543160
research-article2014
IPS0010.1177/0192512114543160International Political Science ReviewMelo et al.
Article
Why do some governments resort
to ‘creative accounting’ but not
others? Fiscal governance in the
Brazilian federation
International Political Science Review
1­–18
© The Author(s) 2014
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0192512114543160
ips.sagepub.com
Marcus André Melo
Federal University of Pernambuco, Brazil
Carlos Pereira
Getúlio Vargas Foundation, Brazil
Saulo Souza
Federal University of Pernambuco, Brazil
Abstract
This paper investigates the determinants of compliance with fiscal rules. Using information from 27
Brazilian state governments, the paper shows that the level of political competition and the degree
of political autonomy of the fiscal watchdogs explain the extent of creative accounting in the Brazilian
federation. Despite hard budget constraints imposed by the much-acclaimed Fiscal Responsibility Law,
state governors retain the strategic ability to undertake fiscal window dressing in response to fiscal stress.
As fiscal watchdogs are not immune to the influence of the legislature and state governor, however,
this paper demonstrates that their level of independence and autonomy is associated with the degree of
creative accounting. In addition, we show that political competition encourages non-compliance through
various causal mechanisms. Building effective and autonomous institutions is a precondition for deterring
fiscal misconduct but this solution is ultimately a by-product of a political equilibrium resulting from selfenforced behavior of politicians.
Keywords
Fiscal rules, audit courts, political competition, creative accounting
Corresponding author:
Professor Marcus André Melo, Political Science Doctoral Program, Federal University of Pernambuco, CFCH 14 Andar,
Recife, 50000, Brazil.
Email: [email protected]
2
International Political Science Review 
Introduction
The implementation of fiscal rules constraining fiscal deficit and government debt has become a
central concern of developed countries in the last decade, as demonstrated by developments in the
European context (Hallerberg et al., 2009b; Irwin, 2012; Posner and Blöndal, 2012). Constraints
on fiscal policy, a long-standing issue, have also reached the top of the governments’ agenda in
several developing countries.1 There has been increasing consensus that fiscal stability is key to
sustainable development. Despite the scholarly interest, we still do not know much about the determinants of compliance with fiscal rules (Ayuso-i-Casals et al., 2009; Hallerberg et al., 2009a,
2009b; Kopits, 2004). There is still much controversy about which institutional and political factors facilitate compliance and what the best enforcement technologies are. Under which circumstances do governments more often resort to ‘creative accounting’? When can enforcement
institutions effectively sanction non-compliance? These questions remain open despite the burgeoning literature about fiscal behavior and fiscal rules.
This paper addresses these issues by investigating the influence of the state political system in
fiscal misbehavior at the subnational level in Brazil. We explore the general hypothesis that the
degree of independence of watchdogs from the governments they monitor influences fiscal behavior decisively. We demonstrate that creative accounting can be understood as a function of the
degree of activism and political autonomy of fiscal watchdogs— the state Audit Institutions
(Tribunais de Contas (TCs)) in charge of enforcing the Fiscal Responsibility Law (FRL). We also
analyze the impact of political competition on the governors’ decision to make use of creative
accounting. While we find a negative effect of institutional independence of audit institution and
window-dressing initiatives (less independent institutions are poor enforcers of fiscal rules), we
demonstrate a positive influence of power alternation (turnover) on states’ levels of creative
accounting (because political volatility creates incentives for non-compliance).
The paper thus advances our current understanding of the political logic underlying compliance
with fiscal rules, a theme of considerable importance in both developed and developing countries.
The paper is organized as follows. In the first section we discuss comparative experiences of creative accounting in reaction to fiscal rules. The second section develops a critical dialogue with the
literature about the extent to which creative accounting practices might be interpreted as a reaction
to fiscal rules. The third section analyzes the determinants of the FRL and the fiscal governance
game with specific reference to the Brazilian case. In the fourth section we focus on the organization of Audit Courts at the subnational level in Brazil and their role as enforcers of the FRL. The
empirical analysis is presented in the fifth section. Following a descriptive analysis of creative
accounting in the states of the federation, we present some preliminary econometric evidence supporting our claims. In the sixth section we conclude the article, pointing out the main contributions
of our research.
Is creative accounting a reaction to fiscal rules?
Based on the experience of some countries, mainstream literature warns us that the imposition of
numerical rules in some countries may encourage the use of dubious accounting practices, thereby
lessening the degree of transparency in public financial management. Kopits (2001), for instance,
observes that at times fiscal rules are ‘fig leaves’ in the sense that, if governments adopt rules, they
are unlikely to strictly abide by them. It follows that, by their very nature, fiscal rules may invite
abuse by inducing nontransparent behavior. For Milesi-Ferretti (2003), fiscal rules are imposed on
‘measured’ fiscal variables, which can differ from ‘true’ variables because there is a margin for
creative accounting. The rules can thus lead to ‘ugly’ outcomes, namely, a considerable amount of
Melo et al.
3
window dressing, but have little overall effect on fiscal policy. The probability of detecting creative
accounting will depend on its size and on the transparency of the budget.
In a similar vein, Koen and van den Noord (2004) stress that if accounting conventions leave
some scope for judgment, creative accounting emerges as an unorthodox treatment of operations.
That is, when fiscal rules threaten to bite, or are biting, governments may be tempted to take advantage of the implied degrees of freedom by putting the best possible gloss on the accounts presented
to the outside world. In the end, governments have incentives to present flattering fiscal accounts
and to report improving fiscal performance. New governments, in contrast, have reasons to audit
the accounts inherited from their predecessors in order to gain credit for fiscal rectitude.
Brazil has made great improvements in fiscal governance in the last 15 years or so and has been
held up as a role model for other developing countries (Giambiagi, 2007; Goldfajn and Guardia,
2004). Indeed, The Economist went as far as suggesting that the European Union (EU) had much to
learn from the Brazilian experience with fiscal rules (‘The history of fiscal federalism may offer the
euro zone some lessons’, February 11, 2012). However, the federal government’s fudging of the
fiscal target by omitting some infrastructure spending from the sums and bringing forward dividends from state-owned firms as a means to keep its reputation for fiscal sobriety clean, has also
made headlines worldwide (‘Brazil’s economy: Wrong numbers’. The Economist, January 18,
2013). The Financial Times also reported that in the first few days of 2013 the government announced
a series of accounting transactions to meet its fiscal primary surplus target of 3.1% of gross domestic
product (GDP; Brazil’s monetary jeitinho, Financial Times, January 15, 2013). This move prompted
the downgrading of two of Brazil’s major public banks: BNDES and Caixa (‘Brazil’s BNDES and
Caixa hit by downgrades’, Financial Times, March 21, 2013). In addition, in 2010, the volume of
the federal government’s ‘unpaid commitments’—a typical strategy of creative accounting—
roughly equaled the total budgeted for investment in 2011 (Almeida, 2011: 6). Despite the overall
good health of the Brazilian economy, creative accounting has been a serious problem (Garcia,
2010). The fiscal situation of the Brazilian states is much less serious now compared to the early
1990s, when creative accounting was seen as a major destabilizing factor; however, there is potential for great imbalance in the future if these problems continue (Pellegrini, 2012).
