025
0.023
0.040
0.030
0.043
0.039
0.051
0.049
0.013
0.015
0.030
0.030
0.020
0.019
0.004
nd
0.000 0.000
0.020
0.020
nd
0.020
0.005
0.020
0.020
0.055
0.060
0.022
0.030
0.041
0.050
0.016
0.020
0.019
nd
nd
0.020
nd
0.023
0.011
0.019
0.026
0.085
0.031
0.019
0.027
0.047
0.040
0.014
0.007
0.024
0.029
0.021
0.011
nd
0.014
0.010
0.020
0.018
0.085
0.065
0.021
0.033
0.029
0.039
0.014
0.007
0.020
0.029
0.021
0.011
1.000
nd
nd
1.000
0.920
0.770
1.000
0.920
0.700
0.430
1.000
0.910
0.360
nd
1.000
0.000
1.000
1.000
0.910
nd
1.000
nd
0.900
1.000
0.920
0.700
0.550
1.000
1.000
0.420
0.000
1.000
nd
1.000
1.000
1.000
1.000
1.000
0.970
0.760
0.920
0.940
0.910
0.700
1.000
1.000
1.000
0.717
1.000
0.000
1.000
1.000
1.000
0.980
0.910
0.960
1.000
0.989
0.920
0.950
1.000
0.990
1.000
0.717
1.000
nd
1.000
1.000
1.000
1.000
0.990
0.970
0.897
1.000
1.000
0.930
1.000
1.000
0.980
1.000
0.597
1.000
nd
1.000
1.000
1.000
1.000
1.000
0.930
0.867
1.000
1.000
0.930
1.000
1.000
0.990
1.000
0.597
1.000
0.844
1.000
1.000
1.000
1.000
1.000
0.933
0.972
1.000
1.000
1.000
1.000
0.991
0.985
1.000
1.000
1.000
0.882
1.000
1.000
1.000
1.000
1.000
0.858
0.972
1.000
1.000
1.000
1.000
1.000
0.985
1.000
1.000
1.000
0.882
1.000
SC
0.026
0.023
0.019
0.014
0.014
0.017
1000
SE
SP
0.110
0.022
0.030
0.060
0.023
0.040
0.017
0.040
0.031
0.060
0.030
0.045
0.028
TO
0.040
0.036
0.030
0.020
0.020
0.020
Reform I AdminCost
STAB Expenditure
-
STA
-epn
__--
-detm
-tt
~
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
0.011
SC
SC
1.000
1.000
1.000
1.000
1.000
1.000
SC
0.043
0.016
0.043
0.016
SE
SP
SE
SP
0.930 0.970 0.980
0.870
0.880
0.981
nd
1.000
0.980
1.000
0.984
1.000
SE
SP
0.026
TO
TO
0.950
1.000
1.000
1.000
1.000
1.000
0.076
1.000
1.000
1.000
1.000
0.960
1.000
1.000
1.000
O
193
Appendix C. Indicators
-T-tal
1997
AC
nd
AL 0.450
AM 0.510
AP 0.770
BA 0.750
CE 0.960
nd
DF
ES
nd
GO 0.530
MA 0.560
MG 0.950
MS 0.920
MT 0.530
PA 0.060
PB 0.500
PE 0.570
PI 0.340
PR 0.510
RJ 0.900
RN
nd
RO 0.280
RR 1.000
RS 0.360
SC 0.960
SE 0.540
SP 0.370
TO 0.620
1998
nd
0.620
nd
0.830
0.720
0.230
0.850
nd
0.350
nd
nd
0.970
1.000
0.150
0.500
0.910
0.439
0.490
0.880
0.370
0.610
nd
0.370
0.950
0.560
0.370
nd
Reform 3 TaxComplaince
Total VAT Returns
axnavi
eBfzl
1999 2000 2001
2002
nd
nd
nd
0.398
1.000 0.964
0.480 0.590
0.300 0.354 0.329 0.329
0.780 0.790 0.720 0.450
0.680 0.690 0.690 0.690
0.910 0.950 0.953 0.905
0.790
nd
0.350 0.350
0.980 1.000 0.960 0.960
0.420 0.421
0.412 0.324
0.520 0.600 0.520 0.970
0.340 0.980 0.980 0.980
0.980 1.000
1.000
1.000
1.000 0.812 0.844 0.745
0.150 0.149 0.853 0.670
0.960 0.520 0.510 0.480
0.830 0.910 0.890 0.770
0.319 0.313 0.380 0.362
0.510 0.540 0.580 0.680
0.880 0.880 0.890 0.990
0.458 0.467 0.677 0.677
0.700 0.880 0.910 0.580
0.610 0.614 0.614 0.992
0.390 0.360 0.720
nd
0.894 0.946 0.946 0.994
0.790 1.000 0.880 0.910
nd
nd
0.828 0.828
0.740 0.280 0.490 0.490
Reform 4 OnlineBudget
Expenditure Budgeted on-line
-
2003
1.000
0.943
0.298
0.587
0.702
0.993
0.403
1.000
0.421
0.861
0.983
0.707
0.745
0.558
0.508
nd
0.323
1.000
0.558
0.680
0.287
0.523
0.472
0.994
0.969
0.766
0.686
-
2004
0.043 AC AC
0.964 AL AL
0.329 AM AM
0.450 AP AP
0.688 BA BA
0.857 CE CE
0.403 DF DF
1.000 ES ES
0.394 GO GO
0.815 MA MA
0.984 MG MG
1.000 MS MS
1.025 MT MT
0.922 PA PA
0.488 PB PB
0.813 PE PE
0.378 PI
P1
1.000 PR PR
0.829 RJ RJ
0.586 RN RN
0.580 RO RO
0.523 RR RR
0.258 RS RS
0.994 SC SC
1.071 SE SE
0.828 SP SP
0.740 TO 1 TO
- Il
1997 1998
nd
nd
nd
1.000
1.000 0.980
nd
0.240
nd
0.980
1.000
1.000
1.000 1.000
nd
1.000
nd
nd
nd
1.000
nd
0.980
0.920 0.970
1.000 1.000
0.940 0.950
1.000 1.000
0.980 0.990
0.930 0.901
1.000 1.000
0.870 0.980
nd
0.000
1.000 1.000
0.000
nd
1.000 1.000
nd
0.810
nd
nd
0.900 0.940
1.000 1.000
-
1999
1.000
nd
0.984
1.000
1.000
0.926
0.990
1.000
1.000
0.570
nd
0.880
1.000
0.960
1.000
0.990
nd
1.000
0.980
0.821
1.000
nd
1.000
1.000
nd
1.000
1.000
2000
1.000
nd
0.980
1.000
1.000
1.000
0.760
1.000
1.000
0.670
0.970
0.840
1.000
1.000
1.000
1.000
nd
1.000
0.980
0.967
1.000
nd
1.000
1.000
nd
1.000
1.000
2001
1.000
1.000
0.985
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.980
0.888
1.000
1.000
1.000
1.000
0.965
1.000
0.980
1.000
0.980
nd
1.000
1.000
nd
1.000
1.000
2002
1.000
1.000
0.985
1.000
1.000
1.000
0.000
1.000
1.000
100
1.000
0.905
1.000
1E000
1.000
1.000
1.000
1.000
0.870
1.000
0.980
0.925
1.000
1.000
nd
1.000
1.000
2003
1.000
1.000
1.000
1.000
1.000
1.000
0.999
1.000
1.000
1.000
0.973
0.961
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.980
1.000
1.000
1.000
nd
1.000
1.000
2004
0.001 AC
1.000 AL
1.000 AM
1.000 AP
1.000 BA
1.000 CE
0.999 DF
0.954 ES
1.000 GO
1.000 MA
0.975 MG
0.961 MS
1.000 MT
1.000 PA
1.000 PB
1.000 PE
0.965 PI
1.000 PR
1.000 RJ
1.000 RN
0.980 RO
1.000 RR
1.000 RS
1.000 SC
SE
nd
1.000 SP
1.000 TO
0
-
-
0
Q
O O
VNO
O
ON
ON
CI
14
AAV
CO
OO-o-ooON-OO-
-C5
C C
=
-
0
00
O
C)
o--
So
C
O
00zI§
-OCD
0OC0
0
C71
00ON
-
00
0
O
1c1e
0n
No-
ON
C
\O
0
o
--
QMAO.Cl
O
O NC',O
0C
ON
c
0
-
-O
'OO
C)
ON
oo
e
O
00
IN
0'O N0C
-o
IN
o
ON0
-NNO
o
cClC-
e
N
1C-C-00
00
00 a- § 0N0N
OC
r
o~oO
C
ri
tn
1=
rn
=
o-oNooNoo-oNo
C-.o lC
C,
oNooN
C> ii
C,4O
t1-
o-
ON
ON0ON C ONOO
0O
O VN0O
O Okn O
00 O rn
O
r
OON 00
ON ONO ON ON ON cCONN O
C"OONlN, O
AC)
~-l--ON--ON--ON-Noo-oN-o-~
O O
~
NNC-O NO
0 ON00O o
OoN
t-
(71
C>
kr
ccl
'IT
ON~-0C
0
r fnONN
nOO
~ONONC-ONO~U-~ONCNCN-CN0l-CNO
NON-1
0ON N~
-oo-ONOO-~ONON
~ON
995
00
~ ~~NNI~ON
~00
O--o-Oo---O-----
.
