Bizim krizlerimize çok benzer Latin Amerika ülkelerinin ve özellikle Brezilya’nın krizleri. Türkiye
ekonomisi ile Brezilya ekonomisi arasındaki birçok önemli farka rağmen krizlerimizin kök nedeninin
iki ülkede de aynı olduğu söylenebilir.
Gerçekten öyle midir?
Dünya Bankası ile İMF’in hazırladığı ülke raporunun Yönetici Özeti ülkenin son krizini nedenleri ve
çözüm yolları ile birlikte her yönü ile analiz ediyor.
1. Brazil spends a lot more than the country can afford and on top of this spends poorly. This
is the key conclusion of this study, which analyses the roots of Brazil’s recurrent fiscal
problems and presents options for their resolution.
2. For the past two decades, Brazil has experienced a consistent increase in current spending
which is now putting fiscal sustainability at risk. In recent years, the impact of this steady
increase in expenditures was aggravated by the drop in revenues resulting from a deep
recession and by the increase in tax expenditures. This has resulted in annual fiscal deficits
of above 8 percent of GDP in 2015-2016 and an increase in public debt from 51.5 percent of
GDP in 2012 to more than 73 percent in 2017. The fiscal adjustment needed to stabilize public
debt in the medium terms is large—about 5 percent of GDP in the primary fiscal balance.
Failure to stop this trend would result in a loss of investor confidence (domestic and
international), exchange rate depreciation, and high inflation, bringing Brazil back to the
type of macroeconomic crisis it experienced in the 1980s and early 1990s. Further, beyond
fiscal stabilization, Brazil needs to create additional fiscal space to increase investment in
infrastructure. Currently, public spending is not only above what Brazil can afford, but it also
does very little to support growth as public investment was almost completely eliminated in
recent years.
3. The constitutional expenditure limit (“teto dos gastos”) adopted in December 2016
introduced a gradual adjustment path for government spending over the next ten years. Its
implementation requires a reduction in spending of around 0.6 percent of GDP every year relative to the current trend for the next decade. This corresponds to a nearly 25 percent
cumulative reduction in federal primary expenditures (as a share of GDP), which would
reduce the size federal budget (as a share of GDP) to the levels of the early 2000s. Finding
these savings will be challenging as budgetary rigidities, ample spending mandates and
budget cuts in the past few years have already greatly reduced
discretionary spending. In other words, the “teto” can only be met through rigorous
prioritization of expenditures.
4. This report aims to demonstrate how such prioritization can be achieved in a way that
protects the poor and vulnerable and minimizes negative impacts on employment and the
provision of public services. To do so, the study takes a deep look at existing programs and
spending allocations and identifies reforms that would make public spending more effective,
efficient and equitable.
5. Reduced spending is not the only strategy to restore fiscal balance but it is a necessary
condition. Some have suggested that instead of cutting spending, Brazil should increase tax
revenues and reduce the high interest payments on its public debt. Certainly there is scope
to expand the taxation of high income groups (e.g. through income, property, and capital
gains taxes) and to reduce the reliance on indirect taxes which places a heavy burden on the
poor. Additional gains in the fiscal balance could be made by reducing quasi-fiscal operations
via public banks, and addressing costs related to the management of public debt and foreign
reserves. Such measures are not discussed in detail in this report but should still be part of
the fiscal adjustment strategy. However, these additional measures can be no substitute to
tackling the fundamental causes of the increase in current spending and reviewing the
associated excessive responsibilities and mandates of the Brazilian state. The failure to do
so will inevitably lead to new fiscal crises and painful adjustment down the road.
