As the Liberals gear up for October’s federal election, there is much talk of reconnecting with Canadians and remaking the party’s policy profile. One of the party’s big ideas is democratic reform. Justin Trudeau ran for the leadership on democratic renewal, including a specific commitment to champion a new voting system: the alternative vote (AV). In a leadership debate in Vancouver in 2013 he declared that AV would “change the tone of politics completely” because politicians would need to reach out to voters of other parties. Since then, other senior Liberals have echoed his views. In August 2014 Deputy Leader Ralph Goodale asserted that AV “requires you to focus on what pulls people together, not just how much you can vilify your opponent. It would have a very interesting impact on the tone of the political debate.”1
In fact, evidence of broad Liberal support for AV preceded Trudeau’s ascension to the leadership. In 2012, AV was overwhelmingly endorsed at the party’s national policy convention. Former Liberal MP Omar Alghabra captured the sentiments of the delegates in a commentary entitled “The Voting System Canada Needs (and Deserves)” in which he argued that AV would prevent MPs from winning their seats with less than a majority of the support in their ridings. He noted that in the 2011 election the Conservatives turned 39.6 per cent of the popular vote into 53.9 per cent of the seats and that “such disproportionality creates a sense that our system is unfair.”2 Alghabra argued that AV was a “practical proposal” that would retain the good features of the existing single-member plurality (SMP) voting system, namely a directly elected local member, while avoiding the problems he and other Liberals associated with more radical proposals like proportional representation (PR), specifically too much party influence and a potential increase in single-issue politics.
It was thus surprising to see more than half the Liberal caucus vote in December 2014 for NDP democratic reform critic Craig Scott’s motion to adopt PR for federal elections. It would appear that Liberals are not entirely united behind AV as their electoral reform option, with much depending on how they understand the electoral dilemmas facing the party. Those confident that the party can vault past the NDP back into government or at the very least official opposition tend to favour AV. Those worried that vote splitting among Liberal, NDP and Green voters could perpetually place the Conservatives in power are inclined to opt for PR. The latter group, which includes former leader Stéphane Dion, flexed their muscle at the federal Liberal policy convention in February 2014, forcing an expansion of the electoral reform platform to include considering PR in addition to AV. Party leader Justin Trudeau has vowed to take an “evidence-based approach to electoral reform rather than an ideological one”3 – presumably meaning that a Liberal government would judge proposed voting systems by how closely they measure up in practice to the claims made about their workings in theory rather than how well they serve the interests of the party introducing them.
I propose to take him at his word, assessing claims made by Liberals about the potential impact of AV, including that it will limit strategic voting and wasted votes, encourage cross-party cooperation electorally and legislatively, increase public engagement with politics and lead to better representation for Canadian voters. I do so by evaluating the experience of states comparable to Canada that have used AV. Good examples include Australia, which has used AV for its lower house elections since 1919, and three Canadian provinces that used it in the early to mid-20th century.4 By examining Australia’s near century of experience with AV and the 17 elections using it in Alberta (1926–55), Manitoba (1927–53) and British Columbia (1952–53), we can assess how well this voting system facilitates strategic coordination among parties and their voters, empowers voters to make choices and represents what they choose with their votes.
The alternative vote and strategic coordination
The alternative vote is considered a majoritarian voting system because it works to assure that the winning candidate in any given district has gained a majority of the votes cast. Voters rank-order their preferences among the candidates to influence the selection of single winner. After the votes are cast, returning officers add up all the first choices indicated on the ballots; if any candidate has gained a majority of those votes, that candidate is declared the winner. But if no candidate gains a majority on the first count, then the candidate with the lowest number of first choices is eliminated from the count and their votes are redistributed on the basis of the second preferences indicated on the ballot. This process continues until a single candidate has secured more than 50 per cent of the votes. (AV can be and has been used in multimember districts as well, but it is the single-member version that is under consideration here.)
At a glance, the appeal of AV to today’s Liberal Party is obvious. Anticipating a vote pool that stretches from the centre-right to the centre-left, the party is best located to benefit from vote transfers from nearly all other parties, while having significant first-choice support itself. AV could be a recipe for Liberal electoral dominance, depending on how competitive the party is across the country.