00 C11 N
OOOOOLi6-ONNNNNN
is113
00
oo
EN'ON
CloNo
r
ON
OOO
ooooo
O
ON
O
O0
ON
o
0
CO
o
ss2@s
o
00
0
ON
C5
pNON
Rs
-
NlC
C)
-O
0
eanE,
I
ON C-O
ON C -O
9
-Co0~
1=1 111.0
dNO
o
N
-o
C
o
-Fo
ONNO
oooo
o-
ON
C N
N
O
0
i=s1
N00
ClO
o6o
Cl)O
e-s
-
O
c>
n
-O
0
ON
ON-S
,T
'n
N
00C
0
-z
o
--
o
i
o
i"
oo
ON: ,d
O
C5
oC
lO
c
0-C
=
O
ooo
0
O~O
~
ON '0OVN
C-
oC
ON
00
ONNO
oo oooo
OO7
o~ioNONo
d
N dd
8o
ON CC CON-)
C>
"dlOON
oo 0
-iN
ON
ooo
o
NNN
oa
ONO
o
-O
IN
ClON
C-iO
ON
'R00C4
ONO -S
~
ON
IN
ONON-ON
~OOO
O
-ON
o
ON
-ON
00 C-ON
o
ON
NI
did oldO
c
o
o0oooo
0N0No: ON0
oOOOOOOOOOOOOO4
o
o
idON
o
-
O
'R
-O
ON
'S 7S
N-51O0
O
o
O
ONOOOON
'So"S"87o
-s4
.,s
oo
C-
id(
-ooo
ONONONONON
O--ON
o
OO
-
-(
CD
-
ONO -OO
~
N
00ON
-
-2~
o
oN
o
7o
ci,
q
-S
O
QNON
-
Sf
o
ONON0
Z
00
-S
qr
Appendix C. Indicators
__TAR
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
P1
PR
RJ
RN
RO
RR
RS
SC
SE
SP
1997
nd
nd
0.500
nd
1.670
0.240
nd
nd
0.370
nd
1.060
0.260
nd
nd
nd
1.840
nd
0.630
nd
nd
nd
nd
0.740
nd
nd
0.110
nd
1998
nd
nd
0.733
0.360
1.060
0.330
0.220
nd
0.000
nd
nd
0.090
nd
0.150
1.140
0.440
nd
0.650
0.450
nd
0.040
nd
0.930
0.500
0.140
0.110
0.220
Reform 9 DebtOutstanding
Outstanding Debt
To
nI
2002
1999
2000
2001
nd
0.382
0.350
0.417
nd
0.001
nd
nd
0.997
0.925
1.010
0.010
0.410
0.300
1.470
0.600
1.067
0.390
0.420
0.290
0.370
0.373
0.478
0.658
0.070
0.370
0.330
0.327
0.150
0.190
nd
nd
2.468
2.167
2.389
2.180
nd
nd
nd
nd
0.990
1.180
1.130
1.080
0.090
0.140
0.175
0.218
0.080
0.034
0.060
0.060
0.323
0.654
0.310
0.326
0.040
0.070
0.210
1.010
0.250
0.220
0.120
0.310
0.069
nd
0.002
0.036
0.870
0.820
0.760
0.870
0.500
0.550
0.470
0.470
0.040
0.038
0.024
0.024
0.030
nd
1.150
1.170
0.505
0.505
0.013
0.670
0.270
1.130
1.220
1.100
0.256
0.378
0.378
0.394
1.030
0.193
0.430
0.808
0.042
0.029
0.034
0.034
0.140
0.200
0.420
0.420
Reform 10+L185 DebtInstallments
Total Debt Negotiated
_
2003
0.411
2.020
1.980
0.773
0.222
1.801
0.385
0.007
4.420
nd
1.136
0.318
0.178
0.837
4.027
0.026
0.077
1.010
0.631
0.014
1.316
0.039
0.000
0.277
2.342
0.001
0.900
_
2004
0.057
0.000
0.000
0.000
0.141
0.000
0.000
0.000
2.281
0.098
0.910
0.000
0.000
0.455
3.744
0.158
0.034
1.108
0.312
0.018
0.000
0.039
1.310
0.021
1.084
0.000
0.011
_
AC AC
AL AL
AM AM
AP AP
BA BA
CE CE
DF DF
ES
ES
GO GO
MA MA
MG MG
MS MS
MT MT
PA PA
PB PB
PE
PE
P1
P1
PR PR
RJ RJ
RN RN
RO RO
RR RR
RS
RS
SC
SC
SE
SE
SP
SP
U TiY
1997
nd
nd
0.060
0.050
0.140
0.020
0.050
nd
0.080
nd
0.040
0.030
nd
nd
nd
0.240
nd
0.160
nd
nd
nd
nd
0.130
0.060
0.140
0.000
0020
_ve
1998
nd
nd
0.029
0.070
0.140
0.020
0.020
nd
0.000
nd
nd
0.040
0.002
0.030
0.010
0.080
nd
0.250
0.040
0.030
0.010
nd
0.070
0.080
0.140
0.000
0.010
196
1997
nd
nd
0 260
0.210
0.120
0.680
nd
nd
0.620
nd
0.120
0.220
nd
nd
nd
0.670
0.060
0.002
nd
nd
0.010
nd
0.040
nd
0.500
0.140
0.020
1998
nd
0.010
0.313
0.020
0.030
0.760
nd
nd
0.000
nd
nd
0.020
0.060
0.440
0.020
0.420
0.059
0.002
nd
0.090
0.130
nd
0.010
0.100
0.580
0.140
0.040
1999
nd
nd
0.430
0.540
0.236
0.872
0.090
1.130
0.260
nd
0.040
0.020
0.035
0.510
0.030
0.320
0.043
0.002
nd
0.063
0.130
0.970
0.010
0.107
0.030
0.028
0.040
2000
0.092
nd
0.381
0.390
0.470
0.840
0.240
1.750
0.571
nd
0.160
0.010
0.138
0.395
0.020
0.190
0.074
nd
nd
0.022
0.360
0.052
0.080
0.113
0.125
0.269
0.009
2001
0.063
nd
0.490
0.460
0.200
0.841
0.030
1.550
0.492
0.300
0.120
0.289
0.051
0.304
0.050
0.210
0.052
nd
nd
0.106
0.260
0.052
0.060
0.113
0.045
0.271
0.030