6. The single most important source of long-term fiscal savings is pension reform. The large
and increasing deficits of the pension system constitute a key source of fiscal pressure. It is
essential to adjust the pension system to the reality of rapid demographic change and to
bring it in line with international standards. Further the current pension system is
inequitable in that 35 percent of pension subsidies (i.e., the deficit of the pension system)
benefit the richest 20 percent of the population, while only 18 percent of this subsidy benefit
the poorest 40 percent of the population. Approval of the reform proposal currently in
Congress would be a significant step towards correcting this imbalance. World Bank
projections indicate that the reform negotiated in Congress in May 2017 would halve the
projected deficit of the RGPS pension system (for private sector workers) over the next few
decades—from 16 percent of GDP to 7.5 percent by 2067. Over the next ten years, the reform
would bring approximately one-third of the required fiscal savings under the “teto”
(reaching 1.8 percent of GDP in 2026). The reform is socially fair as it would mostly reduce
the subsidies to workers who earn above 3 minimum wages.
7. Additional measures will be required to make the pension system financially sustainable
and more equitable. Even with the reform, large pension deficits are projected to remain.
The proposed reform also does not sufficiently address the deficit in the federal RPPS
pensions system (for federal civil servants), which results from the extremely generous—and
highly inequitable—benefits awarded to civil servants who joined before 2003. Similarly, the
reform does not solve the deficits of the RPPS pensions systems of subnational
governments, which are projected to increase sharply in the next 5 to 10 years as many pre2003 civil servants retire. This will jeopardize many states’ fiscal solvency. Therefore,
additional measures will be required to align the benefits of the RPGS and RPPS to the level
of contributions.
8. To address the remaining deficits in the RGPS, the following measures should be
considered:
– Further reducing replacement rates, since even with the adoption of the current reform
replacement rates will still be high by international standards. Reducing the replacement
rate
10
–
by another 20 percentage points would reduce RGPS deficits by 1.8 percent of GDP in the
medium term;
Delinking the minimum pension from the minimum wage (and correcting only for cost of
living increases) would have the most significant impact to reduce the fiscal deficit
compared to other policy actions. Keeping minimum pensions constant in real terms, would
achieve an additional reduction in the RGPS deficit (urban and rural) of 0.5 percent of GDP by
2027 and 2.3 percent by 2067.
Deficits in the RPPS and the fairness of the pension system overall should be addressed
9.
by removing the privileges awarded to pre-2003 civil servants. Civil servants who joined
before 2003 are currently entitled to an extremely generous pension package, in which the
value of benefits
1
far exceeds their level of contributions.
the future, in the short term there is a large deficit remaining due to the pre-2003 system.
This deficit will weigh on taxpayers for the next couple of decades. In 2015, the estimated
cost for the tax payer was 1.2 percent of GDP at the federal level and a further 0.8 percent of
GDP at subnational level. Since the majority of civil servants are among the richest 40 percent of the population, their pension benefits imply a large transfer from current tax
payers, the majority of whom are much poorer than then RPPS beneficiaries. Further, the
benefits of the pre-2003 cohort of civil servants imply a generous transfer from the young to
the older generation. In other words, there is an inherent social injustice in the concept of
acquired rights (direitos aquiridos) as applied to these benefits. Given the depth of Brazil’s
fiscal problems and the huge size of its current and expected future pension deficits, all
generations should contribute to the necessary adjustment. This could be achieved, for
instance, by taxing the pensions of existing high income retirees and recipients of survivor
pensions.
The result is that, while the RPPS post-2003 will balance in
10. A further reform which should be considered is to recognize that the RGPS rural pensions
and social pensions (BPC) are in fact social programs, and to reform them accordingly. Both
of these pension systems are de-facto non-contributory, and their objective is to prevent the
elderly from falling into poverty. However, unlike the Bolsa Família, program which is well
targeted to the poor, rural retirement benefits and social pensions are not well targeted to
the poor. In fact, 70 percent of the beneficiaries of BPC and 76 percent of beneficiaries of
rural pensions do not belong to the poorest 40 percent of the population. This is largely
because the level of benefits provided by these programs is so much higher than that of
other social assistance programs—the maximum benefit provided by Bolsa Família, is about
one-third of that provided by BPC and rural pensions. A reform to consolidate social
pensions with other social assistance programs (by aligning their targeting and levels of
benefits, as discussed below) could generate significant savings through better targeting and
potentially free up resources to spend on urgent social needs such as access to water and
sanitation, early childhood education, or old age care.
11. The civil service wage bill can be significantly reduced. While the size of the Brazilian civil
service is not large by international standards, the level of wages of federal civil servants is
on average 67 percent higher than in the private sector, even after controlling for education
and other relevant factors such as age and experience. The remuneration of state level civil
servants is also very high, and on average more than 30 percent above the relevant private
sector equivalent. In relative terms, the wage gap appears especially large in the judiciary
and legislative branches, as well as at lower levels of the executive branch.
1 For instance, for teachers the net benefit (expected pension payments minus total lifetime
contributions) is roughly equivalent to 300 minimum wages for a pre-2003 contract and
around 30 minimum wages for a post-2003 contract.
11
12. The reduction of these exceptionally high wage premiums for federal civil servants would
also be desirable from an equity standpoint. The vast majority (83 percent) of federal civil
servants are among the richest quintile of the population. Estimates suggest that halving the wage premium with respect to the private sector would provide savings of 0.9 percent of
GDP. These initial findings suggest the need for an in-depth study to compare the
remuneration in the public sector with that in the private sector. However, the evidence is
sufficient to recommend a freeze in nominal wages of federal civil servants in the short term
while a more detailed study is carried out on the adequate level of remuneration for the
diverse set of public service careers.
13. Improving methods for public procurement of goods and services would produce savings
across the government. Public procurement is a critical step in the provision of public
services to the population, be it in health care, education, infrastructure, etc. An analysis of
public procurement by the federal government over three years (2012 to 2014), covering
about R$155 billion (or about 5 percent of the annual federal government budget on
average) indicates that the federal government could save between R$24 billion and R$35
billion in three years (or 0.15 to 0.2 percent of GDP annually; approximately 1 percent of the
federal budget) through the introduction of customized strategies for public procurement.
The lower bound (of R$24 billion) is under a conservative scenario based on the development
of better purchasing strategies to achieve economies of scale, greater competitiveness,
diversification of suppliers, minimization of the effects of seasonality on prices, among other
strategies that can be implemented in the purchasing phase. The higher figure (of R$35
billion) corresponds to an aggressive scenario that would also require demand management,
reducing waste, replacing materials and services, and standardization of products of lower
complexity. All proposed savings do not require modifications to the laws of bidding and
contracts or the drafting of new laws. All proposals are based solely on procurement
planning and strategy following the existing policy framework.
14. Spending on policies to support businesses has grown rapidly, reaching 4.5 percent of
GDP in 2015; yet there is no evidence that the existing programs have been effective and
efficient in their objective to boost productivity and sustainable job creation. On the
contrary, these programs are likely to have negative consequences on competition and
productivity in Brazil. Much of this spending is off-budget, resulting from tax exemptions
and provision of subsidized credit through public banks. It would be crucial to conduct
robust evaluations of these programs to inform the public debate and redesign policies.
Preliminary analysis suggests that up to 2.0 percent of GDP in federal spending could be
saved (or reallocated) over the next decade by eliminating (or reformulating) the least
efficient programs:
– The largest program in this area, SIMPLES, accounts for about 1.2 percent of GDP in lost tax
revenues. It would benefit from reform to make it less costly and more effective in
encouraging the formalization and growth of the most productive firms, and greater job
creation.
– Government proposals to roll back payroll tax exemptions (Deshoneracao da Folha) would
generate savings of up to 0.4 percent of GDP. Several studies indicate that this program has had little to no impact on employment, and that the cost of the few jobs created (or
protected) was very high, at more than 3 times the workers’ salary.
– The Inovar-Auto program does not appear to be effective, it costs a lot to domestic
consumers, and has been found to violate WTO rules. The program should be reformed,
focusing on removing anti-export bias and building the capabilities of suppliers (instead of
focusing primarily on the internal market and the final assembly by large automotive
companies). This
12
would generate large benefits to consumers, and a possible additional 0.03 percent of GDP in
fiscal savings.
The tax exemptions provided to the Zona Franca of Manaus, which costs the equivalent of
0.38 percent of GDP, also appear to be inefficient and should at least be reformulated to
effectively support the local economy.
Some savings are already being realized on subsidized credit, since the PSI program was
stopped in 2015. Studies have found that the PSI subsidized credit program has had little
effect in supporting investment and productivity growth. Instead the program appears to
have introduced distortions, since it largely benefited old and unproductive firms. The costs
of the PSI (associated with existing loans at subsidized rates) will continue to weigh on public
finances for years to come—it is expected to cost approximately 0.4 percent of GDP in 2018,
gradually tailing off to 0.1 percent of GDP by 2026.
–
–
15.
incentives and focus on the most vulnerable populations. The social protection system in
Brazil encompasses almost the full gamut of programs that are offered in high income
countries to help households manage risks across the life cycle. However, this study finds
that similar or even better results could be achieved with lower spending: many programs
with similar functions are deployed as if in isolation, which result in millions of families
being entitled to multiple benefits. Better policy coordination and more integrated
administration and delivery could reduce overlaps, save resources and realign incentives to
increase productivity of the labor force. The overall recommendation is to redesign and
integrate social protection programs (labor market programs, social assistance, and social
pensions) into a coherent system that better leverages Brazil’s capacity to target households
in need and thereby achieves fiscal savings. The study recommends to reshape all social
programs into three complementary parts, generating savings of up to 1.3 percent of GDP
over the next decade:
Social protection and labor programs would benefit from reforms that better align
(a) Consolidated social assistance: This would entail redesigning and integrating all de-facto
non- contributory cash transfer benefits—including social pensions (BPC and rural pensions),
social assistance, and Salário Familia—into a consolidated program modelled along best
practice examples from OECD countries. Building on the well performing Bolsa Família
program, a consolidated social assistance benefit could ensure progressive spending,
internally coherent eligibility criteria, and incentive-compatible benefit withdrawal rules
that stimulate formal labor market participation. Under such a program, all poor individuals
would be eligible to no more than one means-tested transfer. The potential fiscal savings
would depend on eligibility and benefit levels (and the government’s ability to detect errors
by ensuring the interoperability of administrative data), but could reach up to 0.7 percent of
GDP without raising the poverty rate. A short-term measure compatible with this broader
reform objective would transform Salário Família, into a means-tested benefit at the
household level (leveraging the capacity of Cadastro Unico) in order to incentivize Bolsa
Família, beneficiaries to transition into formal jobs. While not generating short-term fiscal
savings, the expected positive effects on labor supply and the productivity of formal workers
would generate long-term economic and social benefits.
(b) Targeted wage subsidies: Abono Salarial could be transformed into a proper wage subsidy
paid to the employer as an incentive to hire (i) the long-term unemployed, or (ii) first-time
job seekers in the formal labor market (mostly young people). This reform would reduce the
number of beneficiaries of the wage subsidies. Some of the fiscal savings could be reinvested
to other, currently-underfunded active labor market programs, such as training programs
and labor market
13
16.
intermediation that are more effective at improving matches and thus foster productivity.
Overall no significant fiscal savings are expected from this measure. However, resources
would be spent more effectively and better target the most vulnerable workers. Also, the
reform would augment Brazil’s capacity to provide labor market intermediation services and
job search support, which could result in productivity gains from employing labor that was
otherwise unengaged.
(c) Integrated income support for the unemployed: Integrating FGTS and Seguro Desemprego
into a sequenced set of income support instruments for the unemployed could reduce public
expenditures on unemployment insurance by more than 95 percent. Following international
best practices, the unemployed would become eligible to Seguro Desemprego only once their
FGTS account balance is exhausted, while payments of the latter would be capped to a
monthly maximum that is aligned with reasonable replacement rates (for instance 70
percent of the past wage). This reform would also generate the fiscal space to lengthen the maximum payment period of seguro desemprego (e.g. up to 7 months) for the minority of
workers that really need it, especially during economic downturns. The reform would
eliminate much of the moral hazard currently leading to excessive turnover, which is
created by how the two programs interact. An increase in remuneration of the FGTS balance
to market rates should be part of this reform. The reform would also stimulate employment
and bring fiscal savings of approximately 0.6 percent of GDP. Moreover, the fines that
employers pay for involuntary separations could be used to increase the funding for labor
market intermediation and job search support services.
Public spending on primary and secondary education suffers from significant inefficiencies,
such that the same level of services could be provided while saving 1 percent of GDP at the
local level. Inter-municipal efficiency analysis shows that the current performance in
education services could be maintained with 37 percent less resources spent on primary and
47 percent less spent on secondary education, corresponding to savings of approximately 1
percent of GDP. Low student-teacher ratios drive a large part of the inefficiency (39 percent).
Increasing the number of students per teacher by 33 percent in primary and 41 percent in
secondary schools would save R$22bn (0.3 percent of GDP) per year. This could be achieved
simply by natural attrition: not replacing every teacher who retires in the future until
reaching an efficient student-teacher ratio. Further improvements can be achieved by
reducing teachers’ absenteeism and increasing the share of teachers’ time spent teaching.
While reducing inefficiencies will generate savings in aggregate, the analysis also shows that
for some municipalities large gains in quality could be achieved with small increments in
spending, while in many others more spending would only imply even greater waste.
17. The constitutional linkage of education spending to 25 percent of municipal tax revenues
may be one of the main causes of expenditure inefficiency. Richer municipalities, with high
fiscal revenues per student, tend to be far less efficient than poorer municipalities. Thus, it
is likely that in order to comply with constitutional rules, many wealthy municipalities are
forced to spend resources that do not necessarily increase learning. This is worrying given
the drastic demographic transition the country is undergoing. With the rapid decline in the
fertility rate to below 1.8, the number of students has been falling rapidly in many
municipalities, especially in elementary school. As this fall in the number of students is not
necessarily associated with a fall in fiscal revenues, this implies that to comply with the law,
many municipalities are obliged to spend more per student, even if this additional
expenditure is not necessary. If this additional expenditure per student does not result in
improved learning, this would explain the greater inefficiency of richer municipalities.
18. Spending on higher education is both inefficient and very regressive—a reform of the
system could save 0.5 percent of GDP in the federal budget. The federal government spends
14
approximately 0.7 percent of GDP on federal universities. Efficiency analysis suggests that
approximately one fourth of this money is wasteful. This is also reflected in the fact that the level of spending per student in public universities is two to five times higher than that
spent in private universities. Limiting the funding allocated to each university based on the
number of students would enable savings of approximately 0.3 percent of GDP. Further,
while students do not pay any tuition fees to attend federal universities, more than 65
percent of the students in these universities come from families among the richest 40
percent of the population. Hence spending on federal universities constitutes a subsidy for
the richer part of Brazilian society. Since there are high private returns from obtaining a
university degree (in terms of higher future salaries), most countries charge students for
attending public universities and offer public loans which can be repaid out of the students’
future income. Brazil already provides student loans for attendance of private universities,
under the FIES program. There is no clear rationale why the same model is not applied to
public universities. Extending the FIES to federal universities could be combined with
providing free tuition scholarships for the students from the poorest 40 percent of the
population (currently 20 percent of all students in federal universities, and 16 percent of all
university students), by expanding the PROUNI program. Together, these reforms would
improve equity and would save at least 0.5 percent of GDP in the federal budget.
19. In the health sector, approximately 0.3 percent of GDP could be saved by efficiency
improvements at the local level, while maintaining the same level of health services, and
another 0.3 percent with the end of income tax deductions for private health expenditures.
Comparing efficiency levels across municipalities, the analysis in this study identifies
possible efficiency gains of 37 percent in primary care (potential to save R$9 billion) and 71
percent in secondary and tertiary care (potential to save an additional R$12 billion). In total,
this amounts to potential savings of 0.3 percent of GDP. The observed inefficiency is mainly
due to the fragmentation of the public health system, and notably the high number of smallsized hospitals, which does not allow for economies of scale in service provision. There is
also a lack of system integration and inadequate incentives for providers and patients to
choose the most cost-effective treatment. Much could be gained, for instance, by identifying
and treating non-communicable diseases before they become hospital cases. In addition, tax
expenditures are large and highly regressive, notably the tax expenditures on income tax
(IRPF) deductions for the cost of private health insurance could be removed leading to
revenue gains of 0.3 percent of GDP. Efficiency improvements in health care will be critical to
ensure the sector can absorb the expected increases in cost associated with demographic
change.
20. In sum, based on in-depth analysis of sectoral policies, this study identifies at least 7
percent of GDP in potential fiscal savings at the federal level by 2026 (Table 1). This includes
1.8 percent of GDP from the pension reform proposal (as negotiated in Congress in May,
2017), 0.9 percent of GDP from reductions in the wage bill of federal public servants, 0.2
percent of GDP from efficiency gains in public procurement, 1.3 percent of GDP from
rationalization of social assistance and labor market programs, 2 percent of GDP from
reductions in subsidized credit and in tax expenditures to support business, 0.3 percent in
tax credits for health, 0.5 percent from reforms in financing for higher education; and a further 1.3 percent of GDP from reforms to improve efficiency in health and education,
which however accrue to subnational governments. The available data allowed us to
quantify potential savings from the implementation of selected policy reforms, but these
are by no means exhaustive. These savings can support fiscal consolidation or be reallocated (within the same sectors, or across sectors).
21. Realizing the savings identified in this study will require changes to current budget rules
and rigidities. The reforms outlined in this report assume that these rules and associated
institutional
15
arrangements can be changed. Indeed, unless they are, Brazil will not comply with the
expenditure ceiling and risks jeopardizing the current incipient recovery and instead return
to macroeconomic and fiscal crisis. Some of the proposed measures can be achieved without
changes in legislation, while others require more complex and comprehensive reforms. All of
them are feasible over the medium- term, but some may require additional technical work
to determine policy details and sequencing of their implementation.
22. Finally, beyond the current fiscal adjustment, improving the quality of public spending
requires the institutionalization of a system of regular and rigorous policy monitoring and
evaluation. Monitoring should provide focus on results and connect outcomes to budgetary
inputs. Evaluation can provide evidence for choosing the most efficient policy solutions to
achieve objectives and therefore should guide decisions on which programs to cut and which
to maintain or expand. The evaluation of spending efficiency cannot be a one-time exercise
but needs to become an institutionalized process that guides policy making. Many countries
have adopted an institutional framework to ensure the ex-ante screening of new program
proposals and the regular monitoring and evaluation of existing programs (see Boueri
Miranda et al., 2015). Beyond the institutional framework, the regular monitoring and
evaluation of spending efficiency will require greater availability of administrative data.
Currently, limited access to data, especially on tax expenditures (from the Ministry of
Finance) and social programs’ incidence (from IBGE and the Ministry of Social Development),
limits the availability of rigorous evidence that is needed to make informed policy decisions.
Providing access to researchers outside of government—which is standard practice in most
OECD countries— could enhance credibility and the quest for efficiency.
16
