It is extremely strongly recommended that you read the frustrated voter and voting system values, vote swap and electoral reform materials before you attempt to read this. The entire article is available from Craig Hubley as is description of approval version BMAV and the allocation vote version B5AV.
The feature BSTV and BMAV share is fixed bioregional borders and fractional top-up seats. This replaces the D'Hondt method used in most countries to round off seats in different regions to whole numbers. Instead of varying the power of the vote slightly between regions, all votes in strictly applied BSTV and BMAV have exactly the same voting power. A fractional seat is allocated by top-up using methods similar to MMP systems, that typically yield 1/7 MMP seats (at average 3 members/district) or as few as 1/11 (if 5 members/district). In Canadian federal elections this means 300 constituencies plus 50 or 30. Not all districts with a fractional seat will receive a local member, but their votes will count towards the election of regional members & actually count more towards the party allocated top-up seats if they don't receive a member. That is, they may determine the party of the representative elected in the other district, bypassing one that is more popular locally for the one from the party that was more popular in the districts that had a higher fraction of votes elect no one. Everyone elected runs in a district and has district responsibilities - see the +C feature. To allow independent slates and looser alliances than parties, Labels not parties are the basis of this second round seat allocation. Label association is voluntary and in some cases can actually cause a candidate to lose who would otherwise win, if others of that label won too many seats elsewhere.
Bioregional Single Transferable Vote with Circuits + Labelled (i.e. party-proportional) Members
optimal electoral reform for Canadian provinces achieving
- stable districts
- bioregional accountability
- community voices
- NO party lists
- no party hacks
- party-independent members
- NO wrongly excluded parties
(excerpted from a report by Craig Hubley whose complete contents are as follows:)
complete contents (not published here)
The proposal 4 Introduction: the purpose and benefits of the BSTV+C+L proposal 10 General advantages of reconciled STV+MMP with fixed districts 10 Detailed purpose of this document and proposal 10 Context: Values, goals and guarantees of electoral reform 13 Recommendation A: establishment of Democracy Commissions 16 Review electoral and voting system 16 Audit administration of political parties 20 Recommend changes to administration of political parties 20 Advocate and criticize of democracy and systems in general 21 Recommendation B. Heavy STV, light MMP, is harder to resist 23 Voting Systems – STV, MMP and party-proportionality 24 B (a) STV for 50-90% of seats 27 How STV works 27 The BC-STV experience 29 A defence of STV (STV and MMP) 30 Population projections (Advantages of STV) 31 District maps 32 Recommendation C: Electoral districts should correspond with bioregional borders and population 32 Advantages of basing EDs on existing political boundaries 33 Circuits 33 District magnitudes 35 Regional differences and proportionality 36 Northern areas 37 Transparent and accessible process 38 Recommendation D. Party machinations and transparency measures to minimize the harm they do 38 Parties should not be able to limit the number of candidates seeking election under their banner in a given district. 39 Rights of candidates and Party members to attach their name to a party/faction label 40 Vote allocation and mandatory vs. voluntary seat surrender 41 No explicit legal status for parties 42 Cross-membership and co-nomination encouraged 42 Special constitutional challenges presented by political parties 43 NO wrongly excluded parties 44 Recommendation E: Aboriginal factions/parties 45 Recommendation F: Federal & municipal electoral reforms use this ballot 46 Appendix A: Voter's-eye view of BSTV+C+L 48 Appendix B: The 2006 Nova Scotia general election under BSTV+C+L 50
Several failed provincial electoral reform referenda let voters choose between a single new voting system, either single-transferable-vote (STV) or mixed-member proportional (MMP), and the status quo first-past-the-post (FPTP) system. While early results were promising especially in BC, dismal recent results strongly suggest that no proposal for a single simple ‘proportional’ system can succeed in such a vote against a well-organized ‘no’ movement, and that supporters of STV or MMP are easily divided with supporters of each easily convinced not to vote or to vote against adopting their second-choice system. Such vote-splitting guarantees that either STV or MMP in their typical simple form would be adopted by the Canadian public. A pattern is set by the repeated failures, and advocates of new referenda (on MMP in BC, on STV in ON, or on a referendum to change the voting system first followed by a choice between systems that did not include the status quo at all) must realize that this form of advocacy is futile. There is no chance of success on this path to electoral reform, and accordingly Fair Vote and reform advocates must substantially change tactics. Futhermore, federal reforms are only likely after one large province reforms, which will set the pattern for analogous federal reforms (although Senate reform should adopt a completely different system, as in Australia, to avoid arbitrary biases within one voting system affecting both houses). This is very well known, ensuring out-of-province actors will influence provincial referenda.
The choices available to the reform movement in Canada are as follows:
1. Give up on complex electoral systems that require voters to mark more than one choice, and adopt an “alternative vote” (also called “instant-runoff”) system using existing single-member electoral districts with the same process of determining borders. This leaves geographically-large less-populated districts with single members (which at the federal level has led to extreme disparities such as Labrador and PEI versus Toronto) and it proves a ballot that is also useful for other single-winner elections such as Mayor. Focus on federal Senate reform to prove STV or MMP ‘proportional representation’.
2. Continue the futile advocacy of ‘the other system’ not tested in one’s own province (MMP in BC, STV in ON, etc.) or a New Zealand type referendum in which the public must first vote to abandon FPTP and then choose between two or more systems to replace it with FPTP no longer on the list (leaving the ‘no’ side the option of attacking every undesirable aspect of every possible system). In practice all of these forms of advocacy have exactly the same result: A continuation of first past the post at least until Senate reform or UK reforms prove themselves, which is likely to be more than one generation.
3. Reconcile STV’s multi-member geographical representation with MMP’s party proportionality into a system that the public can learn to execute itself with paper ballots, and which addresses the deficiencies of simpler systems.
The author considers advocates of (2) to be either ‘useful idiots’ or insincere advocates in favour of no reform at all. This paper is addressed to those strongly preferring option (3).
We propose a bioregional multi-member district system that allows for instant-runoff small rural or isolated single-member districts, fixes electoral boundaries for all districts permanently to match watershed and municipal borders set by more principled criteria. Uniquely among single transferable vote systems, BSTV+C+L deals well with varying population by increasing the contribution of votes in under-represented districts to elect party-proportional members to about 10% of seats (called "+L" seats in this document). Though that number can increase if necessary to as many as a third or half of all seats if desirable, evidence in Canada suggests it isn't: The public mistrusts party insiders and a combination of lack of accountability within parties and overly plaintiff-friendly libel or defamation laws makes it extremely difficult for members or media to restrict power of such insiders. In such an environment the “party list” becomes unacceptable to choose candidates, because there is insufficient scrutiny by the public on its construction. So-called ‘open list’ systems which allocate seats based on popular vote for each candidate disadvantage members from smaller districts, requiring a questionable weighting factor. As the proposed system already weights ‘wasted’ votes in each district to elect ‘missing’ party-proportional members, two separate weighting factors seem definitely undesirable. Another variant employs allocation or approval voting rather than ranking (see below).
[Weight wasted votes to elect under-represented parties] In the variant of this system (BSTV) that we recommend, districts would normally elect four to seven members at once using a preference (ranked) ballot, but urban areas might be larger to ensure they included the entire boundaries of a community and infrastructure, while rural or wild areas may elect fewer (or one) to reduce the district’s geographic size. Uniquely among all systems known to this author, BSTV+C+L permits the number of members elected to vary freely and immediately with population rather than shifting the borders of districts. This is possible because the "+L" seats make up for fractional seat numbers: If a district loses population, it will either lose enough to pass over an integral threshold (from 3.1 to 2.9, for instance) or not (say from 3.5 to 3.3). In the latter case, the ‘wasted’ votes are weighted less (at 30% down from 50%) to calculate which members of the under-represented parties go to the legislature, reducing the chance of a local member. In the former case, one less member is elected, but the odds of electing a local candidate from an under-represented party goes from very low (a 10% weighting) to very high (a 90% weighting). While the numbers of party-proportional seats are fixed by the total province-wide vote in all districts, who represents the smaller under-represented or excluded parties in the legislature can have a profound impact on the evolution of those parties. As these parties are often influential or represent rising political views previously excluded by elites, there’s an argument that this influence on small parties and the chance to moderate their more extreme views with a chance to represent a district helps a polity evolve faster than otherwise. The parties themselves may prefer to send leaders and one or two deputies, but the transparency that should be imposed on them in this case should be strict, including requirements (see below) that some parties would find onerous. Each province must decide how to represent under-represented parties or deal fairly with their leaders and deputies. We recommend that whatever other rules are adopted, to elect a deputy at all should require that they stood for election simultaneously on the same ballot as the leader, and are neither appointed nor elected by a less scrutinized internal balloting.
[Circuits within districts represented by one member] Guaranteed local representation is a feature of FPTP and MMP systems that is diluted greatly in pure STV. To deal with that problem, the system proposed also requires elected members to choose a subset of their district, a "circuit", equivalent in size approximately to an existing first-past-the-post (FPTP) electoral district or "riding" to represent. Those with higher vote totals get first choice of which circuit to represent. They may choose areas that voted for them to consolidate their base or areas that did not to increase their profile there - if recall of representatives is allowed in the electoral system, it is those electors only who have the power to recall their circuit's member.
