Part III, The Electoral System

Derived from §0.5 (a trustworthy mandate is the precondition of the whole model), and Criteria 1 (legitimacy), 7 (representation), 5 (capture-resistance). This is the foundation reform: experts executing a distorted mandate merely deliver the wrong thing competently, so the mandate must be formed honestly.

This is where BIG's founding work, the analysis of First Past the Post and the STV+ proposal, is incorporated and held to the same standard as the rest of the design.

III.1 Why First Past the Post fails (the diagnosis)

FPTP is the clearest example of an inherited, undesigned flaw (§0.1). It fails on seven measurable counts:

Comparators prove it is fixable: Ireland (STV) Gallagher ≈ 3.1; Germany (MMP) ≈ 1.8; New Zealand's switch from FPTP to MMP cut its index from ~12 to 1.8-4.

III.2 What the electoral system must achieve (derived requirements)

From the foundations, the system must be:

1. Proportional, seat share tracks vote share (Criterion 7).
2. Locally accountable, a real link between representatives and places (Axiom 12, subsidiarity).
3. Tactical-resistant, voters can vote their genuine preference (Criterion 1, honest mandate).
4. Gaming- and gerrymander-resistant (Criterion 5).
5. End-to-end auditable and verifiable (Axiom 7).
6. Inclusive, every eligible citizen can vote, online or offline (Criterion 12).
7. Simple enough to use and to trust (Criterion 11).

III.3 The system: STV+ (Single Transferable Vote, enhanced)

The recommended method is STV+, which best satisfies the requirements above (scoring 81/100 against BIG's electoral-systems matrix, versus 39/100 for FPTP, with MMP the strongest alternative at 76).

Core mechanism:

• Multi-member constituencies of 4-7 seats, large enough for proportionality, small enough to preserve a genuine local link.
• Ranked-preference ballots, voters rank candidates; no vote is wasted on a hopeless first choice.
• Droop quota for election: Q = floor( valid_votes / (seats + 1) ) + 1.
• Surplus transfer, votes above quota transfer at fractional value to next preferences; lowest candidates eliminated and transferred, until all seats are filled.
• Greatly reduced tactical incentive. Ranking your genuine preference is almost always the best strategy, and the wasted-vote and lesser-evil tactics that dominate FPTP largely disappear. Honesty requires the caveat (an evidence-first constitution must not assert a false theorem about its own mechanism): no ranked method is strategy-proof (Gibbard–Satterthwaite), and STV is technically non-monotonic — in rare surplus-transfer and elimination orderings, raising a candidate can occasionally hurt them. STV+ minimises tactics rather than eliminating them, which is enough to make the resulting mandate far more honest than FPTP's.

The "+" enhancements (what lifts it above plain STV):

1. Independent Boundary Authority, district lines drawn by an independent body on published, algorithmic, anti-gerrymander criteria (§III.5), never by incumbents.
2. End-to-end verifiable count, round-by-round, machine-readable, publicly published, and cryptographically verifiable (Part VIII), so any citizen can confirm the result. Paper trail always retained (§III.6).
3. Statutory voter education, funded, neutral instruction in how to rank and how to verify (links Part II.6).
4. Mandatory audit sampling, a minimum percentage of ballots independently hand-audited against the electronic count every election (risk-limiting audit).
5. Campaign-finance and integrity rules (§III.7-III.8) that close the money-capture vector.

The "+" is integrity, not a proportional top-up. Unlike MMP/AMS, the "+" adds no compensatory tier; it is the boundary, verifiability, education, and finance hardening around STV. Residual disproportionality is therefore real and stated: a Droop quota gives an effective threshold of roughly 1/(seats+1) — about 12.5% in a 7-seat district, ~20% in a 4-seat one — so a small, geographically dispersed party can still win little. STV+ targets a Gallagher index in the low single digits (Ireland's STV ≈ 3.1), a large improvement on FPTP's ~11.8 but not the ~1.8 of MMP. MMP scores 76 to STV+'s 81 on BIG's matrix because STV+ is preferred on the local link and verifiability at a modest proportionality cost — a trade the matrix weights make explicit, and which is itself a published value choice (§0.4.1), not a neutral fact.

III.4 Election and lot, the bicameral mandate

The model does not rely on a single selection method. STV+ produces the Representative Assembly (elected, accountable, proportional). It is paired with a Citizens' Sortition Chamber (selected by lot, Part II.3 mode C) constituted in Part IX. Each checks the other's weakness:

• Election gives accountability but creates re-election incentives, party discipline, and exposure to campaign money.
• Lot gives independence from those incentives and a statistically representative cross-section, but no electoral mandate.