At the same time, AV appeals to Liberals because they know there is little likelihood of electoral cooperation with, or the elimination of, their historic NDP (and more recent Green) rivals. Thus proponents claim AV would allow likeminded voters and political parties with considerable policy overlap to avoid SMP’s zero-sum competitive dynamic, which basically amounts to a game of electoral “chicken” where parties have to worry about splitting the available pool of votes and voters fear wasting their vote by making the wrong choice. Under AV, they argue, voters would have an improved ability to choose, while different party organizations could maintain their independent existence.
Clearly such a possibility would have appeared attractive to Reform and Progressive Conservative voters and politicians circa 1993 to 2003,5 as it does on the centre-left today. But would AV in fact operate in this way in Canada? Comparative experience suggests that it depends on the presence of factors seldom considered in Canadian advocacy of the reform.
The key point of reference for AV in Canada is the experience with it in the Australian lower house. Briefly, since 1919 AV has allowed the two centre-right parties (now known as the Liberal and National parties) to share the pool of centre-right votes, thus preventing their common electoral foe, the Australian Labor Party, from exploiting a split in their voting support. More recently, Labor and the Green Party have shared the centre-left vote total with the aid of AV.
Should we expect the Australian experience to be reproduced here? The foremost expert on the use of AV in Canada is University of Lethbridge Associate Professor Harold Jansen, whose research has examined the workings of the system in 17 elections in the three provinces that experimented with the system. He was particularly interested in whether AV in Canada encouraged greater party cooperation electorally and legislatively. In comparing Canadian experience with Australia’s, he points out that the Australian parties adopted AV only after they had committed to electoral cooperation.6 Earlier attempts to adopt majority voting in Australia in the absence of such commitments failed. It was only when the two centre-right parties faced a possible Labor takeover resulting from a three-way vote split that they finally settled on AV.7
In contrast, Jansen finds that AV was not adopted to secure electoral cooperation between parties in either Alberta or Manitoba and found little evidence that it encouraged such cooperation over the period of its use. In British Columbia, as in Australia, the co-governing centre-right (Liberal and Conservative) parties adopted AV to allow them to compete separately while preventing the Co-operative Commonwealth Federation (CCF) from “going up the middle” on a vote split. Ironically, however, it was short-lived, since the sponsoring parties were soon eclipsed by a new force, the populist Social Credit Party, which repealed the system after being elected.8
Another factor influencing the long and consistent vote-pooling effect of AV in Australia is the rule that voters cannot simply “plump” for a single candidate, but must mark a preference ranking next to every candidate on the ballot. Hence parties distribute “how to vote” cards to spell out the order in which their supporters should mark their ballots. But where voters are not required to rank all the candidates, they typically do not. This was the case in Canada where a clear majority of voters using AV did not mark more than a first preference on their ballots, with the proportion of voters “plumping” increasing over time. This tendency has also emerged in Australian states that have recently repealed mandatory preference ordering.9
Thus, while Canadian parties might want their voters to take up the strategic possibilities of AV, comparative and historical evidence suggests that, barring exceptional circumstances, most will not. In all the Canadian AV elections, consistent and strategic preference ordering between two opposition parties was exhibited in only one contest in Alberta, none in Manitoba and the two elections that used the system in B.C.10
If AV does not necessarily encourage cooperation among parties, how might it affect competition between them? One consistent claim by both Australian and Canadian proponents of AV is that it increases the influence of smaller parties because voters can now support a smaller party without fear of wasting their votes. Hence smaller parties can exact concessions from large parties by promising to instruct their voters to give their later preferences to them in return for policy commitments or promises of representation.
This increased influence for smaller parties is not direct: if anything, AV reduces the chances of smaller parties gaining legislative representation. While there is considerable research into indirect influence, the quite complex forms of potential preference trading found in Australia make the degree of small-party influence on larger parties difficult to gauge.11 We know of cases of larger parties trading representation for smaller parties in Australia’s proportionally elected upper house for preference support in the lower house (and can imagine Senate appointments serving the same purpose in Canada). But small-party influence on policy is harder to prove. As occurs here under SMP, centrist parties operating under AV may pluck policies from the margins simply because they seem potentially popular.