[Geographically-large less-populated districts get single members] Proponents of status quo first-past-the-post (FPTP) often cite the difficulty of campaigns or representation of very large thinly populated districts. Unlike most STV systems, in the recommended variant the largest mostly-wild regions or those with unique cultural or ecological differences may elect only a single member at a time and are accordingly weighted less in the party-proportional round, to ensure each vote counts the same. For instance, a district that by population is entitled to only 0.7 members would elect one but would contribute nothing to the party-proportional round and have no chance of its local candidates being selected in that round (unless selected as leader or deputy in that round). Also, the elected party would be relatively over-represented in the legislature based on popular vote, and would thus stand less chance of electing another member in the party-proportional round. Where a district is entitled to 0.1 up to 1.9 members, the single transferable ballot acts as an instant-runoff ballot: Marking the ballot can be made as simple as a first-past-the-post ballot by simply interpreting an X as a "1". Since it's not mandatory to identify a second or third choice on any ballot, a simple X clearly intends a rank of 1. Though any ballot including both an X and a 1 would be ambiguous and thus spoiled, other numbers are unambiguous. The ballot is thus relatively robust from some of the challenges of large thinly populated districts where media reach, literacy, native language use, may present barriers to teaching everyone how to use a new ranked ballot.
[Rank one to all candidates] In the recommended variant, electors may rank as many candidates as they wish. As in all single-transferable-vote systems, their vote only counts towards one; If they rank only one or a few unpopular candidates it may not count at all in the proportional round. Some variants could require voters to rank all candidates or as many as will be elected, but this is not recommended. Regardless of the outcome of this STV round, the party (or "label" to avoid assuming that parties are the only form of organizing) affiliation of their top-ranked candidate is recorded separately and counts towards the party's total. This is weighted (or multipled) by the degree to which their district was under-represented in the STV round and used to determine which of that party's unelected candidates will also go the legislature. So, as explained above, a candidate for an under-represented party who barely misses an STV seat in a district electing 2.8 members is twice as likely to go to the legislature than one who misses such a seat in a 4.4 member district. Some variants could weight the party support also but this is not recommended because it is open to attack as being "not really proportional" across the province. If the party-wide popular vote for each party is the sole factor determining which party gets the next party-proportional seat this criticism is reduced, though at the (acceptable) cost of over-representing districts that strategically vote for parties that they know will receive seats in the final PP or L round.
[Natural optimum number of +L party-proportional seats] There is a natural emergent optimal number of PP or L seats – removing the necessity of arbitrarily setting aside a proportion (10%, 15%, 25%, 33%, 50%) of seats ‘for parties’ – avoid this is so desirable that the recommended variant sets aside only the natural number that arises from the number of districts and how they are divided. Because probability of a fixed district being under-represented by 0.1, 0.2, ...0.9 members is approximately equal, an average of 0.5 additional "label" or "party-proportional" members will be elected by votes from each district. Assuming a 100-member legislature that elects its members in 20 districts averaging 5 members each, 10 members (20 x 0.5) or 10% of the total would be party-proportional. Assuming a 210-member legislature that elects its members in 70 districts averaging 3 members each, 35 members (70 x 0.5) or 16.7% of the total would be elected in that second round. A 52-member legislature similar to Nova Scotia's, electing members in 13 districts averaging 4 members each, but one of which was single-member, would elect only 6 members or 11.5% of the total in the second round. Though this number seems small it is sufficient to elect three leaders of small parties, a simultaneously-elected deputy from the largest of these, and two more members for the most under-represented of the major parties in the legislature from the most under-represented districts. This allocation thus achieves very nearly party-proportional and geographically-proportional final results with no need for "party lists" or other undemocratic mechanisms that transfer power or control over the legislature from the public to parties and without so frequently handing the balance of power in the house to persons elected indirectly or affiliated with smaller parties (as in MMP). Examination of recent Nova Scotia results (see below) suggest no controversies arise.
[Leaders and simultaneously-elected deputies of small parties may get first party seat] Because smaller parties gain influence and leverage in such a system, there is a need to somewhat tighten the rules under which they operate, as they are often poorly scrutinized and taken over by cliques or narrow interests. Our very strong recommendation is that while the system may make deliberate accommodation for party leaders (s/electing them first in the second round) it should do so only if those leaders are themselves subject to a relatively open leadership selection process. Deputies should only be eligible if selected on the same ballot by the same voters as the leader and ideally should be selected together on a "ticket" (as the US President/Vice-President are). This prevents internal factional squabbles in small parties from becoming a distraction in the legislature, it reduces small party leader (often a thankless job) leverage to threaten to resign, it reduces the chance of such parties being caught off-guard in snap elections, and (most importantly for the public) it also guarantees that only relatively well-scrutinized races can select +L or PP members. Whether to legally require a non-leader to step down in favour of a leader without a further ratification by the public is an open question in this system; This author advises that no such requirement should be part of the electoral system but some might be enforced as part of a party's own constitution; Accordingly a member may be ejected from a party for not stepping down for the leader or deputy, but would not be deprived of his or her seat, even a "+L" seat, if they stood for that party in the election and were honestly selected by voters who believed they would take the seat. It is not an unacceptable result to the public if a popular elected member from a party that isn’t in accord with the party’s leadership gains influence by acting as its House Leader – very often this forces accommodations and evolution of the small party’s position or role and educates the party in the realities of working as a minority in a parliamentary system. It will also elevate persons popular with the public to potential leadership roles in a party. This general public influence over small parties compensates for the threat of their taking balance of power and forcing unpopular policies. At the very least, unpopular policies of a small party forced through by a legislative wedge must be advanced by popular persons and accordingly have every chance of being heard out in the media before changes occur.
[Democracy Commission educates the public and monitors parties] Because political parties change their roles and structures under any deep reform, this author recommends a permanent Democracy Commission be set up with broad terms of reference to 1. educate the public about civics and best practices 2. audit parties and recommend changes to their rules or regulation 3. improve democracy itself by constantly surveying government performance and response. Only a relatively educated public is likely to choose a more complex hybrid such as BSTV+C+L over less balanced systems. A key role of such a Commission would be auditing party internal governance to ensure that whatever rights they guarantee their members or their members reasonably expect due to the party’s values and traditions are honestly advertised when soliciting members or donors. This may help prevent internal scandals suddenly coming to public attention.
[Use allocation voting for major public consulations, e.g. budgets] The system proposed has a variant in which allocation voting is used, which the authors recommends allow both positive (approval) and negative (disapproval). This variant may be used for electing individuals but may be easier to justify using for genuine allocation problems such as budgets. If used in elections, it provides parties and candidates with much more explicit feedback about their performance and lets them indicate disapproval directly and ranking indirectly. For instance, a voter may allocate 3 positive votes to one candidate they strongly support and one negative vote each to two they disapprove of, for a total of 5 (in the recommended variant). Or, 3 positive to their preferred candidate and 2 more positive to their second choice. Or, if they wish to approve only a general direction of policy, one positive vote to each of four good candidates and one negative to a bad one. Or, if their entire experience of politics is disgust, 5 negative to one disgusting candidate. A post-election assessment of voter sentiments, intentions, ambitions and tolerances would then be much more likely to be accurate than in any other system. Increased expressiveness would attract voters to actually cast a ballot, better feedback would improve responsiveness of candidates and parties to criticism, and awareness of how broad or narrow their support was would change how parties campaigned. Such an allocation voting system however is probably best tested extensively (not in elections) before it is used to elect persons, as no such system is in use anywhere now in the world. The “Borda count” system which turns ranking into weighting is also not recommended.
Introduction: the purpose and benefits of the BSTV+C+L proposal
General advantages of reconciled STV+MMP with fixed districts The major benefits of reconciling STV+MMP into a fixed-district fractional-member two-round system combining both are:
1. district borders can be fixed, ending gerrymandering, if fractional numbers of members can be supported (which any combined STV+MMP system can do, it simply weights the undercounted votes differently from each district)
2. unlike MMP's arbitrary prior threshold of votes to receive a seat, any combined system can arrive at a common threshold for all parties to get a seat organically, that is, one can see after that fact that any party that got more than X% got a seat, but all those that got X%-1 did not.
3. having two rounds lets some decisions that would otherwise be arbitrary such as random vs. fractional transfer, static vs. dynamic thresholds, means of breaking ties, eliminating last-ranked vs. choosing first-rank, be done one way in one round and the opposite way in the other round so there are fewer reasons for supporters of any given method to object
4. unlike MMP, the number of make-up seats in the second round are small, so objectionable "party lists" can be eliminated, members given more or less flexibility to give up seats to leaders elected by the whole party but also not allowed to give them up to a leaders' cronies or party insiders, which were never chosen in a well-publicized election: A vote for a party from someone who doesn’t care who the leader is, can’t be corrected, but it’s fair to say that most people know only the leader and the local candidates and not every candidate that might be on a long list.