These widely publicized episodes of strategic manipulation of fiscal data leading to the downgrading of two major public banks illustrates why the study of creative accounting in Brazil is a matter of
great macroeconomic import. Understanding creative accounting in contexts like these is crucial for
a better understanding of the political economy of Brazil and other emerging economies.
In this paper we focus on the subnational level—the state governments—because this allows us
to observe how the variation in a number of variables of interest influences compliance with fiscal
rules. We develop the argument that notwithstanding the constraints imposed by the fiscal rules,
incumbent politicians have managed to make use of ‘creative accounting’ in order to facilitate
meeting budget deficit ceilings established by the FRL (passed in 2000). That is, governors resort
to fiscal window dressing in response to legal constraints. Given the risk that discretionary fiscal
policies may deviate from what may be desirable for a society, legal or regulatory restraints have
been advocated as a measure to constrain the ability of governments to decide their levels of taxation and spending (Alesina and Perotti, 1994). While the imposition of such fiscal constraints
would reduce government bias, it may also promote, as already mentioned, the use of dubious
accounting practices and government opportunism, which may lead to distortions and economic
costs (Milesi-Ferreti, 2003).
Hard budget constraint legislation, such as the FRL, was originally designed to curb fiscal irresponsibility; however, recent studies have identified extensive creative accounting practices in
Organization for Economic Cooperation and Development (OECD) countries (Ayuso-i-Casals
et al., 2009; Bernoth and Wolff, 2006; Milesi-Ferreti, 2003; Von Hagen and Wolff, 2006). As
4
International Political Science Review 
documented by Irwin (2012), unpaid commitments and deferred spending are recurrent problems
in the EU context. A central theme in this literature is that information and monitoring problems
are key to understanding fiscal behavior. Stéclebout-Orseau and Hallerberg (2009) propose a
game-theoretical model of the ability of both the EU and European voters to sanction deviant fiscal
management. This model is particularly interesting because it focuses on the strategic behavior of
watchdogs. In the EU context qualified majority votes are required to sanction members and there
is significant asymmetry of power among small and large countries in fiscal governance. The
upshot is that the incentive structure facing the EU watchdogs leads to either inaction and/or selfdefeating behavior. The authors conclude, however, that the watchdogs play a key role as providers
of information in a context of severe information asymmetries.
In general the literature on fiscal rules focuses on two problems. The first is related to the incentive structure facing actors. From this perspective compliance with fiscal rules results from the
self-interest of governments: fiscal rules are endogenous. The compliance problem is reduced to a
matter of ensuring that interests are aligned. The second problem is the ability of the enforcer to
secure compliance. In this case, compliance depends on the means at the disposal of the ‘enforcer’,
usually the Central government, the Ministry of Finance or external actors.
The problems of fiscal policy are compounded in federations because of the need to reconcile
the interests of various actors at different institutional levels. A common problem is that the
subnational units have incentives to run deficits if they expect to be bailed out by the central
governments. Fiscal behavior is also shaped by the nature of political competition in the subnational units and in the political relations between the central government and subnational governments. Political bargaining between central and lower-level politicians plays a major role in
the policy-making process involving the creation of fiscal policy rules for subnational governments. The imposition of hard budget constraints on states and municipalities would require the
central government to retain strong agenda powers, whereas political alignment across government tiers can be an important factor leading fiscal rules to become self-enforcing (Braun and
Tommasi, 2004).
Political economy literature also correlates the dynamic of political interactions within legislative bodies to the size and nature of the budget deficit. A large number of representatives from
different parties pursuing their own spending preferences will probably lead to budget imbalance
(Alesina and Perotti, 1994; Persson and Tabellini, 2001). According to this argument party fragmentation is conducive to the proliferation of veto players, which among other things can be associated with added difficulties in implementing fiscal stabilization reforms. Thus, a highly
fragmented party system tends to postpone the government’s response to fiscal crises.
On the other hand, as coalition partners will probably support their allied parties’ budget decisions, incumbents will find themselves in a position to implement fiscal adjustment measures once
a majority is secured for the government’s coalition in the legislature. Webb (2004), for instance,
argues that a strong coalition in favor of the government’s interests contributes to fiscal discipline,
particularly if having a balanced budget improves the reputation of the coalition parties (see also
Liu and Webb, 2011). For Rodden et al. (2003) this will mostly be the case if the proposed fiscal
constraints are expected to bring political benefits to all parties in the coalition. Alt and Lowry
(1994) and Poterba (1994) demonstrate that although a coalition may not have created the deficit,
a divided coalition government tends to put off the adoption of balanced budget measures.
Another political aspect highlighted in the political economy literature associates the proximity
of elections with changes in the government’s fiscal stance, precisely in the sense that the negative
impacts of political variables in the budget are more widely observed in pre-election times. The
prospect of re-election is therefore a variable, which is widely used in analyses of the fiscal opportunism of politicians in office. A hypothesis often tested is whether positive fiscal outcomes would
Melo et al.
5
enhance a candidate’s chances of staying in office for the sake of his or her reputation (Alt and
Lassen, 2003). Conversely, according to models based on a context of soft budget constraints (e.g.
Aizenman, 1998; Cossio, 2001), negative re-election prospects could provide incumbents with an
incentive to increase public spending on purpose as a means of creating fiscal burdens for their
successors, unless the likelihood is high that the incoming politicians will be members of some
allied party.
The Fiscal Responsibility Law and the fiscal governance game in
Brazil
Fiscal stabilization rules in Latin America vary remarkably in terms of format, scope, and performance. Yet in almost all cases, the aim has been to restrict politicians’ discretion in fiscal
policy (Kopits, 2004). Some rules focus exclusively at the national-level (Argentina, Chile and
Peru) or on subnational finances (Mexico), while others apply to all government levels (Colombia
and Brazil). In Argentina, Australia, and India, the federal government approves a FRL only for
itself, establishing a framework for states and provinces to pass their own FRLs voluntarily. In
some rare cases, a province will enact its own FRL before the enactment of the federal FRL,
while in others (e.g. Canada) a province introduces its own FRL for fiscal sustainability
purposes.
Argentina is a case worth highlighting because unlike Brazil, where the central government has
managed to impose fiscal rules on the states, provinces have total discretion over the adoption of
fiscal rules. By the end of the 1990s, the absence of ex ante fiscal controls in Argentina, coupled
with the President’s lack of ideological or party cohesion, enabled a number of provinces to overborrow despite a national framework approved for controlling subnational debt in 2004 (Webb,
2004). As the national level faced a deteriorating budget balance and growing debt payments, the
Argentine Congress approved a FRL in 1999, establishing numeric limits for the central government’s fiscal deficit and implementing transparency measures regarding public finance. Although
the national law did not stipulate conditions for subnational governments, it encouraged the provinces to pass similar laws of their own, which about a third of them in fact did (Liu and Webb,
2011). The provisions of the laws differed across provinces, as did the degree to which they were
adhered to and complied with. The Argentine FRL was revised numerous times to allow for numerical targets during a longer transition period and Congress suspended its application in 2009
(International Monetary Fund, 2012).
The United States is another interesting case. Here there is greater independence to approve
balanced budget legislation at the subnational level. In fact, the implementation of hard budget
constraints in the USA was a voluntary decision taken by subnational unities. Therefore, states with
tight anti-deficit rules also have fiscally conservative electorates. The primary factor explaining
their lower borrowing costs is the electorate, not the fiscal rules.2
The Brazilian states are fiscally and political autonomous subnational entities with their own
constitutions and governments. Their relative importance in the Brazilian political system can be
illustrated by their fiscal size. As of 2011, public spending by state governments amounted to 13%
of GDP, compared to 7.4% by local governments and 28.9% by the federal government. Similar
proportions are observed regarding revenue collection: 12.9% by states, 7.5% by municipalities,
and 27.4% by federal authorities. The Brazilian Constitution itself allows each level of government
to both create and levy taxes. In particular, state governments have control over the single most
important tax in the Brazilian tax system, the Merchandise and Service Circulation Tax (ICMS),
which is a value-added tax imposed on sales of goods and telecommunications services accounting
for about 30% of the overall tax revenue in the country.3
6
International Political Science Review 
However, in the mid-1990s the federal government started de-earmarking revenue that was part
of the pool of taxes forming the basis of constitutionally mandated transfers to the states and
municipalities. This in practice was tantamount to a recentralization of resources into the hands of
the federal government. This and related measures were taken in reaction to the fiscal decentralization of revenue mandated by the new constitutional rules in 1988. The fiscal and tax reform initiatives of the Cardoso and Lula administrations (1995–2008) show that the federal government has
successfully implemented an agenda of reforms for fiscal federalism while also promoting extensive fiscal reforms in the states.
This centralization of decision-making was made possible by the fact that state caucuses in
Congress did not act as collective players, allowing the swift approval of constitutional amendments favoring the Federal Government. However, the Federal Government concentrates the
authority to legislate over most of the policy responsibilities of states and municipalities, converting the federal arena into the main locus of decision-making on federal issues, and consequently a
significant share of bills pertaining to federal and state relations are ordinary bills. These institutional factors limited the veto power of state governments in Brazil.
The FRL was applicable to all government levels, regardless of their prior economic conditions,
as top-down legislation.4 The FRL was introduced in response to the run against the Real and the
concomitant confidence crisis that affected the Brazilian economy in the wake of the Asian and
Russian crises. The currency crisis was triggered by the default of the state of Minas Gerais in a
much-publicized move that made headlines in major economic newspapers worldwide. However,
the FRL has to be discussed within a process of reassertion of federal fiscal authority since 1995
(De Mello and Moccero, 2006; Giambiagi, 2007; Souza, 2007). Along with its companion law, the
so-called Fiscal Crimes Law, the FRL is the culmination of a relatively successful set of measures
designed to constrain fiscal behavior and control the indebtedness of state governments. The struggle to rein in the states’ irresponsible behavior started with Cardoso’s term as Franco’s Finance
Minister and absorbed much of Cardoso’s political capital during his two terms of office (1995–
1998 and 1999–2002).
The FRL illustrates the kinds of policy outcomes that reflect the national executive’s ability to
implement its policy preferences in the political game. In its relations with the state governments,
a powerful president and finance minister have managed to recentralize fiscal authority in the
country, curbing their fiscal autonomy. The executive was able to implement its preferences
because of its institutional prerogatives and because there were gains-from-trade in federal government–state relations. For Lora (2006), the effectiveness of the FRL owes not so much to the direct
effect of the debt ceilings, but rather to changes in power relations in the political arena, particularly between the federal and subnational governments.
Nowadays there is no question about the positive effect of the FRL with regard to the fiscal situation of the states, which has improved considerably since the enactment of this law. Whereas all
state governments had faced a deficit prior to the enactment of the law, the consolidated state
accounts have since systematically presented a surplus roughly equivalent to 4% of GDP (see
Figure 1). A similar success story is that of public debt. A succession of primary surpluses enabled
the government to effectively reduce the GDP/debt ratio. Since 2002, when the net debt peaked at
55% of the GDP/debt ratio, there has been a reduction in net debt measured as a percentage of
GDP, which was estimated to be less than 36% in 2008.
The distinctive features of the Brazilian federation therefore include a strong federal executive
branch vested with the power to propose limits for debt ceilings at all levels. The Senate is also a
key player in fiscal governance. Article 52 of the Constitution stipulates that it is up to the Senate
to approve global debt ceilings for the three tiers of government. Moreover, intense political competition between major political parties helps explain the political dynamics in the federation. This
Melo et al.
7
Figure 1. States’ fiscal accounts 1993–2010.
Source: Brazilian National Treasury.
endogenous perspective allows us to understand fiscal reform initiatives as generating political
benefits for incumbent politicians. At the subnational level, political competition is also intense
overall—despite the varying patterns found across the country—and this has been a precondition
for the political equilibrium in fiscal governance.
Fiscal governance in Brazil has been shaped by the reassertion of federal fiscal authority in the
country. Powerful presidents have had the tools and the incentives to ensure fiscal stability at all
levels. The TCs, as we will discuss later, play a significant role in the enforcement of the law. The
FRL specifies the fiscal rules governing public sector indebtedness, credit operations, and public
account reporting in great detail. It stipulates that the federal government is prohibited from refinancing subnational governments, therefore eliminating the possibility of bailouts as well as any
changes in the financial clauses of the existing debt-restructuring agreement. The FRL imposes
debt ceilings for each level of government on existing debt. Any excess is to be eliminated within
one year, otherwise new financing and voluntary transfers from the central government to subnational levels are prohibited. Other sanctions include withholding federal transfers by the federal
government, denial of credit guarantees and banning of new debt. In addition, the FRL contains a
golden rule provision for capital spending (i.e. annual credit disbursement cannot exceed capital
spending).
