SUBNATIONAL
CIVIL SERVANT
PENSION SCHEMES
IN BRAZIL: CONTEXT, HISTORY,
AND LESSONS
OF REFORM
Social Protection and Jobs Global Practice
Latin America and the Caribbean Region
March 2022
Acknowledgements
The preparation of this report has been led by Asta Zviniene (Sr. Social Protection Specialist),
Rovane Schwengber (Social Protection Specialist), and Raquel Tsukada Lehmann (Consultant),
with important contributions from Rafael Amaral Ornelas (Consultant), Fabiano Silvio Colbano
(Senior Economist), Daniel Ortega Nieto (Senior Governance Specialist), Sara Brolhato de Oliveira
(consultant), Francisco Luis Lima Filho (Consultant), and Felipe Drumond (Consultant). The team
would also like to thank for the comments offered by Robert Palacios (Lead Specialist), Heinz Ru-
dolph (Lead Financial Sector Economist), Allex Albert Rodrigues (Subsecretary for RPPS, Ministry
of Labor and Social Security), Marcos Mendes (Economist, Associate Researcher at Insper), and
members of the National Council for Civil Servants Regimes (CONAPREV - Conselho Nacional
dos Dirigentes de Regimes Próprios de Previdência Social), as well as administrative data and
insights generously provided by multiple researchers and government officials of federal and
state governments. Finally, the team has greatly benefited from the guidance received from Ra-
fael Munoz Moreno (Program Leader) and Anita Schwarz (Lead Economist), and the leadership
of Paloma Anós Casero (Country Director for Brazil), Pablo Ariel Acosta (Program Leader), and
Pablo Gottret (Practice Manager).
Table of Contents
1
Executive Summary
page 6
2
Introduction
page 16
3
Evolution of civil service in
Brazil since 1988
page 20
4
The fiscal consequencesof
civil service expansion and
pension scheme maturation
page 32
5
Contributionof HR policies to
growth in wageand pension
spending
6
page 42
Previous pension policy
response to pension
spending increases
7
page 52
Adoption of Federal 2019
reform by subnational
governments
8
page 64
Lessons and
recommendation
page 78
9
References
page 82
Executive Summary
1
1. Executive Summary
Brazil’s pension system takes up an oversized proportion of its social protection spending. It
comprises of Regime Geral de Previdência Social (RGPS), covering private sector workers, and
over two thousand Regimes Próprios de Previdência Social (RPPS), insuring public civil servants
at federal and subnational levels. While the total membership of RPPS only stands at about 10
percent of RGPS coverage, its spending amounts to almost half of RGPS pension outlays.
This paper attempts to present an integrated view of RPPS pension schemes, their influence
on subnational budgets, and their interaction with human resource policies. After a brief intro-
duction, Chapter 3 starts by documenting the history of civil service and its associated pen-
sion schemes, looking for explanations on how subnational RPPS became so big, dispersed,
and difficult to reform. The fiscal consequences of subnational civil service pension scheme
expansion and maturation, including RPPS role in the fiscal challenges and policies of the last
few years, are discussed in Chapter 4. Chapter 5 attempts to expose important interlinkages
between pension and human resource policies and argues for the need of integrated policy
approach. Chapter 6 describes the history of previous RPPS reform attempts, while Chapter 7
focuses on the effects of federal pension reform of 2019 on subnational civil servant pension
schemes. The paper ends with lessons and policy recommendations for the future.
History of civil service and RPPS. Prior to the adoption of Brazil’s constitution in 1988, the size
of the civil service was limited. The Constitution has established universal access to public
services, which has led to increased demand and a rise in the number of public officials, es-
pecially at the municipal level. Major changes to the way civil servants were contracted were
also introduced, granting workers substantial legal protections and backloaded remunera-
tion structure, culminating in high pension after the lifetime of steep wage increases.
Since then, the overwhelming majority of civil servants exhibit life-time attachment to their
employer (the government) and tellingly, even as pensioners, are referred to as “inactive civil
servants”. The life-time affiliation concept helped create an implicit expectation that retire-
ment should not imply any notable change in the civil servant’s income stream. This meant
continuation of salary and benefits associated with the last position into retirement and sub-
sequent wage increases. The civil servant hiring wave that followed the establishment of the
8
Constitution in 1988 as well as legal protections and generous benefit packages granted to
civil servants at that time are major events that influence fiscal and political landscape of
1 Executive Summary
subnational governments today.
Role of RPPS spending in subnational finances. For the first few decades, the creation of
new generous pension regimes had an extremely limited effect on subnational government
finances, as the numbers of pensioners were still low. However, the beginning of the XXI
century marked the start of a rapid maturation of RPPS, when big previous hiring wave has
led to a new wave of retirements. This has strained subnational finances, compromising ser-
vice provision and investment. The trend is expected to continue, with the full maturation
of most subnational systems not expected before 2040. However, the variability of pension
scheme maturation levels across regions and states is considerable: northern states tend to
lag behind in this process due to later hiring waves, while older states of the south and south-
east have already been struggling with resultant pressure on public finances for a while. In
aggregate, half of state revenues are spent on wages, pensions, and other personnel costs.
Furthermore, personnel spending, including pensions, has increased faster than the stock of
government debt, especially after 2009. This means that some of the state borrowing in the
last decade was used to meet growing personnel spending obligations.
Federal policies have tried to limit the growth of subnational spending on personnel by shap-
ing the overall fiscal trajectory of local governments. The institution of CAPAG imposed lim-
its to subnational government indebtedness and allowed federal oversight of subnational
spending. The Fiscal Responsibility Law, enacted in 2000 in the aftermath of subnational
government bailout, also helped to better manage subnational public finances. It set a cap on
personnel spending - wages and pensions - at 60 percent of state revenue (Receita Corrente
Líquida, or RCL).
As a consequence of the vigorous economic growth between 2008 and 2014, procyclicali-
ty of CAPAG and Fiscal Responsibility Law allowed some flexibilization in borrowing limits,
some of which found its way into the further expansion of civil service wage bill. Brazil, as
a whole, has approached the 60 percent personnel spending limit in 2015-2017 period, led
by the Southern and Southeastern states, where pension bill alone already hovers around
20 percent of state revenues. Midwestern and Northern states, with still immature pension
schemes and free of constraints of Fiscal Responsibility Law, have seen significant wage
bill expansions in the 2000s, now spending 40 percent of state revenues on wages alone,
substantially higher than older states. This spells trouble ahead, once pension schemes of
these states inevitably mature and expose the flaws of Fiscal Responsibility Law. Reliable
municipality data is harder to collect, but in some ways their situation is even more alarming
for the long term.
The suboptimal fiscal adjustment mechanism, observed in 2015-2016, partially stems from
the current fiscal framework, which is mostly conceptualized in relation of current spending
to current revenues and does not encourage states to think ahead. It is also clear that blunt
macro-level policies aimed at controlling the growth of the wage and pension bill have been
9
shown ineffective, which requires more detailed examination of micro-level issues contrib-
uting to the rigidity and elevated growth rate of these expenditures.
Subnational Civil Servant Pension Schemes In Brazil
To summarize, continued maturation of pension schemes has been and will continue further
straining subnational finances at least until 2040 with varied severity across different re-
gions. While Fiscal Responsibility Law, implemented with the help of CAPAG index and other
tools, have had some positive effect, their procyclicality and focus on current expenditures
limit their effectiveness as tool to achieve long-term fiscal prudence. To help subnational
governments face future fiscal challenges, the Ministry of Economy recently launched sever-
al programs to support subnational governments in debt and/or fiscal distress, in exchange
for the commitment to fiscal reforms. Strong commitment to control personnel expenditures
is a major part of all these programs. While Covid-19 has appropriately taken government at-
tention away from these issues in 2020-2021, it also was extremely detrimental to long term
subnational finances and formation of human capital, which in turn reduce long term growth
potential. Therefore, continuation of the pension reform agenda has become even more im-
portant and urgent in the aftermath of the pandemic.
Effect of human resources policies on RPPS. Rapid personnel expenditure growth in recent
years has been primarily driven by pension scheme maturation, but it has also been strongly
influenced by micro-level human resource policies. Average hiring ages, retention of workers
with the right to retire, use of temporary workers, and the number of hours for which civil ser-
vant contracts are issued all contribute to the elevated ratio of retirees to workers. The overall
employment contract is also heavily backloaded with steep wage increases and generous
pensions. In addition, automatic promotions within numerous careers allow employees to
reach the top salary in a relatively short period.
In addition to flawed career design, other characteristics of the public employment system
also make it difficult to control personnel spending. Each of the 5,570 municipalities and
the 27 state governments have prerogatives to regulate their civil service. This gives rise to
a multitude of different rules and schemes, hard to oversee and regulate, which might en-
courage abuse. . A plethora of bonuses and variable remuneration payments further swell
the wage bill. Some of such payments are deemed “pensionable” and get entrenched in the
form of pension increases to be paid for many years into the future. The federal government
and some states have tried to curb these expenses, but oversight is even laxer at municipal
level. It is hard to see how more sensible, efficient, and fairer human resource policies can be
achieved without greater unification and simplification of civil servant employment rules, as
well as increased transparency and accountability for their implementation. It is also imper-
ative that human resource policies at all branches of government are informed by their effect
on pension spending, both short- and long-term.
Previous attempts at civil servant pension reform. Pension policy has long been a focus of
discussions geared towards limiting overall personnel expenditure growth. In sharp contrast
to human resource policies, which have been lightly regulated at the federal level, pension
scheme rules that apply to all RPPS schemes have been enshrined in the federal Constitu-
10
tion. Two of the original principles, integrality and parity, made those schemes especially
expensive and continue to apply to pensions in payment today. Integrality principle ensures
1 Executive Summary
that retirement earnings and survivor benefits correspond to the totality of the servant’s last
salary as an active worker. Parity means that retirees, and their survivors, have a right to all
salary increases, associated with the position from which the former civil servant retired.
There were multiple reform attempts to reduce the generosity of pension provisions. The first
pension reform that somewhat tightened pension eligibility requirements of RPPS took place
in 1998 through the Constitutional Amendment 20. Subsequently, the 2003 reform abolished
both integrality and parity principles, but only for civil servants entering service after 2003.
Crucially, the reform also completely excluded uniformed personnel. Although on the surface
the reform measures might appear mild, they had considerably decreased life-time pension
benefit generosity for post-2003 hires. As a consequence, the reform arrested the growth of
pension expenses for teachers and other civil servants hired post-2003. However, the sav-
ing-generating effects of the reform have been delayed for a few decades at the time of re-
form implementation, and the effects of pre-2003 pension rules were only expected to wain
around 2040. The long-term cost of exempting uniformed personnel from 2003 reform was
still expected to pressure the sub-national governments in the long term
The projected path of pension deficits for ten selected states follows a similar trajectory. On
average, state pension outlays are projected to rise around 50 percent, from current 15% of
the state revenues to 22% in the next 15 years. However, there is a wide dispersion across the
states. For example, pension deficit is expected to reach 35 percent of state revenues in Rio
Grande do Norte, while the state of Amazonas, which has put a lot of effort in fostering fiscal
sustainability of its pension scheme, is projected to peak at the pension deficit of 13 percent
of state revenues.
In parallel to politically difficult legal pension reforms, federal government has long encour-
aged the subnational governments to pre-fund their pension liabilities to prepare for the
expected increase in pension spending. This was commonly done through a policy called
“wage bill segregation” (segregação de massas), which diverted contributions of younger
workers into pension reserve fund. At first, the slow build-up of additional fiscal pressure due
to wage bill segregation was manageable, but as current deficits of the old pension funds
continued to grow, many subnational governments have decided to abandon this policy.
While some early adopters of the policy with still immature pension schemes, like the state
of Amazonas, have clearly been able to benefit, maintaining wage bill segregation might be
too costly for the states that start the process when their pension scheme maturation is al-
ready well advanced. Often, the large current costs of wage bill segregation policy effectively
prohibited the local governments from adopting the more fiscally desirable introduction of
complimentary individual savings accounts, which would have allowed them to reduce their
long-term pension liabilities by capping RPPS pension benefits.
11
Deficit projection of RPPS systems by percentage of current net revenues – selected states.
Subnational Civil Servant Pension Schemes In Brazil
40%
35%
30%
25%
20%
15%
10%
5%
0%
2020 2025 2030 2035 2040 2045 2050
RG do Norte Rio de Janeiro Bahia Santa Catarina Alagoas
Avarage
São Paulo Maranhão Paraná Mato Grosso Amazonas
Source: The World Bank projections using data provided by the state governments.
Recently, pension policy discussion has been somewhat unhelpfully focused on state trans-
fers to cover pension scheme deficits. Unfortunately, concentrating attention solely on defi-
cits allows celebration of “improvements” through mere accounting changes. One idea has
been to require local RPPSes to prepare long term asset and liability management plans,
Planos de Custeio, in order to reduce pension deficits and strengthen the fiscal position of
pension funds in the long term. While the long-term focus of this policy is to be celebrated,
it may have encouraged states and municipalities to transfer state assets to pension funds,
which does not create additional fiscal space in the consolidated public accounts. Increased
employer (government) contributions to deficitary pension funds suffer from the same flaw,
even though such an approach might help politically constrain inefficient state spending in
other areas. Meanwhile, the only true solutions to arrest crowding out of state fiscal space by
personnel costs amount to finding ways of reducing wage and pension expenditures without
jeopardizing provision of public services. This, in turn, involves reducing unfairly high life-
time compensation for public servants and improving service provision efficiency.
In retrospect, pension reforms of 1998 and 2003 have been able to arrest long term growth
of civil servant pension expenditures. However, by applying most reform provisions only to
the civil servants hired after 2003 and by exempting uniformed personnel altogether, that
stabilization of pension expenditure growth is only achieved by 2035-2040. The parallel ad-
ministrative reforms of wage bill segregation and narrowly focused efforts to limit pension
deficits, rather than consolidated government spending on pensions, have had some limited
successes but often distracted from the main focus of reducing pension expenditures in or-
der to open fiscal space for strategic investments and improved public services.
Pension reform of 2019. Relentless fiscal pressure build-up has led to the Federal Pen-
12
sion Reform of 2019. New federal mandates with respect to subnational pensions include
preparation of financial management plans (Planos de Custeio), unifying pension scheme
1 Executive Summary
administration of different government branches under single management unit, exclusion
of short-term benefits from the RPPS benefit rolls, mandatory implementation of compli-
mentary pension savings plans to civil servants that earn wages above mandatory insured
wage ceiling in order to limit RPPS benefit amount, imposing minimum individual pension
contribution rate of 14 percent, and implementing federally prescribed pension arrange-
ments for the uniformed personnel. The compliance with these mandatory requirements will
be treated as a necessary condition to obtain the Social Security Regularity Certificate, which
in itself is a requirement for receiving discretionary Federal Government transfers and per-
missions for credit operations with federal public banks.
The pension reform has also devolved some decision power on pension policy to subnational
governments, but Federal Constitution still retains the right to determine pension scheme
rules applicable to federal public servants, which was immediately exercised. In addition to
improving the fiscal situation of the federal civil servant pension scheme, the changes were
also intended to provide a template for the follow-up subnational pension reforms. These
reforms, including retirement age increase to ages 62 and 65 for females and males respec-
tively, revision of survivor benefit formula and eligibility, increase in retiree contribution base,
and allowance of extraordinary pension contributions, are optional for the subnational gov-
ernments. Local governments are, in principle, allowed to strengthen their parametric re-
forms beyond what was approved for the federal civil servant scheme. However, so far, most
of the first-mover states have opted to copy or weaken down federal civil servant reform,
prioritizing political expediency.
