May 1, 2009

STV: A better democracy and more progressive politics


Dennis Pilon from UVic published this excellent article on STV in the CCPA Monitor, so I am republishing it below:

MAY 12 DECISION DAY FOR B.C. VOTERS: Change to STV system would be helpful to progressives

By Dennis Pilon

May 12, 2009 will be the “make-or-break” day for voting system reform in British Columbia, and perhaps the country. On that day, B.C. voters will decide in a province-wide referendum whether to change the provincial voting system.

Since 2005, there have been three referendums in Canada asking voters if they would like to change the voting system – in P.E.I., Ontario, and B.C. – and all three have failed to pass. Of course, B.C.’s referendum “lost” with 58% of the popular vote, which is why the province is putting the question back on the ballot in the 2009 general election.

If B.C. voters turn down this second chance to discard the existing undemocratic first-past-the-post (FPTP) voting system, it will send a message to political élites that they can ignore the issue of electoral system reform for a long time. The choice is stark and limited: voters can choose to keep the existing FPTP system or opt for the single transferable vote (STV) form of proportional representation (PR), as recommended by the B.C. Citizen’s Assembly.

The problems with the current system are well known by now: phony majority governments, uncompetitive elections, unrepresentative results in terms of party representation and our social diversity, strategic voting pressures, and so on. And yet progressives forces in B.C. appear to be, at best, half-hearted and divided about the decision they have to make.

A considerable academic literature exists detailing the flaws of B.C.’s traditional first-past-the-post (FPTP) voting system. We call it “representative democracy,” but how well does it represent what individual voters want? How well does it represent the collective votes cast for each political party? How well does it represent the social diversity of our society? The answer in all cases is: not very much, if at all.

FPTP does a poor job of translating what individual voters say they want their votes to produce. In every election, roughly 50% of the votes, on average, fail to contribute to the election of anyone. FPTP tends to exaggerate the support for the larger parties, allowing them to turn as little as 40% of the vote into as much as 70% of the seats. And it also inflates the support for smaller parties, such as the Bloc Québecois, whose support is concentrated in one region.

FPTP also fails miserably in representing the diversity of the populace. Its legislatures look completely different from their voters. Few women, people of colour or indigenous people get elected, and working people of all ethnic groups and cultures are left out. The problem is in the all-or-nothing nature of the FPTP system. Because there can be only one winner in each riding, the winners tend to be those with the power and financial clout to win nominations – typically affluent white males.

The FPTP system is patently adverse to free and unfettered political contests. It places many barriers in the way of new political forces (such as the Green party) intruding on traditional party turf. Because it’s a “winner takes all” vote, new parties have to gain at least 40% of the ballots cast to be competitive. With such a high threshold for new parties, voters tend to be conservative, voting strategically for candidates and parties they think can win, even when they might want to vote otherwise. Even many supporters of the bigger parties are poorly served by the present system, in the sense that many of them are “orphaned” when their votes for a losing candidate go for nought. Reminding voters of how constrained their choices are under FPTP might whet their appetite for a voting system that would give them more freedom to choose as they like.

STV and PR myth-busting

On their website, the No-STV forces offer a concise set of reasons to oppose changing the voting system: STV is complicated, confusing, prone to errors and delay; it reduces local accountability, increases the size of ridings, allows MLAs to avoid direct accountability for their decisions, increases party control, and allows special interests to dominate party nominations.

None of these claims is backed up with any credible evidence. Let’s take each in turn to assess how serious these problems really are.

Simplicity vs. complexity

Critics of STV claim it is too complicated. They compare the ease with which votes under FPTP are tabulated with the seemingly Byzantine complexity and time-consuming nature of counting ballots under STV. They claim this complexity may confuse voters at the ballot box, perhaps disenfranchising some of them, or preventing them from understanding how the count is conducted and where votes go, leading perhaps to incidences of fraud. Neither claim can be backed up with evidence.

The present system is admittedly simple when it comes to counting the ballots on election night. But its results are often far from clear. For instance, if the critics of STV asked voters to explain how the Liberals gained 97% of the seats with just 57% of the votes in the B.C. provincial election of 2001, very few would be able to do it. On the other hand, while few would dispute that STV is more difficult to count (though computers make this less of an issue), its results are often much more transparent. If we look at Irish election results under STV, the percentage of the popular vote for different parties is pretty close to the percentage of seats they win. Thus the results make intuitive sense: parties with 10% or 20% of the votes tend to get 10% or 20% of the seats.