2002
0.453
nd
0.490
0.630
0.100
0.841
0.030
1.550
0.495
0.260
0.030
0.169
0.180
0.465
0.070
0.050
0.061
nd
0.200
0.106
0.020
0.668
0.060
0.126
nd
0.271
0.030
2000
0.261
nd
0.012
0.030
0.060
0.020
0.059
0.040
0.038
0.080
0.050
0.020
0.006
0.025
0.010
0.060
0.044
0.130
0.010
0.099
nd
0.020
0.130
0.146
nd
0.094
0.008
I Reve iue
2002
2001
0.292
0.244
nd
nd
0.015
0.015
0.020
0.030
0.027
0.027
0.030
0.017
0.036
0.025
0.030
0.030
0.087
0.080
nd
nd
0.080
0.070
0.025
0.025
0.034
0.036
0.059
nd
0.010
0.020
0.080
0.050
0.024
0.024
0.090
0.070
0.020
0.030
0.042
0.042
0.110
0.130
0.020
0.087
0.090
0.060
0.146
0.143
0.110
0.080
0.038
0.038
0.210
0.210
2003
0.206
0.017
0.013
0.037
0.027
0.018
0.072
0.093
0.143
nd
0.069
0.052
0.014
0.019
0.450
0.037
0.014
0.053
0.135
0.026
0.173
0.040
0.000
0.040
0.131
0.018
0.214
2004
nd
nd
nd
nd
0.068
0.008
0.121
0.105
0.156
0.060
0.082
0.000
0.006
0.016
0.466
0.158
0.004
0.034
0.071
0.022
0.000
0.040
0.147
0.108
0.061
nd
0.119
2003
nd
0.024
0.913
0.264
0.563
0.821
nd
nd
3.617
0.437
0.991
0.057
2.271
0.031
3.414
1.871
0.468
nd
0.621
0.515
nd
0.073
0.982
0.571
1.902
1.506
0.900
2004
0.057
0.000
0.000
nd
0.708
0.497
0.826
nd
1.228
0.430
0.910
nd
1.028
0.129
3.209
1.879
nd
1.018
0.328
0.444
0.000
0.073
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
SC
SE
SP
T0
Reform 12 DebtJudged
Estoque valor em execueao
Reform 11 DebtDisputed
Value under Dispute
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
SC
SE
SP
1999
nd
nd
0.063
0.060
0.044
0.015
0.750
0.010
0.170
0.030
0.040
0.030
0.004
0.010
0.010
0.130
0.043
0.100
0.040
0.066
0.200
0.035
0.060
0.083
0.200
nd
0.010
2003
nd
nd
1.034
0.432
0.036
0.864
nd
nd
0.803
0.231
0.063
0.064
0.362
0.496
0.231
0.057
0.081
0.012
0.256
0.072
0.195
0.028
0.046
0.160
0.250
0.169
0.036
2004
nd
nd
ind
nd
0.070
0.387
0.108
nd
0.467
0.353
0.057
nd
0.104
0.234
0.096
0.036
0.176
0.106
0.128
0.026
nd
0.028
0.063
0.199
0.220
nd
0.107
AC AC
AL AL
AM AM
AP AP
BA BA
CE CE
DF DF
ES ES
GO GO
MA MA
MG MG
MS MS
MT MT
PA
PA
PB
PB
PE
PE
P1
PI
PR PR
RJ RJ
RN RN
RO RO
RR RR
RS
RS
SC
SC
SE
SE
SP
SP
1997
0.060
0.200
0.270
0.010
0.580
0.050
nd
nd
0.230
0.020
0.690
0.350
0.110
nd
0.980
1.040
0.322
nd
nd
nd
0.020
0.110
0.570
nd
0.180
0.000
nd
1998
0.180
0.360
0.391
0.090
nd
0.160
nd
nd
0.000
nd
nd
0.060
0.770
0.060
1.100
1.050
0.378
nd
nd
0.110
0.400
nd
0.770
nd
0.140
0.000
0.120
1999
0.340
nd
0.504
0.090
0.565
0.220
0.400
nd
0.710
0.060
0.780
0.020
0.653
nd
1.010
1.260
0.527
nd
0.370
0.248
0.300
0.590
0.840
nd
0.320
nd
0.250
2000
nd
nd
0.488
0.130
1.090
0.261
0.440
nd
0.393
0.090
0.860
0.520
0.590
0.126
1.130
1.480
0.473
nd
0.380
0.461
0.101
0.030
0.970
0.446
0.330
nd
0.160
2001
nd
nd
0.516
0.120
0.620
0.220
nd
nd
1.616
0.170
0.960
0.427
0.779
nd
1.130
1.660
0.500
nd
0.510
0.485
nd
0.030
0.890
0.446
0.800
nd
0.320
2002
0.387
nd
0.516
0.010
0.620
0.260
0.390
nd
1.894
0.150
0.930
0.401
0.779
0.014
1.520
1.700
0.481
nd
0.380
0.485
nd
0.082
1.010
0.515
0.810
nd
0.320
1.123
0.548
1.002
nd
0.011
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
P1
PR
RJ
RN
RO
RR
RS
SC
SE
SP
Appendix C. Indicators
Reform 13 Training
Number of Trainned Officials
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI1
PR
RJ
RN
RO
RR
RS
SC
SE
SP
'1'0
1997
0.000
0.660
nd
0.000
nd
2.230
0.420
nd
0.270
0.370
1.610
0.730
0.480
0.460
0.640
0.490
0.340
0.350
0.260
nd
0.430
0.000
0.790
0.790
0.220
0.030
0.010
1998
0.290
1.330