Although 25 states have already complied with the requirement to establish complementary
pension regime, 6 of them still lack the implementation. Also, very few states have consol-
idated the management of pensions from all their government branches into a single RPPS
management unit. Some implementation delays regarding this mandatory reform have been
due to the complications in integrating different IT solutions now used by separate branches
to assign and monitor benefit payments. However, even the states that are making steps
towards integration of pension sub-schemes tend to take a narrow view of the task, and rare-
ly attempt a more ambitious integration of some human resource and pension assignment
system functions, including pro-active benefit audits. Recent experiences from Alagoas and
Santa Catarina suggest that a pension record audit, greatly simplified by improved IT solu-
tions, could yield savings of 10 to 20 percent of pension expenditures.
The new pension rules for the uniformed personnel were also set at the federal level in 2019,
with subnational governments mandated to adopt the law in its entirety. This separate law
maintains more generous benefits for the uniformed personnel, including policemen and
firefighters, compared to other civil servants. Their pension benefits also preserve the gen-
erous principles of integrality and parity. Given federally regulated pension rules, one of the
few remaining tools for the local governments to influence subnational uniformed personnel
pension deficits is human resource policy.
13
From the optional reform features, expansion of contribution base for pensioners has the
largest potential to generate immediate reduction in pension deficits. However, only seven
Subnational Civil Servant Pension Schemes In Brazil
states in Brazil have so far lowered the exemption threshold to one minimum wage following
federal government’s example. The reluctance by states to adopt the expansion of contribu-
tory base for pensioners is regrettable. The 2003 reform has introduced big unfair differenc-
es in treatment between the pre-2003 and post-2003 entrants into the public service. The
pensions of pre-2003 entrants are heavily subsidized by the state and the contribution base
extension to them would only claw back a small portion of this subsidy. Another reform with
potentially substantial savings is the adoption of proportional calculation of survivor ben-
efits. The remaining reforms, including an increase in the retirement age and the flexibility
to levy extraordinary contributions for a limited time period are also important, but not as
relevant for the short-term fiscal impact.
Overall, the take up of the reform tools provided to the states through the federal 2019 pen-
sion reform has been underwhelming, possibly due to the short-term improvement to state
finances in 2020 and the overwhelming urgency of COVID-19 response. It is likely that munic-
ipal governments, with lower implementation capacity, are even further behind in the reform
process. Consolidation of record management and proactive audits of pension records of all
government branches is a priority, as is wide expansion of pensioner contribution base and
revision of survivor benefits. However, even with the extra efforts along these lines, it is clear
that the 2019 reform measures are insufficient to arrest the strong growth of pension expen-
ditures projected for the next two decades.
Reform lessons and recommendations. Going forward, lasting solutions would be easier
to find if fiscal, human resource, governance, asset management, and pension policies are
viewed as integral parts of interdependent system. Local governments that feel having ex-
hausted the legal and political pension reform space provided by the federal 2019 reform
should focus on human resource policy, which still offers ample space for increased effi-
ciency and positive spillovers into reduced pension expenses. To address expected growth
in pension expenditures in the next two decades, pension policy itself would benefit from
reexamining the inequities left by exemptions from 2003-reform. Some ways to devise partial
claw-backs of unjustified subsidies to pre-2003 cohorts are highly progressive contribution
rates, extraordinary additional contributions, and some lump sum inducements to voluntarily
accept post-2003 benefit rules.
Throughout the system, there is an urgent need for more transparency, better governance,
and policy-maker friendly IT solutions. Covid-19 has also exposed a need and convenience
of being able to serve clients and manage records remotely. It is neither reasonable nor effi-
cient for all 2154 RPPS schemes to acquire or develop their own IT systems and set up their
own asset management departments. Therefore, IT and asset management services should
be made available, sponsored at the federal or at least state level, that could be used by
smaller RRPS schemes. The requirement for all public employers to report individual level
employee data into a federal eSocial database from April 2022 (already mandatory for private
employers) offers unique opportunity to systematize data in a common format. It is important
14
to ensure that subnational RPPS also have legal authority to access this data for all govern-
ment branches in order to fully perform their record management and audit duties. However,
1 Executive Summary
subnational governments cannot passively wait for that to happen. Temporary solutions are
urgently needed, which could include cloud-based service rentals from private IT compa-
nies, or own-developed software packages that could be shared between cooperating RPPSs
schemes. Sharing of asset management services is already happening and should be further
encouraged with appropriate safeguards.
Introduction
2
2. Introduction
Brazil prides itself for having a comprehensive social protection system, comprising of a
wide array of labor market policies, extensive social assistance net, and a relatively high
pension coverage. However, its social protection spending is heavily skewed towards older
people, especially those who have served as civil servants during their career. Private sector
workers are covered by the Regime Geral de Previdência Social (RGPS) and public civil ser-
vants belong to the Regime Próprio de Previdência Social (RPPS). Unusually, in Brazil, apart
from the federal RPPS regime, each subnational entity – states, the Federal District and mu-
nicipalities – was allowed, until 2019, to create its own individual RPPSs, which resulted in
the existence of over two thousand different RPPSs at the subnational level.
In total, currently there are 2,154 RPPSs in Brazil, as 27 states and 2,127 municipalities run
their own pension schemes1. Together, they account for an insured population of about 8.9
million people (active civil servants, retirees and survivors) at the three levels of government –
federal, states including Federal District, and municipalities. About 58 percent are active civil
servants and the remaining 42 percent are pensioners or survivors (Table 1). While the total
membership of RPPS only stands at about 10 percent of RGPS coverage, its spending amounts
to almost half of RGPS pension outlays.
Considering Brazil’s 5,597 subnational entities (27 states and 5,570 municipalities), 38 per-
cent have their own RPPS. The rest cover their civil servants through the national RGPS pen-
sion system. Out of existing RPPS regimes, 21 are in the process of extinction, also to be re-
placed by RGPS arrangements.2 Several states have large numbers of RPPSs, with many small
municipalities running their own schemes: Rio Grande do Sul has 332 regimes, followed by
Minas Gerais (222), São Paulo (221), Paraná (179), and Goiás (170). A few states with very small
territory and population have disproportionately large number of RPPSs, such as Pernambuco
(149), Alagoas (74) and Paraiba (71). RPPS are in general small in terms of assets under man-
agement: 65 percent hold assets between 1 million and 5 million BRL, and 7 percent of RPPS
1 Source: Indicador de Situação Previdenciária – Seprev (2020), accessed on May 2021. Available at https://
www.gov.br/previdencia/pt-br/assuntos/previdencia-no-servico-publico/indicador-de-situacao-previdenciaria/
arquivos/2020/indicador-de-situacao-previdenciaria-isp-2020-v1-03-10-2020-10h30.pdf..
2 Painel Estatístico da Previdência, accessed on June 17, 2021.
18
manage assets below 1million BRL. Only 1.6% of RPPSs manage more than 1billion BRL.3
2 Introduction
This paper attempts to place civil servant pension schemes in a larger picture and present
its importance for the country’s fiscal standing. It starts by providing a historical context in
which subnational civil service pension schemes were set up, discusses fiscal consequenc-
es born by the maturation of these plans, and considers influences of human resource and
fiscal policies on pension scheme evolution. It further provides historical overview of pen-
sion policy thinking applied to subnational pension schemes to date and surveys the take-up
of 2019 federal pension reform template by subnational governments until April 2021. The
paper attempts to demonstrate the interdependence of fiscal, human resource, governance,
and pension policies and argues for the search of joint solutions in the challenging quest to
arrest subnational pension expenditure growth in Brazil.
Table 1 – Size of the population covered by civil servant pension regimes in Brazil in 2019.
RETIREES AND SPENDING,
GOVERNMENT LEVEL ACTIVE TOTAL
SURVIVORS % OF GDP, 2018
Federal government 688,778 740,997 1,429,775
States & Federal District 2,014,773 2,064,150 4,078,923
Municipalities 2,521,955 925,559 3,447,514
Total RPPS 5,225,506 3,730,706 8,956,212 4.1%
RGPS 58,156,477 29,089,1604 87,245,637 8.6%
Source: SRPPS/SPREV/ME - CADPREV, accessed on July 2019; Boletim Estatistico da Previêência Social.
3 Sample of 1756 systems which have provided timely records of applications and investments of resources
(DAIR- Demonstrativo das Aplicações e Investimentos dos Recursos) retrieved in November 2019 (reference peri-
od March 2019). Source: Seprev/ME.
4 Number of pension and survivor benefits paid in December 2019, including accident insurance.
Evolution of civil
service in Brazil
since 1988
3
3. Evolution of
civil service
in Brazil since 1988
The 1988 Constitution is a cornerstone of current civil service in Brazil. Prior to its adoption,
the rules and mandates governing civil service were not clear and were not consistent across
agencies. Older and larger states have expanded their civil service during early 1980’s, but it was
mostly limited to increasing the capacity of administrative sectors and inspector careers. Only
a handful of institutions responsible for key functions such as taxation, diplomacy, and public
advocacy had already started to professionalize. The rest of the public administration, including
lower levels of government, did not have established permanent corps of public employees.
The 1988 Constitution has also established the universal access to public services, which lead
to increased demand and a rise in the number of public officials, especially at the munici-
pal level. For example, until 1988, the State was not held responsible for providing universal
services such as education, public health, and social assistance.5 However, after the end of
the military dictatorship in 1985, the country was experiencing an unprecedented widespread
sense of optimism, accompanied by high promises of prosperity and expansion of individual
rights. The push for increased access to services, and placement of these major responsibil-
ities on subnational governments, has led to extensive hiring. In addition to teachers, health
professionals, public service providers, and policemen, the creation of numerous new posi-
tions in the judiciary was noticeable, given the momentum that the Federal Constitution had
provided by expanding civil rights, and triggering increases in individual litigation.
Between the 1988 and 2019, the number of civil servants grew by 211 percent at the munic-
5 The 1988 Federal Constitution explicitly shared the responsibility of service delivery with states and munici-
palities. Secondary education, for instance, should be provided mainly by states, pre-primary education delivered
by municipalities, while primarily education should be a shared concern between them (Art.211). Health is also
shared responsibility. Social assistance is planned at the federal level, while implementation is carried out at the
municipal level by the Centros de Referência de Assistência Social – CRAS.
22
ipal and 30 percent at the state level (Figure 1).6 More than a thousand new municipalities were
created between 1984 and 1997, which drove extensive hiring observed in the following decade.
3 Evolution of civil service in Brazil since 1988
The number of public officials in the federal government had remained remarkably stable over this
period. As a result, the profile of Brazilian public service had changed profoundly. If in 1988 federal
civil servants constituted 16.2 percent of all public service (Figure 2), their share has almost halved
to only 8.6 percent by 2019 (Figure 3). On the other side of the spectrum, municipal civil service has
grown from 36 percent to 59.7 percent of all public servants over the same period. Overall, govern-
ment employees of all levels currently constitute 18 percent of all wage employees in Brazil.
Figure 1 – Number of public employees by government level.
7
6
5
4
Millions
3
2
1
0
1988
1989
1991
1993
1990
1992
1994
1995
1996
1997
1998
1999
2001
2002
2003
2004
2005
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2000
2006
2018
2019
Federal States Municipalities
Figure 2 – Division of public employees by Figure 3 – Division of public employees by
government level (1988). government level (2019).
36,0% 16,2 % 31,6% 8,6%
1988 2019
47,8% 59,7 %
Federal States Municipalities
Source: Atlas do Estado Brasileiro (IPEA, 2020).
6 With the 1988 Federal Constitution, municipalities were elevated to the status of federative entities. This en-
couraged the division of large municipalities. Between 1984 and 1997, 1,405 municipalities were created - an
increase of 34,3% compared to the previous period (Gomes and McDowell, 2000). In 1984, Brazil had 4,102 munici-
palities; currently there are 5,570. In 1996, stricter rules for the creation of municipalities were imposed.
23
While the overall growth of state employment was much lower than that of municipalities over the
1988 to 2019 period, there was substantial variation between the states. With the 1988 Constitution,
Subnational Civil Servant Pension Schemes In Brazil
three new states (Amapá, Roraima, and Tocantins) were created in the North of the country, leading
to the restructuring of the public service and substantially contributing to the 116 percent growth
in the number of state public servants in the region (Figure 4). In addition, the North of the country
along with the Midwest (87 percent state civil service growth), were the territories with the highest
population expansion in the last few decades which, not surprisingly, lead to the commensurate
growth in the number of state civil servants. On the other hand, the Southern, Southeastern, and
Northeastern states, with more stable populations and more established institutions, had a smaller
increase in their state public service. The Southeastern region specifically, which still employs 39
percent of all state civil servants in Brazil (Figure 5,) has seen a very limited growth in hiring over the
same period and has even been reducing the number of employees since 2014. Thus, compared to
1988, the overall state civil servant employment in the Southeast only grew by 9 percent.
Figure 4 – Regional growth of state public employees (index 1998 = 100).
260
220
180
140
100
60
1988
1989
1991
1992
1993
1994
1995
1996
1997
1998
1999
2001
2000
2002
2003
2004
2005
2007
2006
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
1990
Midwest North North East South Southeast
Figure 5 – Division of state public employees by region (2019).
13% 13%
Midwest
14% North
Nothe East
39% South
Southeast
21%
Source: Atlas do Estado Brasileiro (IPEA, 2020).
24
The 1988 Constitution has also brought major changes to the way civil servants were contracted.
Although the transition rules ensured the retention of some political appointees, Brazil quickly
3 Evolution of civil service in Brazil since 1988
became the regional reference on public sector meritocracy7, as shown in Figure 6. One of the
newly established pillars of the meritocratic system was the hiring process which, in Brazil, is
based on competitive exams, widely publicized and open to all citizens. In addition to this merito-
cratic access to employment, civil servants were also granted substantial legal protections and
autonomy. The dismissal of a civil servant could no longer be made based on subjective criteria.
The constitution defined very narrow grounds for dismissal and, after 3 years of employment,
positions were considered permanent. The possibility of losing the job due to poor performance,
for example, was only included in 1998, and has been tightly regulated since.
Figure 6 – Civil service meritocracy (2004, 2012-15).
100
2004
2012 - 2015
80
60
40
20
0
BRA
CRI
URY
COL
CHL
ECU
PER
MEX
NIC
PRY
SLV
DOM
GTM
PAN
BOL
HND
LAC
Source: A decade of civil service reforms in Latin America (2004–13) (IADB, 2014)
Since 1988, the overwhelming majority of Brazilian public employees were employed under
the privileged civil servant contract guaranteeing permanent employment (Figure 7). The
rules governing public sector regime contracts are described in civil servant statutes, reg-
ulated by local law. In addition to the Federal government, each of the 27 states and 5570
municipalities have their own civil servant statute. However, since 2002, there was some
increase in the use of temporary workforce which by now constitutes around 8 percent of the
public sector employment (Figure 8). This is largely driven by hires to provide basic educa-
tion, but each government level has its own legislation regulating temporary positions. At the
same time, the number of government contracts regulated by the private sector employment
regime (Consolidação das Leis Trabalhistas - CLT) has been decreasing and now constitutes
only 5 percent of total public employment. This is due to the decrease in the role of state-
owned enterprises, which predominantly use this contracting model.
7 The meritocracy index, developed by Longo (2006) and calculated by the IADB, is a composite index that ranges
from 0 to 100 in evaluating a set of ten objective characteristics of a country’s hiring and promoting system. As
critical points, it includes an assessment of whether hiring to public vacancies are open and widely publicized,
whether there are safeguards against arbitrariness or clientelism in hiring, and if terminations of employment in
the public sector are not motivated by political leanings.
25
Figure 7 – Evolution of public employees by type of contract (general government).
Subnational Civil Servant Pension Schemes In Brazil
14
12
10
8
Millions
6
4
2
0
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Employment Regime Public Regime Temporary
Figure 8 – Division of public employees by type of contract (general government).