In sum, through the FRL the executive has managed to reassert its fiscal authority over subnational governments. While the endogenous component was crucial for reform initiation and its
approval, Cardoso’s administration (1994–2002) was politically strong, as a result of the re-election amendment approved in 1997.5 Braun and Tommasi (2004) point out that for fiscal rules to be
enforced requires self-enforcement by the players (states) or an external enforcer with the power to
ensure compliance. We argue that in the Brazilian case the FRL had both endogenous and exogenous components. That is, in addition to providing the executive with the enforcement technology,
the FRL has also been an effective ‘commitment device’. The executive was able to implement its
preferences because of its institutional prerogatives and because there were gains-from-trade in
federal government–state relations. Governors developed an interest in reforms in the wake of the
approval of the re-election amendment and in view of the compensation mechanisms involved in
the reform process. The law itself was proposed at the end of a protracted process of restructuring
of Brazil’s fiscal federalism, which involved bailouts of heavily indebted states and extensive privatization of public companies in the states.
8
International Political Science Review 
Although we argue that the executive has the upper hand in executive–legislative relations,
particularly in fiscal and budgetary matters, the interests of the states and municipalities are important in explaining the highly successful implementation of the FRL. By saying this we do not mean
to underestimate theories of federalism in the analysis of the compliance of subnational unities to
federal laws. There was no decentralization of policy-making powers with the approval of the 1988
Constitution. In fact, the Constitution decentralized powers to subnational unities at the policy
implementation stage only.
In order to understand the interest of state governors in the FRL one has to consider governors
as rational actors seeking political survival. Governors have an interested in fiscal expansion
because this helps them achieve this goal. However, in the context of highly indebted and fiscally
vulnerable states (as a result of high interest rates) governors were also interested in shifting the
blame for the austerity measures on to other actors. This move is affected by the future electoral
chances of governors, however. In highly polarized states in which governors faced close elections
(or faced a high probability of defeat), governors would have an interest in ‘stacking the deck’ for
the future governor, in curbing her or his ‘fiscal powers’, thereby increasing unpaid commitments
(Souza, 2011). On the other hand, in states in which the incumbent does not face electoral risks, the
probability of creative accounting would be lower because of his or her extended time horizon. The
incumbent would have to bear the brunt of the consequences of fiscal indiscipline. In the case of an
incumbent who is in the second term of office—and therefore barred from competing in the next
election—he/she still would have an interest in the law: he/she would have an excuse to say no to
demands from her/his own constituency, and particularly for pay hikes. Governors and mayors
used the FRL and its companion Fiscal Crime Laws as an excuse to say no to demands for special
interest expenditures and transfers that would violate the FRL. Webb (2004: 8) notes that ‘subnational governments have put up posters listing the penalties, as a reminder of why they are turning
down special requests’. Therefore, the FRL was both a ‘shield’ (permitting some to evade responsibility and shift the blame) and a ‘sword’ (that can be used to constrain political competitors).
Governors blamed the law for the harsh measures they had to take and used it whenever they
could against their opponents. This enabled them to extract benefits from the law. We hypothesize
that the more fiscally vulnerable a governor, the higher the attractiveness of the ‘pass the buck’
strategy. This situation contrasts sharply with the state of affairs in the early and mid-1990s, when
governors had the upper hand in their negotiations with the federal government. In 2000, when
Cardoso passed the law, the states were considerably weaker and could not offer any resistance.
Much of the privatization of state banks and utilities had already taken place. The federal bailouts
had already occurred and the FRL was seen as inevitable. Indeed the FRL was the culmination of
a decade-long adjustment period.
It was the governors’ evaluation of the value of blame shifting versus binding the hands of the
competitors that determined the governors’ preference intensity. Gama Neto (2007) provides extensive evidence that in 1999 and 2000, most governors were in a situation of great fiscal vulnerability
and most of them would not be able to run for re-election again (in 1998, 19 governors had been
elected and would not be allowed to run for a third time). This explains the high level of support for
the FRL by governors. This support was not sufficient or even maybe necessary, as suggested by the
approval of many initiatives taken in the late 1990s that directly impinged on subnational interests.
Nonetheless they help explain the smooth and successful implementation of the FRL.
Tribunais de Contas: an effective enforcement technology?
The Audit Courts in Brazil are constitutionally defined as ancillary bodies of the legislative branch,
whose task is to examine the accounts of the three branches of government in terms of their
Melo et al.
9
compliance with the constitutional principles of public administration (public morality, impartiality,
publicity, efficiency) as well as the specific legal requirements for hiring of personnel, concession
of pensions, procurement, intergovernmental transfers, competitive public bidding, and fiscal
responsibility. Although the Audit Courts are not technically part of the judicial system, in practice
they operate as quasi-independent judicial authorities.6 Not only do the board members enjoy tenure, but they are also appointed until the official retirement age for public servants, currently 70
years old (Speck, 2000).
The TCs are state-level institutions, which enjoy great levels of functional, administrative,
and political independence. The Constitution of 1988 contained a number of changes that
strengthened the audit institutions. In terms of political autonomy vis-à-vis governors, new
provisions restricted the appointment powers of governors to the board, which is made up of
seven members (the Pleno). The latter were granted virtual life tenure and cannot be dismissed
ad nutum by the governors or the legislatures. Governors appoint one member freely (subject
to expertise among other requirements) and they nominate two members from the pool of the
auditors and public prosecutors of the institution, enhancing its technical profile and political
autonomy. These two members are tenured civil servants without any particular loyalty to the
governor. The senior auditors and prosecutors are recruited through a very competitive system
and they enjoy civil service status (tenure) and high salaries (their salaries are set as a percentage of the judges of the Federal High Court). These officials therefore have incentives to be
strict in the application of sanctions. Auditors are very critical of the subordination of the courts
to the whims of political actors. Senior auditors also participate in the judgment sessions with
the right to vote. Introduced in 1988, these features provide important incentives for them to be
impartial.
The constitution vested the TCs with the role of external control of public administration. They
exercise oversight over the execution of budgets. Although they can impose fines on members of
the executive and legislature should the law be breached, it is up to the independent auditors and
especially public prosecutors to press charges against the perpetrators of crimes (in about half of
the states, public prosecutors sit on the board of these institutions). The FRL mandated that the TC
must audit the enforcement of the law, by imposing procedural rules (reporting transparency, etc.)
that the Tribunal checks.
In order to produce their different types of reports a TC can take on three different forms of
activities. Firstly, they can work on the mandatory reports of accounts required by law that are
submitted every year by the state and municipal governments and legislatures, which is largely a
routine task. The second form of action is based on the self-initiative of the auditors and the board
as a result of suspicion, evidence, or information that they themselves gather about the municipalities or other administrative units. Actions in the third form are investigations, which are prompted
by whistle blowing or accusations by third parties. These accusations are usually made by municipal councilors, opposition candidates, trade unions, and citizens.
An active Tribunal is one that does much more than the minimum required. There is a huge variation in the degree of activism among audit institutions in Brazil. Activity can be measured in
terms of the ratio of the number of audit cases performed by each Tribunal and the number of
administrative units under its jurisdiction. An active TC is one that does much more than the minimum required of one report per administrative unit, especially reports resulting from the initiative
of an auditor or of third-party allegations.
Although there is considerable homogeneity in the functioning of the TCs,7 some important differences in their institutional design remain. For instance, there is great variation with regard to
organizational aspects in terms of the number of employees, infrastructure, and equipment available as well as the number of administrative units under their jurisdiction (jurisdicionados).
10
International Political Science Review 
Empirical analysis
Illustrative cases of creative accounting: a survey of practices
In this section we describe the numerous administrative practices and strategies of creative accounting used by state governments to illustrate our analysis. This compilation is based on extensive
administrative documentation, interviews with public officials and experts, along with media
reports. For descriptive purposes, we selected the top seven states in terms of recorded unpaid commitments (restos a pagar): Rio Grande do Sul, Goias, Mato Grosso do Sul, Alagoas, Minas Gerais,
Rio de Janeiro, and Paraiba, ranked according to financial deficit after the 2006 elections. We
deliberately included information on two additional units states that had financial surpluses in their
2006 balance: the state of Sao Paulo, given its political importance and strong impact on Brazil’s
GDP; and Pernambuco, for having reported the country’s highest primary surplus as a proportion
of the current net revenue (RCL) in 2006.
The great majority of these states recorded yearly financial deficits as unpaid commitments
(restos a pagar). These are the kinds of expenses the payment of which is delayed to the subsequent fiscal year, thereby deferring their impact on the primary balance. If inadequately documented over the years, financial deficits, also known as end-of-year negative cash balances, can
obscure the evolution of a state’s underlying fiscal position. For the most part, unpaid commitments have been affected by ‘creative accounting’ and consequently official information concerning the states’ readily observable end-of-term cash balances has been disputed.
To a large extent, the state governments resorted to many accounting tricks in their attempt to
record expenditures within the FRL spending limits, with most attention being given to the limit on
spending on personnel set at 60% of net current revenue. For example, some types of personnel
expenses were wrongly classified as income tax deductions, or pension payments were not reported
as personnel spending. In these cases, governors took advantage of lenient interpretations of the
state TC regarding the expenses they should record as expenses on personnel. Similar inconsistencies of interpretation plagued the calculation of the net current revenue, leading states such as
Pernambuco and Paraiba to artificially raise the indicator as a means of accommodating spending
increases, by adding the proceeds of privatization to budget revenues.
In addition, as a result of the FRL restraining overall spending in the electoral year, the states
resorted to expenditure hikes in pre-electoral years. Likewise, spending sprees at the beginning of
electoral years are a common strategy used by incumbents to circumvent the spending caps applicable for the last eight months of term. In some cases, state governments struggled to curb basic
expenses after elections in order to officially meet fiscal targets and thus avoid financial or even
criminal sanctions, as seen in Rio de Janeiro and Sao Paulo. Other forms of covering up voteearning expenses included improperly using earmarked funds, for example public health funds
shifted to finance infrastructure projects, educational funds diverted to cover campaign expenses,
and so forth.
There are many different accounting devices used as guises in the Brazilian states (a dataset on
these practices is available from the authors). We use for our qualitative analysis information extracted
from state fiscal management reports and from auditing documents released by the TCs, as well as
media exposés and information from interviews. The quantitative information has been checked
against states’ budgetary execution data published by the Brazilian National Treasury. The states
retain some ability to undertake fiscal window dressing in a response to fiscal stress despite the hard
budget constraints imposed in the context of the FRL. It should be noted that a number of Brazilian
states outside this sample, but close to balance or in surplus, also show signals of opportunistic fiscal
behavior. For instance, in Ceará and Pará, where full compliance with headline measures of primary
surplus was reported over several years, the fiscal balance suffered a sharp decline in the electoral
Melo et al.
11
Figure 2. Descriptive distribution of the average of unpaid commitments by state (2000–2002).
year of 2006, and the governor-elect in both states complained about financial deficits when they took
office in 2007. The Northern states of Amapa, Rondonia, and Roraima minimized reported deficits
through fiscal gimmickry as well. Incorrect posting of the proceeds from privatization, for example,
has helped them cover current expenditures, while increases in public works spending have been
underestimated in the approved budget, causing a rise in expenses recorded at a later time as unpaid
commitments. In addition, the expenditure limits of the legislative branch have been consistently
circumvented over the years and therefore budget deficits are understated.
Furthermore, state governors in Sergipe and Parana have taken advantage of the lack of fiscal
transparency. They released multiple versions of their fiscal reports or failed to publish their balance sheet data in the time stipulated in the fiscal legislation. Some of the fiscal tricks used have
been less elaborate than others. In any event, however, these examples demonstrate a propensity of
incumbents to distort the fiscal position as a means to conceal episodes of fiscal profligacy.
Therefore, before jumping to the conclusion that the FRL has turned Brazilian subnational politicians into highly disciplined public managers who neither overspend nor deal with public finances
irresponsibly in election times, we should take a closer look at not only the states’ headline primary
balances, but also at how politicians in office have managed to minimize the impact on the expenditure level of the numerical budget rules. There seems to be a significant margin for questionable
accounting practices in conjunction with lax legislative interpretations to beautify their reported
fiscal position.
Preliminary statistical evidence
In this section we report the result of some empirical exercises (see Table 1). Because of the small
N we are cautious in making generalizations and present the findings as illustrations of causal processes as opposed to hard evidence supporting the hypotheses advanced. The data for the restos a
pagar (unpaid commitments) as a percentage of per capita GDP comes from the state balance
sheets from 2000 to 2002.8 Figure 2 shows that during the period analyzed (but this is also true for
more recent years), there is a huge variation in unpaid commitments from one state to the next. The
state of Bahia had the smallest outstanding debt to be paid in the subsequent fiscal years and the
12
International Political Science Review 
Figure 3. Activism of audit institutions (2000).
state of Parana was the champion of unpaid commitments. What institutional and political factors
can explain this significant variation?
As the TCs must audit the enforcement of the FRL, our key explanatory variable is a measure
of activism of an audit institution at the subnational level. As proxy of the quality of the TCs we
use an index of institutional activism created by Melo et al. (2009). In fact, this variable is the ratio
between the number of audit cases performed by each Tribunal and the number of administrative
units under its jurisdiction. As mentioned before, this is a very good proxy of activism or more
appropriately ‘productivity’ (see Figure 3). The TCs are legally required to carry out audits in their
routine oversight of each unit under their jurisdiction at least once a year. However, there are cases
in which Tribunals do not deliver audits for all units (in which case the ratio would be less than
one). Therefore, the greater the ratio, the more active a Tribunal is. We expect to find a negative
correlation between restos a pagar and institutional activism of the Audit institution. That is, the
more active a TC is, the lower the amount of unpaid budgetary commitments.