0.749
3.130
3.210
2.470
0.480
0.150
0.510
1.430
nd
0.630
0.740
0.520
1.010
0.510
1.719
0.820
0.130
0.140
0.010
nd
2.100
2.190
1.040
0.040
0.340
1999
1.870
0.260
0.192
3.330
2.849
0.911
0.800
0.870
0.940
1.004
0.260
1.780
1.730
0.480
1.130
1.960
0.420
0.570
0.310
0.443
0.250
0.207
0.940
0.738
1.880
nd
1.370
2000
2.209
1.460
0.470
0.950
0.170
0.900
1.030
3.990
0.666
0.390
1.190
0.490
0.440
0.540
1.450
2.330
0.263
0.670
1.910
0.642
0.370
0.230
2.690
1.057
1.430
5.900
1.970
2001
0.768
1.820
0.964
0.930
2.140
0.859
0.400
1.890
0.626
0.450
L.010
1.685
0.700
0.550
1.750
1.600
0.261
0.240
1.530
0.406
0.490
0.230
2.140
1.057
2.410
4.960
0.360
2002
1.017
0.500
0.964
0.960
4.150
0.876
1.630
1.890
0.849
0.620
1.960
2.603
0.724
0.267
1.070
1.860
0.439
0.330
0.560
0.406
0.490
0.223
0.610
0.513
L070
5.297
0.360
Reform 14 LargeTaxpayers
Taxpayers paying > 50% VAT Revenue
2003
1.021
3.150
0.514
1.000
2.100
0.507
0.432
0.329
0.170
0.455
2.951
0.628
0.122
0.675
0.406
1.001
1.156
0.166
0.139
0.679
0.728
0.394
0.279
0.199
0.723
2.891
0.053
2004
0.397
AC AC
nd
AL
AL
nd
AM AM
nd
AP
AP
1.452 BA BA
0.974 CE CE
0.217 DF
DF
ES
0.628
ES
0.384 GO GO
0.268 MA MA
2.624 MG MG
MS MS
nd
0.755 MT MT
PA
0.758 PA
0.476 PB
PB
0.636
PE PE
1.015
PI
PI
0.463 PR
PR
0.325
RJ RJ
RN
0.616 RN
nd
RO RO
0.394 RR RR
0.325
RS RS
SC
0.746
SC
1.610
SE
SE
SP
nd
SP
0.610 1
TO
1997
nd
0.0015
0.0004
nd
0.0002
0.0005
0.0004
nd
0.0001
0.0002
0.0002
0.0008
0.0004
0.0010
0.0015
0.0004
0.0001
0.0001
0.0001
nd
nd
0.0014
0.0000
0.0004
0.0050
0.0002
0.0100
1998
nd
0.0011
0.0006
0.0008
0.0002
0.0002
0.0005
0.0002
0.0003
0.0002
nd
0.0006
0.0002
0.0004
0.0015
0.0006
0.0006
0.0003
0.0001
0.0003
0.0004
nd
0.0000
0.0003
0.0050
0.0002
0.0013
1999
0.0003
0.0010
0.0005
nd
0.0001
0.0002
0.0001
0.0003
0.0002
nd
0.0000
nd
0.0001
0.0005
0.0020
0.0005
0.0008
0.0001
0.0000
0.0006
0.0008
nd
0.0000
0.0002
0.0050
0.0001
10.0010
2000
0.0011
0.0005
0.0005
nd
0.0000
0.0003
nd
0.0003
0.0008
nd
0.0000
0.0000
0.0001
0.0033
0.0030
0.0005
0.0004
0.0001
0.0001
0.0005
0.0010
0.0009
0.0000
0.0001
0.0048
0.0001
0.0040
2001
0.0009
0.0006
0.0004
0.0014
0.0000
0.0002
nd
0.0004
0.0001
0.0002
0.0001
0.0013
0.0001
0.0022
0.0010
0.0002
0.0004
0.0001
0.0000
0.0008
0.0025
0.0009
0.0000
0.0001
0.0097
0.0001
0.0010
2002
0.0009
0.0005
0.0004
nd
0.0001
0.0002
0.0010
0.0004
0.0002
0.0003
0.0000
0.0067
0.0002
0.0063
0.0010
0.0002
0.0002
0.0001
0.0001
0.0008
nd
0.0030
0.0000
0.0001
0.0007
0.0001
0.0010
2003
0.002
0.0003
0.0003
0.0026
0.000
0.0001
0.006
0.0003
0.0000
0.0004
0.0000
0.0001
0.0002
0.0002
0.0006
0.0002
0.0004
0.0000
0.0000
0.0010
0.0005
0.002
0.0000
0.0000
0.0006
0.0001
0.0002
2004
0.002 AC
0.0003 AL
0.0003 AM
0.0026 AP
0.000 BA
CE
0.0001
0.003
DF
0.0003 ES
0.0000 GO
0.0003 MA
0.0001 MG
0.0001 MS
0.0002 MT
0.0005 PA
0.0006 PB
0.0003 PE
0.0004
P1
0.0000 PR
0.0000 RJ
0.0003 RN
0.0000 RO
0.002 RR
0.0000 RS
0.0000 SC
0.0005 SE
0.0000 SP
0.0005 T0
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
is&
-283571.7483 -2184991.9412 -1664912.2854 -1247636.8168 -1838056.9146 -1885372.3086 -1090066.6189
632604.1044 1222588.2694
704525.9303
544202.4282
189845.2526
538629.8638
540323.7568
-2.9803
-0.8916
-2.3091
-3.3775
-1.4937
-4.0566
-2.3632
t
0.0765
0.0836
0.0747
0.0803
0.0805
0.0803
0.0793
Beta - GDP
0.0049
0.0043
0.0053
0.0082
0.0047
0.0038
0.0058
se
15.7182
9.2919
15.2438
18.5033
13.9735
21.2916
16.7170
t
-10072.9796
-78541.2475
63304.2342