8%
87 % 5%
Employment Regime
Public Regime
Temporary
Source: Atlas do Estado Brasileiro (IPEA, 2020).
Hence, civil servants under the temporary contract (regulated by Law 8745/1993, see art. 8)
and those under the labor regime are enrolled in the RGPS system (Law 6647/1993), while
those under the public sector regime are enrolled in the RPPS if there is such a regime at
their local level employment. If a municipality does not run its own RPPS, its civil servants are
enrolled in the RGPS. Although temporary civil servants do not enjoy the same rights, often,
through legislative or bargaining process, they are converted to permanent staff retroactively.
As well as providing a specific employment model for permanent government staff, the 1988
Constitution also defined an exclusive pension system for them. This system was further de-
tailed in the 1990 federal law of Regime Jurídico Único (RJU). The schemes that federal gov-
ernment, as well as the states and municipalities, were allowed to create under this law are
collectively known as the Regimes Próprios de Previdência Social (RPPS), or “own regimes
of social security”, as opposed to the “general regime of social security” (RGPS) that covers
private sector workers and some public sector workers that are not enrolled in the RPPS.
26
In effect, these new pension arrangements made civil servant career truly permanent as, in
contrast to the private sector employees, civil servants now remained affiliated with their life-
3 Evolution of civil service in Brazil since 1988
time employer even after their retirement. This strong affiliation is apparent from the term “in-
active civil servants” commonly used to describe public sector retirees. It could even be argued
that life-time affiliation concept also helped to create an implicit expectation that retirement
should not imply any significant change in the civil servant’s income stream, even if they stop
discharging their civil servant duties. Therefore, many civil servant statutes foresee continu-
ation of benefits associated with the last job into retirement, including promotion in rank for
the uniformed personnel, increases in pension that mirror wage increases for the last retiree’s
position (continued to be called “salary” of a retiree), and monetary benefits associated with
the discharging of previous duties, for example a bonus for teaching in a remote school.
For the first few decades, creation of these new generous pension regimes had very limited
effect on subnational governments. Most of the civil servants hired in the first years of public
service expansion were young, and therefore were not eligible to retire for at least a couple
of decades. Subnational government finances were arranged accordingly, as budget was
fully allocated between service provision, investment, and payment of wages for active civil
servants. Pension payments were fully covered by pension contributions of a large young
employee pool and did not need budget support.
However, the beginning of the XXI century marked the start of a rapid maturation of civil ser-
vant pension schemes, as the first large cohorts of civil servants started to retire. Given that
the increased hiring period in the 80s lasted for about a decade, and average pension draw
period due to early retirement ages tends to last about 30 years, the full maturation of most
subnational systems is not expected before 2040. Reaching full maturation of the pension
schemes of younger states and municipalities might take even longer, as their hiring wave
dates in the 1990s and even 2000s.
The examples of Pernambuco and Sergipe states demonstrate how hiring spikes can stress
state finances and service provision (Figure 9). Both states had large hiring intakes in 1980s,
with last workers hired in that period about to retire. Pernambuco seems to have had to re-
place most of the retirees from the first hiring wave by new increased hiring between 2005
and 2018, taking on additional cost of wages, while most of the retirees from the first wave
were still drawing pensions. In contrast, Sergipe struggled to replace the retiring servants,
inevitably jeopardizing provision of public services. The absence of replacement hiring, like
in Sergipe, can also be seen in the age distribution of active and retired teachers in Espírito
Santo and Ceará, as shown in the Figure 10 in the right-hand graphs. Mato Grosso and Rio
Grande do Norte seem to have managed to stay closer to Pernambuco path, and have man-
aged to retain a more age diverse teaching workforce.
Even though the pension scheme maturation process is not yet complete, in general, the
average state’s pension system currently is almost as mature as that of the federal gov-
ernment. Figures 11 and 12 demonstrate that, on average, the number of retirees is already
exceeding the number of active workers at both of these government levels. Meanwhile, the
27
municipalities, due to the more recent expansion of their public service, still have a much
lower retiree to active worker ratio (the obvious exceptions are big municipalities of older
urban centers like Rio de Janeiro and Sao Paulo).
Subnational Civil Servant Pension Schemes In Brazil
Even though state and federal governments seem to have similar retiree to active employee
ratios of 1.07 and 1.14 respectively, the variability between the states in considerable (Figure
13). The proportion of retirees to active workers in the northern states is very close to the
municipal average of 0.4. Note that the RO (created in 1981) and AP, RR and TO (1988) are
the youngest Brazilian states, elevated from the status of federal territory to state during the
1980’s, what thus explains a much younger civil service labor force and fewer retirees. On the
other hand, eight states in the older South, Southeast, and Northeast regions that have seen
a much more modest recent civil service expansion have a higher proportion of dependents
on the system than that of the federal government itself. The states of Rio Grande do Sul and
Minas Gerais stand out most, with almost 1.7 retirees and pensioners for each active profes-
sional, in comparison to 1.14 ratio for the Federal Government scheme.
Figure 9 – Hiring spikes of public servants during the 80’s – Pernambuco and Sergipe states.
Pernambuco
20000
16000
12000
8000
4000
Number of civil servants
0
1971
1995
2019
1959
1965
1983
2007
1989
1953
1977
2001
2013
1956
1962
1968
1980
1992
2004
2010
2016
1950
1986
1998
1974
Sergipe
12000
8000
4000
0
1971
1995
2019
1959
1965
1983
2007
1989
1953
1977
2001
2013
1956
1968
1980
2004
2016
1950
1962
1986
1992
1998
2010
1974
Teachers Active Uniformed Personnel Active Others Active
Teachers Retired Uniformed Personnel Retired Others Retired
Source: The World Bank, based on state government data.
28
Figure 10 – Varied success in maintaining age-diverse teaching staff under pension cost pressure in
3 Evolution of civil service in Brazil since 1988
various states.
High-hiring policy Low-hiring policy
High groth rate of wages
1200
Minimum of 13% adjustments in
Mato Grosso Ceará
2020, by federal legislation
800
400
0
21
29
33
53
61
69
85
89
93
22
32
37
62
82
25
37
41
45
57
65
73
77
27
42
47
52
67
72
92
49
81
57
77
87
97
number of people
idade idade
1200
Low groth rate of wages
Santa Catarina Espírito Santo
800
400
0
24
34
39
64
84
22
32
37
62
82
29
44
49
54
69
74
94
27
52
72
92
59
79
89
99
42
47
57
67
77
87
97
idade idade
Contributors male Contributors female Pensioners male Pensioners female
Source: The World Bank, utilizing data provided by the state governments
Figure 11 – Proportion of retirees and Figure 12 – Total number of retirees and
pensioners to active employees by pensioners as well as active employees by
government level. government level.
1.4 3.5
1.2 3.0
pensioners and survivors/active
1.0 2.5
0.8 2.0
employees
millions
0.6 1.5
0.4 1.0
0.2 0.5
0.0 0.0
Federal States Municipalities Federal States Municipalities
Active employees Retirees and pensioners
(total beneficiaries)
Source: Anuário Estatístico de Previdência Social do RPPS (ME, 2020).
29
Figure 13 – Proportion of retirees and pensioners to active employees in State governments.
Subnational Civil Servant Pension Schemes In Brazil
2,0
1,69
1,63
1,48
1,46
1,42
1,6
1,41
1,37
1,33
1,10
1,07
1,05
1,2
0,96
0,95
0,95
1,01
0,86
0,85
0,82
0,78
0,76
0,76
0,61
0,8
0,56
0,38
0,26
0,4
0,06
0,05
0
GO
RS
MG
RJ
ES
PB
BA
RN
SC
CE
SE
PI
SP
PE
AL
PR
DF
MA
MT
AC
PA
TO
RO
AP
RR
MS
AM
Midwest North North East South Southeast
Source: Anuário Estatístico de Previdência Social do RPPS (ME, 2020).
Further disaggregation of the maturity8 level per worker category within a state reveals that
imbalances are usually more prominent in specific careers. Amazonas offers a typical ex-
ample of a pension system with diversified maturity levels across categories, where teachers
belong to a most mature member distribution (Figure 14). Other categories of civil servants,
and especially military, reveal still immature age distributions, as they do in most states, with
larger concentration of members below the average pension age. Relative immaturity of uni-
formed personnel schemes is especially common across Brazil. Given that their young age
structure and associated increases in future pension liabilities is often not well documented
and known to the executive branch government, and that these sub-schemes do not yet gen-
erate big pension expenditures, the building up of this unfunded fiscal liability is not yet fully
recognized by most subnational governments.
8 Maturity level of a pension system refers to the relative sizes of retiree and contributor cohorts. The scheme
is considered mature when its oldest pensioner cohorts, adjusted by survival probability, are similar in size to the
average contributor cohort.
30
Figure 14 – Different maturity level of pension scheme membership by career - Amazonas.
3 Evolution of civil service in Brazil since 1988
Capitalized Fund Financial Fund
1200
Teachers
800
400
0
1200
Uniformed Personnel
800
Individuals
400
0
1200
Other civil servants
800
400
0
24
28
32
36
40
44
48
56
60
64
68
72
76
80
84
88
20
52
92
96
idade
Contributors male Pensioners male Contributors female Pensioners female
Source: The World Bank using data provided by the state government.
The fiscal
consequences of civil
service expansion
4 and pension scheme
maturation
4. The fiscal
consequences
of civil service
expansion and
pension scheme
maturation
The expansion of subnational civil service and maturation of pensions schemes that fol-
lowed a few decades later have had a strong impact on subnational government finances. In
aggregate, half of state revenues are spent on wages, pensions, and other personnel costs
(Figure 15). Generally, state level personnel spending, wage and pension spending included,
increased just below the growth rate of state revenue over the last two decades (Figure 16).
However, personnel spending has increased faster than the stock of government debt, espe-
cially after 2009. This suggests that some of state borrowing in the last decade was used to
meet growing personnel spending obligations.
In 2015, the Brazilian economy went into the crisis, with GDP falling 3.5 percent in 2015 and
3.6 percent in 2016. Tax revenues have decreased, for the first time in recent history. The
governments needed to adjust their public accounts, but it soon became clear that personnel
spending could not meaningfully adjust due to its rigidity. This was because an overwhelm-
ing majority of employment contracts were permanent, and important seniority-related wage
increases were largely automatic, governed by complicated and legally insulated civil servant
statutes. Pension system also lacked any discretionary adjustment mechanisms apart from
pension indexation, which was legally linked to the growth of senior position salaries from
which former civil servants have retired. Therefore, the only remaining personnel cost ad-
justment tools were reduction of the base wage growth and hiring freeze.
34
Figure 15 – Aggregate state revenues and spending, 2020, in BRL billion
4 The fiscal consequences of civil service expansion and pension scheme maturation
1000
900
800
700
600
500
400
300
200
100
0
1995
1996
1997
1998
1999
2001
2000
2002
2003
2004
2005
2007
2006
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Wage Bill Pension Other Personnel Outsourcing
Other Current Expenditures Investments Total Revenues
No Outsourcing
Source: The World Bank using data from Plano de Ajuste Fiscal (PAF), National Treasury Secretariat
Figure 16 – Comparison between the growth rates of aggregate state revenue, debt, and personnel
expenditures in Brazil.
280
(index: 1996=100)
260
240
220
200
180
160
140
120
100
1996
1997
1998
1999
2001
2002
2003
2004
2005
2007
2008
2009
2010
2011
2012
2013
2014
2015
2000
2006
2016
2017
2018
2019
Debt Wage Bill Pension Bill State Revenue (RCL)
Source: The World Bank using data from Plano de Ajuste Fiscal (PAF), National Treasury Secretariat.
Since all wage structures were partially determined by base wage, and pensions in turn de-
pended on salary growth of senior employees, it was assumed that the freezing, or at least
reduction, of base wage will translate into large wage and pension bill savings. However,
overall wage growth component attributable to base wage increase turned out to be rather
small. In addition, base wage was in itself not fully insulated from union influences and civil
servant statute rules. For example, in some states, the statute stipulates that base wage
35
growth cannot be lower than last year’s inflation, and unions often sign multi-year pacts with
the government, fixing future base wage increases a few years into the future. In nominal
Subnational Civil Servant Pension Schemes In Brazil
terms the salaries still grew, powered by automatic promotions and related automatic salary
increases. Therefore, overall effect of the reform on wage growth was muted (Figure 17).
Figure 17 – Growth of payroll, average salary and active employees, state public sector (2019).
198
200
Until 2014 – payroll: 6.4%; salary: 5.6%; employees: 0.8%
2014 to 2016 – payroll: -5.4%; salary: -2.8%; employees: -2.8% 186 187
2016 to 2018 – payroll: 1.2%; salary: 3.5%; employees: -2.1% 182
181 179
178
180 176
170
177
177
171
162
166
160 155
154
141 147
140
139
130 135
127
120
120 116
121
111 109
115 116 109
107 110 105
111 103
100 108 104
105 106 105
100 104
2003
2004
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2005
2017
Payroll (real values) Average salary Active employees
Source: Gestão de pessoas e folha de pagamentos no setor público brasileiro (Word Bank, 2019).
Hiring freezes could not provide a significant dent in personnel expenditures either, as wages
of newly hired still low-paid employees constitute a very small part of personnel budget. At
the first glance, hiring freezes were somewhat successful, as the number of active servants
did decrease somewhat through attrition. However, given that the reduction of employees
over this period is almost exclusively driven by retirements, the overall pension and wage bill
could not have changed by much due to this policy. Overwhelming majority of 2014 and 2015
retirees were hired before the pension reform of 2003 and have been retiring with pensions
equal to 100 percent of their last salary.
In the absence of the ability to meaningfully cut personnel spending, cuts in investment, fol-
lowed by some reduction in other current expenditure, mainly health and education, took the
brunt of the adjustment (Figure 18). While outsourcing, a rather new hiring modality, could be
expected to absorb some reduction in revenues in 2015-16, it did not do so (see Figure 15). This
demonstrates that outsourced services in Brazil also contain some real or perceived rigidities,
stemming from legal and/or political processes. Outsourcing has been growing in popularity
since 2000 and has been replacing some of the public servant hiring, mainly in health and edu-
36
cation, with great promise of increased flexibility. Instead, in a crisis, subnational governments
primarily adjusted by cutting investments, jeopardizing economic growth and future tax reve-
4 The fiscal consequences of civil service expansion and pension scheme maturation
nue. Given how little Brazilian states already spend on investment, a deeper or more protracted
crisis could not have continued to be handled this way. Thus, spending on public services such
as education and health, which is generally indexed to revenues, was next in line for expendi-
ture cuts, partially as an automatic consequence of lower revenues.
Looking back, federal policies have historically played a big role in shaping subnational gov-
ernment fiscal trajectory at a macro level. The period between the signing of the new Federal
Constitution of 1988 and 1995 was marked by the centralization of tax revenues in the federal
government and a lack of mechanisms to constrain the fiscal behavior of subnational gov-
ernments. Big part of state expenditure in that period was financed by explosive increase in
external and internal debts.
Figure 18 – Cumulative growth of aggregate state spending on personnel, investment, and other
current expenditure in Brazil.
360 334
(index: 1995=100) 326 330 326
314 318
303
310 288
274 309
260 244 248
225 232 235 231
214
202
210 187 221 220
213
172
151 156 189
160 146 140 179
134
113 122 127 161 165
118
100 108
110
116 120
112 116 119
108
100 106
60
1997
1995
1998
1999
2001
1996
2000
2002
2003
2007
2004
2005
2006
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Personnel spending Other current expenditure Investments
Source: The World Bank using data from Plano de Ajuste Fiscal (PAF), National Treasury Secretariat.
The institution of CAPAG, in 1997, introduced by the Ministry of Finance through Senate Res-
olution 78/1998, imposed limits to subnational government indebtedness, and allowed fed-
eral oversight of subnational spending, effectively allowing spending to increase at the pace
of increases in state revenues. The analysis of the payment capacity (CAPAG) ascertains the
fiscal situation of subnational entities that aim at taking out new loans with the guarantee
of the federal government. CAPAG unveils whether a new indebtedness represents a credit
risk for the National Treasury, based on three indicators: indebtedness, current savings, and
liquidity ratio. Hence, by assessing the degree of solvency, the relationship between current
income and expenditure and the cash position, a diagnosis of the fiscal health of the state or
37
the municipality is made. Increased percentage of net current revenue spent on personnel
expenses worsens both the liquidity indicator and the current savings indicator of a state,
Subnational Civil Servant Pension Schemes In Brazil
leading to lower CAPAG grades and, consequently, difficulties in obtaining loans. In addition,
the debt refinancing associated with a fiscal adjustment plan and a credible rule to avoid de-
fault in debt payment (direct withdrawal from the debtor’s account) seem to have also been
decisive in controlling subnational indebtedness.
The Fiscal Responsibility Law, enacted in 2000 in the aftermath of subnational government bail-
out, constituted another ‘milestone in the management of public finances in Brazil’ (Mendes
2020). It set a cap on personnel spending at 60 percent of state revenue (Receita Corrente Líqui-
da, or RCL). If total expenditure on personnel exceeds 95 percent of that limit, the law prohibits
any human resource action that implies an increase in expenditure, such as hiring or career pro-
motions, providing a counterweight to some provisions of civil servant statutes. The law also
established unprecedented provisions forbidding inter-government financial support, including
from state banks, which was an instrument widely used by states to finance fiscal debts.
Under the new law, elevated personnel spending also threatened the state’s eligibility to re-
ceive discretionary federal transfers, as these were suspended in case the subnational entity
fails to comply with minimum investments in health and education. The Brazilian Constitu-
tion determines that states, the Federal District, and municipalities must invest at least 25
percent of the revenue resulting from taxes (including transfers received from the federal and
state governments) in maintaining and developing education and qualifying education pro-
fessionals, as well as spending a minimum of 15 percent in public health. Failure to comply
with these regulations might ultimately even result in the intervention by the Federal Union.
As a consequence of the vigorous economic growth between 2008 and 2014, CAPAG and
Fiscal Responsibility Law allowed some flexibilization in borrowing limits, some of which
found its way into the expansion of civil service wage bill. Moreover, there was also a gradual
relaxation of rules and creation of exceptions for contracting debt through both Senate reso-
lutions and National Treasury regulations. While state governments have not experienced an
explosive growth in the number of civil servants comparable to that of municipalities, their
payroll still grew substantially, almost doubling in real terms in the decade from 2003 to 2014.
In the same period, the average salary increased by 82 percent in real terms and the number
of personnel grew by 9 percent.
The regional effects of Fiscal Responsibility Law and 2008-2014 credit expansion can be
seen in Figure 19. Brazil, as a whole, has approached the 60 percent personnel spending limit
in 2015-2017 period, led by the Southern and Southeastern states, where pension bill alone
already hovers around 20 percent of state revenues. However, Fiscal Responsibility Law has
clearly had an effect in this region, keeping its wage bill to around 35 percent of state reve-
nues over the last 15 years. Northeastern states have also felt maturation of their pension
schemes with progressively increasing share of revenues spent on pensions, which, through
fiscal pressure, has contributed to the stability on wage bill spending. However, Midwestern
and Northern states, with still immature pension schemes and free of constraints of Fiscal
38
Responsibility Law, have seen significant wage bill expansions in the 2000s, now spending
40 percent of state revenues on wages alone, substantially higher than older states. This
4 The fiscal consequences of civil service expansion and pension scheme maturation
spells trouble ahead, once pension schemes of these states inevitably mature and exposes
the flaws of Fiscal Responsibility Law.
To overcome immediate delicate fiscal situation born by 2015-2016 crisis, and to avoid break-
ing any fiscal responsibility rules, the harder hit states also adopted some temporary solutions.
They included accumulation of payment arrears and engagement in debt renegotiations with the
federal government. However, rolling of debts is nothing else than a postponement of the fiscal
problem in the absence of reforming the underlying structural fiscal drains. The economic crisis
of 2015-2016 showed the necessity to improve the processes of managing and financing respon-
sible subnational spending in general and pension scheme spending in particular.
Reliable municipality data is harder to collect, but in some ways their situation is even more
alarming for the long term. There is no reason to believe that explosive doubling of salaries,
seen at the state level over the 2003-2014 period, would not be replicated at the municipal
level. This, combined with the doubling of municipal labor force between 1988 and 2003, and
additional 50 percent growth after that, spells an explosive recent growth of the wage bill at the
municipal level. Most municipalities, especially younger ones, may have handled the crisis of
2014-2015 a little bit easier than the states, given that they were not yet burdened by high pen-
sion payments at the time and therefore were not strongly restrained by Fiscal Responsibility
Law. However, about half of current active municipal employees are hired before 2003 and will
retire with extremely generous pension benefits, rapidly swelling pension expenditures in the
next couple of decades. The municipalities will need much more flexibility than the current civil
servant statutes and pension rules allow to be able to navigate this fiscally challenging period.
39
Figure 19 – Wage and pension spending as proportion of state revenues by region.
Subnational Civil Servant Pension Schemes In Brazil
60
Brazil North
50
40
30
20
10
0
1995
1996
2004
2008
2012
1995
1996
2004
2012
1998
2000
2006
2016
1998
2000
2006
2008
2016
2002
2010
2014
2018
2002
2010
2014
2018
1997
2005
2001
2009
2013
1997
2005
2007
2017
2001
2009
2013
1999
2007
2017
2003
2011
2015
2019
1999
2003
2011
2015
2019
60
Northeast Midwest
50
40
30
20
10
0
1995
1996
2004
2012
1995
1996
2004
2008
2008
2012
1998
2000
2006
2016
1998
2000
2006
2016
2002
2010
2014
2002
2010
2018
2014
1997
2005
2009
2013
1997
2005
2018
2001
2007
2017
2001
2009
2013
1999
2003
2011
2015
2007
2017
2019
1999
2003
2011
2015
2019
60
South Southeast
50
40
30
20
10
0
1995
1996
2004
2012
1995
1996
2004
2008
2016
2008
2012
1998
2000
2006
1998
2000
2006
2016
2002
2010
2014
2010
2005
2018
2002
2014
2018
1997
2001
2009
2013
1997
2005
2001
2009
2013
1999
2007
2003
2011
2015
2017
2007
2017
2019
1999
2003
2011
2015
2019
Pensions (%RCL) Wage bill (%RCL)
Source: The World Bank using data from Plano de Ajuste Fiscal (PAF), National Treasury Secretariat.
The suboptimal fiscal adjustment mechanism, observed in 2015-2016, partially stems from
current fiscal framework, which is mostly conceptualized in relation of current spending to
current revenues and does not encourage states to think ahead. When spending increase deci-
sions affecting personnel spending are made, the impacts of these decisions are generally not
assessed for the long term. For example, Fiscal Responsibility Law only demands three-year
long projections. Thus, low current pension expenditures are taken as an indicator that further
hiring and wage increases are fiscally responsible as long as increased personnel costs can
40
be paid for the next three years. Unfortunately, highly backloaded civil servant contract, with
high seniority-related wage increases and generous pensions, translates current human re-
4 The fiscal consequences of civil service expansion and pension scheme maturation
source policy decisions into large rigid fiscal liabilities in future years and decades. Younger
subnational entities with low pension bills can already see the future by observing older states,
struggling to keep their fiscal health and flexibility under the high pension cost burden.
By 2019, state revenues were starting to increase, and fiscal authorities could afford to turn
their attention from immediate crisis management to ensuring longer term fiscal sustainabil-
ity. The procyclical nature of fiscal rules, which force fiscal adjustment during recessions but
permit fiscal flexibilization during periods of growth, also allowed governments to start looking
ahead. Federal pension reform of 2019 was an important expression of that shift in focus. It
gave the states an important policy tool to adjust their long term pension expenditures, and
a few pioneering states have done so in late 2019 and early 2000. However, in a few months,
COVID-19 crisis has forced federal and subnational governments back into the crisis mode.
Looming COVID-19 health and economic crisis has shocked Brazilian authorities into quick
action. Federal government has rolled out a well-funded emergency cash transfer program
Auxílio Emergencial, which helped to maintain economic activity during the crisis and con-
tributed to the tax revenue increases. The subnational governments have also received direct
federal transfers to combat the COVID-19 pandemic in the amount of R$ 97 billion, or 1.4
percent of GDP. Finally, the second half of 2020 have also seen increasing commodity prices,
which significantly contributed to improved economic outcomes and state revenue in com-
modity producing states. In all, paradoxically, subnational governments have experienced a
fiscally strong 2020. As result, a somewhat backward looking CAPAG indicator is projected to
reach A or B level for 20 states in 2021, compared with only 10 states in 2020.
However, positive outcomes in 2020 do not change the fact that the subnational fiscal out-
look is continuing to deteriorate since before the pandemic. Prior to the COVID-19 crisis,
more than half of the 27 Brazilian States were in a critical situation and have delayed pay-
ments to public servants and providers. In 2020, public investment was approximately 15%
lower in comparison to 2015, even after 15% increase in 2020 is factored in. The emergency
cash transfer program has already been significantly curtailed in 2021, and it is unlikely that
direct federal emergency transfers will continue. Therefore, fiscal pressures, experienced be-
fore the pandemic, are likely to return to the forefront of subnational government agenda.
To help subnational governments to face the fiscal challenges, the Ministry of Economy recently
launched several programs to support states in debt and/or fiscal distress, in exchange for the
commitment to fiscal reforms: (i) Fiscal Recovery Program (Regime de Recuperacao Fiscal ), to
support the states in serious fiscal distress; (ii) Fiscal Sustainability Program (Plano de Equilib-
rio Fiscal ), which aims to help subnational governments in fiscal distress to recover CAPAG B
rating; and (iii) Expenditure Management Program (Pro-Gestao), which aims to contribute to the
improvement of the fiscal, budgetary and patrimonial management of the Brazilian public admin-
istration, increasing the efficiency and effectiveness of the expenditure. Strong commitment to
control personnel expenditures is a major part of all three of these programs.
41
To summarize, growing personnel expenditures are unbalancing the fiscal framework and
creating a vicious cycle of fiscal unsustainability and constrained economic growth in the
Subnational Civil Servant Pension Schemes In Brazil
Brazilian subnational entities, which are forced to reallocate spending from other sectors.
The subnational entity is left with a very narrow fiscal space for investments and a limited
decision-making power over the remaining budget, which eventually leads to non-compli-
ance with the Fiscal Responsibility Law the constitutional minimum spending requirements,
or low CAPAG score. It makes it nearly impossible for states and municipalities in such situ-
ation to invest in credit operations, hire new civil servants, or improve infrastructure. Finding
ways to reduce personnel costs could allow them to refocus and redirect resources where
they are really needed for growth and improved public service delivery. During the 2015-2016
crisis episode, blunt macro-level policies aimed at controlling the growth of the wage and
pension bill have been shown ineffective, which requires more detailed examination of mi-
cro-level issues contributing to the rigidity and elevated growth rate of these expenditures.
Contribution of HR
policies to growth in
wage and pension
5 spending
5. Contribution
of HR policies to
growth in wage
and pension
spending
While rapid personnel expenditure growth in recent years has been primarily driven by pen-
sion scheme maturation after the hiring waves of 80s and 90s, it has also been strongly
influenced by micro-level human resource policies. Apart from the obvious influence of
these policies on the growth of wage bill, these decisions are also key in determining pen-
sion spending trajectory. For example, average hiring ages, retention of workers with right to
retire, use of temporary workers, and the number of hours for which civil servant contracts
are issued all contribute to the elevated ratio of retirees to workers. Similarly, remuneration
policies suffer from excessive fragmentation which hinders equal treatment and oversight,
inviting abuse and complicating control of wage bill growth. Furthermore, while the appro-
priateness of average salary level of civil servants can be debated, the overall employment
contract is heavily backloaded with steep wage increases and generous pensions, clearly
making it by far a safer and financially superior option to private sector employment in many
cases. This hinders job mobility between private and public sectors, harming economic
growth. Big variation in remuneration among different government levels and branches adds
to the complicated picture in need of reform.
Current rules of public hiring favor a late start of civil servant career. Given meritocratic approach
to hiring and high competition for public posts that offer exceptional job security, the process
tends to require a lot of effort from job applicants. Preparation for public examinations can take
years, as can gathering of all the credentials that might help secure the coveted position. Figure
20 for Mato Grosso shows a typical situation, where on average people are hired in their early 30s.
44
The high proportion of retired RPPS members in subnational pension schemes can also in part be
explained by low retirement ages. Even though, until recently, civil servants had to make 35/30 years
5 Contribution of HR policies to growth in wage and pension spending
of contributions for men/women respectively before reaching retirement age eligibility, some of those
years could be accumulated in private sector employment before joining public service. The require-
ment for teachers is 5 years shorter, advancing retirement eligibility even more. Uniformed personnel
have even more generous requirement reductions. Finally, while most states allow postponed retire-
ment in return to suspended pension contributions, given the very generous pensions and limited fi-
nancial inducement to prolong careers, many civil servants tend to retire at the first opportunity and
continue to enjoy their pre-retirement income. For example, average retirement age in Mato Grosso, as
in most states, has recently hovered around age 55 (Figure 20). Combined with a rather late average
hiring age of 30, early retirement implies that the length of public servant working life is similar to the
length of his or hers pension receipt, while pension receipt can continue even after pensioner’s death
in a form of survivor benefit. Simple math suggests that a scheme that requires a member to pay con-
tribution rate of 11 to 14 percent of wage for 25 years and then promises to pay back 100 percent of
retiree’s last wage for an even longer period is bound to run into financial trouble once it fully matures.
If each permanent civil servant position generates a new pension recipient every 25 or so years,
it would be logical to hire some personnel under the general labor contract, which until recently
offered much less generous pension benefits. Such approach would be particularly sensible
when needed skills are not government-specific, and where the need for services might be
temporary. However, the percentage of temporary staff in the Brazilian public administration is
low in terms of international perspective. While the federal government’s use of temporary staff
is close to that of countries with consolidated public civil services (Figure 21), it stands out in
comparison to Argentina and Chile, both of which heavily depend on temporary staff.
Figure 20 – Age structure of Mato Grosso civil servants and retirees.
14%
Avarege hiring age Avarege retiring age life expectancy
12%
Proportion of Mato Grosso population
10% expected
maturation
8%
6%
4%
2%
0%
19 24 29 34 39 44 49 54 59 64 69 74 79 84 89
Contributors Temporaries Pensioners
Source: The World Bank using data provided by the state government.
45
Figure 21 – Percentage of temporary workforce in central governments (2019).
Subnational Civil Servant Pension Schemes In Brazil
100%
80% 72%
60%
45%
40%
20%
20% 14% 15%
12%
9%
2%
0%
UK USA Brazil Portugal Canada France Argentina Chile
Source: Atlas do Estado Brasileiro (IPEA, 2020).
Some studies9 have pointed out that the use of state and municipal temporary contracts
has been hampered by legal uncertainty. Subnational governments have been exposed to
lawsuits by temporary workers seeking the recognition of rights such as maternity leave,
stability of employment for pregnant women, and the 13th salary, among others. In some sit-
uations, lawsuits have even resulted in the judicial interruption of the use of such contracts.
