Case study — Brussels deliberative committees (commissions délibératives)
Jurisdiction: Brussels-Capital Region, Belgium (regional + community institutions). Updated: 2026-10-04. Evidence grading follows RESEARCH-METHOD.md. Labels: Observed (official/primary), Claimed (organiser self-description), Evaluation (external/academic), Analysis (Cicada synthesis), Open question.
1. Identity and jurisdiction
Deliberative committees are hybrid parliamentary bodies in which randomly selected residents deliberate together with elected members of parliament, inside two Brussels assemblies (SRC-0029, SRC-0031):
- the Parliament of the Brussels-Capital Region (PRB) and the United Assembly of the Common Community Commission (ARCCC): committees of 45 citizens + 15 parliamentarians;
- the French-speaking Brussels Parliament (Parlement francophone bruxellois, PFB — Assembly of the Commission communautaire française): committees of 36 citizens + 12 parliamentarians.
A fixed ¾-citizens / ¼-parliamentarians ratio applies; official rationale (Claimed, SRC-0030) is that this ratio avoids hierarchical dynamics and turns participating MPs into carriers of the recommendations inside their permanent committees. The mechanism is institutionalised in parliamentary rules, not one-off.
Context: Brussels also operates a separate permanent citizens’ assembly for climate (organised by Bruxelles Environnement) — a distinct mechanism not covered by this record (SRC-0034).
2. Dates and legal/institutional basis
Observed: the mechanism was created by amendments to the assemblies’ rules of procedure — PRB-ARCCC adopted 13 December 2019 (parliamentary documents A-100/1-2 2019–2020) and PFB adopted 20 December 2019 (document 13 (2019–2020) nos 1–2) (SRC-0031, SRC-0033); the rule sits at art. 28bis of the PRB rules per SRC-0034. Political origin: a 2017 citizen panel run by the Ecolo party recommended mixed citizen–MP committees; an earlier regulatory proposal had failed partly because the original MP/citizen power-sharing arrangement was found to violate the Constitution (SRC-0036, secondary/interview).
Observed: a dedicated data-protection law followed — decree of 11 February 2022 on processing personal data for the constitution of deliberative committees (PFB), published 8 March 2022 (SRC-0032).
3. Purpose / question-setting
Observed: two trigger routes (SRC-0029, SRC-0033):
- Suggestion citoyenne: any Brussels resident 16+ may propose a topic; at 1,000 supporting signatures the competent Assembly must analyse it and decide whether to constitute a committee. Refusal must be publicly justified; non-retained suggestions may be resubmitted within 12 months.
- Parliamentary initiative: one or more political groups may request a committee.
The Bureau élargi of the competent assembly proposes constitution; the assembly votes (SRC-0031). The threshold of 1,000 was deliberately set low versus an initial 5,000, on expert advice, to let unstructured groups reach it (SRC-0030).
4. Eligible population
Observed (SRC-0032): persons aged 16+ registered in the population or foreigners registers of a Brussels-Capital municipality, regardless of nationality; excluded are persons under a conviction or decision causing exclusion/suspension of municipal voting rights. Conflict-of-interest exclusions are decided case-by-case by the Bureau élargi (SRC-0030).
5. Invitation frame
Observed: the Belgian National Register of natural persons. The Assembly asks the Register to draw a sample of eligible residents; the 2022 decree specifies the data categories consultable (name, sex, date of birth, main residence, non-elector status) (SRC-0032). 10,000 invitation letters are sent per committee, signed jointly by the assembly president, the relevant permanent-committee president and the secretary-general, accompanied by a public communication campaign (SRC-0031, SRC-0034). Observed (practitioner figure): ~8% response rate, per parliament services (SRC-0035).
6. Recruitment and sortition method
Observed: double draw (SRC-0029, SRC-0033):
- First draw — 10,000 randomly selected residents receive the invitation; interested people register within ~30 days via the democratie.brussels platform or a freephone 0800 number, providing sociodemographic information (gender, age, residence, language, education) for the second stage;
- Second draw — the 45 (PRB) or 36 (PFB) citizens are drawn from respondents under pre-set representativeness criteria.