One may argue that an active audit institution does not necessarily mean an institution independent of the influence of politicians. Melo et al. (2009) have already demonstrated that the key
explanatory variable explaining the degree of activism of an audit institution in Brazil is the presence or lack of a senior auditor and/or a public prosecutor on the Audit Board. That is, tribunals
with auditors on their boards are more prone to action. Therefore, we included in the model a
dummy variable with the value of one if the audit institution has a senior auditor on the board and
zero otherwise.9 We expect that the presence of an auditor on the board may deter governors from
making use of unpaid commitments.
In line with our discussion above of the role of political competition in fiscal behavior, we
included in the model the variable government turnover and legislative party fragmentation in the
State Assembly. The former variable consists of an index of elite instability in power, which was
built by taking as a reference three consecutive elections for state governors in Brazil: 1994, 1998,
and 2002.10 In line with the literature, we predict a positive coefficient for government turnover
and unpaid commitments. That is, the greater the electoral risk, the greater the incentives governors
will have to pass deficits on to their rival successors via unpaid commitments. To investigate the
role of legislative fragmentation in fiscal behavior, we have built an index of the number of
Melo et al.
13
political parties per seats in the 1999–2002 period. In line with the literature, we assume that the
greater the fragmentation, the higher the coordination cost governors face while managing their
legislative coalition. Therefore we expect fragmentation to be positively associated with creative
accounting.
We also control for the state per capita GDP because richer states tend to be in a worse fiscal
situation for a variety of reasons, including the ability to contract loans and issue debt papers.
There is, however, little evidence that the use of dubious accounting practices is influenced by the
fiscal situation of each state at the time the FRL was introduced. A correlation performed between
the domestic and external debt-to-RCL ratio in the years prior the FRL and the level of unpaid
commitments-to-RCL ratio after 2000 is not statistically significant at conventional levels (Pearson
correlation = 0.255). Therefore, states with larger deficits and debts do not seem more inclined to
adopt dubious accounting practices, contrary to what might be initially presupposed.11
Although the results generated by our empirical exercises should be interpreted with caution
given the small number of cases, we find empirical evidence of the correlation between restos a
pagar (unpaid commitments) and the levels of activism of TCs. As expected, the more independent
a TC (measured by the presence of a senior auditor on its board) is, the smaller the incentives governors will have to rely on window-dressing mechanisms (measured by unpaid commitments).
Political competition, both in the executive branch as well as in the legislative sphere, also matters
for creative accounting. That is, the higher the government turnover and legislative fragmentation
within state assemblies, the greater the incentives governors will have to make use of unpaid
commitments.
Even a fiscally constrained governor (with an active audit court that publicizes fiscal wrongdoing) might be tempted to behave opportunistically in the present, pushing potential fiscal imbalances to the future in a very competitive political environment. That is, overspending in an electoral
year while covering it up as unpaid commitment might tip the balance in close gubernatorial races.
Spendthrift governors derive utility from increased spending in their last year in office. On the
other hand, a governor facing weak political competition will be more inclined to follow fiscal
rules under the check of an active audit court, as the incentives to overspend are weaker. This substantive interpretation can be derived from a closer look at Models 1 and 2 (see Table 1). The latter
suggests that the inclusion of turnover substantially increases the goodness of fit of the model,
shown by a large increase in the r-square. Specifically, an increase of one audit is expected to
decrease, on average, creative accounting by 2.3%. Given that the average of restos a pagar in the
Brazilian states is about 17% of their total budget (see descriptive statistics in Table 2), the substantive impact of activism is not trivial. The marginal effect of turnover on unpaid commitments is
even higher. That is, one single power alternation at the state level increases the percentage of
unpaid commitments by about 5.4%. If a state, for instance, alternates its political elites in power
twice within the same time frame analyzed, creative accounting measured as a percentage of total
budget increases by about 11%.
In other words: ‘bad policies may lead to good politics’: overspending and creative accounting
may increase voter support for incumbent parties.12 The political forces working in favor of ‘creative accounting’ are stronger than the effects achieved by an efficient audit system as demonstrated
by the increase in the explanatory power when we move from Model 1 to Model 2 (see Table 1),
which includes government turnover.
On the basis of these preliminary empirical exercises we reach two major conclusions. Firstly,
although not uniformly distributed among Brazilian states, there is ample evidence of creative
accounting in the states, which in itself indicates that the influence of the TCs is binding and that
there are costs for breaching the law. Otherwise, why take the trouble to hide irregular practices
and decisions? Secondly, because the TCs are not immune to the influence of the legislators and
14
International Political Science Review 
Table 1. Determinants of unpaid commitments.
Models
(1)
Activism
(2)
–.0309**
(.0128)
–.0263***
(.0143)
.0568***
(.0174)
Turnover
Auditor
(3)
–.0229**
(.0114)
.0699***
(.0145)
-.0818**
(.0439)
Fragmentation
(4)
(5)
–.0205**
(.0103)
.0571***
(.0146)
-.0666* (.0383)
1.2362* (.7361)
GDP p/capita
Constant
N
R2
.0128***
(.0388)
26
0.0886
.1208***
(.0458)
26
0.3242
.1428**
(.0512)
26
0.3993
-.9701
(.6463)
26
0.4839
–.0235**
(.0118)
.0542***
(.0173)
-.0623*
(.0384)
1.1210*
(.7450)
5.87e-06
(.00001)
-.8841
(.6542)
26
0.4900
*Notes p < 0.05; ** p< 0.01; *** p< 0.001
Table 2. Descriptive statistics.
Variable
Observation
Mean
Standard
Deviation
Unpaid
Activism
Turnover
Auditor
Fragmentation
GDP p/capita
26
26
26
26
26
26
.1693057
1.145306
1.384615
.5384615
.9056923
4023.308
.1402841
1.351282
1.202561
.5083911
.0352247
2198.409
Minimum
.027194
.0048309
0
0
.8106667
1421
Maximum
.5282267
5.668421
4
1
.9463333
11272
state governors, there is evidence that the institutional fragilities of the TCs are associated with
more creative accounting. More independent and active institutions constrain the use of creative
accounting at the state level. Therefore, the quality of subnational institutions and the degree of
political competition influence compliance. Considering that states have seen their influence dramatically diminish in the last two decades in Brazil and the relatively small scale of creative
accounting, the overall fiscal performance of the states has not been compromised. However, the
analysis suggests that there is a potential problem looming large, especially with regard to the sustainability of the FRL in the long run. This could potentially threaten macroeconomic stability as a
whole by distorting expectations and the credibility of the fiscal conditions of the country.
