What countries with high human development can teach us
The welfare state, prosperity and public debt. What seven countries did, how they did it and what Brazil can learn.
7 countries · 24 chapters
Read the full studyCountries
Good practices from other countries: how they work, what they deliver and where their limits lie.
Good practices in focus
Choose a country
- EstoniaConnected services and traceable data use.
- SingaporeFewer forms and rules for preserving reserves.
- South KoreaEvaluate before building and monitor procurement.
- CanadaProposals with explicit costs and housing with support.
- United StatesFund prototypes and compare policies using evidence.
- PortugalStarting a company without visiting multiple offices
- NigeriaCan a grant help small businesses hire?
- TogoLearning to act when obstacles arise
- BeninDo more registrations mean better businesses?
- BangladeshEscaping poverty takes more than credit
- ItalyEconomic activity with an inclusion mission
- FranceLocal work for people facing long-term unemployment
- UruguayComputers delivered or learning demonstrated?
- ChileOpening public procurement to smaller firms
- Costa RicaA team accountable for care in each area
- SwitzerlandLearning at school and at work
- SpainCompeting with worker participation
- United KingdomProtecting the mission as a business grows
- Swedensocial protection keeps pace with economic transformation
- Denmarkproductive capacity, security and local delivery
- Norwayturning oil into lasting wealth without displacing the real economy
- Finlandbuilding skills and a public administration that reaches people
- Icelandrights require reserves and protection against crises
- New Zealandopening the books, holding managers accountable and correcting management
- Australiacombining services, a public safety net and lifelong saving
An applied Brazilian case
Ceará: when incentives, technical support and evaluation work together
How can municipal learning outcomes improve without turning a single indicator into automatic policy?
Ceará offers a useful case because it combines a fiscal incentive rule, technical assistance, learning assessment and municipal autonomy. The aim here is not to copy a formula, but to reconstruct the mechanism, the evidence, the costs that still need to be established and the conditions that could change the decision.
A study of an existing policy. This is neither a ready-made OpenGov3 proposal nor a recommendation to reproduce the rule in another state.
2007–2009
Law No. 14.023/2007 linked part of the distribution of ICMS revenue to municipalities to educational quality; the financial effects began in 2009. The formula was changed later.
From 2007 onwards
PAIC organised cooperation between the state and municipalities, with training, materials, assessment and support for school systems.
Expansion
The initial focus on literacy was extended to other stages, and SPAECE began supporting monitoring in the 2nd, 5th and 9th grades.
Today
The policy continues to evolve. IPECE publishes indices and legal changes, so the historical rule should not be presented as the current formula.
The architecture: four mutually reinforcing parts
1. Fiscal incentive
The historical design redistributed part of existing municipal revenue according to educational outcomes. This changes incentives among municipalities, but does not automatically create a new state education allocation.
2. Delivery capacity
The state provided technical support, training, materials and cooperation structures. Incentivising an outcome without providing the means to achieve it can widen capacity gaps.
3. Regular measurement
SPAECE and diagnostic assessments help identify learning outcomes and guide interventions. Testing serves management only if participation, comparability and distributional consequences are examined.
4. Municipal autonomy
The provision of basic education remains strongly municipal. The case combines local autonomy with state-level coordination, rather than replacing all municipal decisions with a central operation.
Source 1: Ceará: Law No. 14.023/2007 (opens in a new tab)Source 2: IPECE: municipal share of ICMS (opens in a new tab)Source 3: SEDUC Ceará: PAIC Integral (opens in a new tab)Source 4: SEDUC Ceará: SPAECE (opens in a new tab)
What the evidence allows us to conclude
Boundary-based evaluation
An evaluation summarised by the World Bank associates the results-based financing mechanism with a gain of around 0.15 standard deviations in Portuguese and Mathematics in the 9th grade for pupils exposed to the incentive. The report also points to larger effects when incentives and pedagogical support are combined.
Limit: The design identifies effects at specific margins and in specific periods. It does not show that any reward formula would produce the same result elsewhere.
Synthetic control
Ponne (2023) finds results consistent with an increase of around 12% in Portuguese in the early years of basic education and 6.5% in its later years when incentives and technical assistance are considered together. In Mathematics, the estimated increases were not statistically significant.
Limit: Test percentages and standard deviations are different metrics. They should not be added together or converted into a single “effect score”.
Literacy in the early school years
Costa and Carnoy (2015) use triple differences and find a positive effect of PAIC on Portuguese and Mathematics, but no reduction in the gap associated with prior attendance in early childhood education.
Limit: A positive average effect can coexist with persistent inequalities. The analysis needs to track who improves less.
Source 1: World Bank: results-based financing and Ceará (opens in a new tab)Source 2: Ponne (2023): Better Incentives, Better Marks (opens in a new tab)Source 3: Costa & Carnoy (2015): early grades literacy intervention (opens in a new tab)Source 4: World Bank: From Bad to Best / Ceará education reforms (opens in a new tab)
Three alternatives for thinking through the mechanism
Reward results alone
This can create focus and accountability, but municipalities with lower capacity may fall behind. The Ceará case suggests that incentives without technical support are an incomplete version of the mechanism.
Provide technical support alone
This can strengthen capacity, but the evidence here does not provide a clean test of “technical support alone” that would establish this alternative's superiority.
Combine incentives, support and evaluation
This is the architecture closest to the observed case. It still requires weights, equity safeguards, costs, data quality and a way to revise the rule when perverse incentives emerge to be defined.
Who does what, and which costs still need to be established
- State
- Sets state-level rules, coordinates cooperation and provides technical assistance and assessment tools within its remit.
- Municipalities
- Manage municipal school systems and implement local teaching and management policies.
- IPECE
- Calculates and publishes allocation indices under the applicable legislation.
- SEDUC / SPAECE
- Produces assessments and technical support that inform diagnosis and management.
Fiscal redistribution
The historical incentive changes how part of the municipal share of ICMS is distributed. It is incorrect to treat it alone as either new state expenditure or fiscal savings.
Ongoing operation
Training, regional teams, assessments, materials, systems, visits and support have recurring costs. The sources consulted do not provide a current consolidated account of the arrangement's total cost.
Transition and adaptation
Another state would incur costs for data, legal design, training, implementation and corrections. Importing Ceará's percentages or weights without that calculation would be a hypothesis, not a budget.
When should an adaptation be reconsidered?
- Do not reward gains in averages achieved by excluding pupils with greater difficulties or reducing participation in assessments.
- Separate the effect of incentives, the effect of technical support and contextual changes whenever the evaluation design allows.
- Track inequality between schools and municipalities, not just the state average.
- Review weights and indicators when the formula starts encouraging unwanted behaviour.
- Preserve basic funding and continuity of service; variable rewards do not replace the duty to provide education.
If learning gains are not broadly distributed, participation falls, inequality increases, operating costs exceed plausible alternatives or the formula starts rewarding behaviour that does not match the desired outcome, the design needs to be revised. The goal is learning with equity; the indicator is a tool.
The case provides a concrete basis for OpenGov3's next step: comparing how different institutional rules would select, fund, implement and challenge an intervention of this kind.
Sources for this case
- 1. Ceará: Law No. 14.023/2007 (opens in a new tab)
Historical rule linking 18% of the discretionary portion of the municipal ICMS share to the Municipal Educational Quality Index, with financial effects from 2009.
- 2. IPECE: municipal share of ICMS (opens in a new tab)
Index series, legislation and methodological notes. The formula changed over time; the case does not treat the 2007 rule as the current formula.
- 3. SEDUC Ceará: PAIC Integral (opens in a new tab)
Programme of cooperation with the 184 municipalities, training, management and monitoring of learning outcomes.
- 4. SEDUC Ceará: SPAECE (opens in a new tab)
Large-scale external assessment covering the whole relevant pupil population, including literacy and the 5th and 9th grades of basic education.
- 5. World Bank: From Bad to Best / Ceará education reforms (opens in a new tab)
Overview of reforms, state–municipal cooperation and the complementary roles of incentives, technical assistance, assessment and autonomy.
- 6. World Bank: results-based financing and Ceará (opens in a new tab)
Summarises an impact evaluation using the state boundary and effects associated with incentives and technical support.
- 7. Ponne (2023): Better Incentives, Better Marks (opens in a new tab)
Synthetic-control evaluation of fiscal incentive and technical assistance policies in Ceará.
- 8. Costa & Carnoy (2015): early grades literacy intervention (opens in a new tab)
Triple-difference impact evaluation of PAIC, comparing Ceará, neighbouring states, grades and participating/non-participating schools.
How to read these experiences
Social democracy is a political tradition; the welfare state is a set of institutions; the HDI measures health, education and income. They are not synonyms. The cases presented show different choices and trajectories, not proof of the superiority of a single school of thought.
What the HDI measures (opens in a new tab) · Social democracy and political cooperation in Norway (opens in a new tab)