Projeto Brasil 2030 (Projecto B30): evidence and evaluation protocol
Research edition: 17 October 2026. Status: proposed research protocol, not an adopted government programme.
Purpose and scope
Projeto Brasil 2030 (Projecto B30) is an independent, nonpartisan reference library about Brazil's development challenges, international outcomes and constitutional institutions. It does not endorse a candidate, party or electoral platform. The initial collection is a research foundation; it is not 100 completed impact evaluations or a comprehensive article-by-article legal commentary.
“First world” and “global power” are not sufficiently precise evaluation criteria. For this library, analysts should separately measure material living standards, health, learning, safety, equal opportunity, institutional quality, environmental resilience and international scientific/economic capacity. This is an editorial measurement choice, not an official definition or a claim that these dimensions can be combined objectively into one score.
Five research questions
- Which development problems are documented, how large are they, and whose experiences disappear in national averages?
- Which countries achieve strong outcomes on comparable measures, and what evidence explains those outcomes?
- Which constitutional rights, competences, procedures and financial rules govern each policy domain?
- How can changes in outcomes be measured without confusing correlation with the effects of a policy?
- What evidence is still missing before an intervention's costs, benefits and feasibility can be assessed?
What the initial collection establishes
The challenge catalog identifies 100 distinct research areas, not an empirically ordered ranking. Sources vary in depth: a measured national baseline is stronger evidence than a relevant source portal. The country comparison is a purposive sample of useful comparators, not a global league table. The constitutional review maps institutions and major provisions; it does not certify every statute, amendment, court precedent or proposed reform.
No numerical post-2031 result is asserted without a model. A target is a value judgment; a projection is a conditional calculation; an observation is a measurement. These must remain separate in the database and on the website.
Evidence record
Each substantive statistic should eventually have the following fields. A blank field is an unresolved research task, not a zero.
| Field | Required interpretation |
|---|---|
| Indicator and definition | Exact numerator, denominator, population and unit |
| Observation period | When the underlying event or measurement occurred |
| Publication date | When the producer released this edition |
| Retrieval date and edition | When the project checked it, with archived file or checksum where available |
| Geography | Country, state, municipality or other explicitly defined coverage |
| Value and uncertainty | Point estimate plus interval or source caveat where available |
| Method and series break | Survey, administrative register, model; changes in definitions or population estimates |
| Source location | Primary URL and table, page or article when verified |
| Evidence status | Extracted and checked; corroborated; provisional; missing |
| Comparable series | Which earlier years and other countries use genuinely comparable definitions |
| Distribution | Income, region, age, sex, race/colour and disability where lawful and statistically supported |
| Reviewer | Named review responsibility and last correction |
Evaluation, not promises
Brazil's federal guidance treats evaluation as part of planning, execution and budgeting and explicitly links it to costs, benefits and accountability. The guidance is a methodological reference rather than a blanket binding rule for all evaluations. Casa Civil, evaluation guidance
The Ipea ex-ante guide addresses assessment before an intervention is implemented. Its existence is evidence of an established evaluation framework, not evidence that a particular intervention works. Ipea, ex-ante guide, 2018
The following is this project's proposed analytical template, informed by that framework:
- Define the problem and document the baseline, including existing initiatives and recent improvements.
- Write a falsifiable causal hypothesis and specify rival explanations.
- Compare the status quo with multiple alternatives, including administrative changes and non-government provision where relevant; do not assume legislation is always necessary.
- Identify responsible institutions, constitutional authority, statutory requirements, procurement constraints and implementation dependencies.
- Estimate incremental capital and recurring expenditure, staffing, maintenance, financing and displacement of other spending. Record the price year and discount-rate assumptions.
- Specify distributional impacts, unintended harms, operational failure modes and rights implications.
- Assess causal evidence, external validity and uncertainty; prefer systematic reviews and credible evaluations to appealing country anecdotes.
- Predefine monitoring measures, evaluation design and the evidence that would justify continuing, modifying or ending an intervention.
A useful budget identity is incremental cost = additional service volume × unit cost + setup + administration + maintenance − genuinely avoidable expenditure. It is only an accounting scaffold. It does not estimate fiscal space or establish that savings are cashable.
A research timetable through and beyond 2031
This schedule concerns the library's evidence work, not an election campaign or a promised national reform programme. Dates are proposed checkpoints, conditional on available people and resources.
| Period | Research output | Completion criterion |
|---|---|---|
| First 30 days | Source inventory and indicator dictionary | Every challenge has a responsible data producer, defined measure and explicit missing-data status |
| Days 31–90 | Checked baseline series | Each published value has a reference year, definition and reproducible extraction; unresolved contradictions remain visible |
| Months 4–6 | International comparability review | Country comparisons document scale, income, institutions, selection effects and measurement differences |
| Months 7–12 | Independent methodological and constitutional review | Reviewer comments, responses and corrections are published |
| 2027–2030 | Annual evidence editions | Time-series revisions and changes in the law are tracked, with dated releases |
| 2031 | Benchmark edition | Refresh observations and distinguish them from any scenario assumptions |
| 2032–2035 | Retrospective tracking | Compare observed trajectories with prior conditional scenarios, documenting shocks and attribution limits |
| 2040 | Long-horizon review | Reassess structural outcomes and whether definitions still serve the public |
The repository does not automatically collect new releases or implement this schedule. It contains no background monitor.
Rules for future scenario analysis
A mathematical illustration, not a Brazil forecast: a baseline index of 100 growing at 2% annually becomes 110.4 after five years; at 3% it becomes 115.9. The difference is sensitivity to an assumption, not an intervention's measured effect. Use the same caution for projected income, crime reduction, learning and service access.
For any future scenario, publish the baseline year, horizon, model equations, assumptions, uncertainty ranges, comparison case and sensitivity tests. Never extrapolate a national success story as a causal effect transferable to Brazil. Never add overlapping benefits across 100 challenges. Household poverty, schooling, employment, crime and health can be causally related and statistically correlated.
Quality control and limitations
- Primary administrative data can undercount events; surveys can have sampling error; international datasets can harmonize or revise national numbers.
- Improvement and remaining need can coexist. A declining homicide rate is neither proof that violence is solved nor evidence that every policing strategy caused the decline.
- A corruption perception index is not the share of officials who commit crimes or the amount of money stolen. Case counts also reflect enforcement and reporting.
- GDP per capita is not median household disposable income. PPP and market-exchange-rate comparisons answer different questions.
- Output delivery is not outcome improvement. Money spent and projects launched are not substitutes for measured benefits.
- Human review should precede claims that a document is a definitive legal interpretation or a fully costed policy evaluation.
These are methodological cautions adopted by the project. They are not findings about the magnitude of any particular Brazilian problem.
Methodology of this edition
The requested deep-research skill was read. Firecrawl and Exa were unavailable in the exposed tools; the available web search/open tools were used instead. Work was divided into three parallel research strands plus this synthesis. Each strand records its own sources and limitations. No systematic-review completeness claim is made, and no invented total of uniquely read sources is presented. Search snippets were used for discovery; factual confidence depends on the underlying materials actually inspected. Some official pages could not be opened reliably.
All claims are bounded by information checked for this edition, dated 17 October 2026. Dynamic websites can later change. This initial edition does not contain archived copies of every cited source, a completed causal review for all 100 topics, or independently audited fiscal forecasts.