Another way of judging influence is more direct – what impact have vote transfers had on the actual results of the elections? More specifically, how often have transfers from smaller parties changed the results in any given riding, shifting control of the seat from one major party to another? On examination of the evidence, the results are stark: not much and not often. In Australia more counts are proceeding to multiple stages (i.e. fewer candidates are being elected on the first count with an outright majority of the votes); however, the transfers are not having much effect on the outcome. The percentage of seats shifting from one major party to another as a result of vote transfers has remained consistently small. In their 2006 study of the Australian electoral system, David M. Farrell and Ian McAllister found that the effect was minimal: “The proportions remain firmly in single figures.” The Canadian evidence shows even less impact. Out of the total pool of 17 elections, vote transfers changed the results in just 2.6 per cent of the ridings in Alberta, 1.6 per cent in Manitoba and 12.5 per cent in B.C.12
Such influence could be important in a close race between major parties. In Australia, such transfers blocked Labor from acceding to power three times over the last century. However, Jansen reports no such impact in Alberta and Manitoba. Only in B.C.’s first election with AV in 1952 can such an impact be detected, a factor that no doubt contributed to its ultimate repeal.13 Hence, on the whole, proponents of AV in Canada who expect a decisive impact from its use may be disappointed.
AV’s strategic coordination appears to amount mainly to funnelling support back to the major parties. Yes, it allows voters to take a chance on supporting new or smaller parties, but the evidence suggests that such chance-taking seldom results in added representation or influence on the dominant parties. In his simulation of the possible impact of AV on the 1997 federal election results in Canada, Antoine Bilodeau concluded that the likely scenario was that the Conservative and Liberal parties would have benefited at the expense of Reform and the Bloc Québécois.14
Taking a longer view, there is reason to expect that Canadian experience with AV would differ from Australia’s. The regional strengths of our traditional three major national parties would produce a different dynamic than the left-right dualism of the Australian party system. For instance, in some parts of the country AV might bolster the NDP, and in some the Liberals, as the main challenger to the Conservatives.
Additionally, recent simulations suggest that the pattern of potential transfers among Canadian parties may not reproduce the clearer lines of party alliance found in Australia. A 2013 survey of Canadians by Abacus Data found that Conservative, Liberal and NDP supporters were divided about which other party they would rank second. While 57 per cent of Liberals ranked the NDP as their second choice, 28 per cent chose the Conservatives.15 Such results challenge the typical media characterizations of the Conservatives as Canada’s right-wing party with the Liberals and NDP occupying the left. They also demonstrate that voting under AV could be a lot more chaotic in Canada that its proponents may have bargained for. In sum, the evidence, at best, suggests that electoral cooperation between parties is possible under AV, but that the voters’ ability to use the system strategically is doubtful.
The Alternative Vote and the democratic deficit
Some Liberals have argued that AV would increase citizen engagement in politics and better represent what Canadians say with their votes. Such claims resonate with a broader public discourse on the democratic deficit in Canada. But how well does AV respond to the concerns identified with the democratic deficit? The available evidence is not encouraging. Put bluntly, AV does not foster a more competitive political environment, higher degrees of party proportionality, improved voter turnout or better electoral representation of our social diversity.
As a majority voting system, it may come as little surprise that AV performs poorly at fairly representing party choices. Though some researchers, such as Donald Horowitz, claim that AV is a more proportional voting system than Canada’s present SMP system, evidence from Australia and western Canada does not bear this out. The research shows that SMP and AV are virtually indistinguishable in terms of disproportionality of electoral results.16
Another way to gauge AV’s impact is to measure its effect on the competitive dynamic of the party system. Does AV encourage more parties to run and win representation (even if such representation is not proportional)? There is some evidence that AV may encourage more parties to run for office (though Jansen suggests that historical period effects may have produced this in Canada, noting that in the same period Saskatchewan witnessed an increase in parties without AV). But when it comes to results, use of AV did not lead to an increase in the number of parties gaining legislative representation in either Australia or Canada.17
AV helps produce majority results within a district but it should be underlined that this does not necessarily mean that the governing legislative majority represents an actual majority of voters. Indeed, in the Australian case there have been four elections (1954, 1960, 1969 and 1990) where the winner had fewer votes in total than the opposition after preference transfers but still gained more seats. As Clive Bean notes, this can occur because “although under preferential voting the winning candidate in each seat must secure majority support, as with first-past-the-post voting one party may achieve a number of narrow victories while another gains a smaller number of large victories so that when the votes are totalled the party with fewer votes has more seats.”18 Thus, as with SMP, a lot depends on the geographic dispersal of votes: both privilege more geographically proximate pools of votes over more dispersed ones, regardless of the parties’ voting strengths.
AV, like SMP but unlike PR systems, tends to produce legislative majority governments. Since 1919 Australia has had only two minority federal governments (1940 and 2010) while the western Canadian use of AV led to the election of three minority governments (two in Manitoba and one in B.C) out of a possible 17. Has AV reduced executive dominance, as some suggest? In the Australian case, party leaders and even governing prime ministers have found themselves challenged – and sometimes replaced – while holding office.19 But this seems to be a matter of political culture more than political institutions, since Canadian provincial experience with AV produced few comparable phenomena.
Turning to voter turnout, testing AV’s possible impact is more difficult. Australian experience is not terribly informative as the country uses both compulsory registration and compulsory voting, making the voting system’s impact on turnout difficult to parse. Harold Jansen did compare voter turnout rates from before, during and after the use of AV in the three western Canadian provinces, finding that it coincided with no significant change in levels of electoral participation.20
Finally, AV’s track record in facilitating the representation of social diversity in the legislatures that have used it does not appear to be any better than Canada’s current SMP system. Indeed, Australia and Canada occupy nearly the same spot on the international rankings of women’s representation in the lower houses.21 The electoral system does matter, however. The significantly different levels of women’s representation in the upper and lower Australian houses puts the lie to efforts to reduce such differences to political culture. While Australia’s lower house uses AV, its upper house, the Senate, uses the proportional single transferable vote (STV). Since the 1970s, Australian electorates have returned significantly more women to the Senate than to the House of Representatives. Such results have been reproduced at the subnational level in Australia as well.22 Overall, there is no good reason to expect that shifting to AV will result in more diverse representation.
Hardly worth the effort
Manitoba Liberal MP Kevin Lamoureux was refreshingly frank in the aftermath of the parliamentary vote on Craig Scott’s PR motion when he admitted that political parties tend to support the voting system that benefits them most. His own political career demonstrated this nicely: as one of two Liberal MLAs in the Manitoba legislature he had supported PR, but when he was elected federally he switched to supporting AV. As he put it, “There does seem to be a bit of what is in the parties’ best interest going on here.”23 On that basis, Liberal support for AV is understandable. But if we are to take seriously Justin Trudeau’s promise to take an “evidence-based approach to electoral reform,” we have to conclude that the evidence for AV is not compelling.
The alternative vote is often pitched in the Canadian context as a pragmatic, accomplishable reform that offers both political parties and voters more strategic leverage over political outcomes. Voters are promised more latitude to support new or smaller parties while political operatives are provided with a competitive framework that avoids zero-sum vote splitting. But the concrete results of AV as practised in Australia or historically in western Canada are less impressive. The evidence suggests that voters can support new or smaller parties without “wasting” their vote, but the effort does not produce much in terms of results. On the other hand, AV is a boon to the major parties, reducing the danger and uncertainty they face from new party competition.
October’s election could very well result in the NDP and Liberals placing a concrete electoral reform proposal before the Canadian people. One can only hope that they will do better than AV. For voters, the payoff with AV hardly seems worth the effort of attempting to reform the voting system. AV will do little to address the broader issues associated with Canada’s democratic deficit – better representation of political opinion, increased public engagement with politics, increasing our social diversity in the legislature – or to substantively increase the power of voters to affect the outcomes of elections. If anything AV might produce less representative outcomes compared to SMP, in which a popular but dispersed political view (like support for the Green Party) can occasionally gain a plurality of votes in a riding or two. Indeed, on reviewing this experience with AV, one must face the conclusion that to substantially increase voter power and address the broader democratic reform issues, a more root-and-branch approach to reform, like some form of proportional representation, would need be considered.