As a result, it is still unclear what is the set of minimum labor rights of these employees. Po-
litical pressures by organized labor is another hurdle for subnational governments seeking
greater flexibility of labor contracts. For example, Mato Grosso has been increasingly relying
on temporary teacher contracts, accumulating 20 million of temporary teachers by 2018.
However, the government had to agree to a mass conversion of 5 million of these contracts to
permanent positions from 2020, retroactively recognizing their past service as temporaries
for the purposes of salary determination and civil servant pension eligibility.
Mato Grosso also brings another interesting example of how human resource policies can be
used to limit pension spending in the medium run, without the changes to pension policy.
Through the Complementary Law 662/2020, the state has extinguished more than 4,000 civil
servant positions reducing the current wage bill as well as future pension liabilities. The policy
has built in the strategy of rearranging the structure of service provision through permanent
civil servants already in office, compensating for the dispensing of additional functions by a
gratification smaller than the full wage paid to the former positions. By canceling existing ca-
reers and hiring new civil servants only into newly created careers, wage adjustments given to
active personnel do not apply to current pension beneficiaries retired from the old roles.
Senior management positions are dominated by political appointees. In general, any leadership
position across government agencies can be filled through a political appointment, despite the
9 The study of Brazilian Public Law Society is the main example of temporary work legal analyses. Available on:
https://movimentopessoasafrente.org.br/nossos-temas/matriz-de-vinculos-e-seguranca-juridica/
46
provision that a portion of these posts need be occupied by career public servants. As a result,
Brazil does not have a professional senior management system like, for example, the United
5 Contribution of HR policies to growth in wage and pension spending
Kingdom. This has resulted in a high management position turnover in Brazil to the order of
30 percent per year, elevated further during political transitions (Lopez & Silva, 2020). This
practice also substantially elevates pension expenses, as highly paid pre-retirement positions
generate a new pensioner with all the position-specific additional benefits every few years. The
have also been known practices of senior level pacts, where with the understanding that this
is done with the expectation of prompt retirement, so that the lucrative pre-retirement position
is bequeathed to the next member of the pact. Since 2019, there have been some initiatives to
change this reality. On a small scale, the federal government and some states have initiated
competency-based selection processes for senior management positions.10
Another human resource policy greatly affecting pension expenses is related to part-time
contracts. Among teachers and, to the lesser extent, health professionals, partial employ-
ment is often a reality of service provision, for example in a one-shift school. It is customary
in such cases to issue a 20-hour per week contract, rather than a 40-hour one. However,
sometimes such hiring can be rigid. For example, if a school has a 25-hour per week shifts,
but only 20-hour and 40-hour contracts are available, the school would often issue a 40-
hour contract, which would later convert into 40-hour based pension calculation. The system
would also sometimes encourage gaming, where a 20-hour contract is converted into 40-
hour contract a few years prior to retirement, so that 40-hour based pension is generated.
It is also not uncommon that some pensioners receive multiple pensions, all from the same
or from different subnational governments, where all of the pensions have originated from
long-time partial contracts.
Although the admission to the public service is generally meritocratic, there is room for im-
provement. Partly due to the fear that any flexibility will make political capture possible, the
selection process in the Brazilian public sector can be extremely bureaucratic. The main hir-
ing method are objective tests which only allow the assessment of hard skills (Fontainha et
al., 2014). Processes such as interviews, curriculum analysis, and practical tests are scarcely
conducted. Thus, behavioral and soft skills competences are not considered. In addition,
the selections happen at an unstable frequency, even during periods of economic growth,
emphasizing the lack of strategic workforce planning. Workforce planning is generally input
and process-driven, and largely focused on responding to new current needs rather than
strategic and long-term planning. Civil servants are hired in one job category and must pass
competitive public examinations to change into another career stream. Many job categories,
even for similar functions, are ministry-specific, and sectoral pressures have resulted in dif-
ferentiated work conditions across different bodies even for similar functions.
Turning to remuneration policies, there is a big civil servant wage variation between Brazilian
government levels and branches, including a substantial public sector wage premium at the
federal government and the judiciary (Figure 22). Overall, the average remuneration at the
10 This issue is being discussed in the National Congress.
47
federal level is almost twice that of the states and more than three times that of the munici-
palities (Figure 23). As a result, the salary premium paid to the federal government is almost
Subnational Civil Servant Pension Schemes In Brazil
100% higher than that of the private sector (Figure 24). The state level public sector wage
premium is close to 40% and there is no premium for municipal public sector employees. The
average premium for the entire Brazilian public administration is approximately 20%, close to
the average premium observed internationally.
Figure 22 – Average salary by government level and branch (2019).
R$ 18,000
R$ 15,274
R$ 15,000
R$ 12,000 R$ 10,196
R$ 9,439 R$ 9,298
R$ 9,000 R$ 7,685
R$ 6,000 R$ 4,811
R$ 4,239
R$ 2,970
R$ 3,000
R$ 0
Federal
Judiciary
States
Judiciary
Federal
Executive
Federal
Legislative
States
States
Executive
Legislative
Municipalities
Legislative
Municipalities
Executive
Source: Atlas do Estado Brasileiro (IPEA, 2020).
Figure 23 – Average salary by government level (2019).
R$ 12,000 R$ 10,233
R$ 9,000
R$ 6,000 R$ 5,240
R$ 3,002
R$ 3,000
R$ 0
Federal
States
Municipalities
Source: Atlas do Estado Brasileiro (IPEA, 2020).
The civil service is organized in a career system with high fragmentation of salary rules. In
the public sector, in general, each career has its own pay scale and particular criteria for sal-
ary progression. For example, it is estimated that the federal government has approximately
300 remuneration schemes. Although there have been some attempts to simplify, there are
wage differences even between careers with similar job descriptions, e.g. in the case of ana-
lyst careers shown in Figure 25. This lack of standardization favors an unequal remuneration
policy, in which some specific groups are privileged.11
11 For example, in the last two decades careers close to the government center and the security area are more
benefited more than the others.
48
Figure 24 – Public sector wage in relation to the private sector by country.
5 Contribution of HR policies to growth in wage and pension spending
120%
BR A– Fed
100%
Prêmio salarial do setor público (com controles)
80%
BWA
COL
PAK
60% SLV
CRI
UGA HND ECU
TGO PHL GTM
TZA BR A– Sta
40%
GHA MEX
RWA MDV DEU
MLI BOL PRY PAN GRC
CAN
20% URY
ETH ITA IRL
GMB PER BR AZIL CHL
BGD CHN GBR
KAZ
0% USA
DOM EST
TJK VNM MNG BRA – Mun FIN
RUS
EGY
-20%
7 8 9 10 11
Log do PIB per capita, 2015
Latin America and Midle East
Brazil Sub-Saharan Africa
Caribbean North Africa
Latin America and
South Asia North America East Asia and Pacific
Caribbean
Source: Worldwide Bureaucracy Indicators (Word Bank, 2019).
Figure 25 – Salary range of different analyst Figure 26 – Average service time until the
careers in federal government (2019). last career salary level is reached (2019).
25,000 35
20,000
30
15,000
10,000 25
5,000
Number of Years
20
0
Analista Adm-Agências
Analista Adm-PREMC
Analista Técnico-Adm-PEC Suframa
Analista Adm-Meio Ambiente
Analista Adm-DNIT
Analista Adm-DNPM
Analista Adm-INCRA
Analista Técnico-Adm-PGPE-Ceplac
Analista Técnico-Adm-PGPE-Inmet
Analista Técnico-Adm-PEC Cult
Analista Técnico-Adm-PEC Faz
Analista Técnico-Adm-PGPE-SPU
Analista Técnico-Adm-PGPE
15
10
5
0
Careers
Source: Nova Administração Pública - Diagnostico Source: Gestão de pessoas e folha de pagamentos
Força de Trabalho Governo Federal (ME, 2019) no setor público brasileiro (Word Bank, 2019)
49
In addition, automatic promotions within numerous careers allow employees to reach the top
salary in a relatively short period, with two thirds of careers in federal civil service topping
Subnational Civil Servant Pension Schemes In Brazil
out in less than 20 years (Figure 26). States and municipalities generally follow federal pub-
lic administration rules, but have also introduced many different salary tables and specific
progression rules for each career. While progression policies vary across states, there are
also important commonalities. Workers are typically not only issued automatic bonuses ev-
ery three years (trienniums,) but can also submit various continuing education certifications
which can move them into a higher salary bracket, not necessarily with a commensurate in-
crease in responsibilities. Uniformed personnel commonly experience especially steep wage
growth with seniority (Figure 27), which in many states culminates in one last automatic
promotion in rank on the day of retirement, so that the pension benefit equals to that of an
immediately superior position. These automatic seniority-related wage increases make it
difficult to control the wage and pension bill, as replacements with less experienced and less
expensive workers are short-lived. Replacements by young better educated workers might
even increase costs, as often happens in education sector, which started hiring teachers with
higher credentials and correspondingly higher wages in many states. In those cases, higher
pensions should be expected in the future.
Figure 27 – Recent real wage growth experience in three states with longest time series.
12%
9%
8%
5%
4%
3% 3%
3%
2%
Total Teachers Military
Amazonas Rio Grande do Sul Tocantins
(2010 - 2019) (2011 - 2019) (2013 - 2018)
Source: The World Bank using data provided by the state governments.
Due to remuneration policies described above, pensions in Brazil often are higher than wages,
as is illustrated with Alagoas data in Figure 28. The graph on the left of the panel shows that av-
erage retired teacher incomes at ages 70 to 80 are almost twice as high as those of presumably
already experienced teachers at age 50, which is likely an outcome of part-time contracts con-
verted into full-time working arrangements just before retirement. Other civil servants, shown
in the middle of the graph, achieve high pension to average wage ratio by means of especially
steep wage progression due to seniority. Finally, uniformed personnel enjoy both a steep career
progression and pre-retirement boost in income due to pre-retirement increase in rank.
50
Figure 28 – Average wage and pension income by civil servant age cohort and category of workers –
5 Contribution of HR policies to growth in wage and pension spending
teachers (TCH), Other civil servants (CS) and uniformed personnel (MIL) in Alagoas.
Avg. Wage and Pension – TCH Avg. Wage and Pension – CS Avg. Wage and Pension – MIL
15000
10000
5000
0
20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100
Average Wage Average Pensions
Source: The World Bank using data provided by the state government.
In addition to the problems with career design, other characteristics of the public employment
system also make it difficult to control personnel spending. Each of the 5,570 municipalities
and the 27 state governments have prerogatives to regulate their civil service. This gives rise to
the multitude of different rules and schemes, generally more favorable than the private sector,
hard to oversee and regulate, which might encourage abuse. Some examples include poorly
accounted leave permissions and holidays, which are often easy to claim second time in the
form of monetary payments. A plethora of bonuses and variable remuneration payments fur-
ther swell the wage bill. Some of such payments are deemed “pensionable” and get entrenched
as increased pensions to be paid for many years into the future. The federal government and
some states have tried to curb these expenses, but oversight is even laxer at municipal level. It
is hard to see how more sensible, efficient, and fairer human resource policies can be achieved
without greater unification and simplification of civil servant employment rules, as well as in-
creased transparency and accountability for their implementation.
Previous pension
policy response to
pension spending
6 increases
6. Previous pension
policy response
to pension
spending increases
Rapidly increasing civil servant pension costs at subnational level have long made pension
policy a focus of reform discussions geared towards limiting overall personnel expenditure
growth. In sharp contrast to human resource policies, which have been lightly regulated at
the federal level and, therefore, hard to reform at scale, pension scheme rules that apply to all
RPPS schemes have been enshrined in the federal Constitution from the date of its signing
in 1988. Two of the original principles, integrality and parity, made those schemes especially
expensive. Integrality principle ensured that retirement earnings and survivor benefits cor-
respond to the totality of the servant’s last salary as an active worker, often with bonuses
associated with the particularities of the last job, for example a bonus for working in a remote
school. Parity meant that retirees, and their survivors, had a right to all salary increases, as-
sociated with the position from which they retired.
The combination of these two original principles results in extremely generous pensions still
in payment today, regardless of numerous pension reform efforts. This could be illustrated
by the ratio of retiree benefits to salaries, which is particularly elevated for some careers and
states, and is especially prominent among the uniformed personnel (Table 2). However, there
is also remarkable diversity in outcomes, underscoring how different human resource policies
for various careers in different states can influence overall pension expenses under essentially
the same federally prescribed pension rules. For example, the overall pension to salary ratio
stands at 67 percent in Ceara and Pernambuco but rises to 125 percent in Mato Grosso do Sul.
Diversity of outcomes is also in part due to the fact that each branch of government at each
government level commonly run parallel semi-independent pension schemes, often only
using executive branch pension scheme as a payment agency. For example, an internal de-
partment of the judiciary or legislative branch of government would have sole access to em-
54
ployee data, decide when an employee is eligible to retire, would calculate his or her pension,
and would only ask the executive branch RPPS administrator to pay the stated amount to a
6 Previous pension policy response to pension spending increases
certain person from a certain date. This sometimes results in lack of uniformity of benefit
assignment rules, when some government branches would include aggregation of certain
career-related special benefits into the pension amount, while others would not. Since the
data on contributors would not typically be shared with the executive branch, it would not be
possible to assess full pension liabilities of the state or audit pension assignment processes
and records of other government branches.
Table 2 – Ratio of average benefits and average wages in select states and careers.
UNIFORMED OTHERS TOTAL OTHERS NON-
STATE TOTAL TEACHERS
PERSONNEL EXECUTIVE EXECUTIVE EXECUTIVE
CE 0.67 1.32 0.54 0.68 0.72 0.96
ES 0.84 1.65 0.9 0.96 0.92 1.32
MS 1.25 1.49 1.06 1.45 1.3 1.51
PB 0.67 1.09 1.14 0.72 0.84 -
PR 0.99 1.27 0.99 0.97 1 1.33
RN 1.04 1.35 1.31 1.05 1.13 1.12
RR 1.21 0.75 0.75 1.94 1.3 -
SC 1.08 1.49 1.22 1.12 1.19 -
Note: The ratios are calculated for different groups of workers and states for which data was available. Source:
The World Bank using data from SPREV/Ministry of Economy.
Even though the RPPS pension schemes since their inception were managed separately
at each level of government, most of the civil servant retirement rules, including benefit
eligibility, amount, and indexation, were initially specified in the Federal Constitution. The
entities directly sponsoring specific RPPSs were only allowed, within limits, to set member
contribution rates and set some rules for managing the scheme assets. The first pension
reform affecting the RPPS took place in 1998 through the Constitutional Amendment 20. The
amendment somewhat tightened benefit eligibility conditions, but generous benefit formula
was maintained in the Constitutional text. The amendment also determined that eligibility
for special pension and survivor benefits would be regulated by ordinary federal level laws,
however these did not materialize. However mild, this reform opened the way for further civil
servant pension scheme revisions in 2003.
The 2003 reform, through Constitutional Amendment 41, abolished both integrality and par-
ity principles for civil servants. However, the new rules only applied to civil servants that en-
tered public employment after the year 2003. Crucially, the reform also completely excluded
uniformed personnel. The major changes in pension rules were 1) to institute the minimum
retirement age of still very generous 55/60 for women and men respectively, with 5 year re-
55
duction for teachers, 2) to apply the 100 percent replacement rate not to the last salary, but
to the average of 80 percent of the highest annual salaries earned over a person’s career, in-
Subnational Civil Servant Pension Schemes In Brazil
dexed to past inflation, 3) switch to inflation indexation of benefits, 4) to levy pension contri-
bution of 11 percent at a minimum on wealthier pensioners who receive benefits that exceed
the national insured salary income ceiling threshold, and 5) to allow the establishment of a
said insured salary ceiling for civil servants, as long as individual pension savings accounts
(Fundos Complementares) were offered to civil servants with earnings above that ceiling.
Although on the surface the reform measures might appear mild, they had considerably de-
creased life-time pension benefit generosity for post-2003 hires. Figure 29 illustrates this ef-
fect for a typical civil servant, showing average wage growth indexed life-time salary income
in dark blue, 11 percent individual pension contribution plus 22 percent employer pension
contribution in a black line, and expected life-time pension benefits in red and light blue for
pre- and –post 2003 hires respectively. While for pre-2003 hires life-time pension contribu-
tions under the black line in dark blue were clearly significantly lower than life-time retire-
ment benefits, these two areas for post-2003 hires are more comparable in size.
The gains and drawbacks of 2003 reform can be seen in pension expenditure and revenue pro-
jections for Mato Grosso generated prior to the latest pension reform of 2019 (Figure 30). The
projections separate pre- and post-2003 civil servant pension expenditures, clearly showing
that the reform did arrest the growth of these expenses for teachers and other civil servants
hired post-2003. However, it is also clear from the graph that the saving-generating effects of
the reform have been delayed for a few decades at the time of reform implementation, and that
the effects of pre-2003 pension rules will only start to wain around 2040. The long-term cost of
exempting uniformed personnel from 2003 reform is also plainly visible from the graph.
Figure 29 – Pension benefits of a typical civil servant before and after the 2003 reform.
Pre-2003 Post-2003
5000
4500
4000
3500
salary/pension
3000
2500
2000
1500
1000
500
0
25 30 35 40 45 50 55 60 65 70 75 80 25 30 35 40 45 50 55 60 65 70 75 80
Age
salary pre-2003 pension 33% contribution post-2003 pension
Source: The World Bank.
56
The projected path of pension deficits for other states follows a similarly shaped trajectory (Figure
31). On average, increase in state pension outlays is projected to rise around 50 percent, from current
6 Previous pension policy response to pension spending increases
15% of the state revenues to 22% in the next 15 years. However, there is a wide dispersion across the
state. For example, pension deficit is expected to reach 35 percent of state revenues in Rio Grande do
Norte, while the state of Amazonas, which has put a lot of effort in fostering fiscal sustainability of its
pension scheme, is projected to peak at the pension deficit of 13 percent of state revenues.
In parallel to politically difficult legal pension reforms, federal government has long encouraged the
subnational governments to pre-fund their pension liabilities to prepare for the expected increase
in pension spending. The hope also was that, by diverting part of government revenues and assets
into RPPS asset build-up, the state will be too fiscally constrained to continue with the pace of prof-
ligate hiring and wage increases, which would also indirectly limit pension liability growth. Mechan-
ically, this policy of “wage bill segregation” (segregação de massas) worked as follows: new hires
would migrate to a new capitalized pension system (Fundo Previdênciario), where, due to the initial
absence of older members, contribution revenues would exceed expenditures for a few decades,
allowing for pension asset build-up to guarantee the future benefits. As the bulk of pensioners re-
mained enrolled in the old system (Fundo Financeiro) of the pay-as-you-go type, without a steady
influx of new contributors, it was inevitable that, at some point, contributions will no longer be able
to cover the benefit payments. The ensuing deficits will then have to be covered by the state, fiscally
constraining its expansionary policies. The additional decades-long fiscal effort required from the
state to implement this policy of RPPS asset build-up is commonly called “transition cost” and is
considered fully paid only after the death of the last member of Fundo Financeiro.
Figure 30 – Projection of pension expenditures and revenues, in terms of state revenue (RCL), for
the state of Mato Grosso, based on 2017 data.
40%
35%
30%
25%
20%
15%
10%
5%
0%
2022
2024
2026
2028
2030
2038
2021
2023
2025
2027
2032
2034
2035
2036
2040
2041
2042
2043
2044
2046
2047
2048
2050
2052
2053
2055
2058
2059
2060
2029
2031
2033
2037
2039
2045
2049
2051
2054
2056
2057
Uniformed Personnel Other Pre-2003 Servants Other Post-2003 Servants
Pre-2003 Teachers Post-2003 Teachers Contribution Revenues
Source: The World Bank using data provided by the state government.
57
Figure 31 – Deficit projection of RPPS systems by percentage of current net revenues – selected states.
Subnational Civil Servant Pension Schemes In Brazil
40%
35%
30%
25%
20%
15%
10%
5%
0%
2020 2025 2030 2035 2040 2045 2050
RG do Norte Rio de Janeiro Bahia Santa Catarina Alagoas
Avarage
São Paulo Maranhão Paraná Mato Grosso Amazonas
Source: The World Bank projections using data provided by the state governments.
At first, the slow build-up of additional fiscal pressure due to wage bill segregation was manage-
able, but as deficits of Fundo Financeiro continued to grow, many subnational governments have
decided to abandon this policy. They have reassigned the resources accumulated in Fundo Prev-
idenciario for the payment of Fundo Financeiro benefits, which would now cover all civil servants.
It usually took just a few years to completely use-up earlier accumulated pension assets, which
provided a short-term relief for the state budget. Recently, the decision to abandon asset accu-
mulation in Fundo Previdenciario has been made by Minas Gerais (in 2013), Rio Grande do Norte
(2014), Santa Catarina (2015), Distrito Federal (2017), and Sergipe (2017). Goiás has extinguished
its Fundo Previdenciario in 2017 and has recreated it in 2020, with slower pace of accumulation.
The total assets accumulated in Fundos Previdenciarios by 2017 are shown in Figure 32.
Some states, like Amazonas, have decided to slow asset accumulation in Fundos Previden-
ciarios based on actuarial assessments, which deemed those funds to be overfunded. The
complicated process of shifting some soon-to-be paid liabilities from Fundo Financerio to
Fundo Previdenciario was termed “buying of lives”, or Compra de Vidas. However, the actuarial
assessments used to justify this policy are notoriously dependent on a few key assumptions,
including the ability of the state to reduce future wage growth of civil servants, and main-
tain its historically high profitability of investments, which have benefited enormously from
the increases of government bond prices in recent past. Just as importantly, Compra de Vidas
was motivated by extreme budget pressures felt by the states that attempted to maintain their
adherence to wage bill segregation policy in the fiscally constrained post-crisis environment.
State of Amazonas remains one of the earliest and most dedicated adherents to the wage
bill segregation policy, since year 2003. Still low level of pension scheme maturation in 2003
undoubtedly was one of the important factors that allowed the state to sustain this policy,
58
as pension financing obligations were still relatively low at the time, and transition costs as-
sociated with wage bill segregation were not as difficult to accommodate compared to older
6 Previous pension policy response to pension spending increases
states. Having maintained pre-funding discipline for 18 years, Amazonas is close to starting
to reap the rewards. As retirement flow starts shifting from Fundo Financeiro to Fundo Prev-
idenciario in a few years, the pension pressure on the state budget will start to decrease. As
shown in Figure 31, the peak of the deficits is projected for the year 2032.
Figure 32 – Asset composition of state level pension funds in Brazil – year 2017.
8000
7000
6000
5000
In BRL millions
4000
3000
2000
1000
0
SP
AC
GO
PB
RN
SC
RS
MG
PR
SE
BA
PI
AL
CE
AP
MS
PE
MT
RJ
RO
RR
ES
AM
TO
PA
DF
MA
Other goods and rights (not including “other goods and Investments Cash and cash equivalents
rights” for RJ)
Source: Relatório Resumido da Execução Orçamentária (RREO).
However, maintaining wage bill segregation might be too costly, especially for the states that start
the process when their pension scheme maturation is already well advanced. For example, the 2003
pension reform legislation stipulates that the establishment of complimentary individual savings ac-
counts (Fundo Complementar) allows the state to impose the ceiling on civil servant wages insured
through Previdenciario or Financeiro funds (because the wage above that ceiling could be insured
through Fundo Complementar), which later leads to capped RPPS benefits. Similarly to wage bill
segregation policy, establishment of Fundo Complementar also involves transition costs, as some
contributions of higher paid civil servants are diverted from Fundo Financeiro or Previdenciario to
complimentary individual accounts. However, crucially, this policy is combined with the cutting of fu-
ture pension liabilities to those same higher income civil servants. Unfortunately, high transition costs
associated with maintaining wage bill segregation policy have previously precluded a few states, in-
cluding Amazonas, from taking on additional costs of also transitioning to individual pension savings
accounts, at the expense of forgoing an opportunity to reduce their pension liabilities.
The states that have adopted individual pension savings accounts under Complementary
Pension Funds usually establish them for newly appointed employees, but a few have also
59
made it possible for old employees to migrate to the complementary fund. Federal civil ser-
vant regime has been one of the entities to offer such switch to employees hired before the
Subnational Civil Servant Pension Schemes In Brazil
establishment of the complimentary fund. In order to entice the workers to switch, it has
offered a Special Benefit, to be paid in the form of additional pension, to compensate switch-
ing employees for the contributions made before the switch. The Special Benefit has been
calculated by prorating generous pre-2003 pension promises to the proportion of contribu-
tions paid before the switch. While this still allowed to reduce accumulation of new pension
liabilities towards the switchers, liabilities accumulated to date were fully honored. This has
established a precedent for the other subnational governments. However, the state of Goias
is now preparing a proposal for the Special Benefit based on post-2003 promises, which
would be offered to all switching employees, including those hired before 2003. Depending
on the take-up rate of this offer, the new approach might allow Goias to cut some of its pen-
sion liabilities. Another incentive for the older employees to switch to the new arrangements
is the possibility of the difference in contribution rates, with non-switchers required to pay
higher extraordinary contributions to finance the growing deficits of Fundo Financeiro.
The 2019 federal reform of the pension system has included a mandatory implementation
of a complementary regime for all subnational entities, as very few have adopted them fol-
lowing the 2003 reform. The 2019 Subnational Entities Finance Bulletin from the Nacional
Treasury Secretariat shows the adoption of wage bill segregation as well as complimentary
funds by each state in Brazil (Table 3).
Table 3 – Efforts at RPPS pre-funding: Years of institution and abandonment of wage bill segregation,
and year of creation of the complementary pension regime.
EXTINCTION OF MASS COMPLEMENTARY
STATE MASS SEGREGATION
SEGREGATION REGIME*
AC No - Law authorized
AL 2009 No Yes
AM 2001 No Yes
AP 2005 No Law authorized
BA 2007 No 2017
CE 2013 No Yes
DF 2008 2017 2019
ES 2004 No 2013
GO 2013 (recriated in 2020) No 2017
MA No - Law authorized
MG 2002 2013 2015
MS 2012 No 2020
MT No - Yes
60
PA 2016 No 2016
6 Previous pension policy response to pension spending increases
PB 2012 No Law authorized
PE 2013 No- Law authorized
PI 2012 No 2020
PR 2012 No Law authorized
RJ 2012 No 2013
RN 2005 2014 Law authorized
RO 2012 No 2018
RR Yes No No
RS Yes No 2016
SC 2008 2015 2017
SE No 2017 2019
SP No - 2013
TO - - No
Source: 2019 Subnational Entities Finance Bulletin - National Treasury Secretariat. Note: *Authorized: the
regime was authorized by local law, but has yet to be implemented.
Note: Year of creation of complementary regime not available for all applicable states.
Policy discussion on containing state pension costs has recently been somewhat unhelpful-
ly focused on state transfers to cover pension scheme deficits, called Aportes (see Table 4).
These payments amounted to a record BRL 111,6 billion in 2019 and are paid in addition to the
resources already allocated to pension schemes in the form of employer contributions and
state asset transfers. Between 2016 and 2019, aportes grew at the annualized rate of 11 percent,
compared to an average annual inflation rate of 4.3 percent. However, concentrating attention
solely on aportes, or pension deficits, allows celebration of “improvements” through mere ac-
counting changes as is suggested by several pension financing options shown in Figure 33.
Table 4 – Social security cost: unpaid expenses with resources linked to Social Security.
COST OF THE SOCIAL SECURITY SYSTEM TO THE STATE TREASURY (IN MILLIONS)
STATE 2016 2017 2018 2019
AC R$ 287.57 R$ 408.93 R$ 423.00 R$ 616.67
AL R$ 1,140.07 R$ 1,285.24 R$ 1,347.44 R$ 1,442.86
AM R$ 998.82 R$ 1,154.49 R$ 1,252.99 R$ 1,435.36
AP R$ 34.39 R$ 13.64 R$ 0.82 R$ 0.38
BA R$ 2,536.86 R$ 3,223.94 R$ 3,656.52 R$ 4,239.64
CE R$ 1,448.96 R$ 1,576.49 R$ 1,463.65 R$ 1,530.20
DF R$ 1,211.60 R$ 564.00 R$ 232.94 R$ 643.34
61
ES R$ 1,802.43 R$ 1,993.89 R$ 2,283.80 R$ 2,443.30
Subnational Civil Servant Pension Schemes In Brazil
GO R$ 2,220.46 R$ 2,613.11 R$ 2,622.55 R$ 3,415.53
MA R$ 763.46 R$ 1,137.08 R$ 838.01 R$ 993.08
MG R$ 13,401.81 R$ 15,321.64 R$ 17,363.29 R$ 18,099.53
MS R$ 1,135.72 R$ 1,658.24 R$ 897.18 R$ 1,208.09
MT R$ 1,104.92 R$ 1,395.85 R$ 1,619.54 R$ 1,740.05
PA R$ 2,227.59 R$ 1,423.78 R$ 1,265.90 R$ 1,936.75
PB R$ 1,135.03 R$ 1,302.74 R$ 1,418.30 R$ 1,501.53
PE R$ 2,132.31 R$ 2,562.64 R$ 2,650.04 R$ 3,079.16
PI R$ 573.36 R$ 456.91 R$ 1,405.95 R$ 1,715.28
PR R$ 2,299.45 R$ 4,449.74 R$ 4,915.45 R$ 6,021.94
RJ R$ 10,821.08 R$ 13,063.00 R$ 12,312.85 R$ 13,390.06
RN R$ 1,398.06 R$ 1,502.36 R$ 2,295.50 R$ 2,333.85
RO R$ 8.20 R$ 8.19 R$ 6.77 R$ 92.13
RR R$ 26.77 R$ 5.27 R$ 5.30 R$ 0.56
RS R$ 9,748.63 R$ 10,699.11 R$ 11,089.54 R$ 11,880.51
SC R$ 3,070.13 R$ 3,665.12 R$ 3,820.84 R$ 4,164.67
SE R$ 896.84 R$ 946.42 R$ 762.93 R$ 986.32
SP R$ 19,796.65 R$ 21,339.70 R$ 24,081.00 R$ 26,302.35
TO R$ 0.04 R$ 214.47 R$ 548.74 R$ 455.26
Total R$ 82,221.20 R$ 93,986.01 R$ 100,580.85 R$ 111,667.29
Source: 2020 Subnational Entities Finance Bulletin - National Treasury Secretariat.
One idea has been to require state RPPSes to prepare long term asset and liability manage-
ment plans, Planos de Custeio, in order to reduce pension deficits and strengthen the fiscal
position of RPPS pension funds in the long term. While the long-term focus of this policy is
to be celebrated, it may have encouraged states and municipalities to transfer state assets
to RPPS pension funds. For example, Rio de Janeiro state has transferred the ownership
of some state oil assets to state RPPS, and in so doing has “diminished” civil servant pen-
sion scheme deficits to be financed by the state. While the asset transfer has formally kept
the state in compliance with Fiscal Responsibility Law, the state annually transfers large
share of the petroleum royalties directly into the state RPPS budget (seen as other revenue
in Figure 33), having directed over BRL 9,7 billion in first four months of 2021 alone.12 While
from a narrow perspective of RPPS’s fiscal sustainability this approach might be tempting,
it only amounts to accounting exercise and does not create additional fiscal space for state
investments and provision of public services. Increased employer contributions to deficitary
12 http://www.fazenda.rj.gov.br/petroleo/despesa/2020.html, accessed on May 27 2021.
