This article was downloaded by: [The University of Manchester Library]On: 10 October 2014, At: 05:17Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

RepresentationPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/rrep20

Australia's experience with thealternative voteClive Bean aa Senior Lecturer in the School of Social Science , QueenslandUniversity of TechnologyPublished online: 06 Jul 2007.

To cite this article: Clive Bean (1997) Australia's experience with the alternative vote,Representation, 34:2, 103-110, DOI: 10.1080/00344899708522996

To link to this article: http://dx.doi.org/10.1080/00344899708522996

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information(the “Content”) contained in the publications on our platform. However, Taylor& Francis, our agents, and our licensors make no representations or warrantieswhatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions andviews of the authors, and are not the views of or endorsed by Taylor & Francis. Theaccuracy of the Content should not be relied upon and should be independentlyverified with primary sources of information. Taylor and Francis shall not be liablefor any losses, actions, claims, proceedings, demands, costs, expenses, damages,and other liabilities whatsoever or howsoever caused arising directly or indirectly inconnection with, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden.Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Australia's long experience with the alternative vote suggests thatit is an improvement on first-past-the-post voting but it too canproduce perverse results.

Australia's Experience with theAlternative VoteClive Bean

Clive Bean is SeniorLecturer in the School ofSocial Science,Queensland University ofTechnology and haswritten extensively onelections and votingbehaviour in Australia.

103

Australia is unique in its use of the alternative vote in democraticelections over a period of many decades. Preferential voting, as thealternative vote is commonly called in Australia, was introduced at thefederal level in the election of 1919, having been brought into use forelections in some of the individual states before then. By the early 1960s,preferential voting in single-member constituencies was in effect in allAustralian states save Tasmania, where the single transferable vote (STV)is used. In national elections, preferential voting is used in the House ofRepresentatives, while the Senate is elected via STV.

Preferential voting in AustraliaThe system of preferential voting requires voters to mark their ballotpapers with consecutive numbers starting with '1' to indicate the orderin which they favour the various candidates standing for election in theirelectoral division. In House of Representatives elections and most stateelections voters must indicate a preference for every candidate on theballot paper in order to cast a valid vote (although this provision has nowbeen relaxed in New South Wales state elections). If any candidateachieves a majority when the first preference votes are counted thenpreferential voting works just like first-past-the-post (simple majority orplurality) voting, which it replaced in the House of Representatives. If noone secures a majority on the first count, the least popular candidate isexcluded and the relevant votes are redistributed according to the secondpreferences of the candidate's supporters. If there is still no majority, theremaining candidate with the fewest votes is excluded and his or hervotes reallocated according to the voters' next preferences, and so on,until one candidate eventually does gain a majority of the votes.

The alternative vote is thus a majoritarian system because nocandidate can be elected without securing a majority of the votes in theelectoral division, even if that majority may come partly in the form ofsecond, third or fourth preferences. Unlike first-past-the-post voting, acandidate who ultimately has only a minority of support within the

Representation/Volume 54 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

constituency and whom a majority of voters objects to, cannot winelection. This characteristic of preferential voting is related to the reasonfor its introduction in Australia. Historically, the non-Labor parties inAustralian politics have tended to be less organisationally cohesive thanthe Australian Labor Party and preferential voting was introduced bynon-Labor governments seemingly intent on keeping the Labor Partyfrom winning seats in which the non-Labor vote was split among two ormore candidates.

It is almost certainly true that the largely urban-based Liberal Partyand its coalition partner the rural-based National Party would not havebeen able to survive as separate organisations if the first-past-the-postsystem had continued. Under the alternative vote, however, when theLiberal and National parties are unable to reach agreement on which ofthem should field a candidate in a particular seat, in most cases they areable to put competing candidates up against each other safe in theknowledge that as long as their combined first preference vote is greaterthan Labor's, one of them will ultimately win the seat through thesecond preferences of the other candidate's supporters. In this way, formost of its period of operation, preferential voting has tended to work tothe advantage of the coalition parties rather than Labor.

The alternative vote in practiceIn many respects, however, preferential voting behaves much likeplurality voting, principally because both systems are based on single-member electorates.1 Plurality is certainly the system with which thealternative vote is most comparable and best evaluated against. The twoshare many of the same strengths and weaknesses. For example, bothsystems tend to produce decisive results, even when the popular votemay be very closely divided. On the other hand, neither ensures that theparty winning the most votes nationwide, or even a party that wins anoutright majority of the national vote, will win the relevant election.Although under preferential voting the winning candidate in each seatmust secure majority support, as with first-past-the-post voting oneparty may achieve a number of narrow victories while another gains asmaller number of large victories so that when the votes are totalled theparty with fewer votes has more seats.

In the Australian political context, with its strong two-party system(counting the Liberal and National coalition partners as one party forpractical purposes), the calculation of votes versus seats can be done intwo different ways, either by adding up the first preference votes for eachparty or by using the 'two-party preferred vote'. The two-party preferredvote is simply a calculation of which of the two major parties each voterprefers over the other and is obtained by fully distributing the voterpreferences.2 The principal justification for using this measure is that inAustralian elections the final result is always a question of which of the

1. The magnitude ofelectoral districts ratherthan the voting formulais the most importantfactor in distinguishingthe performance ofdifferent electoralsystems. See Douglas W.Rae, The PoliticalConsequences of Electoral

Laws, revised edn (NewHaven and London, YaleUniversity Press, 1971),p.108.

2. The AustralianElectoral Commissionhas done this routinelysince 1984. Prior to then,the two-party preferredvote was estimated byelectoral analysts such asMalcolm Mackerras inseats where preferenceswere not fully distributedduring the course ofproducing the result.

104 Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

two major parties will win and this is also true in the vast majority ofindividual electoral divisions.

There are numerous examples in Australian electoral history of whenthe party winning a larger share of the primary (first preference) votethroughout the country has not won the election and even one case(1954) when the Labor Party received an absolute majority of the primaryvote and still failed to win the election. In the 1950s and 1960s the LaborParty always found itself on the losing end of these situations, but duringthe late 1980s and early 1990s the tables were turned and in the electionsof 1987 and 1990 Labor won with a smaller share of the first preferencevote than the Liberal and National parties' combined total. The logic ofthe alternative vote system should mean that the two-party preferred votewould more accurately reflect the final result than the simple firstpreference vote, but even using this figure there are four examples sincethe second world war of the minority party winning the election: 1954,1961 and 1969, when the coalition partners were the beneficiaries of thedistortion, and 1990, when Labor won the election with less than 50% ofthe two-party preferred vote (see Table i).3

Table 1: The two-party preferred vote in the Australian federalelections of 1954,1961,1969 and 1990

1954196119691990

Laborvotes %50.750.550.249.9

Liberal-Nationalvotes %49.349.549.850.1

Source: See footnote 3.

Minor parties and the flow of preferencesNotwithstanding the ongoing benefits that the Liberal and Nationalparties derive from preferential voting, the more general reversal offortunes under this system from the 1950s to the 1980s has much to dowith the flow of preferences from minor parties. By comparison withplurality voting, preferential voting is said to benefit minor parties,although Australian experience suggests that it does not do so either byhelping them to attract more votes than they might in a first-past-the-post environment or to win seats in their own right.4 Rather, it makesminor parties influential in determining which of the major parties winsclosely fought seats, through the medium of'how to vote' cards. How tovote cards suggest how the voters should order their preferences and areroutinely distributed by Australian political parties outside pollingbooths to their potential supporters. In a tight race, how the minorparties choose to rank the major parties on their how to vote cards canthus be crucial.