57455.6024
-4893.5929
-3132.6143
Beta - Reform
85467.4081
68480.4916
85151.6738
1628.4988
se
9217.5356
150324.3412
-0.1471
-1.9236
0.7434
0.6723
-0.5309
-0.5225
t
0.0034
-0.0001
0.0001
0.0024
0.0029
Beta - Reform*GDP
0.0032
0.0001
0.0016
0.0017
0.0008
0.0010
0.0016
se
1.7444
2.1220
-0.0643
1.4951
3.1223
3.5885
t
Beta - dev level 1
216.0000 2148966.4232 1758606.9933 1465617.4424 1867008.1744 1928765.6100 1047963.5940
se
541809.8628
549342.5442
639657.6293 1236171.5037
0.9217
543284.6380
711892.5620
2.4703
2.6679
3.0153
0.8477
3.9555
3.4459
t
Beta - dev level 1 * GDP
-0.0299
-0.0104
0.0019
-0.0173
-0.0137
0.0013
0.0147
0.0131
0.0090
0.0129
se
0.0152
0.0145
-2.2888
0.1319
-0.8991
-1.3436
0.0876
-1.1555
te
Be - devevl2
-0.0013
0.0040
0.0003
-0.0030
0.0025
0.0000
0.0014
0.0002
0.0041
0.0128
0.0067
se
0.0057
2.1744
-0.9123
0.3130
-0.7425
0.3806
t
-0.0017
2247820.4010 1694602.9614 1377131.8706 1716782.4216
1817253.5173 1031585.8857
Beta - dev level 2
581476.3882
583601.3845
676631.8092 1260274.4330
se
750540.5503
565667.4084
2.3683
3.8516
2.2578
2.6857
t
3.0350
0.8185
Beta - devlevel2*GDP
-0.0047
-0.0063
-0.0011
0.0045
-0.0050
0.0048
0.0086
se
0.0087
0.0131
0.0101
0.0102
0.0138
-0.1252
-0.4802
-0.5400
-0.4946
t
0.4487
0.3458
Beta-dev level2- Reform-GDP
-0.0037
0.0000
0.0024
-0.0091
0.0004
-0.0023
se
0.0010
0.0069
0.0001
0.0059
0.0036
0.0035
t
-3.8585
0.7209
0.3463
0.1162
-1.5355
-0.6575
obs
191.0000
197.0000
207.0000
187.0000
197.0000
199.0000
R-squared
0.9577
0.9390
0.9681
0.9671
0.9518
0.8698
Constant
se
Table 29. Results of the Base Model Estimation, Reform Indicators and State Economic Development Levels.
-532458.8948 -1245495.7968 81131.2277
480086.9324 493702.5790 970554.2674
0.0836
-2.5228
-1.1091
t
0.0714
0.0780
0.0712
Beta - GDP
0.0063
0.0059
0.0046
se
11.2665
13.1729
15.3626
t
Beta - Reform -49200.1742 -80657.3209 -506463.0052
61427.0181 193514.3985 666380.4993
se
-0.4168
-0.7600
-0.8010
t
-0.0021
0.0168
Beta - Reform*GDP
0.0010
se
0.0012
0.0249
0.0202
-0.1032
0.8546
0.6753
t
Beta - dev level 1 601815.0195 1352403.5390 -109624.6887
se
477962.5436 489560.9404 963129.5226
-0.1138
2.7625
1.2591
t
0.0108
Beta -devievel 1 GDP
-0.0017
-0.0131
0.0118
0.0116
se
0.0104
-1.2519
0.9173
t
-0.1465
Bet-devievl
I-Refom
*GDP
0.0043
-0.0027
0.0452
0.0185
0.0243
0.0073
se
2.4435
0.5950
-0.1095
t
Beta - dev level 2 387411.5079 1415827.0210 -217163.3886
se
498836.7879 514210.8617 968203.0815
0.7766
2.7534
-0.2243
t
Beta- devlevel2 GDP
-0.0064
0.0069
0.0143
se
0.0076
0.0088
0.0093
-0.7282
1.8786
0.7430
t
Beta8de 12
GDP
-0.0008
0.0640
-0.0113
se
0.0010
0.0235
0.0220
-0.7562
t
-0.4812
2.9037
161.0000
obs
180.0000
163.0000
0.9796
R-squared
0.9570
0.8343
Constant
se
-1702503.1570 -1362144.0261 -569963.0096 -1118278.1617
557499.0157
-3.0538
0.0833
0.0048
17.3655
90714.4187
87800.7381
1.0332
-0.0018
0.0019
-0.9114
1747257.1542
553363.1689
3.1575
-0.0101
0.0100
-1.0132
-0.0079
0.0065
-1.2201
1652919.9855
588633.4355
2.8081
-0.0038
0.0093
-0.4055
0.0009
0.0027
0.3279
177.0000
0.9586
399189.4622 338412.5987 329530.8680
-3.3935
-1.6842
-3.4123
0.0757
0.0705
0.0769
0.0022
0.0048
0.0033
34.9738
14.6304
23.1096
-119266.3144 -87872.4033 -51627.0085
88271.0204 101077.7813 43675.7392
-1.1821
-0.8694
-1,3511
0.0016
0.0225
0.0067
0.0076
0.0003
0.0029
5.4914
2.2977
2.9608
1454482.6142 677889.5339 1221171.9602
401466.9466 326466.3418 329269.3560
2.0764
3.7087
3.6229