7. Stratification criteria
Observed (SRC-0031): gender, age, geographic distribution, language (PRB-ARCCC) and education level. Complementary topic-linked criteria may be added — sociodemographic, political (e.g. vote/abstention), or behavioural — identified by the Comité d’accompagnement after an exchange of views with the relevant permanent committee; they must be objectively linked to the topic and avoid illegitimate bias (SRC-0031). The 2022 decree additionally requires that criteria be limited in number and prohibits criteria based on GDPR special-category data (arts 9–10) (SRC-0032).
Observed: a standing methodological body exists — the Comité d’accompagnement of 8 members (4 effective + 4 substitutes), recruited by public call every two years with expertise/experience requirements in participation/deliberation (SRC-0031).
8. Size and term
Observed: 60 members (PRB) / 48 (PFB) per committee; at least 4 deliberation days (typically 4–5 across weekends); committees are not organised in the run-up to elections since deliberation + six-month follow-up needs ~9 months (SRC-0030, SRC-0031). The committee dissolves after its follow-up phase.
9. Compensation / accessibility / support
Observed: the 2022 decree authorises a flat-rate allowance (défraiement) and travel-expense indemnity under the law of 3 July 2005 on volunteers’ rights; Comité d’accompagnement members may receive attendance tokens (SRC-0032). Rationale (Claimed, SRC-0030): material compensation enables participation by less well-off people. Accessibility: a “buddy” pairing system and table composition accommodate Brussels’ multilingual population (FR/NL and others); the pre-meeting information session explicitly covers publicity/anonymity and individual support needs (SRC-0029).
10. Evidence and expert process
Observed: a three-phase process (SRC-0029): (1) informative phase — an information fiche and hearings of resource persons (experts, witnesses), with an optional preparation session before each meeting; (2) deliberative phase (see §11); (3) recommendations phase. The external operator’s role and the agenda are fixed with the Comité d’accompagnement after topic selection (SRC-0031).
11. Facilitation / deliberation
Observed: participants are spread across tables of 5–10, recomposed each meeting day so citizens and MPs mix as widely as possible; external facilitators ensure speaking turns (SRC-0029). The small-table deliberative phase is not public (except derogation) — official rationale: open, unconstrained speech (SRC-0030).
12. Decision / recommendation method
Observed: recommendation proposals are drafted by volunteers from the tables in collaboration with parliament services, then put to the committee’s vote — citizens vote anonymously and need not justify their votes, while parliamentarians’ positions are public (SRC-0029, SRC-0030). Rapporteurs produce a report containing the debate synthesis, adopted and rejected proposals with motivations, and vote results (SRC-0029).
13. Mandate and recipient institution
Observed: recommendations go to the permanent parliamentary committee covering the topic — i.e. back into the normal legislative track, where follow-up instruments are interpellation of the minister(s), a resolution, an ordonnance (PRB-ARCCC) or a decree (PFB) (SRC-0029). The same MPs who sat in the committee carry the follow-up (design intent, SRC-0030).
14. Required response / follow-up
Observed: within 6–9 months after the recommendations vote, the committee delivers a follow-up report; citizens are reconvened within six months to be told and given reasons for the follow-up accorded; citizens may opt to be kept informed beyond six months via the platform (SRC-0029, SRC-0030). A refused citizen suggestion must also be publicly justified (SRC-0029).
15. Transparency and participant privacy
Observed: the democratie.brussels platform publishes each committee’s documents, videos and reports and offers a “follow” subscription (SRC-0029). The 2022 decree is a full privacy frame (SRC-0032):
- independent, bias-free draw technique required; selection algorithms must be made public where the Assembly can access them;
- retention limits: personal data of non-selected invitees ≤3 months after the committee’s mission ends; for participants ≤12 months;
- citizen-suggestion signatures verified against the National Register and destroyed ≤3 months after verification;
- the Assembly is data controller.
Citizens’ votes are anonymous (SRC-0030); deliberation phases non-public (SRC-0029).
16. Outcomes / implementation evidence
Observed / secondary-monitored (SRC-0034, cross-checked against parliament records): seven committees held to date —
| Topic | Period | Recommendations |
|---|---|---|
| 5G in Brussels | Apr–Jun 2021 | 43 |
| Homelessness | Jun–Jul 2021 | 97 |
| Citizens’ role in crises | Oct 2021 | 22 |
| Urban biodiversity | Apr–May 2022 | 21 |
| Dual training | Jun–Jul 2022 | not published (exceptional 36+12 composition) |
| Urban noise | Apr–Jun 2023 | 21 |
| Public cleanliness | 2026 | 26 adopted 30 May 2026 |
Cumulative figure “>220” reported by RTBF; primary-source-verified sum 204 (SRC-0034). Estimated cost ~€90,000 per committee (Brussels Parliament via RTBF, SRC-0034). Implementation flows through ordinary parliamentary instruments; recommendation-level implementation rates are not yet systematically catalogued (open question).
17. Critiques / limitations
- Observed: the dual-training committee ran at 36+12 (PFB-size) composition inside the PRB track “without public justification” of the deviation (SRC-0034) — a documented inconsistency in applying the composition rule.
- Observed: ~8% invitation response rate means substantial self-selection upstream of the stratified draw (SRC-0035).
- Evaluation: the Belgian academic teams that helped design the mechanism also study it (G1000/UCL researchers were co-designers of G1000, the Ostbelgien model and these committees) — evaluations should be read with that dual role noted (SRC-0037 context).
- Open questions: MP attendance/attrition per committee; what share of recommendations reach ordonnance/decree stage; whether citizen-only anonymity in voting has measurable effects.
18. Software / operational tools
Observed: the democratie.brussels participation platform is built on the free-software Decidim (footer attribution), with CSAM electronic-identity login for citizen suggestion signatures (SRC-0030 footer). The platform handles suggestions, committee documentation and follow-up notifications; recruitment additionally uses postal letters and a freephone channel (SRC-0031). The 2022 decree mandates public selection algorithms where available — an unusual statutory audit requirement (SRC-0032).
19. Sources
- SRC-0029 — democratie.brussels FAQ: committees process (official).
- SRC-0030 — democratie.brussels FAQ: “pour en savoir plus” rationale (official).
- SRC-0031 — Vade-mecum et glossaire, PFB (official methodology doc).
- SRC-0032 — Decree of 11 Feb 2022 on personal-data processing (legislation, eJustice).
- SRC-0033 — PFB rules-change proposal, doc 13 (2019–2020) no. 1 (official).
- SRC-0034 — Brussels Governance Monitor, citizens’ assemblies dossier (independent monitoring, secondary).
- SRC-0035 — participation.brussels practice page (institutional portal; response-rate figure).
- SRC-0036 — OECD OPSI innovation profile (secondary).
- SRC-0037 — Democracy R&D case study on institutionalisation (practitioner/academic).
20. Cicada relevance / open questions
Analysis (Cicada synthesis — not established design):
- Brussels is the key hybrid case: sortition and election fused in one body with a fixed ratio. It tests whether mixing MPs into deliberation improves follow-through — the designed answer to the perennial “recommendations ignored” problem.
- The citizen-suggestion trigger (1,000 signatures → mandatory analysis + justified refusal) is a concrete agenda-setting contract Cicada could model.
- The 2022 decree is directly relevant to Cicada’s privacy/audit workstream: statutory retention limits, prohibition on special-category stratification data, independent/unbiased draw requirement, and a legal duty to publish selection algorithms.
- Anonymous citizen voting next to public MP voting is an interesting anti-pressure mechanism inside a mixed body.
- The Comité d’accompagnement is a reusable governance pattern: an expert oversight body for methodology, decoupled from the parliament’s own services.
- Open questions: implementation conversion rates; MP attrition; how the second-stage stratified draw is technically executed and by whom; whether the model survives majority change (institutionalisation depth vs actual usage frequency).