The question is not about simple or not simple, but which trade-off in terms of simplicity makes the most sense, given what voters need. Is it more important to have easy counting but suffer with less than transparent election results? Or should we risk complicated vote counting but have results that match up with voter expectations? Given that few voters are actually involved with the vote counting process but nearly all want to know the results, clearly simple results should trump simple vote counting. Irish voters tend to see it this way, and have twice voted to keep the STV system when politicians have tried to get rid of it.

Critics of STV argue that its relative complexity will create confusion and chaos at the polls. The evidence from the jurisdictions using STV refutes this claim. In Ireland, where STV has been used for nearly a century, people seem to grasp how it works, and this is also true in Australia, which has used the STV system for its Senate elections since 1949 with no problems. Even Canadians have used it – for about 35 years at the provincial level in Alberta and Manitoba – and again the evidence is strong that voters have few problems voting by STV.

Complaints about corruption are also baseless. What such claims overlook is that the public is not the key watchdog over any electoral procedure. Voting systems are effectively watched over by the competing political interests. If any funny business were going on with STV in Ireland, for instance, clearly the main parties affected would notice and be loud in their opposition to it. The same would be true if STV was introduced in B.C.

STV not anti-party

STV critics claim that the system will ruin parties because it forces members of the same party to run against each other. Because each member of the same party will want voters to rank them first, they will have an incentive to slag each other and this will lead to legislative parties that cannot work together effectively. If we turn to STV-using jurisdictions, however, this is not what has come to pass. Ireland has maintained a stable party system, and strong party discipline has characterize its politics for most of its history as an independent state. Two nationalist parties have traditionally dominated the legislature, with a small Labour party gaining some support. Since the 1970s, two new parties have emerged, a centrist party and a Green party, but the system remains stable and anchored by the two traditional large parties.

Australia, too, has seen STV produce stable party dynamics in Senate elections, with some openings for new parties. So, on the whole, the evidence of what happens in countries using STV is that stable party systems emerge, but the systems are very competitive, which means that, when voters do want an alternative, they can get one.

Local links

Whether one accepts the criticism that STV would weaken the local link between an MLA and local constituents depends on how we understand the strength of that link in B.C.’s current FPTP system. Most “local” people in B.C. do not feel represented by their local MLAs because they did not vote for them. Of course, under STV, many more voters might find a local representative who shares and supports their views. But a bigger problem with the “local representation” argument is simply that it is a lot of romantic nonsense. Academic studies of voter behaviour regularly report that people do not vote on the basis of local identities or issues, but rather vote on the basis of party identification and the broad issues that the parties represent.

Critics are saying we should keep a system focused around a feature – local representation – that no one really cares about politically and, in doing so, sacrifice the one thing – accurate party representation – that we do know people want.

The local link rationale does not even stand up within the context of how our present voting system has worked. The facts are that this alleged concern for good local representation has not stopped politicians from ritualistically increasing the size of these “local ridings” over time. B.C. ridings have doubled in population size since the 1950s, hardly the mark of a system that values a close link between electors and the elected.

Another line of argument here focuses on the geographic size of the STV ridings in B.C. and the challenges that will arise in campaigning across vast territories, particularly in the province’s interior and north. Some say this will particularly hurt female candidates. But the critics overestimate the degree to which campaigns happen locally. Most of the costs in elections are province-wide, since that is how messaging and advertising is focused. Furthermore, the increased size of STV ridings will be offset by the lowering of the threshold to win. A winner under STV just needs to get a proportion of the total, rather than more than any other single candidate. This will alter the campaign strategies of candidates and their parties.

Will it hurt women? Some parties have adapted to make running for office more viable for women under the current system and they will do the same under STV, aided by the way PR systems in general promote diversity. Basically, more women and visible minorities tend to get elected in PR systems in Western countries.

Critics claim that STV is bad for diversity and the representation of women, and here, at least, they do have some evidence, citing Ireland, Malta and Australia as examples. Some who get into the details point out that STV ridings are too small. Good PR systems, they claim, have multi-member ridings of 10-20, whereas STV typically has just 3-5 members elected in each riding. Others simply look at the percentages, and point out that the number of women elected to national legislatures using STV is not impressive. But this is a misreading of the evidence. Women have actually made progress with STV since the 1970s in all the jurisdictions where it is used. Their numbers still fall well short of matching male representation, but this is not because of any barrier inherent to STV, any more than it is in the FPTP system.

No serious student of representation and voting systems has ever claimed a direct causal relationship between using one system and getting a specific result. Experts note how social factors combine with the incentive structures of voting systems to create openings. Culture matters. It should be noted, in that context, that many STV-using countries are culturally if not politically conservative. A PR system is structurally more democratic and competitive, but it still requires political action and pressure to be appropriately responsive to social needs such as diversity. Given the influence of the women’s movement in B.C., the introduction of STV in this province would make a fascinating laboratory to see how these social and institutional factor would interact.

Not PR, not PR enough

Recently, some have tried to argue in B.C. that STV is not actually a form of PR at all. They look at the election results in Ireland and Malta and note that there is not a perfect correspondence between popular votes and seat totals. Of course, anyone who studies voting systems seriously knows that none of them translates votes into seats in an absolutely proportional manner. All systems tend to over-represent the largest parties, for instance. But we still call them PR. Thus this charge that STV is not PR is clearly unfounded, but others make a more serious case that STV is simply not proportional enough. Here they reason that, with multi-member constituencies of just 3-5 members, the threshold to get a seat could be very high – that in a 5-member riding, for instance, the quota to get a seat would be roughly between 17% and 20%. In B.C., with existing voting patterns, that would mean that the Green party would be shut out, despite getting between 8% and 12% of the provincial total.

Again, however, an examination of the facts tells a different story. In Ireland, despite ridings of just 3-5 members, the results have been roughly proportional. The Labour Party gets about 10% of the national total and regularly gets 10% of the seats. The Progressive Democrats regularly poll 5% and get 5% of the seats. Even the Greens, with just 2%-3% of the national popular vote, gain 2%-3% of the seats.

What the critics forget is that support for smaller parties, while small at the national level, can be concentrated regionally, thus allowing them to win seats in particular areas. This would probably be the case in B.C. as well, with the Greens benefiting from strong pockets of support in different multi-member ridings.


The claim that STV or any other form of PR leads to political instability tends to focus on anecdotal evidence from just two countries: Italy and Israel. Italy is alleged to have had “50 governments in 50 years,” while Israel is accused of excessive fragmentation of its party system and weak coalition governments.

Of course, the story is more complicated in these two cases. In Italy, many of the so-called different governments were more like cabinet shuffles in the British parliamentary system, as many of the same people were in successive “governments.” In fact, some argue that Italy’s problems involved too much stability, rather than the other way around, as one party (the Christian Democrats) dominated the polity from 1948-1992.

Israel, too, is poorly understood if people think the country’s instability is just the product of its voting system. Indeed, the voting system itself was a product of a society that thought it required maximum political inclusion to weather the challenges it perceived to its existence from without.

But the real problem with such instability arguments involves the narrowness of the evidence. Most Western countries use PR, not just Italy and Israel, and they have produced very different results that those two countries. If we use the frequency of elections as a measure of stability or instability, we discover that PR countries have actually had fewer elections, on average, than plurality-using countries. So PR countries, on average, have enjoyed very stable government.


Voting system reform in Western democracies is very rare, and public input into the choice is almost unheard of. Until the recent spate of Canadian referendums, there were just three historical examples (Switzerland in 1918, France in 1946, New Zealand in 1992-93). So anyone thinking that the defeat of STV in this second B.C. referendum will still leave a choice in the future for some other kind of PR had better think again.

An unusual set of circumstances created this opportunity for electoral reform in B.C. — wonky election results in 1996 and 2001; a Premier with a one-track mind for pet projects; a Citizens’ Assembly keen on reform — and, as is often the case in history, voters have to make the most of the rare opportunities that arise.

The vote before us is STV or the status quo. There are no good reasons to stick with the existing voting system and many good reasons to embrace STV. It may be the last chance for voting system reform in B.C., and indeed for the whole country.

It would be tragic to miss this chance to make history and create a political system that is more democratic and thus more conducive to success in the struggle for a better province, a better country, and a better world.

Dennis Pilon is an assistant professor in the Department of Political Science at the University of Victoria, a national council member of Fair Vote Canada, and the author of “The Politics of Voting: Reforming Canada’s Electoral System.”