62
pension funds suffer from the same flaw, even though such approach might help politically
constrain inefficient state spending in other areas. Meanwhile, the only true solutions to ar-
6 Previous pension policy response to pension spending increases
rest crowding out of state fiscal space by personnel costs amount to finding ways of reducing
wage and pension expenditures without jeopardizing provision of public services. This, in
turn, involves reducing unfairly high life-time compensation for public servants and improv-
ing service provision efficiency.
Figure 33 – Financing sources of Fundos Financeiros of state RPPS schemes, 2017.
60
Financing Sources (% of net current revenues) – Financial Funds
50
% of current net revenues
40
30
20
10
0
SP
RN
SC
SE
BA
PI
AC
GO
PB
AL
CE
MS
PE
MT
RS
MG
AP
RJ
RO
RR
ES
AM
TO
PA
DF
PR
MA
Total revenue
Aportes Other revenues Employer contributions Individual contributions
(Prev.+Fin.)
Source: Relatório Resumido da Execução Orçamentária (RREO).
Adoption of Federal
2019 reform
by subnational
7 governments
7. Adoption of
Federal 2019 reform
by subnational
governments
The political consensus for 2019 Pension Reform in Brazil grew out of the realization that Bra-
zil’s pension spending was unfairly distributed and was one of the major contributors to the in-
creasing indebtedness of the country, both at the federal and subnational levels. Subnational
governments were also increasingly frustrated by various spending limitations imposed by the
Fiscal Responsibility Law on one hand, and federally mandated generous pension parameters
on the other. Therefore, federal pension reform of 2019 has devolved some decision power to the
states and municipalities, but also included provisions to promote fiscal sustainability and better
governance of pension schemes. The list of the most important reform elements is presented
in Table 5. While expected fiscal impact of these measures will vary across subnational entities,
some qualitative assessment is possible, and is discussed below in greater detail.
The Federal Constitution text now only defines limits within which state and municipal enti-
ties must issue their own local laws in order to organize their civil service pension schemes.
These limits also include the requirement that subnational entities implement certain manda-
tory changes, including preparation of management plans (Planos de Custeio) to inform how
pension liabilities will be financed in future years, unifying pension scheme administration of
different government branches under single management unit, exclusion of short term ben-
efits, like maternity and sickness, from the RPPS benefit rolls, offering complimentary pen-
sion savings plans to civil servants that earn wages above mandatorily insured wage ceiling,
imposing minimum individual pension contribution rate of 14 percent13 , and implementing
federally prescribed pension arrangements for the uniformed personnel14.
13 The exception only applies to subnational entities without actuarial deficits.
14 Constitutional Amendment 103/2019: (a) art 1 (alters art.40 §20 of the CF88); (b) art. 9 §4 and art. 11; (c) art. 9 §6; (d)
art. 10 §1; (e) art 1 (alters art.201 §10 of the CF88) and art. 9 §2 and §3; (f) art 1 (alters art. 149 §1-A of the CF88); (g) art. 23.
66
Table 5 – Most important elements of 2019 pension reform:
7 Adoption of Federal 2019 reform by subnational governments
EXPECTED CONSOLIDATED FISCAL IMPACT
MANDATORY REQUIREMENTS FOR ALL SUBNATIONALS
Preparation of Planos de Custeio None, but allows better planning
Unifying administration of pensions of all Enables benefit audits, potentially saving
government branches up to 20% of expenditures, according to
the experience in the states of Alagoas
and Santa Catarina
Exclusion of short term benefits from RPPS None, but adds transparency
Instituting complimentary pension funds Some cost short term, savings long term
Minimum contribution rate of 14% or equiv- Increases revenues by 27% for entities
alent progressive rates that had 11% contribution rate
Uniformed personnel pension reform Costly short term, limited savings long term.
Involved rolling back some of the fiscal gains
due to earlier reforms in some states.
MEASURES IMPLEMENTED FOR FEDERAL CIVIL SERVANTS AND OPTIONAL
FOR THE SUBNATIONAL ENTITIES
Retirement age increase to 62/65 Minimal savings short term, savings in
very long term
Revision of survivor benefits Important immediate and lasting savings
Increase in pensioner contribution base Strong immediate and lasting increase in
revenue and improvement in equity
Allowance for extraordinary contributions none now, but easier to raise revenue
when needed
However, Federal Constitution retains exclusive power to determine pension scheme rules applica-
ble to federal public servants, which was immediately exercised in 2019 with a list of important para-
metric changes. In addition to improving fiscal situation of the federal civil servant pension scheme,
these changes were also intended to provide an optional template for the follow-up subnational
pension reforms. The changes, including retirement age increase to ages 62 and 65 for females and
males respectively, revision of survivor benefit formula and eligibility, increase in retiree contribution
base, and allowance of extraordinary pension contributions, are referred to as optional, as they can
be adopted by the subnational entities at their own discretion, and should be implemented through
the local laws (Table 6). Subnational governments are, in principle, allowed to strengthen their para-
metric reforms beyond what was approved for the federal civil servant scheme. Some older states
in weaker fiscal position might need to exercise that option. However, so far, most of the first-mover
states have opted to copy or weaken down federal civil servant reform, prioritizing political expedien-
cy. The adoption of mandatory and optional reforms to date is summarized in Table 6.
67
By January 2022, all states had adjusted their pension contribution rates or already had
rates appropriate to the provisions of Constitutional Amendment 103. In addition to federal
Subnational Civil Servant Pension Schemes In Brazil
civil servant scheme (see Box 1), seven states have adopted progressive contribution rates:
Bahia, Maranhão, Minas Gerais, Rio Grande do Norte, Rio Grande do Sul, Roraima and São
Paulo. Piauí state has set a flat rate for active civil servants and progressive rate for the
pensioners. The Federal District chose a linear contribution rate for active civil servants and
progressive rate for pensioners and survivors. Finally, Goiás has adopted a flat rate of 14.25%,
higher than the contribution rate of federal civil servants (see Box 2).
Table 6 – Adoption of the parameters of the Constitutional Amendment 103/2019 by states (by January 2022).
Mandatory Recommended
Single Exclusion of Age of Increase of con- Legal pro- Propor-
UF man- Contribution rates at 14% Complemen- short term retirement tribution base: vision for tional
age- (or progressive) tary regime benefits from (female/ Pensioners’ limit extraordi- survi-
ment RPPS male)2 for exemption of nary contri- vors
unit1 contribution bution benefits
law autho-
AC no linear 14% yes 62F/65M RGPS ceiling no yes
rized
AL no linear 14% yes yes 62F/65M 1MW no yes
AM yes linear 14% yes yes RGPS ceiling no
law autho-
AP no linear 14% no RGPS ceiling no
rized
BA no progressive 14%, 15% yes yes 61F/64M 3 MW% no yes
CE no linear 14% yes yes 62F/65M RGPS ceiling no yes
14% (active
civil servan-
linear ts); 11% from
(active CS), 1 MW up to
pregressive RGPS ceiling
DF yes yes yes 1 MW no
(pensio- and 14%
ners, survi- above RGPS
vors) ceiling (pen-
sioners and
survivors)
ES no linear 14% yes yes 62F/65M RGPS ceiling no yes
GO no linear 14.25% yes yes 62F/65M 1 MW no yes
7.5%, 9%,
progressive 12%, 14%, law autho-
MA no yes RGPS ceiling no
14.5%, 16.5%, rized
19%, 22%
11%, 12%,
progressive 13%, 14%,
MG yes yes yes 62F/65M 3 MW% yes yes
15%, 15.5%,
16%
MS yes linear 14% yes yes 62F/65M 1 MW yes yes
MT yes linear 14% yes yes 62F/65M RGPS ceiling no yes
PA no linear 14% yes yes 62F/65M RGPS ceiling no yes
law autho-
PB no linear 14% yes 62F/65M RGPS ceiling no yes
rized
law autho-
PE no linear 14% yes RGPS ceiling no
rized
14% (linear
actives)
PI no linear 11%,12%,13% yes yes 62F/65M RGPS ceiling no yes
e 14% inac-
tives
PR yes linear 14% yes yes 62F/65M 3 MW no yes
RJ no linear 14% yes yes RGPS ceiling no
11%, 14%, law autho-
RN yes progressive 15%, 16%, yes 60F/65M RGPS ceiling no yes
rized
18%
RO no linear 14% yes yes RGPS ceiling no
11%, 11,5%,
progressive 12%, 12,5%,
RR yes no no RGPS ceiling no
13%, 13,5%,
14%
68
Mandatory Recommended
Single Exclusion of Age of Increase of con- Legal pro- Propor-
man- Contribution rates at 14% tribution base: vision for tional
7 Adoption of Federal 2019 reform by subnational governments
UF Complemen- short term retirement
age- (or progressive) tary regime benefits from (female/ Pensioners’ limit extraordi- survi-
ment RPPS male)2 for exemption of nary contri- vors
unit1 contribution bution benefits
7.5%, 9%,
progressive 12%, 14%,
RS no yes yes 62F/65M 1 MW no yes
14.5%, 16.5%,
19%, 22%
SC yes linear 14% yes yes RGPS ceiling no
SE no linear 14% yes yes 60F/65M 1 MW no yes
progressive 11%, 12%,
SP yes yes yes 62F/65M 1 MW yes yes
14%, 16%
TO yes linear 14% no yes RGPS ceiling no
Notes: (1) Declared by the entities or in the official webpages of the subnational pension system. (2) States that did not
reform the age limit remain at the age thresholds of 55 for women and 60 for men. (3) States that did not reform maintain
the exemption limit at the RGPS ceiling (BRL 6,433.57). Source: The World Bank using data from SEPREV/Ministry of
Labor, data provided by the state governments and data available in official webpages of the subnational entities.
Very few states have consolidated the management of pensions from all their government
branches into a single RPPS management unit. Some implementation delays regarding this
mandatory reform have been due to the complications in integrating different IT solutions
now used by separate branches to assign and monitor benefit payments. However, even the
states that are making steps towards integration of pension sub-schemes tend to take a
narrow view of the task, and rarely attempt a more ambitious integration of some human
resource and pension assignment system functions. Nevertheless, such integration would
be crucial in order for 1) human resource policy decisions to be informed about their pension
cost implications in the short, medium and long term; 2) pension liability monitoring to be
more precise by early forecasting of new retiree inflows and their expected benefits.
The difficulties at technological integration are also likely to sometimes become scapegoats for the
reluctance of different government branches to share power and information. Nevertheless, integrated
IT solutions are fundamental in improving efficiency and fairness of the pension scheme. The require-
ment for all public employers to report individual level employee data into a federal eSocial database
from April 2022 (already mandatory for private employers) offers unique opportunity to systematize
data in a common format. It is important to ensure that subnational RPPS have legal authority to ac-
cess this data for all government branches in order to fully perform their record management and audit
duties. Recent experiences from Alagoas and Santa Catarina suggest that pension record audit, greatly
simplified by improved IT solutions, could yield savings of 10 to 20 percent of pension expenditures.
69
Box 1 – Federal RPPS contribution rate progressivity implemented by the 2019 reform.
Subnational Civil Servant Pension Schemes In Brazil
While the 2019 Reform brought a progressive contribution rate varying between
7.5 percent and 14 percent to private sector workers, rates applied to federal civil
servants are even more progressive (art.11 §1º). From January 2021, federal civil
servants started contributing between 7.5 percent and 22 percent of their sala-
ries. Rates are applied on the portion of the salary that fits into each contribution
range. The exception to these ‘higher level’ rates were made for the federal ser-
vants hired before 2013 who contribute to the complimentary individual pension
accounts. For these, their contribution remained limited to the 14 percent rate, to
incentivize the take up of complimentary funds.
SALARY RANGE RATE
Up to 1 minimum wage (R$1,100.00) 7.5%
From R$1,100.01 to R$2,203.48 9%
From R$2,203.49 to R$ 3,305.22 12%
From R$3,305.23 to the national insured wage ceiling (R$6,433.57) 14%
From R$6,433.58 to R$ 11,017.42 14.5%
From R$ 11,017.43 to R$22,034.83 16.5%
From R$22,034.84 to R$ 42,967.92 19%
Above R$ 42,967.93 22%
Source: INSS and Agência Brasil. Note: Salary ranges are subject to adjustments at the same
date and by the same index used to adjust the RGPS benefits, which are themselves currently
indexed to inflation; except for the minimum wage threshold, which follows specific legislation.
Twenty-five states have already established a complementary pension regime or approved the leg-
islation for its implementation. The creation of complementary pension funds had been previously
allowed by the 2003 pension reform. However, they were not implemented by the subnational enti-
ties until early 2010’s. The oldest complementary pension funds were started in 2013 and 2014 (SP,
RJ and ES), and most of the funds were created very recently, between 2019 and 2020, thus fostered
by the 2019 federal reform. The Constitutional Amendment 103/2019 determined that subnational
entities should establish complementary pension regimes by November 2021.
Regarding the last mandatory item of the 2019 pension reform, 22 states removed short term ben-
efits from RPPS benefit rolls. The excluded benefits consist of maternity benefits, occupational
accident insurance benefits, and similar, which will be paid by the same entities that administer
payments of worker wages. The aim of the reform is focusing RPPS exclusively on the payment of
retirement pensions and survivor benefits (Constitutional Amendment 103/2019 art.9 par.2).
70
To enforce the adoption of the mandatory rules by the subnational entities, ordinance
n.18,084 of July 29, 2020 determined that states, the Federal District, and municipalities
7 Adoption of Federal 2019 reform by subnational governments
should report to the federal Social Security Secretariat on the compliance with these require-
ments. The compliance with these rules will be treated as a necessary condition to obtain the
Social Security Regularity Certificate, which in itself is a requirement for receiving discretion-
ary Federal Government transfers and permissions for credit operations with federal public
banks. The summary of mandatory reforms adopted so far by the states is shown in Figure
34. The progress of reform by municipalities is discussed in Box 3.
Box 2 – State of Goiás - the first adopter of the 2019 federal pension reform.
By January 2019, Goiás’ RPPS had 67,249 retired civil servants compared to 62,940
active workers, a 73 percent growth in retirees over the last 10 years. The 2019
payroll was about BRL 2.9 billion and was projected to reach BRL 6 billion within
a decade, in the absence of a pension reform. According to the state’s Secretariat
of Economy, spending on civil servants’ pensions and survivor benefits was larger
than the state’s investments in education and health. In late 2019, the state gov-
ernment sent to the Legislative Assembly a Pension Reform bill for public sector
employees, which was quickly analyzed. In December of that year, immediately
after passage of the federal law that allowed the state to implement its own reform,
the State Constitution Amendment 65/2019 was enacted, taking advantage of the
political momentum generated by the federal reform discussion.
Goiás adhered almost entirely to the federal pension reform, except for the
adoption of progressive contribution rates and allowance of the extraordi-
nary contribution rates. The state has determined that, given lower salaries in
Goiás in comparison to federal civil servant wages, a progressive contribution
rate schedule identical to that of the federal government would result in lower
overall contribution revenues. In fact, the contribution rate in Goiás was set to
14.25%, the only state in Brazil to have set a higher contribution rate than the
required floor of 14% determined by the Federal Constitution.
At the same time, Goiás has also implemented complimentary individual account
system with accumulated assets, for efficiency purposes, managed by an already
operational Sao Paulo scheme. Furthermore, Goiás is proactively organizing its
municipalities to join its complimentary fund, reducing the overall costs of asset
management. This sets a great example for other states to follow.
Source: https://www.goiasprev.go.gov.br/, accessed on May 2021.
71
The pension rules for the uniformed personnel have also been reformed through a separate Law
13954 in 2019 but, in this case, parameter changes were set at the federal level, with subna-
Subnational Civil Servant Pension Schemes In Brazil
tional governments mandated to adopt the law in its entirety. This separate law maintains more
generous benefits for the uniformed personnel, including policemen and firefighters, compared
to other civil servants, due to two distinct rules: 1) after 35 years of work, uniformed personnel
can request retirement regardless of their age; the eligible length of service can include up to 5
years of work in private sector and, in some states, can be augmented by fictitious time, where
unutilized vacation time is multiplied by two and added to the length of service; and 2) pension
benefits preserve the principle of integrality and parity, setting retiree pension to equal to the last
years’ salary indexed to the salary growth associated with the last position.
The Law 13.954 has also expanded pensioner contribution base to the whole pension amount and re-
duced contribution rate to 9.5 percent for 2020 and 10.5 percent for 2021 onwards for uniformed per-
sonnel. Survivor pensions were also made subject to contributions for the first time, starting in 2021.
The federal law will set new contribution rates after 2025. There are, however, states that obtained
court decisions ensuring that the rates can be determined by local laws and not by the federal legisla-
tion (e.g. Rio Grande do Sul) Thus, the reform of the military will have strong and varied fiscal impacts
on the subnational entities, requiring them to cover the deficits of the scheme, while at the same time
taking away decision power to set pension rules. In the case of Mato Grosso, the pension deficits of
this sub-scheme have increased immediately after the reform, as shown in the right panel of Box 4.
One of the few remaining tools to influence uniformed personnel pension deficits through subnational
policy is human resource policy. For example, a state could potentially hire all soldier level recruits as
temporaries, and only convert them to permanent personnel if and when they are promoted in rank.
72
Figure 34 – State level implementation of mandatory federal reform features.
7 Adoption of Federal 2019 reform by subnational governments
Contribution rate Single Management Unit
14% Yes
No
Progressive rate
14,25%
Complementary Pension Regime Exclusion of Short-Term Benefits
Yes Yes
Law authorized No
No
Source: The World Bank, using data from SPREV/Ministry of Economy, data provided by state governments and
consultation to official webpages.
From the optional reform features (see Figure 35), expansion of contribution base for pensioners
has the largest potential to generate immediate reduction in pension deficits. Contribution as-
sessed on the benefits of retirees and pensioners was already established as part of the reform
of 2003, but it only applied to the portion of the pension that was above the national insured
wage threshold. An important measure of 2019 federal civil service reform allowed to expand
that contribution base to pension payments above minimum wage. However, only seven states
in Brazil have so far lowered the exemption threshold to one minimum wage (about USD 200).
The immediate effect of this reform on pension deficit can is illustrated in Box 4 for Mato Grosso.
73
Figure 35 – State level implementation of optional federal reform features.
Subnational Civil Servant Pension Schemes In Brazil
Contribution over the benefits of
Retirement age retirees and pensioners
62 female
65 male RGPS ceiling
61 female
64 male 1 minimum wage
55 female
60 male 3 minimum wage
60 female
65 male
Legal provision of introducing
extraordinary contribution Proportional calculations for survivor benefits
Yes Yes
No No
Source: The World Bank, using data from SPREV/Ministry of Economy, data provided by state governments and
consultation to official webpages.
74
Box 3 – The adoption of the 2019 reform measures by municipal governments.
7 Adoption of Federal 2019 reform by subnational governments
By 2020, thirty eight percent of Brazil’s 5,570 municipalities run their own RPPS
regimes; 74% of these 2,151 RPPSs have equated their contribution rates to the
14% threshold or implemented an equivalent progressive contribution rate sched-
ule, and eliminated short term benefit payments from their books, following the
determination of the Constitutional Amendment 103. While many municipalities
maintain that their actuarial balances are positive and 14 percent contribution is
not necessary, the validity of these actuarial assessments can often be disput-
ed. The assessments can be substantially influenced by optimistic assumptions
about the ability of municipal governments to contain future wage growth and
maintain profitably of its investments, which has recently greatly benefited from
the run-up of government bond prices.
75
Subnational Civil Servant Pension Schemes In Brazil
QUANTITY CONTRIBUTION EXCLUSION OF
OF RPPSS RATES AT SHORT TERM
STATE % %
IN THE 14% (OR BENEFITS FROM
STATE PROGRESSIVE) THE RPPS ROLL
AC 2 1 50% 0 0%
AL 74 5 7% 3 4%
AM 4 0 0% 0 0%
AP 27 0 0% 0 0%
BA 37 3 8% 4 11%
CE 65 3 5% 4 6%
DF 1 1 100% 0 0%
ES 35 9 26% 6 17%
GO 171 18 11% 9 5%
MA 47 0 0% 1 2%
MG 107 28 26% 13 12%
MS 51 3 6% 4 8%
MT 222 22 10% 22 10%
PA 30 1 3% 1 3%
PB 71 4 6% 1 1%
PE 179 22 12% 17 9%
PI 149 9 6% 4 3%
PR 70 5 7% 3 4%
RJ 80 7 9% 2 3%
RN 40 2 5% 2 5%
RO 332 92 28% 54 16%
RR 30 8 27% 3 10%
RS 2 0 0% 0 0%
SC 70 19 27% 12 17%
SE 221 43 19% 17 8%
SP 4 0 0% 0 0%
TO 30 0 0% 0 0%
Total 2151 305 14% 182 8%
Source: 2020 Subnational Entities Finance Bulletin - National Treasury Secretariat.
76
Box 4 – Immediate results of the Pension reform adopted by the state of Mato Grosso.
7 Adoption of Federal 2019 reform by subnational governments
The pension reform in the state of Mato Grosso was implemented from the 2nd quar-
ter of 2020 through the adoption of four items of the federal reform: (i) the contribu-
tion rate was raised to 14% (Complementary Law 654, of 02/19/2020; (ii) retirement
age was increased to 62 (female) and 65 (male) years old, (iii) 2019 federal civil ser-
vice reform of survivor benefits was adopted (State Constitutional Amendment 92,
of 08/18/2020); and (iv) complementary regime was established (Complementary
Law 670, of 09/04/2020). The fiscal impact of the reform was felt within the same
fiscal year, almost exclusively from expanding the contributory base.
Financial deficit – civil employees Déficit financeiro – militares
R$ 1,200,000
R$ 1,000,000
R$ 800,000
R$ 600,000 R$ 600,000
R$ 550,000
R$ 500,000
R$ 450,000
R$ 400,000 R$ 400,000
R$ 350,000
R$ 300,000
R$ 250,000
R$ 200,000 R$ 200,000
R$ 150,000
R$ 100,000
R$ 50,000
R$ 0 R$ 0
2018 2019 2020 projeção 2018 2019 2020
2021
Source: MTPrev, May 2021.
The reluctance by states to adopt the expansion of contributory base for pensioners is a loss
of opportunity for a more effective adjustment. The 2003 reform has introduced big unfair
differences in treatment between the pre-2003 and post-2003 entrants into the public ser-
vice. The pensions of pre-2003 entrants are heavily subsidized by state and the contribution
base extension to them would only claw back a small portion of this subsidy. The importance
of the measure is also underscored by the fact, that state pension costs will be rapidly in-
creasing for the next two decades – exactly the timing of pension payments for the pre-2003
entrants. More creative solutions for reducing treatment differences between the highly sub-
sidized pre-2003 entrants and their younger colleagues may also be possible. They would
involve designing some inducements, possibly including up front lump sum payments, to
entice pre-2003 entrants to accept post-2003 pension benefit package, which under some
conditions might reduce overall subnational pension liabilities.
The increase in the retirement age has a potential to result in strong long term fiscal impact on
subnational finances, if it is accompanied by the slowdown in hiring of new civil servants. These
two combined policies would allow to reduce the frequency at which each civil service position
generates a pensioner, reducing long term pension deficit. The federal reform substantially in-
creased the retirement age to 65 for males and 62 for females, from the previous 60/55 thresh-
olds. However, given the slow phase-in of the new policy and the fact that delayed retirement
decisions have a very limited fiscal effect on overall personnel costs in the civil servant setting,
where both workers and pensioners cost the government similarly, the overall fiscal effect of this
policy in the next couple of decades is not projected to be very significant. Furthermore, until April
2021, only 17 of the 27 states adjusted the age limit in line with the federal reform.
The federal civil service reform has also introduced the legal provision of allowing an extraor-
dinary contribution for a limited time period. The increase is not to exceed 20-year period and
is only applicable in case of negative actuarial balances (Constitutional Amendment 103/2019
art.9 par.8). Adopting this provision by local legislature would not necessarily mean immediate
increase of contributions but would allow pension scheme administrators to increase contribu-
tion rates in the future as needed without a cumbersome legal process. This legal arrangement
was incorporated into Sao Paulo state’s RPPS legislation, although not widely adopted by other
states. In addition to being a tool in reduce pension deficits, the extraordinary contributions,
that would potentially be applied only to Fundo Financeiro, can act as a tool to iron out some
unfair treatment differences between pre-2003 and post-2003 entrants to public service, as
well as act as an inducement for the older civil servants to switch to the Fundo Complementar.
Finally, several states in Brazil have followed federal government in adopting proportional
calculation of survivor benefits. Under the old formula, survivor benefits stood at 100 per-
cent of the deceased person’s full retirement benefit or, in the case of active worker’s death,
full permanent disability benefit. With the reform, the proportional benefit is now composed
of a benefit of 50 percent plus 10 percent per each dependent, up to a 100 percent in total
for a family (Constitutional Amendment 103/2019 art. 23). The benefit is now also limited in
duration, based on characteristics of the dependents. Furthermore, the reform restricts the
possibility of combining full survivor benefit with own pension benefits.
Overall, the impact of the federal 2019 reform on state finances is likely to be mixed. On
one hand, the federal government has handed subnational governments some useful tools
to achieve moderate short term relieve from fiscal pension pressure, especially by allowing
them to extend pensioner contribution base, reduce survivor pensioner expenditures, and
unify their different pension schemes for greater control and transparency. However, so far,
the take up of these tools by states has been underwhelming, possibly due to the short term
improvement to state finances in 2020 and the overwhelming urgency of COVID-19 response.
It is likely that municipal governments, with lower implementation capacity, are even further
behind in the reform process. It is also clear that the 2019 reform measures are insufficient
to arrest the strong growth of pension expenditures projected for the next two decades, that
is driven by strong forces of continued maturation of the pension system and generous pre-
2003 pension benefit rules.
Lessons and
recommendations
8
8. Lessons and
recommendations
Pensions are the fastest growing large item in subnational government budgets and have
received a lot of attention from federal and state policy makers. The federal reforms of 1998,
2003, and 2019 have required to achieve difficult political consensus and have yielded some
important solutions. However, some of the most important measures of the last reform are
yet to be approved and implemented by subnational governments. It is expected that the
return to pre-pandemic fiscal pressures in 2021 and beyond will accelerate the adoption rate
of these reforms. Making full use of the optional parametric reform package that includes ex-
pansion of the contribution base for pensioners, revision of survivor pension rules, increase
in retirement ages, and introduction of a possibility of extraordinary contribution rates is es-
pecially important. However, while these reforms are critical for containing short term growth
in pension expenditures, they alone are unlikely to arrest pension expenditure and, therefore,
overall personnel expenditure growth in the next two decades.
Going forward, lasting solutions would be easier to find if fiscal, human resource, gover-
nance, asset management, and pension policies are viewed as integral parts of interdepen-
dent system, requiring close coordination between different agencies and adoption of smart
IT solutions, both at the federal and at the subnational level. From macro-fiscal management
standpoint, rules designed to ensure fiscal sustainability of subnational entities would ben-
efit from adopting a more forward-looking perspective. Simplification and standardization of
atomized career structures and strategic approach to hiring are necessary for more efficient
public service provision, as well as control of pension expenditure growth. The complicated
system of automatic promotions and additional benefits, currently blind to long term impli-
cations for wage and pension expenditure growth, also needs to be revised. Greater use of
temporary workers, insured under the same contract as private sector workers, would not
only decrease pension expenses, but also introduce more desirable flexibility for migration
between private and public sector.
To address expected growth in pension expenditures in the next two decades, pension policy
would benefit from reexamining the inequities left by exemptions from 2003-reform. Some
ways to devise partial claw-backs of unjustified subsidies to pre-2003 cohorts are highly
progressive contribution rates; extraordinary additional contributions for Fundo Financeiro
80
with possibility to “escape” by switching to post-2003 rules and complimentary funds; and
some lump sum inducements to accept post-2003 benefit rules.
8 Lessons and recommendations
Throughout the system, there is an urgent need for more transparency, better governance,
and policy-maker friendly IT solutions. These, on their face mundane changes, can lead to
surprisingly big savings. As the experience of Alagoas and Santa Catarina record audits have
shown, the savings can reach 10-20 percent of pension spending. Covid-19 has also exposed
a need and convenience of being able to serve clients and manage records remotely. It is
neither reasonable nor efficient for all 2154 RPPS schemes to acquire or develop their own
IT systems and set up their own asset management departments. Therefore, IT and asset
management services should be made available, sponsored at the federal or at least state
level, that could be used by smaller RRPS schemes. In addition to technological solutions,
legal basis for RPPS access of integrated employee databases for all government branches
is also needed. However, subnational governments cannot passively wait for that to happen.
Temporary solutions are also urgently needed, including cloud-based service rentals from
private IT companies, or own-developed software packages that could be shared between
cooperating RPPSs schemes. Sharing of asset management services is already happening
and should be further encouraged with appropriate safeguards.
Subnational Civil Servant Pension Schemes In Brazil
81
References
9
9. References
Fontainha, F. C. et al. (2014) Processos seletivos para a contratação de servidores públicos :
Brasil, o país dos concursos? Relatório de pesquisa. Rio de Janeiro: FGV Direito Rio.
IMF – International Monetary Fund (2019) “2019 article IV consultation—press release; staff
report; and statement by the executive director for Brazil”, IMF Country Report No. 19/242.
IMF – International Monetary Fund (2020) “2019 article IV consultation—press release; staff
report; and statement by the executive director for Brazil”, IMF Country Report No. 20/311.
Longo, F. (2006). “Analytical Framework for Institutional Assessment of Civil Service Systems.”
In: “Informe sobre la situación del servicio civil en América Latina.” Washington, DC: IDB.
Lopez, F. G., & Silva, T. M. D. (2020). O Carrossel burocrático nos cargos de confiança: análise de
sobrevivência dos cargos de direção e assessoramento superior do executivo federal brasileiro
(1999-2017). (Link)
ME, Ministério da Economia (2020). Anuário Estatístico de Previdência Social do RPPS, Se-
cretaria de Previdência. (Link)
ME, Ministério da Economia (2019). Nova Administração Pública - Diagnostico Força de Tra-
balho Governo Federal. (Link)
Mendes, Marcos (2020) “CRISE FISCAL DOS ESTADOS: 40 anos de socorros financeiros e suas cau-
sas”, INSPER Working paper, available at https://www.insper.edu.br/wp-content/uploads/2020/08/
Crise-fiscal-dos-estados_40-anos-de-socorros-financeiros-e-suas-causas.pdf.
Schettini, dos Santos and Pires (2019) “Militares estaduais no contexto da Nova Previdência”,
IPEA Carta de Conjuntura N.43.
STN – Secretaria do Tesouro Nacional (2017) Aspectos Fiscais da Seguridade Social no Brasil.
Word Bank (2019). Gestão de Pessoas e Folha de Pagamentos no Setor Público Brasileiro: O
que os dados dizem. (Link)
84
Word Bank (2019). Worldwide Bureaucracy Indicators (WWBI). (link)
9 References