Admittedly, other factors can potentially influence fiscal behavior in the case of the Brazilian
federation. The first is ideology and the partisan affiliations of governments. The second is the
presence of independent media and social capital in the states. Melo and Pereira (2013) found that
partisan affiliations played no significant role in developmental outcomes in the Brazilian federation, but that media independence and civic community was a significant factor in explaining fiscal
outcomes in the states (along with corruption and social policy outcomes). These variables were
not included in our empirical exercise because of the small number of observations and consequent
small degrees of liberty in the relevant tests.
Melo et al.
15
Although the contextual elements differ significantly, our findings are consistent with StécleboutOrseau and Hallerberg’s (2009) view of watchdogs as key players in the EU fiscal compliance
game. The main difference is that in the case of the Brazilian federation there is a very strong
external enforcer with the power to sanction fiscal misbehavior: the National Treasurer.
Conclusions
In this article, we examined some political determinants of the budget deficits that can account for
empirical regularities in the use of unpaid commitments in the Brazilian states. We found evidence
for the influence of political competition (political uncertainty proxied by turnover in the party
controlling the state government) and party fragmentation on creative accounting. Furthermore, we
also found that the quality of the fiscal watchdogs influences the degree of fiscal gimmickry. We
tested whether the prospect of not being re-elected provides state governors with an incentive to
overspend and then register the deficit as unpaid commitments. According to the literature
(Aizenman, 1998; Alt and Lassen, 2003), this would create a situation of soft budget constraints.
Our data show that a rise in the turnover rate can result in higher spending and the use of windowdressing maneuvers by state governors. The prediction is, therefore, that in the presence of electoral uncertainty, the incumbent shows above-average spending for electoral purposes and thus
needs to resort to creative accounting.
In addition, we investigated the impact of party fragmentation models on fiscal behavior
(Alesina and Perotti, 1994; Persson and Tabellini, 2001). We found that for a sample of Brazilian
states the magnitude of financial liabilities registered as unpaid commitments was positively correlated with the number of relevant political parties in the legislature. The results are in line with
the literature, in the sense that the logic of the common-pool resource problem clearly applies here,
namely opportunistic politicians in a party will seek to maximize the electoral benefits from overspending and externalize the overall cost to all other parties. Therefore, all else being equal, there
is evidence that highly fragmented state assemblies are associated with greater unpaid commitments than less fragmented ones.
The quality of the audit institution and its independence from political influence are also important factors in a governor’s decision to rely on window-dressing mechanisms. Audit activism and
political independence seem to keep opportunistic fiscal behavior at bay at the subnational level in
Brazil.
Acknowledgements
The authors thank the Brazilian National Council for Scientific Research (CNPQ) for support and Mark
Hallerberg and Carlos Scartascini, as well as participants at the Annual Meeting of the Society for New
Institutional Economics, San Francisco, 2009, for comments and suggestions on previous versions of the
manuscript.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit
sectors.
Notes
1. For Latin American experiences see Lora (2006) and Scartascini and Hallerberg (2011).
2. Poterba and Rueben (2001: 538) recognize this problem by saying that ‘the potential endogeneity of state
fiscal rules is an empirical challenge to any study of fiscal institutions and borrowing costs. It is possible
that states with tight anti-deficit rules also have fiscally conservative electorates and that the primary
factor explaining their lower borrowing costs is the electorate, not the fiscal rules’.
16
International Political Science Review 
3. Up until the mid-1990s, when state banks were privatized, the states also enjoyed great ability to bypass
central bank control because the states’ treasuries could issue debt that was purchased by the states’
banks.
4. Article I of the FRL clearly states that the law ‘is applicable to the Federal Government, the Federal
District, and the Municipalities’. It is important to acknowledge, however, that the National Fiscal
Council, which would oversee the implementation of these provisions at the federal level, has not been
implemented, and the Federal Government has been unconstrained.
5. Cardoso managed to gain approval for a constitutional amendment in 1997 allowing re-election for one
additional term of office for mayors, governors, and presidents. Most incumbent governors and Cardoso
himself ran for re-election in 1998.
6. These institutions have a number of features typical of judicial bodies, such as the right of reply, strict
procedural rites, collegial decision-making, security of tenure of their board members, and civil service
status for their employees.
7. Not only do the board members enjoy tenure security but they are also appointed until the official retirement age for public servants, currently 70 years old.
8. Unfortunately, the dataset for the activities of state audit courts covers just one year and a panel estimation was not possible. The cross-section regressions presented here, however, provide interesting evidence supporting our hypotheses.
9. We also tested for the presence of a public prosecutor on the audit board as well. However, because of
collinearities with Auditor, the variable Public Prosecutor was dropped in our econometric exercises.
10. This index varies from zero (when the same coalition was the winner in all three consecutive electoral
episodes) to four (when no one single coalition was able to win two elections). The intermediate values
of the vulnerability index refer to situations when an electoral coalition won two consecutive elections
but lost the third one (index equal to one), or when the first elite coalition in power is defeated and a new
elected elite wins the following two elections (index equal to two) and finally a situation in which an elite
group has its electoral dominance interrupted by a second electoral elite that gains power for just a single
electoral period, after which the original elite returns to power (index equal to three).
11. We are grateful to an anonymous reviewer for calling our attention to this potential correlation.
12. We appreciate the suggestion by an anonymous reviewer.
References
Aizenman, Joshua (1998) Fiscal discipline in a union. In: Federico Sturzenegger and Mariano Tommasi (eds)
The Political Economy of Reform. Cambridge, MA: MIT Press, pp.185–208.
Alesina, Alberto and Roberto Perotti (1994) The political economy of budget deficits. NBER Working Paper
4637.
Almeida, Mansueto (2011) Nota Técnica: Restos a Pagar e Artifícios Contábeis. Brasília: IPEA.
Alt, James and David Lassen (2003) Fiscal transparency and fiscal policy outcomes in OECD countries.
EPRU Working Paper 03–02.
Alt, James E and Robert C Lowry (1994) Divided government, fiscal institutions, and budget deficits:
Evidence from the states. American Political Science Review 88(4): 811–828.
Ayuso-i-Casals, Joaquin Servaas Deroose, Elena Flores and Laurent Moulin (2009) Policy Instruments for
Sound Fiscal Policies: Fiscal Rules and Institutions. London: Palgrave Macmillan.
Bernoth, Kerstin and Gumtram Wolf (2006) Fool the Market? Creative accounting, fiscal transparency, and
sovereign risk premia. Economic Studies Discussion Paper 1(19).
Braun, Miguel and Mariano Tommasi (2004) Fiscal rules for subnational governments: Some organizing
principles and a case study of Argentina. Paper presented at the IMF/World Bank conference, Oaxaca,
Mexico.
Cossio, Francisco A (2001) O Comportamento Fiscal dos Estados Brasileiros e seus Determinantes Políticos.
