British Columbia’s Citizens’ Assembly on Electoral Reform (2003 to 2009)

British Columbia’s Citizens’ Assembly on Electoral Reform was a body of 160 randomly selected residents, plus a chair, that spent 2004 studying electoral systems and recommended a single transferable vote system called BC-STV. On May 17, 2005, 57.69% of valid votes supported it and a majority voted yes in 77 of 79 districts, but the law required 60% province-wide, so it failed; a second referendum on May 12, 2009 gave it 39.09%. The case shows a citizen body with a guaranteed route to the ballot, and how thresholds, information and campaign rules shaped what happened next.

In British Columbia, a group of ordinary people picked at random were asked to study how votes become seats and to recommend whether to change the system. They worked for most of a year, held public hearings, and voted almost unanimously for a new system. Their proposal went straight to voters. Most voters said yes, but not enough to meet the bar the government had set. Four years later, after another vote, the old system stayed.

For a short overview of other assemblies, including Ireland’s, see comparative cases and Ireland’s citizens’ assemblies.

Defined terms

The British Columbia case uses several terms that have precise meanings.

• First past the post (FPTP), or single member plurality: each district elects one member, and the candidate with the most votes wins, even without a majority.
• Single transferable vote (STV): voters rank candidates in multi-member districts; candidates who reach a quota are elected and surplus or eliminated votes transfer according to voters’ next rankings. The assembly’s version, BC-STV, used districts of two to seven members.⁠¹
• Mixed member proportional (MMP): a system that combines local single-member districts with list seats that make overall party shares more proportional.
• Supermajority threshold: a requirement that a proposal win more than a simple majority, here 60% province-wide plus majorities in a set share of districts.⁠²
• Citizens’ assembly: a body of residents selected by lot that studies, deliberates and recommends. See the glossary.
• Referendum advertising sponsor: a person or group that registers to advertise for or against a referendum question.⁠²

What happened

The British Columbia Citizens’ Assembly on Electoral Reform moved through a dated sequence of steps.

• May 8, 2003. The legislature unanimously appointed Jack Blaney, a former president of Simon Fraser University, as chair.⁠¹
• September 2003. Selection began with letters to randomly chosen people.⁠¹
• January 11 to March 26, 2004. Six weekend learning sessions.⁠¹
• May and June 2004. Fifty public hearings; the assembly received 1,603 written submissions from over 1,430 individuals.⁠¹
• September to November 2004. Deliberation.⁠¹
• October 23 and 24, 2004. Members chose STV over MMP by 123 to 31, voted 142 to 11 against keeping first past the post, and voted 146 to 7 to recommend BC-STV to voters.⁠¹
• December 2004. The final report, Making Every Vote Count, was published.⁠¹
• May 17, 2005. Referendum held with the provincial general election. BC-STV won 57.69% of valid votes and majorities in 77 of 79 districts; the 60% province-wide threshold was not met.⁠²
• September 2005. In the Speech from the Throne, the government acknowledged significant support for reform and said voters had been missing information about how the electoral map would look under BC-STV.⁠³
• May 12, 2009. A second referendum, held under a new act, gave BC-STV 39.09% of valid votes and majorities in 8 of 85 districts.⁠³
• October 22 to December 7, 2018. A third, separate referendum by mail, not linked to an assembly, asked whether to keep first past the post or move to proportional representation; 61.3% chose first past the post.⁠⁴

How it worked

The British Columbia assembly’s mechanics combined random selection, demographic balance and a statutory path to a public vote.

Selection. The final report states that “the members of the Citizens’ Assembly were chosen at random from the province’s 79 electoral districts,” one woman and one man from each district, drawn at selection meetings.⁠¹ After a review found no First Nations members, the government amended the terms of reference so that one man and one woman were added from the Aboriginal community, “bringing the number of members to 161, including the chair”.⁠¹ The design guaranteed geographic and gender balance; it did not guarantee balance on age, income or political views, and the sources consulted here do not report those distributions.

Learning, hearing, deciding. The assembly separated learning (six weekends), public input (50 hearings and written submissions), and deliberation (fall 2004).⁠¹ Members first compared the two leading alternatives, then asked whether to keep the existing system, then voted on the final recommendation.⁠¹ That order matters: the 142 to 11 vote against first past the post came after members had already designed the alternative they would offer.

Statutory path. The Electoral Reform Referendum Act required a referendum with the 2005 general election if the assembly recommended a system different from first past the post.⁠² The assembly, not the government, fixed the ballot question: “Should British Columbia change to the BC-STV electoral system as recommended by the Citizens’ Assembly on Electoral Reform? Yes/No”.⁠²

Threshold. For the result to bind the government, the act required at least 60% of valid votes province-wide and more than 50% in at least 48 of the 79 districts.⁠² The Chief Electoral Officer reported that the Act’s threshold was “higher than the simple majority required under the Referendum Act”.⁠²

Campaign rules in 2005. “Legislation did not establish a role for ‘Yes/No’ groups,” and “no spending limits were in place for referendum advertising sponsors”.⁠² Elections BC did not explain the systems, to keep its neutral role; a government Referendum Information Office mailed a brochure to every household and ran a website and phone line.⁠²

Campaign rules in 2009. The 2009 act and regulation designated an opponent group (No STV) and a proponent group (British Columbians for BC-STV) and gave each $500,000 in public funds.⁠³ The 2009 ballot also changed form: it asked which system British Columbia should use, listing “The existing electoral system (First-Past-the-Post)” and “The single transferable vote electoral system (BC-STV), proposed by the Citizens’ Assembly on Electoral Reform”.⁠³ The thresholds became 60% province-wide and majorities in at least 51 of 85 districts.⁠³

Evidence: measured results

The British Columbia referendums produced official results with clear denominators.