Used together, elected chamber proposes and governs; sortition chamber scrutinises, can delay, and deliberates on value trade-offs, they form a mandate that is both accountable and incorruptible by campaign incentives. (Full powers and interaction: Part IX.)

III.5 Boundaries and district magnitude

• Drawn by the Independent Boundary Authority, never by political actors.
• Published, algorithmic criteria: equal population per seat (within a tight tolerance), respect for community and geography, compactness, and a gerrymander-detection metric appropriate to multi-member STV — seats–votes proportionality within and across districts, not the two-party efficiency gap, which is built for single-member districting and does not transfer. The weighting of the competing criteria (equal-population vs community vs compactness) is itself a value choice, so it is set through the §V.4 value route and published, never presented as a neutral algorithmic given (choosing the weights is where a subtle gerrymander would hide).
• The drawing process, data, and code are open and reproducible (Axiom 7), anyone can re-run it and check.
• District magnitude (4-7 seats) is set to balance proportionality against local linkage; rural/urban variation is allowed within the range on transparent criteria, and the resulting difference in proportional power between a 4-seat and a 7-seat district is published and bounded, so variable magnitude does not quietly reintroduce the vote-weight inequality FPTP is condemned for (§III.1).

III.6 The count, transparency, and verifiability

• Open count, round by round, published in machine-readable form with every transfer shown.
• End-to-end verifiable (E2E-V): each voter can confirm their vote was cast as intended, recorded as cast, and counted as recorded, without being able to prove to anyone else how they voted (receipt-freeness, to defeat coercion and vote-buying). Mechanism in Part VIII.
• Paper record always retained, even where voting is digital, enabling a full manual recount.
• Risk-limiting audits mandatory: an independent statistical hand-audit of paper against the electronic tally, every election.
• Dispute resolution: a defined, fast, transparent process (Court-backed) for challenges, with the paper trail as ground truth.

III.6a Vacancies, recall, ties, and roll maintenance

The recurring mechanics an electoral system must specify (their absence is itself attack surface, §XIX.0):

• Casual vacancies in a multi-member district are filled by countback of the original ballots — re-examining the departing member's ballots to elect the next-placed candidate — preserving the original mandate's proportionality, with a by-election fallback where countback is impossible. Co-option or a single-seat by-election is rejected because it distorts the district's proportional balance mid-term.
• Recall of a representative exists as a high-threshold, court-supervised citizen power for serious misconduct (§XIX.4, §XVIII D4), never for mere unpopularity.
• Ties in elimination or surplus order are broken by a published, seeded, publicly-verifiable random draw (the §VI.3a beacon), so a chance decision over who holds power is auditable rather than opaque.
• Roll maintenance is continuous, not "verified once": movers, deaths, and status changes update the roll, and the identity layer de-duplicates across multi-member districts so no one is registered twice (§VIII.2).

III.7 Candidate eligibility and integrity

• Eligibility is broad (low barrier to standing, Criterion 7, inclusiveness), with disqualifications limited to clear, Charter-defined integrity grounds.
• Funding transparency: every donation above a low threshold is published in real time to the transparency ledger.
• Donation caps and a ban on foreign and anonymous money (capture-resistance, Criterion 5).
• Spending limits per candidate/party, equalising the contest so money cannot buy outcomes.

III.8 Campaign integrity and information

Balancing free expression (§I.3) against an honest mandate:

• No state censorship of political speech. Integrity is pursued through transparency and provenance, not bans.
• Mandatory provenance for paid political advertising: who paid, how much, and (for synthetic/AI media) clear labelling, enforced as a transparency duty, not a speech ban.
• Independent fact-published rebuttal: the Public Information function (§II.5) publishes evidence-based corrections; it persuades, it does not prohibit.
• Foreign-interference defence: detection and disclosure of foreign-origin influence operations (Part VIII), again by exposure rather than censorship.
• Honest about provenance. Content-provenance credentials (C2PA-style) authenticate genuine material; they do not reliably detect fakes, survive poorly when content is re-uploaded, and can be forged or used to unmask a source. Provenance is therefore a help, never the primary defence. The weight is carried instead by the end-to-end-verifiable count, a fabricated video cannot change a result anyone can check (§III.6, §VIII.3), by rapid independent rebuttal, and by an informed electorate (§II.6). Early large-scale evidence supports this ordering: in recent elections synthetic political media circulated widely but reached few and changed little; the verifiable result, not the viral clip, decided power.

III.9 Failure modes and safeguards

Part III ends. The foundation tier (Charter → Citizen → Electoral) is complete: the people are sovereign, rights are inviolable, and the mandate is now formed honestly. Next: Part IV, The Expert Execution Layer, which receives that mandate.