3. Colin A. Hughes, 'TheRules of the Game' inClive Bean, IanMcAllister and JohnWarhurst (eds.), TheGreening of AustralianPolitics: The 1990 FederalElection (Melbourne,Longman Cheshire,1990), p.139.

4. Clive Bean, 'ElectoralLaw, Electoral Behaviourand Electoral Outcomes:Australia and NewZealand Compared',Journal of Commonwealthand Comparative Politics, 24(1986), pp. 57-73, esp.pp.65 & 68.

105 Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

In the 1950s and 1960s the principal minor party was the DemocraticLabor Party and the latter preferences of the voters who chose it as theirfirst option always heavily favoured the Liberal and National parties. Bythe 1980s, the Australian Democrats had taken over the mantle of mostpopular minor party and the preferences of its supporters - and to aneven greater extent those of various environmental candidates - havetended to favour Labor. For example, in the four federal elections since1987, approximately 60% of Democrat preferences have gone to Laboron average (and only 40% to the coalition), the imbalance in Labor'sfavour peaking in 1990 when the party received over 63% of Democratpreferences.5

Indeed, the election campaign of 1990 was the occasion of one of themost blatant attempts to manipulate the preferential voting system,when the Labor Party adopted what came to be known as the 'secondpreference strategy' in which Labor advertising acknowledged that manyvoters would probably give their first preference vote to the AustralianDemocrats or another minor party and pleaded for the secondpreferences of these voters.6 The tactic was widely credited with winningthe election for Labor.

Preferential versus plurality votingUnder a first-past-the-post system, of course, it would not be possiblefor either parties or voters to adopt such strategies. And, despite the factthat preferential voting works similarly to first-past-the-post in broadterms, there are various other indicators of differences between the twosystems. For one thing, the evidence suggests that preferential votingleads to higher levels of informal (invalid) voting thanfirst-past-the-post, because it is more complicated for the voter.7

Another phenomenon, also linked to the fact that voting is compulsoryin Australia, is what is known as the 'donkey vote', whereby a smallproportion of electors simply write successive numbers from the top tothe bottom of the ballot paper without any concern for which candidateis which.

Whenever preference distribution is required to achieve a result, thepotential for the alternative vote to produce a different outcome fromplurality voting is engaged. In House of Representatives elections, theneed to distribute preferences has tended to occur in only about 30% ofseats or less on average. However, the proportion varies considerablyfrom one election to the next, depending on whether the election givesrise to a tight contest or a landslide victory. It is also a phenomenon thatappears to be on the rise. Many more seats have required preferences tobe distributed in the elections of the 1990s than in the previous twodecades (see Table 2), which is consistent with various other indicatorsof a slight reduction in the dominance of the major parties in recenttimes.8

5. Malcolm Mackerras,'Election Results' inBean, McAllister andWarhurst, The Greening ofAustralian Politics, p.199.

6. See Bean, McAllisterand Warhurst, TheGreening of AustralianPolitics.7. Bean, 'Electoral Law,Electoral Behaviour andElectoral Outcomes',p.63.8. Colin A. Hughes,'Individual ElectoralDistricts' in Clive Bean,Scott Bennett, MarianSimms and JohnWarhurst (eds.), ThePolitics of Retribution: The1996 Australian FederalElection (Sydney, Allen &Unwin, forthcoming1997). pp. 166-7.

106 Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

Table 2: Seats in which preference distribution was requiredand results changed, 1972-96

seats requiring preference seats in which resultdistribution changed after preferencenumber % distribution number %

19721974197519771980198319841987199019931996

Average 1970sAverage 1980sAverage 1990s

4933244640314454896258

384270

39.226.018.937.132.024.829.736.560.142.239.2

30.330.847.2

141074621349127

969

11.27.95.53.24.81.68.82.76.18.24.7

6.94.56.3

Sources: Colin A. Hughes, A Handbook of Australian Government and Politics 1965-1974 (Canberra,Australian National University Press, 1977); Colin A. Hughes, A Handbook 0/Australian Government andPolitics 1975-1984 (Sydney, Australian National University Press, 1986) and see footnote 8.