In 2005, 46,607 referendum ballots were rejected⁠² and 24,930 people voted in the general election but chose not to vote in the referendum.⁠² In 2018, of those choosing among proportional systems, mixed member proportional led with 63.05% after two rounds of counting.⁠⁴

On the assembly itself, the major comparative study is When Citizens Decide: Lessons from Citizen Assemblies on Electoral Reform by Patrick Fournier, Henk van der Kolk, R. Kenneth Carty, André Blais and Jonathan Rose (Oxford University Press, 2011), which examines the British Columbia, Netherlands and Ontario assemblies and asks whether ordinary citizens can handle complex institutional choices and how participation affected members.⁠⁵ The book’s detailed findings were not verified for this page beyond its published description.

Failures and criticisms

The British Columbia process drew criticism from reform supporters, from defenders of the existing system, and from people concerned with process design.

The supermajority. Supporters of reform point to the official record: a proposal that won 57.69% and carried 77 of 79 districts did not take effect.⁠² On their view, a 60% bar let a minority of voters keep a system that a majority rejected. Defenders of the threshold argue that electoral rules decide how every later election is counted, so changing them should require broad agreement across regions, not a narrow majority on one day. Both positions are coherent; the official report records only that one threshold was met and one was not.⁠²

Information. The government’s own September 2005 statement held that voters had been missing “a critical piece of information”: how the electoral map and distribution of members would look under BC-STV.⁠³ Reform supporters can read the 2005 vote as strong despite that gap; skeptics can read it as a vote on the idea rather than the concrete map. When the map and funded campaigns were provided in 2009, support fell to 39.09%.⁠³ The sources consulted here do not establish why, and several explanations (more information, less novelty, a funded opposition, a different ballot format) are possible.

Ballot form. The 2005 question named the assembly as the source of the recommendation,⁠² and the 2009 ballot again described BC-STV as “proposed by the Citizens’ Assembly”.⁠³ Critics of the process could argue that naming the assembly invited voters to defer to it; others could argue that disclosing the proposal’s source is basic transparency. The page takes no position.

Unequal campaign resources in 2005. With no designated sides and no spending limits for advertising sponsors,⁠² neither side had guaranteed resources. The 2009 change to equal public funding for designated groups.⁠³ answered that concern, though it did not raise support.

Representativeness. The selection design guaranteed one woman and one man per district and later added two Aboriginal members⁠¹ That initial omission, corrected only after review, shows how a design that balances on some characteristics can miss others.

Competing interpretations

Scholars and participants read the British Columbia case in at least three ways.

The assembly succeeded and the threshold failed it. On this reading, randomly selected citizens learned a technical subject, reached a near-unanimous recommendation (146 to 7) and persuaded most voters in nearly every district.⁠^1,2 The failure was a rule set by politicians.

The public made a considered choice. On this reading, three votes over 13 years, two of them with growing information and one with funded campaigns on both sides, show a stable preference for keeping first past the post.⁠^2,3,4 The 2005 majority was a first impression, not a settled view.

The case shows the limits of borrowed trust. On this reading, a citizens’ assembly can lend credibility to a proposal among voters who know about it, but that credibility does not automatically reach voters who do not, and it can be offset when an organized opposition appears. This interpretation is consistent with the official results but is not proven by them.

What transferred and what did not

The British Columbia model influenced later assemblies, and some of its features did not travel.

Transferred. The model of a randomly selected assembly with a reform mandate was used in the Netherlands and Ontario, which together with British Columbia are the three cases compared in When Citizens Decide.⁠⁵ Ireland later used citizens’ assemblies before constitutional referendums; see Ireland’s citizens’ assemblies.

Transferred in a different form. British Columbia’s own 2018 referendum used a ranked question, mail voting and electronic tabulators, and was the province’s “first referendum with a ranked choice question”.⁠⁴ It did not use an assembly. Its 50% threshold was lower than the 60% used in 2005 and 2009.⁠⁴

Did not transfer. The guaranteed route from a citizens’ body straight to the ballot, with the assembly writing the question, was rare. Many later assemblies advised governments or parliaments, which then decided what to put to voters.

Application to this framework

The British Columbia case informs several design choices in My Vote Forward.

• Thresholds are design decisions that need public reasons. A threshold changes the outcome as much as the vote does. The framework asks that any decision rule be stated before participation; see collective priorities and governance.
• Who writes the question matters. British Columbia let the assembly write the 2005 question.⁠² The framework treats question authorship as a governance issue; see question design.
• Information has to be concrete. The 2005 map gap⁠³ illustrates the framework’s rule that a proposal must state its effects and dependencies before an opinion on it means much; see proposals.
• Report distributions, not just a winner. Reporting both the 57.69% and the 77 of 79 districts⁠² shows the public more than “failed.” See measurement.
• Separate the advisory body from the binding vote. The assembly recommended; voters decided. That separation mirrors the framework’s distinction between priorities, opinions and private voting.

Unresolved questions

• Why did support fall from 57.69% in 2005 to 39.09% in 2009? Information, campaign funding, ballot format and novelty are candidate explanations; the official reports do not settle it.
• What is the right threshold for changing electoral rules, and should it depend on how the proposal was developed?
• Did voters who knew about the assembly vote differently from those who did not? Published research addresses this, but it was not verified for this page.
• How representative was the assembly on age, income, education and prior political views?
• Would a standing, recurring assembly have produced a different trajectory than a one-time body?