The line is question is from La patrie et le partis, 1865, p14-15. The original: “Dans un État démocratique, le droit de décision appartient à la majorité, mais le droit de représentation doit appartenir à tous.”

Washington, D.C. is currently mired in some controversy over the need for post-census redrawing of our Ward boundaries. I’ve also been reading some analysis of the new congressional district plan out of Democratic-controlled Illinois, and I’m anticipating looks of shock and horror when progressives see the maps Republicans come up with in Pennsylvania, Ohio, and Michigan. Then out will come the goo-goo brigades aghast to learn that members of Congress are representing funny-shaped districts.

At times like these, I always recall a conversation I once had with a law professor who told me she teaches proportional representation last in her class about district-drawing issues because it solves the problem too easily and the students need to learn.

Here’s how it would work. The state of, say, Minnesota would need to elect eight House members. So everyone who wants to represent Minnesota in the House will run. Then the voters will rank as many candidates as they want, in order of preference. Anyone who gets more than 1/8th of the first choice votes wins. If that gets us fewer than 8 members of congress, we start doing two things. One is we redistribute “overvotes” from candidates who got more than 1/8th of the first preferences. The other is we start striking the candidates with the fewest votes off the list and looking at their second choices. Eventually, we’ll have eight members of congress. Parties and pressure groups will endorse not just specific candidates, but specific preference orderings. So the Minnesota AFL-CIO and the Minnesota Sierra Club might rank their choices in different order even if they have the same eight favorite candidates. In this system, there’s no such thing as a “safe seat” and no such thing as a gerrymander. If African-Americans show a strong inclination to cluster their votes and back a black candidate, then we’ll have black members of congress without needing to conjure up special majority minority districts.

Really big states will still need some districts, of course, but the details of dividing Texas or California up into four or five big regions wouldn’t have nearly the same scale of consequences.

Semi-proportional methods, such as the limited or single non-transferable vote (SNTV) or cumulative voting, sometimes proposed to encourage minority representation, have the serious flaw that they achieve proportionality only when groups of voters can cooperate to implement voting strategies. To the extent that any of these systems are simpler to implement than true PR (and we contend that they are not), that simplicity is more than offset by the requirement for complex voting strategies.

The new page explains why STV PR is simpler for the voter as well as being more effective in achieving proportional minority representation.

• STV PR gives voters, in aggregate, complete control over the election of individual candidates, while List PR gives them little or no control over the choice of candidates, and relatively little control (often none) over which candidates from a chosen list are seated.
• Lists are created by parties, leaving independent voters out of the nomination process and independent candidates off the ballot. As a practical matter, a candidate cannot appear on more than one list, making crossover voting impossible .
• Because lists are associated with parties, List PR isn’t useful for the non-partisan local elections that are common in the US.
• STV PR district size is limited by the practicalities of having many candidates on a ballot, and of voters being able to meaningfully rank dozens of candidates.

The basic idea of List-STV PR: a voter votes for a list of candidates, much as in a conventional closed-list election. When the votes are counted, rather than allocating seats to lists proportionally to the votes each list receives, the ballots are treated as votes in an STV election: each voter is deemed to have cast an STV ballot corresponding to the list voted for.

The benefit so far is modest but significant: because of the nature of an STV count, lists can overlap, so candidates can appear on more than one list and benefit from each vote for those lists. But the benefits multiply as we look at a series of simple enhancements.

1. Groups of voters can put forward lists by petition. The group might be associated with an issue (the environment, development, taxes), or might be a subset of a party (progressive or Blue Dog Democrats, Tea Party or Log Cabin Republicans). These lists could add new candidates to the mix, but could also reorder candidates from the main party list.

2. Voters can rank individual candidates, optionally followed by a list. The list would be used to fill out the voter’s preference ranking, ignoring candidates that the voter had already ranked higher. Voters can express their individual preferences while still giving a complete preference order when the number of desired rankings is large.

3. Voters can rank multiple lists, one after the other. This becomes useful when an issue-oriented group, perhaps local, strongly advocates a short list of candidates.

4. Lists can refer to other lists. The Sierra Club, strongly associated with the Democratic Party, might rank several environmentally oriented candidates first, and then finish with a reference to the Democrats’ list. A local Sierra Club chapter could rank a few local candidates first, and finish with a reference to the state chapter, or to the national Sierra Club list.

5. Petition lists could be local, with the number of petition signatures required proportional to the size of the locality. Thus, in a statewide election, a petition list could be put forward within one city, or within a county, and only appear on ballots within that area. This would encourage local lists, with a lower bar, while not cluttering all the ballots across the state.

What are the benefits?

Compared to List PR, the List-STV PR hybrid gives the voters (and groups of voters) much more power and flexibility in influencing the order of election of candidates: it’s a very individualized, very open list, with characteristics of conventional open lists and petition lists (eg Norway), and more.

Compared to STV PR, List-STV PR makes larger districts and longer candidate lists practical. Voters are unlikely to be able to usefully rank more than a handful of individual candidates, but parties and larger groups have the resources to vet and rank longer lists.

Moreover, the number of rankings available to a voter is typically limited, for practical reasons relating to ballot-reading equipment or ballot-space considerations. The ability to incorporate an entire list with a single ranking would alleviate this restriction considerably.

The voter experience under List-STV PR is straightforward. A voter can simply rank individual candidates, as with STV PR, or rank a list, as with List PR. Or a voter can rank one or more individual candidates, followed by a list. Finally, a voter might rank individual candidates, then a local list, and finally a larger list.

Because lists under List-STV PR are not tied exclusively to parties, List-STV PR works well for non-partisan elections, and indeed works fine with no lists at all, becoming simply STV PR.

A matter of perspective

List-STV PR can be viewed as an extension of List PR or of STV PR, depending on one’s perspective.

From the perspective of List PR, List-STV PR:

• allows candidates to appear on more than one list
• allows voters to create their own lists by ranking candidates
• allows voters to rank multiple lists

From the perspective of STV PR, List-STV PR:

• allows voters to rank lists as well as candidates
• …which makes it easier to rank many candidates

On the other hand, it is worth noting that under PR/MMD, should a constituent need to contact a representative, she will have a choice among the representatives of her region and may take some comfort in the fact that a member of her party serves as one of her representatives (if that’s the case).

It’s a good point: a single representative is replaced by a group of representatives, with the advantage that at least one, and likely more, are likely to be politically aligned with the constituent.

We’ll have more to say about the proposal later, but for now, here’s a summary and a link to a downloadable copy of the paper.

The State of California elects a bicameral legislature through a first-past-the-post electoral system. Beginning in 2012, the election of California state legislators will be through an open primary, two-round system. Neither of these electoral systems provides for proportional representation in the legislature. Accordingly, neither system leads to a truly representative democracy. The author recommends that the state adopt a proportional representation electoral system — specifically, a mixed-member system — as a more democratic alternative plurality/majority systems now in place or projected to be implemented in 2012. In addition, the author explains why the state legislature should be unicameral and increased in size from the current 120 legislators to 320.

John Quiggin is on the scene:

Habemus PM

The Australian election three weeks ago turned out about as close as possible. The two main parties (Labour and the permanent Liberal-National-Liberal National coalition) each ended up with 72 seats (out of 150) and almost exactly 50 per cent of the two-party preferred vote, the relevant measure of support in our preferential (=IRV/AV) system. That left six remaining seats: one Green, one non-coalition National, one leftish independent and three country independents, all formerly associated with the conservative National party). Because the Parliament has a Speaker, 76 supporters are required for a stable government.

Unsurprisingly, things took a while to sort themselves out. Because of postal voting and the need for recounts, the final determination of seats took more than a week. Then there was another week of haggling and jockeying. The Green MP declared for Labor first, followed by the leftish independent (Labor) and the dissident National (Coalition). No surprises there. That left the three country independents. It was expected they would move as a bloc, but in the end, one announced support for the Coalition, and the other two for Labor (the last of them spending half an our of explanation before finally stating what had been obvious from the moment his ally went that way). So, after 17 days, it was 76-74, and Julia Gillard retained the office she had snatched from Kevin Rudd only weeks before the election.

Overall, it was a startlingly good outcome. Any democratic system is going to have trouble when the vote is as close as this, but compared to the US in 2000, or Belgium/Holland right now, things went relatively smoothly. And, startlingly, to get the independents on board, Labor actually had to promise better government, rather than pork-barreling for those electorates fortunate enough to have a pivotal vote. By contrast, the Liberal leader Tony Abbott, came with an open chequebook and was rebuffed. It’s true that the effect will be to give much more favorable treatment to rural and regional areas in general, but the independents have a fair enough basis for the claim that these areas have been neglected (complex and competing calculations of the relative treatment of urban and rural areas are a staple of Australian policywonkery).

Even better, when the newly elected Senate takes its place (not until July 2011 thanks to the marvels of our electoral system) Labor’s dependence on the Greens will be enhanced by the existence of a Labor-Green majority in the Upper House. Going into the election, Labor had dumped the commitment to action on climate change that gave it victory in 2007 (how this happened is too depressing to relate. I think George Monbiot covered it a while back). But now, with the government dependent on Greens and greenish independents, the issue is back on the agenda.

It’s often said that a country gets the government it deserved. Going into the election, with two competing leaders who had seized power without any real popular support, and policy platforms derived entirely from particularly dimwitted focus groups, I wondered what we Australians had done to deserve this. Now, I wonder how we merited such good fortune. I only hope it will last.

While STV itself is much simpler than most people think (if they think about that kind of thing, anyway), getting a specific rule exactly right is a subtle matter, and many organizations and governments have gotten it wrong. To that end, we’re putting together a set of “reference” or “model” STV rules, so far one each for the Weighted Inclusive Gregory Method (WIGM) and for Meek’s method.

The Droop STV-counting software will include implementations of all the reference rules; right now it supports the Meek and WIGM reference rules, named meek-prf and wigm-prf in Droop.

Besides the core counting rules, we’ll have language for related matters, such as validating ballots, handling casual vacancies, and election audits.

Stay tuned.

The answer, of course, is that (nearly) everybody does. Below is a list of countries that use PR for national elections (source: Wikipedia). If we included countries that use PR for at least some local elections, the list would be considerably longer, and would include the United States.

The single biggest reason that the US doesn’t use PR nationally is that the US was invented a century before proportional representation (see the PR Timeline). Having grown up with single-member congressional districts, we find it hard to envision alternatives, and harder yet to actually make a change.

Here in the US, it’s perhaps most practical to introduce PR for local elections, get used to the idea, and move on to state and eventually federal elections. And before that happens, we need education on the subject.

That’s what the Proportional Representation Foundation is all about.

Algeria
Angola
Australia
Austria
Argentina
Aruba
Belgium
Bolivia
Brazil
Bulgaria
Burkina Faso
Burundi
Cambodia
Cape Verde
Colombia
Costa Rica
Croatia
Cyprus
Czech Republic
Democratic Republic of the Congo
Denmark
Dominican Republic
Equatorial Guinea
Estonia
Finland
Germany
Greece
Guinea-Bissau
Guyana
Hungary
Iceland
India
Indonesia
Iraq
Ireland
Israel
Italy
Japan
Latvia
Lesotho
Liberia
Liechtenstein
Luxembourg
Malta
Mexico
Moldova
New Zealand
Namibia
Netherlands
Netherlands Antilles
New Caledonia
Nicaragua
Northern Ireland
Norway
Paraguay
Peru
Poland
Portugal
Romania
Russia
San Marino
São Tomé and Príncipe
Scotland
Slovakia
Slovenia
South Africa
South Korea
Spain
Sri Lanka
Suriname
Sweden
Switzerland
Taiwan
Thailand
Turkey
Ukraine
Uruguay
Venezuela
Wales
Wallis and Futuna

Like the US, Australia has a two-house legislature, with a Senate with 12 seats per state (and four territorial seats), half of which is elected at time, and a House of Representatives with 150 seats, all elected at once.

The Senate uses proportional representation (STV-PR), with six senators at a time elected from each state. The House of Representatives uses IRV with single-member districts: not proportional.

In the current preliminary results, the Greens received 11.4% of the first-place House votes, but elected only 1 of 150 seats: 0.67% of the seats.

In the Senate, the Greens received about 13% of the first-place votes, and ended up with about 12% of the seats (9 of 76).

An eighth of the population is left with virtually no voice in the non-proportional House, but with an eighth of the seats in the proportional Senate.

It’s really as simple as that. As Ernest Naville wrote in 1865, “In a democratic government the right of decision belongs to the majority, but the right of representation belongs to all.”