Even when preferences do need to be distributed to produce a result,in many cases the outcome is still as it would have been under first-past-the-post because the leading candidate on the first count gains amajority at the final count. It is only when the distribution of preferences'changes the result' that the alternative vote is truly distinctive in itsoperation from plurality voting. On average this happens in only 20% to30% of seats where preference distribution is required, which works outto only about 5% to 10% of all seats, although again the figure variesconsiderably from election to election (see Table 2). It is evidence of thiskind - that in up to 95% of seats the result would be the same - thatleads analysts to conclude that preferential and plurality voting are muchthe same in practice.

Yet, while the differences between the two systems may occur at themargins, they can still prove decisive in close elections. One exercise, inwhich results were recalculated under plurality rules, led to theconclusion that, all other things being equal, the outcome of both the1961 and 1969 Australian elections would have been reversed hadfirst-past-the-post been in operation instead of preferential voting.9 The1961 election has a particular place in Australian political folklore, for itwas won and lost on a small number of preference votes in theQueensland electoral division of Moreton. Among these preferences

9. Bean, 'Electoral Law,Electoral Behaviour andElectoral Outcomes',pp.65-6.

107 Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

were votes that had originally been cast in support of the CommunistParty and these not only helped the Liberal Party to win the seat, but indoing so secured the return of a conservative government.

Preferential voting at the constituency levelThe most recent federal election, of 1996, also provides some interestingexamples of the workings of preferential voting at the constituency level.A record number of five independent candidates gained election to theHouse of Representatives and in at least one instance the preferentialsystem was instrumental to the candidate's success. In the WesternAustralian seat of Curtin, the disendorsed Liberal incumbent AllanRocher, standing as an independent, trailed the endorsed Liberalcandidate Ken Court by 29% to 39% on the count of primary votes.Rocher was still well behind Court after the votes for three minor partycandidates had been redistributed and it was only when the Labor Partycandidate was excluded for the final count that Rocher overtook Court,due to the fact that over 90% of the Labor supporters' preferences weredirected to him.

In the New South Wales electorate of Calare, another (moregenuinely) independent candidate, Peter Andren, was ahead of the Laborcandidate by only 300 votes after first preferences were counted. Boththe National and Liberal parties fielded candidates in this mixed rural-urban seat and the initial vote count was 29.4% for the independent,29.0% to Labor, 20.8% to the National candidate and 15.7% for theLiberal. Little changed with the exclusion of four minor candidates andwhen the Liberal votes were redistributed, enough of these preferenceswent to Andren to keep the independent ahead of the Labor candidate,who was also overtaken by the National, the latter having received thelarge majority of the Liberal voters' preferences (see Table 3). On thefinal count, the redistributed Labor votes split heavily in Andren's favour,giving him a substantial victory over the National candidate in the end.I0

Table 3: Preference distribution in the seat of Calare, 1996Independent(Andren)votes %

Labor(Allen)votes °y

National(Toole)votes %

Liberal(Fardell)votes %

First count...Fifth countSixth countFinal count

29.4

32.035.163.3

29.0

29.930.2Excluded

20.8

21.834.736.7

15.7

16.3ExcludedExcluded

Sources: Australian Electoral Commission and see footnote io.

10. Hughes, 'IndividualElectoral Districts',p.169.

io8 Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

Compulsory marking of preferences and exhausted ballotsOne objection that is often raised to the alternative vote in the Australiancontext concerns the compulsion to mark a preference for everycandidate in the electoral division. Combined with compulsory votingitself (although strictly speaking the compulsion is only to attend apolling booth), this requirement is seen as anathema to those whobelieve that it is anti-democratic to compel individuals to cast a vote(even as the last preference) for one or more political candidates whosepolicies they could never support.

Of course, any change to this requirement would change the natureof the system, however subtly. Meanwhile, there is a provision in theCommonwealth Electoral Act which allows voters to frustrate the spiritof the compulsory marking of preferences and still cast a valid vote. Avote will be valid if it is marked with a ' i ' beside the preferred candidatefollowed by a series of'2's along side all the other candidates (or a ' i ' ,plus a '2', plus a series of'3's). If the voter has chosen a candidate orcandidates who are then excluded from the count, the vote is deemed tobe 'exhausted', since further preferences cannot be determined. Amongother things, if sufficient voters opt for this procedure then it creates thepossibility of an exception to the majority status of a winning candidate,who may end up only with a majority of the non-exhausted votes.

In practice, very few voters take the option of casting their ballots inthis manner, partly because although it is legitimate to do it, at the sametime it is illegal to advocate its use publicly. During the 1996 electioncampaign, Albert Langer, a long-time political activist ended up in jailfor attempting to persuade voters to send a message of discontent to themajor parties by giving their first preference to a minor candidate andthen exhausting their ballots. Whether it was the effect of Langer'scampaign, or the publicity surrounding the court case brought by theAustralian Electoral Commission in an attempt to silence Langer, ormerely a coincidence, history will record that the number of exhaustedballots in 1996 was nearly seven times the number it had been at theprevious federal election in 1993. Even so, the total number in 1996 stillrepresented less than half of 1% of the valid vote.

How to vote cards and voter preferencesIn discussion of Australian elections it is common to hear talk of the'parties' preferences', a term which overlooks the fact that thepreferences actually belong to the voters, not to the parties orcandidates. This terminology is partly just convenient shorthand, but italso stems from the use of how to vote cards (as referred to above). Theunderlying assumption is that the how to vote cards have a major impacton voter behaviour in the polling booth and that there is very little'slippage' from the ordering of preferences suggested by the voter'spreferred party. Yet hard evidence on the extent to which voters do follow

log Representation/Volume 34 Number 2

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4

the party how to vote cards rather than deciding their own preferences isscarce.

The 1996 Australian Election Study11 asked respondents directly aboutthis aspect of their behaviour. Some 55% of all voters in the samplereplied that they followed a party how to vote card, while 45% said thatthey decided their own preferences. The tendency to do so, however,differed markedly between major and minor party voters. Around six inten major party supporters (58% for Labor and 63% for theLiberal-Nationals) followed the party's how to vote card, but only aquarter (27%) of Australian Democrat voters reported doing so, as dideven smaller proportions of Green voters (12%) and supporters of otherminor parties and independent candidates (20%). In other words, thelarge majority of minor party voters claim to make their own decisionsabout which parties to give their preferences to rather than simply fallinginto line with the suggestions of their first choice party. On thisevidence, even the proportions of major party supporters reporting thatthey decide their own preferences are larger than many would haveassumed.

To round off this brief overview of Australia's experience with thealternative vote, it needs to be said that despite having weaknesses andidiosyncrasies - as all electoral systems have - preferential voting hasserved Australia well enough in practice over many decades. It may notimprove on simple majority or plurality voting to any great degree, butthe two do occasionally produce divergent results at the margins, mostnotably in the sense that a candidate with minority support overallcannot get elected under preferential voting. Ultimately, the workings ofthe Australian electoral system are probably best understood through ajoint appreciation of the country's unique combination of preferentialplus compulsory voting.

11. Roger Jones, IanMcAllister and DavidGow, Australian ElectionStudy, 1996: User's Guidefor the Machine-ReadableData File (Canberra,Social Science DataArchives, ResearchSchool of SocialSciences, AustralianNational University,1996).

Representation/Volume 34 Number 2110

Dow

nloa

ded

by [

The

Uni

vers

ity o

f M

anch

este

r L

ibra

ry]

at 0

5:17

10

Oct

ober

201

4