-0.0118
-0.0089
-0.0087
0.0077
0.0074
0.0093
-1.2140
-1.5299
-0.9308
-0.0092
-0.0014
0.0011
0.0234
0.0035
0.0081
0.3064
-0.3924
-0.1686
1327194.8645 537005.8222 1098841.3056
446850.0067 368607.5269 361792.3593
3.0372
1.4568
2.9701
-0.0026
0.0029
0.0026
0.0092
0.0073
0.0083
-0.3130
0.3147
0.3498
-0.0080
-0.0015
0.0005
0.0081
0.0034
0.0025
-0.9814
0.2025
-0.4453
165.0000
203.0000
171.0000
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
C'S
6
0
~C)
0
C)
6
0
0
-49284.729 -282655.523
1021.091 118887.602
Constant -283571.748 -330520.895 -481394.326 299332.121 136307.779
108669.317 131539.779 115084.352 98357.924
94836.634 112950.291
189845.253 228543.451 2748246.818
se
1.209
0.009
-2.149
-0.454
1.207
3.156
-0.175
-1.446
-1.494
t
0.065
0.080
0.068
0.077
0.068
0.064
0.082
0.076
0.079
Beta - GDF
0.004
0.003
0.002
0.003
0.004
0.003
0.014
0.005
0.005
se
23.620
14.905
31.450
29.388
18.693
6.095
20.641
14.980
16.717
t
399194.441 93495.216 -497891.731 -77991.892
-1180.267 -16199.255 -501461.877
-26661.238
Beta - Reform
1118.896 59083.799 282768.727 379638.142 72498.631 126843.317 69971.781
12014.308
se
-1.115
1.290
-3.925
1.052
-1.773
-0.274
-1.055
-2.219
t
0.004
-0.003
0.036
-0.023
0.023
0.000
0.000
0.003
Beta - Reform*GDP
0.003
0.016
0.003
0.003
0.010
0.001
0.000
0.001
se
1.456
-1.460
-1.238
11.467
0.217
2.319
1.050
5.214
t
-1751.459 -122990.802 -488893.714
745182.565 160759.778 -179782.631 -5659.683
14068.012
216.000
Beta - Refc
907.394 58994.153 212920.907 423520.615 132069.906 86298.282 92127.697
0.922 20234.617
se
-0.061
1.217
-2.083
-2.296
1.759
-2.085
0.695
-1.930
t
0.002
-0.005
0.014
-0.051
0.029
0.002
0.000
-0.001
Beta - Reform*GDP Lag
0.004
0.007
0.002
0.011
0.024
0.001
0.002
0.000
se
-0.767
6.208
0.552
-2.140
2.681
2.319
2.109
-0.441
t
8927.401 -241755.217 -1828443.252 233221.930 -111030.336 166063.509
6017.125
268.190
Beta - Reform Lag 2
1318.421 73783.492 236985.271 751704.312 114650.143 95136.360 91187.640
18920.274
se
1.821
-1.167
-2.432
2.034
-1.020
0.121
0.318
0.203
t
-0.002
-0.007
0.003
0.063
0.016
0.000
0.000
0.000
Beta - Reform*GDP Lag'
0.004
0.002
0.021
0.007
0.002
0.032
0.000
0.001
se
-1.059
1.270
-0.449
2.002
0.145
0.773
-0.021
-0.712
t
117.000
118.000
112.000
100.000
97.000
105.000
104.000
132.000
obs
0.976
0.984
0.934
0.984
0.982
0.944
0.794
0.961
R-squared
LR test
K
259327.462
104119.906
2.491
0.065
0.002
28.507
-62914.645
28867.235
-2.179
0.001
0.000
3.020
-39227.725
33041.575
-1.187
0.002
0.001
2.946
-35983.151
25325.427
-1.421
0.001
0.001
0.909
145.000
0.987
-244057.413
220519.507
-1.107
0.080
0.005
16.649
844579.080
686401.019
1.230
-0.066
0.051
-1.300
742206.010
582208.751
1.275
-0.067
0.047
-1.424
558158.882
300278.076
1.859
-0.057
0.042
-1.353
132.000
0.932
Acronyms
For the Brazilian States:
AC
AL
AM
AP
BA
CE
DF
ES
GO
MA
MG
MS
MT
PA
PB
PE
PI
PR
RJ
RN
RO
RR
RS
SC
SE
SP
TO
Acre
Alagoas
Amazonas
Amapi
Bahia
Ceari
Distrito Federal
Espirito Santo
Goiis
Maranhio
Minas Gerais
Mato Grosso do Sul
Mato Grosso
Pari
Paraiba
Pernambuco
Piaui
Parani
Rio de Janeiro
Rio Grande do Norte
Rond6nia
Roraima
Rio Grande do Sul
Santa Catarina
Sergipe
Sio Paulo
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 Politica Fazendiria)
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 sao os Servidores Pn6blicos
Brasileiros?" Sio Paulo Universidade de Sdo 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. ButterworthHeinemann.
Bonfim, Washington (1999). "A Reforma Fiscal no Ceari 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). "Restrucuturagao DEAT." Sao Paulo Departamento de Administratagao Tributiria,
SEFAZ (Secretaria da Fazenda do Estado de Sao 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, Mircio Goes da Cruz, et al. (1998). "Indicadores de G6stion para El Area Tributiria."
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, Andr6a Gomes and Jos6 Roberto Rodrigues Afonso (2001). "e-Governo no Brasil:
Experincias 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 Aranjo 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-Garcia, J. Ram6n and Gabriel Pur6n 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 1012.
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 Tributagao: Fatores de Exito na Nova
Modelagem Institutional da Gestdo Tributiria do Rio Grande do Norte." In Proceedings of I Pleniria
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 ITenabled Public Sector Reform. Routledge Research in Information Technology and Society. London;
New York, Routledge.
Hirschman, Albert 0. (1958). The Strategy of Economic Development. New Haven. Yale University
Press.
Hirschman, Albert 0. (1963). Journeys Toward Progress: Studies of Economic Policy-Making in Latin
America. New York, NY. The Twentieth Century Fund.
Hirschman, Albert 0. (1981). Essays in Trespassing : Economics to Politics and Beyond. Cambridge
[Eng.] ; New York. Cambridge University Press.
212
References
Hirschman, Albert 0. (1987). "The Political Economy of Latin American Development: Seven Exercises
in Retrospection." Latin American Research Review, 22 (3): 7-36.
Hirschman, Alberto 0. (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 Estatistica),
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). Servigos Piblicos no Brasil.
Mudangas e Perspectivas. Concessdo, Regulamentacio, Privatizacdo e Melhoria da Gestdo Pnblica.
Sao 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. Dilulio (1995). Inside the Reinvention Machine: Appraising Governmental
Reform. Washington, D.C. Brookings Institution.
Kettl, Donald F. and John J. Dilulio (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.
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). "Motivag o e Satisfagao do Servidor PIblico." 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 Luis Lemos and Aldemir Freire Moreira (1999). "Esforgo Fiscal e Carga Tributiria
Potencial Dos Estados do Nordeste." CAEN, Universidade Federal do Ceari. Texto para Discussio
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 Jinior, Wilson Silva do (2000). "0 Auditor do Conhecimento e as Tecnologias de Auditoria
Informitica." In Proceedings of I Pleniria 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
danda
J. and JoAnne Yates (2002). "It's About Time:
Temporal Staturing in
zations." Organization Science, 13 (6): 684-700.
is
Osborne, David and Ted Gaebler (1993). Reinventing Government : How the Entrepreneur iahprt
Transforming the Public Sector. New York, N.Y. Plume.
Osborne, David and Peter Plastrik (1991). Banishing Bureaucracy: The Five Strategies for Reinveting
Government. Reading, MA. Addison-Wesley.
Peled, Alon (2000). "First-class Technology--third-rate Bureaucracy: the Case of Israel." Informatin
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). "Relat6rio do PNAFE." Programa Nacional de Apoio i Administragao Fiscal para os
Estados Brasileiros.
PNAFE (N.D.). "PNAFE 6 Realidade: A Nova Face da GestAo Fiscal." PNAFE, Programa Nacional de
Apoio i Administra95o 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.
lamos, Alberto (2001). "Medium and Long-Term Fiscal Sustainability in Brazil: 2000-10." In Brazil Selected Issues and Statistical Appendix. Washington, D.C. IMF.
N-Ackerman,
Susan (1997). "The role of the World Bank in controlling corruption." Law and Policy
International Business, 27.
217
Jlerences
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 Sao Paulo) and (Escola Fazendaria do Estado de Sio 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 0 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 P
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
Rejerences
SRF (2002). "Anilise da Arrecada~ao das Grandes Empresas em 1999." Secretaria da Receita Federal,
Minist6rio da Fazenda, Brasil. Estatisticas Tributarias 09.
SRF (2007). "Carga Tributiria no Brasil 2006." Secretaria da Receita Federal, Ministerio da Fazenda,
Brasil.
Stanback, Thomas M. and Jr. (1987).
Computerization and the Transformation of Employment.
Government, Hospitals, and Universities. Boulder and London. Westview Press.
Stem, 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
0. 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
0. 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, Jos6 A. and Jos6 A. Pagin (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) (200 1). "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, Jos6 Romeu de, Marcelo Piancastelli Siqueira, et al. (2005). "Esforgo Fiscal dos Estados
Brasileiros." In Finangas Pndblicas: IX Pr6mio Tesouro Nacional - 2004. Brasilia, Brazil. Editora
Universidade de Brasilia. 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.