Economia 2(1): 209–258.
De Mello, Luiz and Diego Moccero (2006) Consolidating macroeconomic adjustment in Brazil. ECO
Working Paper 531.
Melo et al.
17
Gama Neto, Ricardo (2007) Refém, Cenoura, ou Porrete: Federalismo, Comportamento Fiscal e Ciclo
Político. PhD Dissertation in Political Science, Federal University of Pernambuco.
Garcia, Marcio (2010) Brazil: Creative accounting and fiscal risk. Economonitor. Available at: http://www.
economonitor.com/blog/2010/10/brazil-creative-accounting-and-fiscal-risk/
Giambiagi, Fabio (2007) Dezessete Anos de Política Fiscal no Brasil: 1991–2007 (seventeen years of fiscal
policy in Brazil: 1991–2007). Texto para discussão. Brasília: IPEA, p.1309.
Goldfajn, Ilan and Eduardo Guardia (2004) Fiscal rules and debt sustainability in Brazil. In: George Kopits
(ed.) Rules-Based Fiscal Policy in Emerging Markets: Background, Analysis and Prospects. London:
Palgrave Macmillan, pp.114–130.
Hallerberg, Mark, Carlos Scartascini and Ernesto Stein (2009a) Who Decides the Budget: The Political
Economy of the Budget Process in Latin America. Cambridge, MA: IADB and Harvard University Press.
Hallerberg, Mark, Rolf Strauch and Jurgen von Hagen (2009b) Fiscal Governance: Evidence from Europe.
Cambridge: Cambridge University Press.
International Monetary Fund Fiscal Affairs Department (2012) IMF fiscal rules dataset. Washington, DC:
IMF.
Irwin, Timothy (2012) Accounting devices and fiscal illusions. IMF Staff Discussion Note, March 2012.
Koen, Vincent and Paul van den Noord (2004) Fiscal gimmickry in Europe: One-off measures and creative
accounting. OECD Economics Department Working Paper 417.
Kopits, George (2001) Fiscal rules: Useful policy framework or unnecessary ornament. IMF Working Paper
01(145).
Kopits, George (ed.) (2004) Rules-based fiscal policy in emerging markets: Background, analysis and prospects. London: Palgrave Macmillan.
Liu, Lili and Steven B Webb (2011) Laws for fiscal responsibility for subnational discipline international
experience. World Bank Policy Research Working Paper 5587.
Lora, Eduardo (ed.) (2006) The State of State Reform in Latin America. Stanford, CA: Stanford University
Press.
Melo, Marcus, Carlos Pereira and Carlos Figueiredo (2009) Political and Institutional Checks on Corruption:
Explaining the Performance of Brazilian Audit Institutions. Comparative Political Studies 42 (9): 1217–
1244.
Melo, Marcus André and Carlos Pereira (2013) Making Brazil Work: Checking the President in a Multiparty
System. New York: Palgrave Macmillan.
Milesi-Ferretti, Gian (2003) Good, bad or ugly? On the effects of fiscal rules with creative accounting. Journal
of Public Economics 88(1–2): 377–394.
Pellegrini, Josué (2012) Como evoluiu a dívida estadual nos últimos dez anos. Instituto Braudel Technical
Note.
Persson, Torsten and Guido Tabellini (2001) Political institutions and policy outcomes: What are the stylized
facts? CESifo Working Paper Series 459.
Posner, Paul and Jón Blöndal (2012) Democracies and deficits: Prospects for fiscal responsibility in democratic nations. Governance: An International Journal of Policy, Administration, and Institutions 25(1):
11–34.
Poterba, James (1994) State responses to fiscal crises: The effects of budgetary institutions and politics.
Journal of Political Economy 102(4): 798–821.
Poterba, James M and Kim S Rueben (2001) Fiscal rules, state budget rules, and tax-exempt bond yields.
Journal of Urban Economics 50(3): 537–562.
Rodden, Jonathan, Gunnar Eskeland and Jennie Litvack (2003) Fiscal Decentralization and the Challenge of
Hard Budget Constraints. Cambridge, MA: MIT Press.
Scartascini, Carlos and Mark Hallerberg (2011) Economic crisis and fiscal reforms in Latin America. Working
Paper Series IADB-WP 235.
Souza, Celina (2007) Coalizões eleitorais e ajuste fiscal nos estados brasileiros. Revista Brasileira de Ciencias
Sociais 22(63): 31–53.
Souza, Saulo (2011) The Cold Austerity of Fiscal Rules: Can they really resist the Elections Heat? Saarbrücken:
Lambert Academic Publishing.
18
International Political Science Review 
Speck, Bruno Wilhelm (2000) Inovação e rotina no Tribunal de Contas da União: O papel da instituição superior de controle financeiro no sistema político-administrativo do Brasil. São Paulo: Editora
Konrad Adenauer.
Stéclebout-Orseau, Eloise and Mark Hallerberg (2009) Who provides signals to voters about government
competence on fiscal matters? The importance of independent watchdogs. In: Joaquin Ayuso-i-Casals,
Servaas Deroose, Elena Flores, et al. (eds) Policy Instruments for Sound Fiscal Policies: Fiscal Rules
and Institutions. London: Palgrave Macmillan, pp.241–255.
Von Hagen, Jurgen and Guntram B Wolff (2006) Fiscal rules and fiscal performance and the EU and Japan.
GESY discussion paper 147.
Webb, Steven B (2004) Fiscal responsibility laws for subnational discipline: The Latin American experience.
World Bank Policy Research Working Paper 3309.
Author biographies
Marcus André Melo is Professor of Political Science at the Federal University of Pernambuco (Brazil). He
was a Fulbright scholar at MIT and a Visiting Professor at Yale University. He is a co-author of Against the
Odds: Politicians, institutions, and the struggle against poverty, Making Brazil Work: Checking the president
in a multiparty system, and Beliefs, Development and Critical Transitions, Brazil 1964–2014. His articles
have appeared in scholarly journals including the Journal of Democracy, The Journal of Comparative
Economics, Legislative Studies Quarterly, International Political Science Review, and Comparative Political
Studies.
Carlos Pereira is Professor of Political Science and Public Policy at the Getulio Vargas Foundation (Brazil).
He has held visiting research fellowship positions at the Brookings Institution and Oxford University, and was
an Assistant Professor of Political Science at Michigan State University. He is a co-author of Making Brazil
Work: Checking the president in a multiparty system and of Beliefs, Development and Critical Transitions,
Brazil 1964–2014. His articles have appeared in journals such as the Journal of Politics, Legislative Studies
Quarterly, Journal of Latin American Studies, and Comparative Political Studies.
Saulo Souza is a senior researcher at the Federal University of Pernambuco and an Adjunct Professor at Asces.
He is the author of The Cold Austerity of Fiscal Rules.