LSE - Small Logo
LSE - Small Logo

Blog Admin

April 12th, 2011

Some Australian state elections using the Alternative Vote show that voters are casting fewer second preferences, or even none at all. Would the same happen in the UK?


Estimated reading time: 5 minutes

Blog Admin

April 12th, 2011

Some Australian state elections using the Alternative Vote show that voters are casting fewer second preferences, or even none at all. Would the same happen in the UK?


Estimated reading time: 5 minutes

Australian state elections using the Alternative Vote hold the key lessons for how AV might operate in the UK. Colin Rallings and Michael Thrasher have argued that they show many or most voters not using the opportunity to cast a second preference at all, but just ‘plumping’ for their top candidate alone. The leading Australian expert Antony Green explores whether they are right to draw this likely lesson for AV in the UK.

Last November Colin Rallings and Michael Thrasher noted on this blog that some of the Australian experience with AV shows many voters choosing not to express second or subsequent preferences. Instead they cast only a first preference vote (also called ‘plumping’), just as they would under first-past-the-post elections. At each AV election in New South Wales and Queensland, the proportion of these incomplete or even single preferences has grown over time.

I argue that the effective plumping rate is not as great as Rallings and Thrasher state. The data they rely on is drawn from ballot paper research conducted by The Electoral Commission of Queensland (ECQ). This includes the winning and second placed candidates whose preferences were never examined as part of the formal distribution of preferences. This inflates the overall plumping rate when determining the winner depends on the plumping rate of the candidates whose preferences were distributed.

The table below is taken from the ECQ survey of ballot papers in a sample of 10 electorates at the 2009 state election.

Percentage of Ballot Papers with
Party                                      ‘1’ only vote       Partial Preferences    Full Preferences
Labor Party                                   56.5                          14.5                          29.0
Liberal National Party                 74.8                           2.3                           22.9
The Greens                                    45.6                           6.1                           48.4
Daylight Saving Party                  32.2                           10.3                        57.5
Family First                                   48.2                           4.6                           47.2
Independents/Others                  55.9                           4.4                           39.7

The Liberal National Party recommended a ‘1’ only vote in every constituency and had the highest rate of ‘plumping’. The Labor Party recommended a ‘1’, ‘2’ vote in some electorates, resulting in Labor’s vote having the highest proportion of partial preferences. Family First was the only party to recommend full preferences in every seat, while other parties had mixed preference recommendations.

Liberal National preferences were not distributed in any of the electorates, and Labor’s in only one. So the important plumping rate is that for the minor parties, not the overall rate including the undistributed major parties.

When AV was re-introduced in Queensland in 1992, Labor and the Liberal-National Coalition continued with their past practice of recommending full preferences. This was abandoned by Labor at the 2001 election after the rise of insurgent right-wing party One Nation. Labor’s slogan of ‘Just vote 1’ influenced One Nation voters not to direct any second preferences, damaging conservative prospects.

More recently the conservative parties have adopted the same tactic to cut preference flows to Labor from the Greens. At the most recent New South Wales election, the deep unpopularity of the Labor government saw the Greens adopt a ‘1’ only strategy, except for a few seats where support for the Greens and left-wing Labor candidates were roughly equal, and where Liberal candidates could have been delivered victory if left voters did not use all their preferences.

So the rise in ‘plumping’ recommendations by major parties is not relevant of itself, but is relevant as a tactic to influence the plumping rate of minor parties. The rise of plumping recommendations by Australian major parties is an attempt to reduce AV contests into first past the post races.

The ECQ research examined only the mechanics of ballot paper number sequences, not their political impact. After the 2007 New South Wales election, I undertook research looking more at the eventual destination of preferences and the impact of party recommendations on these flows. The research showed that third party recommendations have an impact, and also that major parties were likely to abandon plumping tactics in contests where their candidates were likely to be excluded.

The research did not use ballot papers, but used of indicative preference data collected on election night to assist the media in reporting the election count. It can be found here (pages 55-84).

In 43 electorates where the Greens recommended preferences for Labor, the average flow of preferences was 46% to Labor, 11% to the Liberal-National Coalition and 43% exhausted. When the Greens offered no recommendation for either Labor or the Coalition, the numbers were 33% to Labor, 14% to the Coalition and 53% exhausted. So preference recommendations by a significant third party can matter.

In contests where Labor was likely to have its candidate excluded and distributed as preferences, Labor departed from plumping and recommended preferences. In 13 contests where Labor preferences were distributed between Coalition and Independent candidates, Labor’s preferences flowed 8% to the Coalition, 44% to the Independent and 48% exhausted. These were in contests where Labor’s first preference vote was very low and many Labor supporters may not have been aware of any recommendation.

The opposite tactic was taken by the Liberal Party in contests between Labor and Green candidates, recommending no preferences with 73% of Liberal ballot papers exhausting.

These rates of exhausted preferences were recorded in a state where voters must still provide full lists of preferences at Federal elections, and where preference recommendations form part of the active campaigning that occurs outside polling stations.

If the proportion of Australian voters giving preferences has declined over time under AV, sceptics might ask what hope is there that large numbers of UK voters would start numbering preferences on their ballot when they have never done so before?

The Australian research would indicate that even with extensive education, British voters will not spontaneously start giving preferences on the ballot paper. Party recommendations will be critical in determining how voters use the new system.

The behaviour of Australia’s largest parties indicates that it is in their interests to minimise the influence of preferences, to encourage plumping to keep the contest a first past the post race. This is especially the case with any party likely to lead on first preferences.

Rather than UK parties trying to attract preferences from third parties, it is more likely the largest parties in each contest will continue with tactical voting campaigns aimed at driving down support for third parties in each contest. The difference is that under AV voters will have another option, to still turn up and vote for their preferred party, but then also to direct preferences.

Education may help voters choose to exercise their preference option, but more likely it will be the tactics of the parties that will have more impact on how AV would work at UK elections.

For more on the Australian experience of AV, please see Antony Green’s blog article from last week.

Please read our comments policy before commenting.

Print Friendly, PDF & Email

About the author

Blog Admin

Posted In: Alternative Vote (AV) referendum | Electoral and constitutional reform


Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported
This work by British Politics and Policy at LSE is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported.