evidoria

← Back to browse

Good practice Imported

GRAS — the World Bank's Governance Risk Assessment System Screening Brazilian Public Spending for Corruption

Brazil · Brasília · See the Brazil profile

The World Bank's GRAS tool has screened Brazilian public spending since a November 2023 pilot across federal, state and municipal levels, identifying ~200 possible fraud/collusion red flags and flagging hundreds of suppliers for suspected collusion or strawman registration.

~200 red flags
Distinct red flags of potential fraud GRAS can identify in public expenditures (reported Sept 2024)
850+ suppliers
Suppliers showing strong indications of collusive behaviour (pilot results)
~450 suppliers
Suppliers suspected of being registered under strawmen (pilot results)
~500 cases
Cases of conflict of interest involving suppliers owned by public servants (pilot results)
~4,500 companies
Companies found to have connections to political campaigns (pilot results)

Details

Maturity
Scaling
Promoter
World Bank (Governance Global Practice), piloted with federal, state and municipal audit/control bodies in Brazil
Period
Piloted from around November 2023; results reported through September 2024; scale-up project since documented
Keywords
anti-corruption, public expenditure analytics, procurement integrity, audit

Context

The Governance Risk Assessment System (GRAS) is a data-analytics tool built by the World Bank's Governance Global Practice to help government auditors and anti-corruption bodies screen large volumes of public-expenditure data — electoral registers, social-programme records, payroll data and blacklisted-firm lists — for signs of fraud, collusion and improper political influence in public contracting.

Objectives

GRAS was piloted in Brazil starting around November 2023, running across multiple levels of Brazilian government (federal, state and municipal), including several subnational administrations, to screen expenditure data for corruption risk.

Activities

According to the World Bank's own account (11 September 2024), GRAS can identify around 200 distinct red flags of potential fraud in public expenditures.

Results

Reporting on the pilot's results describes concrete findings: over 850 suppliers showing strong indications of collusive behaviour, roughly 450 suppliers suspected of being registered under strawmen, around 500 cases of conflict of interest involving suppliers owned by public servants, and close to 4,500 companies found to have connections to political campaigns.

Conclusions

GRAS is explicitly designed by the World Bank as a replicable methodology rather than a one-off Brazilian tool, and the Bank has documented a follow-on 'Brazil - Governance Risk Assessment System Scale-Up Project.' It is used internally by government auditors rather than published as a public-facing tool, so its screening criteria and full result sets are not independently and publicly auditable, and no independent evaluation of prosecution or conviction outcomes from its flags has yet been published.

Implementation
Indicative cost
Medium (€50k–€500k) — No specific budget figures disclosed in the source; delivered as World Bank technical assistance to Brazilian audit and control bodies, with a documented follow-on Scale-Up Project.
Time to results
Medium (1–3 years) — Piloted from around November 2023; results reported through September 2024; a scale-up project has since been documented.
Staffing & skills
World Bank Governance Global Practice (tool developer), Federal, state and municipal audit/control bodies in Brazil (implementing users)

Conditions for success

  • Access to linked administrative datasets (electoral registers, social-programme records, payroll, blacklisted-firm lists)
  • Institutional buy-in from audit and control bodies across multiple levels of government
  • A replicable methodology design rather than a one-off, country-specific build

Common failure modes

  • Used internally by auditors rather than published as a public-facing tool, so screening criteria and full result sets are not independently and publicly auditable
  • No independent evaluation of prosecution or conviction outcomes from its flags has yet been published

Where it fits

Governance type
multi-level government (federal, state, municipal) with a multilateral technical-assistance partner
Scale
national, across multiple levels of government
Income level
upper-middle-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful