Text of my speech to the Australasian Study of Parliament Group, Qld Chapter – 15 April 2013

This is the text of my speech to the Australasian Study of Parliament Group, Queensland Chapter, in the  Queensland Parliament House on 15 April 2013. I’m publishing the text for anyone who may be interested in reading it.


Tweedledee, Tweedledum – Not all electoral systems are created equal

Sacha Blumen

Speech to the Australasian Study of Parliament Group, Queensland Chapter, 15 April 2013

I’d like to acknowledge the traditional owners of the land, the Jagera and Turrbal people, and pay my respects to their elders past, present, and future.

Electoral reform is in the air in Queensland. The Department of Justice and Attorney General published a discussion paper on the topic in January this year which has engendered much discussion.

Electoral reform has been a topic of interest for many Queenslanders and Australians over the last few decades. Across the country, there have been moves towards one-vote one-value voting systems (most recently in WA), regular redistributions, party registration and placement on ballot papers, and increasing the number of ways in which people can cast their vote (e.g. introduction of electronic voting in the last NSW election).

There have also been reforms to make it easier for people to enrol and for automatic enrolment in NSW elections, in addition to electoral funding and disclosure reforms. In my state of NSW, the Government recently gave redistribution commissioners more flexibility in drawing boundaries, which they may use in the current State redistribution.

In short, experience tells us that electoral reform is never a done and dusted business. And nor should it be – we should always learn from experience, and electoral systems should change in line with societal attitudes. In 1905, laws were changed to enfranchise women in Queensland elections and in 1974 the voting age was dropped from 21 to 18. Some people are now proposing the voting age be dropped to 16.

If you were to believe everything in the newspapers, you would think that the prime purpose of electoral reform is to benefit the party proposing it. The Australian’s recent coverage of Bronwyn Bishop’s floating of optional preferential voting for Federal elections is a case in point. This coverage focussed on which party it was thought would benefit – the Coalition. There was little discussion beyond partisan impacts.

However, I feel confident saying that most electoral reforms are consistent with the interests of the Government introducing them. One notable example that wasn’t was South Australian Liberal Country League Premier Steele Hall introducing much more equal electorates for the South Australian House of Assembly for the 1970 election. That reform substantially rebalanced a very weighted electoral system to the clear detriment of his party – which didn’t win office again until 1979. This example is so notable because it is so rare.

In the zero sum game of politics put forward in the media, a change to the electoral system will likely benefit Party A to the detriment of parties B and C. While this approach might be interesting to the media-political class, it focusses narrowly on partisan interests and ignores the larger question about whether the changes would have wider systemic impacts.

In this talk I will focus on the likely electoral impacts of changing to different electoral systems in Queensland elections, and also their potential wider systemic impacts on the political culture.

These wider impacts may include:

  • individual accountability of MPs to voters and the power of factions within parties;
  • the legitimacy of government and parliament;
  • partisanship;
  • competition and cooperation between parties;
  • campaigning, and the resources needed to run them; and
  • the stability of public policy.

I’ll be looking at a couple of systems. To get our minds into gear, let us first look at a straightforward case — optional preferential voting.

Optional preferential voting

Optional preferential voting is very familiar in Queensland and NSW, and can be thought of as a system somewhere in between compulsory preferential voting and first past the post.

Optional preferential voting was introduced in Queensland for the 1992 state election following an EARC review — EARC standing for the Electoral and Administrative Review Commission. Before this, it had been introduced in NSW in 1980 by Neville Wran. In both states it probably helped the ALP due to splits on the Conservative side of politics.

Now, it is probably benefitting the Coalition – both because of the rise of the Greens, and because the Coalition has a very high primary vote. The Australian recently published stories of Labor MPs calling on OPV to be abandoned in NSW and Qld. The federal Coalition has indicated that it may introduce OPV in federal elections.

Antony Green thinks that optional preferential voting would have benefitted the Coalition in the last Federal election. This is because, compared to compulsory preferential voting, it tends to advantage the candidate with the highest primary vote. The Coalition received many more primary votes than Labor at the last federal election.

In addition to helping candidates with the highest primary vote, OPV would tend to disadvantage parties who have to compete with other parties from the same ideological pool of voters. This is because some voters will exhaust their preferences.

In Queensland, the ALP is competing with the Greens and KAP for the same pool of voters. The LNP is also competing with KAP for a pool of voters. Depending on what happens in the next two years, the LNP might be competing with new parties on the right as well.

Personally, I support OPV as people should only be obliged to preference candidates they want to.

It does, however, make things more complicated for parties and candidates compared to compulsory preferential voting. It might also make election results more volatile, negatively impacting parliamentary government.

OPV makes things more complicated for political parties as: 1. it makes preference negotiations more difficult because parties don’t need to preference each other, and 2. voters need not exercise preferences.

I have personal experience of trying to negotiate a preference deal with the Greens under optional preferential voting. Then, they chose not to do a deal. They probably would have under compulsory preferential voting.

Clearly, one impact of OPV is that it encourages like-minded parties to work together or to merge.

OPV may also result in more volatile election outcomes compared to compulsory preferential voting. Election landslides might be even larger under OPV. This is for the reasons outlined by Antony Green – OPV tends to advantage the leading candidate.

Antony Green thinks the LNP’s landslide in the 2012 Qld election would have been smaller under compulsory preferential voting. I’lI talk about that later.

Now there is nothing inherently wrong with volatile election outcomes – particularly if they reflect the will of voters. However, they can have some negative consequences for parliamentary democracy. They can make it difficult for Oppositions to undertake policy development, hold the Government to account, and present a credible alternative at the next election.

In NSW, the Opposition has 20 MPs in a lower house of 93. In Queensland the Opposition has 7 from 89. It is challenging to see how the Qld Opposition will effectively present an alternative Cabinet to voters at the next election.

On the other hand, I know that the NSW Opposition is using extra-Parliamentary processes to improve its development of policy. These include a Policy Forum comprising directly-elected party members, MPs, and union members, that is examining different policy areas through discussion papers. So perhaps parties can adapt to election wipeouts.

It is interesting to note that parties may or may not tend to bounce back quickly after a big loss. In Queensland they havn’t – in 2001, the Qld Nationals and Liberals won a combined 15 seats at the state election, which gradually rose to 34 seats at the 2009 election.

By contrast, the South Australian Labor Party came close to winning the ‘96 election after the ‘92 wipe out. Perhaps it is difficult to generalise as each of these cases reflects the particular environment in those states at that time.

The systemic impacts of OPV, compared to compulsory preferential voting, may include:

  • more volatile election results – i.e. larger majorities for Government and weakened Oppositions; and
  • greater pressure on like-minded parties to merge or agree not to run against each other.

Let us turn to first past the post voting

First past the post voting

First past the post voting is somewhat familiar to Australian voters through its use in Westminster elections and in the US, and previously in NZ. Queensland used it between 1860 and 1892 and also between 1942 and 1962.

It is usually thought that first past the post supports a two party system.

But the experience in UK, the US and NZ is mixed.

The UK has a vibrant multi-party system with regional variations spanning left-right economic issues and unionist-nationalist axes. The Conservatives, Liberal Democrats, and Labour Party are competitive in large sections of the UK, while Scotland, Wales, and Northern Ireland all have competitive regional parties. In addition, there are smaller parties – such as the UK Independent Party and the Greens – that tend to do better in by-elections, European elections, and local elections.

While there are many parties in the UK, people often vote strategically, turning large parts of the UK or individual seats into two-party contests — such as the Conservative vs Lib Dem battle in the South East and South West of England. Many people will also recall the vote swapping website for Labour and Lib Dems supporters in the 1997 and 2001 elections — a way for those voters to strategically vote against the conservatives.

This competitive and regional multi-party system is a clear evolution from the stable two party system that existed between 1945 and the ‘70s.

Let us turn to the US. Almost all legislative and executive elections in the US use first past the post voting or the runoff election if no candidates receives 50 per cent of the vote.

At the national level, the US has only two broad political parties – the Republicans and the Democrats. No other national parties have existed for more than a handful of election cycles. At the state level, there has sometimes barely been a two party system – with one party being dominant for decades, as the Democrats were in the south between the Civil War and the 1964 Presidential election.

While Ross Perot received substantial voter support in the 1992 and 1996 Presidential elections – 19% in ’92 and 8% in ’96, his Reform Party – created in 1995 – only had substantial success with the election of Jesse Ventura as Governor of Minnesota in 1998.

The Libertarian party, formed in 1971, has gained no more than 1.1% of the Presidential vote since 1976. It currently has 138 elected or appointed officials throughout the US – all or almost all at the local level.

The US Greens Party, founded in 1991, has gained no more than 2.7% in Presidential elections. As of October 2012 they were 134 elected Greens.

And lastly, when first past the post was used in New Zealand – up to the 1993 election – there were usually two major parties, although there had occasionally been a smaller party such as Social Credit or NewLabour.

What would be the impacts of first past the post voting in Queensland?

First post the post decisively benefits parties with the most number of votes — generally.

A famous example is the British 1983 election, when the left vote split between the Labour Party and the SDP-Liberal Alliance:

  • the Conservatives’ won 397 seats out of 650 even though their vote declined slightly to 42%;
  • Labour won 209 seats from 28%; and
  • the SDP-Liberal Alliance won 23 seats from 25% of the vote.

Closer to home, the NZ left was split between the Labour Party and the Alliance in the 1993 New Zealand election.

  • The National Party was re-elected with a bare majority of 50 seats on 35% of the vote;
  • the Labour Party won 45 seats on 34.7%;
  • the Alliance won 2 seats on 18.2%; and
  • the NZ First Party won 2 seats on 8.4%.

The 1993 Canadian election is another example – the Progressive Conservatives dived from 154 seats to 2 seats – with their vote declining from 43% to 16%.

In Australia, some Coalition supporters grumbled after the last federal election that they should have won it as the Coalition received the most primary votes.

Antony Green thinks that if first past the post had been used in that election and people voted tactically (as they do in the UK), Labor would have won 70 seats and the Coalition 77 seats. Tony Abbott would have been elected Prime Minister in a majority Government.

First past the post voting may have changed the 1995 Qld election – the ‘Koala Tollway’ election. Labor was re-elected with 45 seats on 43% of the vote, while the Nationals and Liberals won a combined 43 seats on 49%.

First past the post may have led to a majority Coalition Government. This is because the ALP came second in Mundingburra and Whitsunday on primaries, and the winning margin in both seats was very small after preferences — 12 votes and 52 votes. It seems possible that the Nationals and Liberals would have won both seats under first past the post. Rob Borbidge would have become Premier in a majority Government.

Going back in Qld history, it’s interesting to speculate what would have happened under FTPT when the Nationals and Liberals were not in Coalition.

It’s challenging to estimate how people would have voted. Because people would have voted tactically. But applying first past the post to the observed results in 1986 and 1989 we have the following.

In 1986, with FPTP and no tactical voting, the ALP would have benefitted substantially from the split conservative vote in Brisbane:

  • Nat: 42 seats (instead of the 49 it actually won)
  • Lib: 3 seats (instead of 10)
  • ALP 44 seats (instead of 30)

Sir Joh Bjelke-Petersen would have had to lead a Coalition government.

In 1989, there was no Coalition. Assuming FPTP and no tactical voting, the ALP would have benefitted.

  • ALP: 68 seats (instead of the 54 actually won)
  • Nat: 19 seats (instead of 26)
  • Lib: 2 seats (instead of 9)

This is an extreme of the estimate of what would happen under first past the post. More likely is that a couple of extra seats would have fallen to Labor in each election — perhaps requiring the formation of a Coalition in 1986. It is interesting to consider how Queensland history may have been different if that had occurred.

What are the systemic impacts of first past the post?

It advantages larger parties, and disadvantages those parties who have to compete with another for the same ideological pool of voters.

Parties winning a plurality of votes often win a majority in Parliament, although regional factors can complicate matters — e.g. as in the UK last election when Labour did well in Scotland. Governments tend to have legitimacy even if they don’t win 50% of the vote.

People tend to vote strategically — many make a judgement about how best to optimise their vote and act accordingly.

Election results can be volatile — with the decimation of a ruling party (as in Canada in 1993) or when there is a split (as in Queensland in 1957). Parties can rise and fall quickly. This may lead to less stable public policy.

In Qld, we can speculate that first past the post voting would have dramatically reduced three-cornered contests, and possibly a much earlier merger of the National and Liberal parties. Alternatively, increased competition between the Liberals and Nationals may have led to introduction of compulsory preferential voting.

Compulsory preferential voting

Compulsory preferential voting needs no introduction.

It means that every valid vote is likely to either end up with a major party candidate or a popular independent after preferences.

What would be the impact of compulsory preferential voting on Queensland elections?

It is likely to reduce the volatility of election results.

Antony Green has examined the likely impact of CPV on the 2012 Qld state election. He thinks that CPV would have benefitted the ALP, Katter Australian Party and Independents. The LNP would have won 74 seats instead of 78, the ALP would have 9 instead of 7, KAP winning 3 instead of 2, and there would be 3 Independents instead of 2.

Broadly speaking, it’s challenging to forecast how compulsory preferential voting would impact future Queensland elections – as you would have to balance out the likely impacts of the Greens and Katter Australia party, and any new party on the right that emerges during this Parliament.

In NSW, compulsory preferential voting would benefit the ALP as many left-leaning voters vote Green and do not give preferences. In the 80s and 90s, it would have helped the National and Liberal parties, as they sometimes engaged in three-corner contests.

What are the wider impacts of compulsory preferential voting?

Compulsory preferential voting means that major parties don’t have to work as hard to earn votes.

It tends to benefit parties who are competing for voters from the same ideological pool as they can exchange preferences.

It reduces the range of preference deals parties can do compared to optional preferential voting. This means, for example, that the ALP might treat the Greens worse at the federal level than at the state level, as the Greens will always preference Labor at the federal level but might not at the state election.

Compulsory preferential voting allows major parties to run a second preference strategy, as the ALP did in the 1990 federal election.

It allows new parties to emerge on the right or the left without a landslide to the other side.

We saw what can happen in a non-preferential system with numerous parties on one side of the ideological fence in the first round of the 2002 French Presidential election.

The combined left vote – among Socialists, Greens, Communists and leftist parties – was 36%. However, its split among those parties allowed the National Front’s Jean Marie le Pen to go through to the second round of the Presidential election on 16.9% of the vote, with Jacques Chirac on 19.9%. Le Pen would have had no chance under compulsory preferential voting.

Mixed-member Proportional

Mixed-member Proportional will be familiar to many people due to its use in New Zealand since the 1996 NZ election and also in Germany. Wikipedia tells me it is used in other countries including Bolivia, Japan, Mexico, Thailand, and Venezuala.

For those not familiar with MMP, it comprises a proportional representation system overlaid on single member electorates.

  • Everyone votes for their MP in a single member electorate.
  • The number of MPs in Parliament for each party depends on the separate party vote. The electorate MPs are ‘topped up’ with list MPs from the separate party vote.
  • Parties are only represented in Parliament if they win more than a threshold of votes or electorates. In NZ the threshold is 5% or one electorate.
  • Party votes are ignored in calculating the overall makeup of Parliament if they win less than the threshold – those votes are effectively thrown away.

What has been the impact of MMP on NZ elections?

Since 1996, NZ elections have led to coalition governments led by the National or Labour parties. No single party has won a majority of seats in any election – the closest was Nationals winning 59 of 121 seats in 2011. Six parties were elected in 1996, and eight in 2011. The party with the highest primary vote has always ended up leading a government – although that needn’t happen.

One feature of MMP elections in NZ has been strategic voting. Many parties have used this. The aim has been to win one electorate MP to ensure that their party list MPs also appear in Parliament even if the party wins less than the 5% threshold.

Major party voters have also voted for minor party electorate MPs – such as the free-market ACT (Association of Consumers and Taxpayers) – to obtain additional MPs that would likely support their side of politics.

The clearest example of this has been in the seat of Epsom in Auckland – a blue ribbon conservative seat.

In Epsom, conservative voters have returned an ACT MP in 2005, 2008 and 2011 while voting for the Nationals in the party vote.

In 2011, the Nationals won 65% and ACT won 3% in the party vote. Labour and Greens won a combined 28%.

In the electorate vote, ACT won 44%, the Nationals 38%, and Labour and Greens a combined 16%.

John Key, the National Party Prime Minister encouraged a vote for the ACT candidate. Some Labour and Greens supporters undoubtedly voted for the Nationals candidate. But this didn’t ultimately help ACT in 2011, as they won no additional list MPs. It did however help ACT in previous elections – winning an additional four seats in one election, and an extra seat at another election.

NZ probably provides the closest parallel to what would happen in Queensland if a proportional electoral system was introduced. But what happened in NZ depends dramatically on its history.

  • NZ had traditionally used first past the post voting before introducing MMP. National Party governments had been elected in 1974 and 1981 on less than 40% of the vote despite winning fewer votes than Labour. The Social Credit party had obtained 21% of the vote in 1981 and 2 seats.
  • The Lange Labour government, elected in 1984, was a reformist government that dramatically liberalised the New Zealand economy.
  • This economic liberalisation was unpopular with the Party’s more left-wing supporters and caused political stress, with the Labour party fracturing in the late 80s. The Roger Douglas faction wanted ongoing economic liberalisation, while Jim Anderton formed the NewLabour party to represent traditional Labour values.
  • In the 1990 election, the National Party won 67 seats from 97 on 48% of the vote. Labour won 29 seats on 35%, the Greens won 7% but no seats and NewLabour won one seat on 5%.
  • The new National Party Prime Minister had promised to hold a referendum on a new electoral system. Referenda were held in 1992 and 1993, leading to the introduction of MMP for the ’96 election.

There are parallels between NZ and Australia and Queensland in having reforming governments that alienated their traditional supporters. In Australia, the Democrats, One National, Greens, Katter Australia Party gained some of this alienated support.

What would be the impact of adopting MMP in Queensland?

While MMP could easily be introduced in Queensland, there has been little clamouring for it except from the Greens.

It could be adopted in Queensland with slight variants – perhaps optional preferential voting for the electorate seats and also for the party vote.

While forecasting the results of a Queensland election under MMP is sheer speculation, it’s fun to try.

We can guess what the MMP election results might have looked like. Applying actual results through the MMP algorithm gives the following:

  • 1995 – likely majority Coalition Government;
  • 1998 – hung parliament, with One Nation holding the balance of power. Minority Coalition Government;
  • 2009 – minority Labor government supported by the Greens;
  • 2012 – majority LNP government with 46 seats, Labor 25 seats, KAP 11, the Greens 7, and 2 independents (total 91 seats).

But it’s difficult to know what would have actually happened – as MMP makes it easier for small parties to form and survive.

This might have happened in the latter days of the Goss Government, there was some disquiet from left and small-l liberal supporters about its policies. It is entirely possible that a splinter labor party could have been established and contested the 1995 election together with the Greens and Democrats. I know, because I was at a meeting at which the potential establishment of such a splinter party was discussed.

In addition, it’s entirely possible that major party MPs or independents would have formed their own parties and won a few seats as occurred in New Zealand. A few major party MPs have split from the party they were originally elected under in the last few Queensland Parliaments. As an interstate observer, I would say could well happen under MMP in the current Qld Parliament.

I would speculate then, that MMP in Queensland would broadly lead to similar results as seen in New Zealand.

What are the systemic impacts of MMP?

  • supports there being a collection of parties
  • parties will differentiate themselves
  • large parties can still exist and win a near-majority of the vote as the Nationals have in NZ
  • politics is more fluid
  • however governments may have less legitimacy if the Governing party isn’t supported by most voters

A key element of MMP is having single member electorates:

  • this helps improve accountability of individual MPs as they are directly responsible to their voters – not to a central party committee
  • In addition, this can help decentralise power geographically by putting candidate selection in the hands, at least in part, of local party members
  • voters can sack an electorate MP by voting for an opponent – while still voting for their party
  • single member electorates also help improve information flows from voters to government – as each resident has an identifiable person and office to approach if they need to.

 Hare Clark

Hare Clark is familiar to Australians through its use in the ACT and Tasmania. It is also used in the Republic of Ireland, Malta, some UK elections, and some local elections in New Zealand, Massachusetts and Minneapolis.

Hare Clark is a form of proportional representation using multi-member electorates in which people vote for individual candidates instead of parties.

Hare Clark elections have two dimensions — firstly a contest between parties to win seats in Parliament, and secondly a contest between candidates within each party. Candidates in a single party compete vigorously against each other to be elected. It is common for some MPs to be replaced by other candidates from the same party.

Hare Clark has been used in Tasmania since 1909. According to Antony Green, it has worked well there due to the highly stable population, the clear regional divisions reflected in electoral boundaries, and long experience of using Hare Clark.

Successful candidates are often already well known within their communities through local government, business or union activity. The compact size of the electorates may support this.

Majority governments in Tasmania have usually been elected since the ’59 election, at which 7 MPs were returned from each electorate. Only in four elections since ‘59 have minority Governments been elected: in ‘69, ’89, ’96 and 2010.

Since 1998, five state MPs have been elected from each federal electorate. The reduction from 7 MPs to 5 was an attempt to reduce the chances of a minority government being elected with the Greens holding the balance of power. This succeeded until the 2010 election.

Some impacts of Hare Clark in Tasmania and the ACT have been:

  • small Government majorities in the Assembly, which potentially makes Governments less stable – although this does not appear to have been obviously borne out in practice;
  • party leaders can have high personal votes – this was used strategically by the Liberals in the last ACT election to elect another MP on their leader’s coat-tails;
  • personal votes for individual MPs, and regular turnover in MPs from the same party;
  • smaller parties can be represented, but micro-parties tend not to be elected; and
  • regional factors can be reflected in Parliament.

One challenge with introducing Hare Clark in a mainland state would be the size of the electorates, potentially reducing familiarity with individual candidates.

In Queensland, you would probably have between 11 and 17 electorates:

  • 17 electorates each returning 5 members (total 85 members);
  • 11 electorates each returning 9 members (total 99 members).

Each electorate would be about the size of about 5-8 current Qld state seats. Having 17 electorates might mean each electorate is small enough for individual candidates to be known.

What would be the impact of Hare Clark on Queensland elections?

Firstly, candidate selection would probably be different. Each party would nominate a handful of candidates who would have to campaign against each other. This might reduce the ability of factions within major parties to choose candidates in winnable seats. Or, factions may simply fund and promote their preferred candidates at the general election.

Secondly, election results are more likely to be closer and there are likely to be a wider range of parties in Parliament.

I modelled the 2012 Queensland election result under Hare Clark with 17 electorates.

With 17 electorates, 5 MPs would be elected from each electorate to give a total of 85 MPs.

  • The LNP would have won a majority with 46 MPs;
  • Labor would have won 27;
  • Katter Australia Party would have won 8;
  • The Greens 3; and
  • there would be one Independent.

With 11 electorates each returning 9 MPs, we obtain a similar result.

  • The LNP wins a majority with 52 seats;
  • Labor won 29;
  • KAP won 10;
  • Greens won 6;
  • and 2 Independents.

In modelling, the quota effectively acts as a threshold preventing micro parties being elected. It also helps to produce small majorities for parties receiving close to 50 per cent of the vote.

What are the wider impacts of Hare Clark?

  • much less volatile election outcomes;
  • small government majorities in Parliament;
  • potentially increased cohesion in the parliamentary party as you don’t tend to have a very large backbench; and
  • greater emphasis on individual candidates and their own mini-campaigns — there may be both central and candidate campaigns.

List proportional representation

For completeness, let us also mention list proportional representation.

This involves candidates being elected from party lists:

  • an open list system allows voters to preference individual candidates;
  • a closed list only allow people to vote for a party.

Something to note about Queensland history:

  • In the last 30 years, the winning part/ies have usually come close to 50% but only once did they get more than 50% — in 1989;
  • Independents and third parties have obtained between 3% in 1989 and a combined 31% in 1998;
  • with similar voting patterns, proportional representation would usually have resulted in third parties or independents holding the balance of power or a government with a small majority.

However, in practice proportional representation may well have encouraged the formation of new parties. Small parties could be elected depending on whether there’s a threshold (e.g. 5 per cent).

What would be some wider impacts of list proportional representation:

  • it refocuses campaigning — there are no marginal geographic seats — but there may be marginal demographics or other sections of the electorate;
  • it would reduce the accountability of individual MPs to voters, as they might only be able to be sacked by their party committee or by voters if the party’s support reduces significantly;
  • it may result in local party members being less likely to have input into the selection of any one candidate, and increase factional influence inside parties;
  • it may create incentives for MPs to split from the party they were elected under, especially if they would then have the balance of power, as they might be elected under their new party next election;
  • it may also reduce information flow from residents to government, as there is no one identifiable office for residents in an area to approach.

Some of these things are familiar to us through Senate elections and NSW upper house elections. In addition, City of Sydney elections have some characteristics of list proportional representation.

It is difficult to argue that Senate or NSW Upper House members can be sacked by voters, except if they are at the end of their party ticket and their party is unpopular. In that case, they’re unlikely to be well known. Instead, NSW MLCs are responsible to those who preselected them — the Labor state conference, or Liberal party members in notional geographic electorates. This is not good for political accountability.

It’s interesting to also look at City of Sydney elections as it may have parallels with list proportional representation.

The City of Sydney comprises the CBD plus surrounding suburbs. It has only 180,000 residents, and 8 per cent of Australia’s GDP is produced within its area. There are 41 local councils in the Sydney metropolitan region.

In the City of Sydney, we directly elect a Lord Mayor using OPV, and also elect 9 Councillors using a Senate style system. There are no wards. The Lord Mayoral candidates are all obliged to also stand as candidates for councillor.

This has led to the following impacts:

  • elections are campaigns between Mayoral candidates – with Councillors effectively providing a supporting role
  • people vote for teams rather than individual councillors, and there is no scope for Councillor candidates to run their own campaigns – to do so they would have to run for Mayor
  • the Mayor is the dominant political personality on Council, and is likely to hold a majority or plurality on Council
  • the accountability of councillors is reduced – Councillors on the Mayoral ticket are effectively responsible to the Mayor – the current Mayor sacked one of her Councillors before the last election;
  • the lack of wards may make information flow from residents to Council more challenging – no specific councillor is ‘responsible’ for knowing about the issues in any particular part of the LGA

This system emphasises the legitimacy and profile of the Mayor, and de-emphasises the role of Councillors.

Brisbane City Council, with its 26 wards and Lord Mayor, is quite different – Councillors are generally known in their local area, and can be more or less popular than the mayoral candidate from the same party.

Voluntary voting

The Queensland Government discussion paper touches on voluntary voting. The Rudd Government asked a similar question in 2009 in a Green Paper on electoral reform.

As we know, compulsory voting was introduced in Queensland in 1915 and 1924 at the Federal level.

The evidence suggests that Australians generally support compulsory voting. In the 2010 Australian Election Study, 72% of respondents supported compulsory voting, and 28% supported voluntary voting. 8.4% said they probably or definitely wouldn’t have voted if voting had not been compulsory.

The question for us is: what would be its impact if it was introduced?

Traditionally, it was thought that voluntary voting benefitted the conservatives.

Professor Ian McAllister from the ANU said that voluntary voting might have given the conservative parties a 2 per cent benefit some decades ago but that the gap is much tighter now.

Professor Clive Bean of QUT said that voluntary voting would benefit the conservatives in the short term – based on the 2010 Australian Election Study. In that study, nearly 90 per cent of Liberal voters said that would ‘definitely’ or ‘probably’ vote if voting was voluntary, while 85% of Labor voters said the same. It’s unclear what the margin of error was.

A uniform one per cent benefit to the conservatives would be enough to change some election results – the 1995 Qld election, and the 1990 and 2010 federal elections.

But we need to do more to understand the likely election impacts. What matters is which party would receive a net benefit in marginal seats — which we don’t know the answer to. The net impact might even help Labor in marginal seats.

In addition, we’d have to know who would benefit in inner-city, suburban, outer-metropolitan, regional, and rural seats. How would it differ across demographics? How would it vary between major parties, the Greens, and the Katter Australian Party?

We don’t know. The survey results are too broad. And they are hypothetical. People do not necessarily behave the way they say they’ll behave.

What we can be certain of is that parties would immediately attempt to work out how voluntary voting would impact them.

We can also say that political campaigning and activities would be likely to change.

Parties would have to put a lot of effort into convincing people to vote. There was a 47 per cent National response for the 1998 voluntary postal ballot for the Constitutional Convention. Presumably a state election would have a larger turnout than 47 per cent.

Evidence from low turn outs at some Council elections across Australia, even with fines for not voting, suggest that people may not vote if they think it isn’t important enough. The turn out for the last Qld local Govt elections was 80% while it was 91% for the state election. In some Melbourne Councils it was 50%.

Political parties would have a much bigger job getting people out to vote, particularly in marginal seats. They would make more effort to contact and target voters in the lead up to an election. On polling day, parties might attempt to track people who havn’t voted and try to get them to a polling booth.

They are already starting to target individual voters much more than previously, and this would receive greater emphasis under voluntary voting.

Getting out the vote would be very difficult for political parties given their small membership. The Qld ALP seems to have about 6,000 members, while the Qld LNP had 13,000 in 2008.

Election results would depend to a large extent on which side could get their supporters out to vote.

Election results would likely be more volatile as people may stay home if they were unenthused about the options. Some people have wondered how many seats Labor would have won in the last Qld and NSW elections if voting had been voluntary.

Senator Barnaby Joyce has argued that introducing voluntary voting would pave the way for radical political movements to push racist and extreme green agendas.

His point is that voluntary voting may lead to election results that don’t reflect most people’s views.

This is possible, but unlikely. We don’t tend to see it in NZ or UK general elections.

While some US candidates have had fringe views, e.g. on abortion, they were generally defeated in the last Senate elections even in conservative states. The primary election process was at fault in the selection of those candidates.

Many people express concern that voluntary voting might lead to the hyper-partisanship seen in the US.

It must be pointed out that the UK and NZ also have voluntary voting, and they don’t display the hyper-partisanship seen in the US. Other factors particular to the US must come into play.

Would partisanship increase in Australia under voluntary voting? Possibly, if parties sought to campaign to their base supporters, or other opposing groups, such as anti-abortion and pro-choice groups, tried to get their supporters out.

Campaigning would almost certainly change. Currently, campaigning targets the marginal voter — but under voluntary voting, parties would have to energise both their base as well as marginal voters. This increases the incentives for greater spending.

Energising the base is key in voluntary voting.

We saw how Gordon Brown targetted Scottish Labour voters in the 2010 UK election — and Labour actually improved its vote in Scotland.

We also saw over the last decade how the Republicans and Democrats in the US targetted their core voters in general elections by concurrently running ballots on gay marriage or the minimum wage. These techniques worked. No doubt similar techniques would be used to get the vote out in Queensland elections.

Voluntary voting may also allow voter intimidation. It is much easier to try to prevent people voting if it is not compulsory. There are numerous examples of alleged voter intimidation in the US.

Let’s look at arguments around voluntary voting.

Firstly, the obvious one: why should people be forced to vote if they don’t want to? As blogger Paula Matthewson (also known as Drag0nista) wrote on The Drum:

“What’s the point of forcing voters begrudgingly to the polls? If we engaged them in the political process, they’d do so voluntarily.”

My civil libertarian instincts support voluntary voting. Why should people be forced to do something they don’t want to? How does not voting harm others?

NZ and the UK seem to survive quite well with voluntary voting. There isn’t necessarily a spiral of election expenditure. The UK has expenditure limits of £30,000 per constituency contested — while marginal state seat campaigns in Australia can easily cost $150,000 per seat. NZ also has expenditure limits of about $1 million plus $25,000 per seat contested.

Expenditure caps in Australia may restrict any spending spiral.

An argument in favour of voluntary voting is that election results should reflect the opinions of the adult citizenry who are interested enough to vote. A low turnout may reflect dissatisfaction with the political process. Any other impacts from people not voting must just be borne and addressed.

Another argument is that voluntary voting would also mean that parties would have to engage with voters much more than currently.

One benefit of compulsory voting may be an almost philosophical point that it ensures election results reflect the opinions of most adult citizens.

How about arguments against compulsory voting?

One argument is that compulsory voting is an imposition on liberty. It IS an imposition — as is jury service, seat belts, and road rules. Compulsory voting is a trivial imposition. Filling out a tax return usually takes longer.

Another argument opposes compulsory voting on the basis that it requires people who are uninterested or know very little, or are badly-motivated, to vote. This is true — but with the mass franchise, the only qualification to vote is being an adult citizen. Not some level of property, income or education. This argument against compulsory voting falls over.

In addition, this argument would support people who are very interested, or particularly knowledgeable about politics, to have a larger say in elections. Maybe they should each receive two, three, or ten votes, perhaps graduated according to a scale of worthiness determined by notable elders. I’ve heard some Australian right-wing economists suggest, maybe tongue in cheek, that people with economics degrees should get two votes as ‘they better understand how the world works’. Some right-wing commentators have said that you should only be able to vote if you pay tax. This type of argument falls over as it is inconsistent with a mass franchise.

Weighing up these points does not lead to an obvious answer. Perhaps the most substantial points are that while voluntary voting would enhance the freedom to vote or to not vote, it may also allow for intimidation to attempt to discourage people from voting.

In practice, it is likely that Parliamentary support or opposition to voluntary voting will not depend on principled notions of liberty or impacts on political culture, but pragmatism — 1. whether it is likely to benefit the party in government, and 2. whether it would make campaigning by political parties more difficult.

This is no doubt on the minds of MPs.


Electoral systems are, to quote Dr Paul Reynolds from a 1992 lecture, ‘the calculus for turning votes into seats’. Different systems will lead to different party compositions in Parliament.

But while parties will think of the impacts of different systems on themselves, it is at least as interesting to consider their wider systemic impacts.

  • Do they promote greater volatility in election results and reduce the ability of the opposition to do its job?
  • Do they ensure that political views of sufficient support are represented in Parliament?
  • Do they promote party splits and minority governments?
  • Do they emphasise the role of the leader?
  • Do they promote the power of factions within parties?
  • Are MPs directly accountable to voters or to a central party committee?

I trust and hope that these systemic issues are touched on in the current focus on electoral reform in Queensland.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: