BELOW-THE-LINE VOTES AT THE
2006 VICTORIAN STATE ELECTION
BELOW-THE-LINE VOTES AT THE 2006 VICTORIAN STATE ELECTION
Introduction
A new voting system was introduced for Victoria’s Legislative Council (Upper House) at the 2006 State election. The system changed from preferential voting for a single vacancy (the same system as for the Lower House) to a Senate-style form of proportional representation. Instead of numbering all the squares on a small ballot paper, voters now had a choice of either voting “1” above the line across the ballot paper for a party or group, or of voting below the line for individual candidates. Unlike Senate elections, in which electors who vote below the line are instructed to number every square, Victorian Legislative Council elections are optional preferential –electors only have to vote “1” to “5” for their vote to be counted.
There was considerable interest in the operation of the new system. The media, parties and political scientists studied how voters’ preferences translated into the seats won by the various parties. The VEC examined informal ballot papers, measuring the incidence of types of informal voting and trying to explain them. It appears that the new Upper House system did affect the informal vote. (See the VEC’s Report to Parliament on the 2006 Victorian State election, pages 90-96)
Only some 5% of the voters chose to vote below the line. However, these votes were potentially very important for the result of the election. Preference flows for above the line votes are determined by the parties and groups of candidates, and follow publicly available group voting tickets lodged by the parties and groups. Above the line preference flows are thus predictable. In contrast, preferences on below-the-line votes are decided by the voters themselves. These unpredictable preferences could decide who won the fifth seat in each region.
Because the VEC data-entered the below-the-line votes as part of the election count, the pattern of preferences on each ballot paperis available for study. Such study does not infringe the secrecy of the vote, since the ballot papers cannot be traced back to individual voters. The VEC engaged a team at the University of Melbourne to conduct programming work on the database of below-the-line votes and to produce statistics. The VEC is grateful to Vanessa Teague, Kim Ramchen and Lee Naish for their work in January 2008.
Incidence of below-the-line votes
Below-the-line votes comprised 5.16% of the total votes for the Legislative Council. Reasons for this small proportion appear to be that it is easier to vote above the line than below the line, and that all the party how-to-vote cards urged their supporters to vote above the line. On the other hand, the proportion of below-the-line votes in the State election was more than two times higher than the 2.05% of Victorians who voted below the line at the 2007 Senate election. Clearly, the fact that voters only had to number five squares encouraged some Victorians to vote below the line at the State election.
The rate of below the line voting varied according to parties. Broadly, the larger and more established a party was, the lower its rate of below-the-line voting. Thus, only 2.52% of Liberal voters voted below the line, while 22.91% of People Power voters did so. The highest rate of below-the-linevoting was among supporters of Northern Victoria Region’s Group H (headed by Laurie Whelan), 48.46% of whom chose to vote below the line. It is significant that the supporters of the major parties, who nearly all voted above the line, also tended to conform with their party’s how-to-vote cards (See the VEC’s Report to Parliament on the 2006 Victorian State election, page 98). The reasons appear to be the same in both cases. Firstly, a high proportion of voters for the major parties would be lifelong supporters, who would tend to heed their party’s directions. Secondly, the major parties were better able to get their message out to their supporters, through stationing members at every voting centre to hand out how-to-vote cards.
The Australian Labor Party (ALP) had the second lowest rate of below-the-line voting, at 3.36%. The party did have the most prominent case of below-the-line voting in the State election: supporters of Elaine Carbines, who was the third-placed Labor candidate for Western Victoria Region, urged Labor voters to vote “1” for Carbines below the line, and 2,825 voters did so. This support was not enough to elect Ms Carbines. The highest rate of Labor below-the-line votes was 4.61% in Southern Metropolitan Region, without any issues being apparent.
The Australian Greens were the largest source of below-the-line votes, outnumbering Labor and Liberal below the line votes and making up 28.2% of the total. Below-the-line Green voters made up 13.76% of all Green voters, with the proportion varying little across the State. The below-the-line voting rate was considerably higher for the Greens than for two smaller parties, Family First (8.5%) and the DLP (7.78%). It is difficult to explain this high rate of voting below the line. Possibly some Green voters, having broken with their previous voting habits, were more inclined to make up their own minds about the direction of their preferences.
Geographically, the below-the-line vote rate ranged from 3.96% in Eastern Victoria Region to 6.86% in Southern Metropolitan. The highest proportion of below-the-line votes was in the two regions covering the inner suburbs (Northern and Southern Metropolitan, 6.72%), while the three regions covering the outer suburbs (Eastern, South Eastern and Western Metropolitan) had distinctly fewer below-the-line votes at 4.6%. Part of the explanation for the high numbers of below-the-line votes in Northern and Southern Metropolitan could be that these regions were the strongholds of the Greens. However, supporters of other parties were also more disposed to vote below the line in these regions: in Southern Metropolitan Region, below-the-line votes for the Labor, Liberal, Family First and People Power parties were the highest in the State, and the same was the case for Democrat voters in Northern Metropolitan Region. In Northern Metropolitan Region, the Greens had both their highest vote in the State (18.06%) and their highest proportion of below-the-line votes (15.34%). These figures suggest that there is something in the nature of these regions that made their voters more likely to vote below the line. From the VEC’s Information Kit for the 25 November 2006 State Election, one factor that these regions have in common is a high proportion of the workforce in professional occupations.
Types of below-the-line votes
Although a huge number of combinations of below-the-line votes is mathematically possible, in practice voters tend to follow patterns. The VEC analysis was designed to tease out the numbersof voters who followed likely patterns at the 2006 State election.
Mistakes
Some voters make mistakes when completing their ballot paper,skipping or duplicating numbers. Such votes are formal if the voter has voted “1” to “5”, but preferences cannot be counted from the point of the mistake.
Very few voters made mistakes when voting below the line. Across the State, only 1,437 voters made mistakes on their ballot paper – less than 1% of the total. There was no particular point at which voters were more likely to get it wrong; mistakes were few in number and randomly scattered from the 6th square on the ballot paper to the 38th. In terms of parties, the incidence of mistakes varied from 0.62% for the ungrouped candidates to 2.07% for the Socialist Alliance. Supporters of the larger parties tended to be less likely to make mistakes, probably because most of these voters simply voted “1” to “5”.
Keeping within group
Voters have the option of either keeping to one group of candidates for their first few preferences, or switching from one group to another – for example, voting Labor1, Greens 2, People Power 3, Labor 4 and so on. It is easier to vote within a group, though a well known candidate might attract preferences from supporters of other parties.
In fact, 83.17% of below-the-line voters kept within one group for their first preferences. Although below-the-line voters exercised their own judgement, the great majority of them appear to have been committed to one party or group, or at least found it simpler to vote within that group before going on to other groups. There was little variation by region, but much greater variation by party, from 74.14% for Group E in Northern Victoria (headed by Stefano de Pieri) to 91.19% for the Socialist Alliance. The rate for the major parties tended to be slightly lower than that for the smaller parties. One possible explanation for this is related to the number of candidates in each group. While the smaller parties fielded two or three candidates per region, the larger parties mostly had five candidates in each region. There is a correlation of -.688 between each party’s average number of candidates per region and rate at which each party’s voters kept within the group. This indicates that the more candidates there were in a group, the more inclined voters were to stray from that group.
Having voted for their favoured group of candidates, did voters keep within groups in completing their further preferences or did they switch between groups? The proportion of voters who kept within groups while filling out their ballot papers was 78.1%. Nearly all voters who had voted first for all the candidates in a group continued to vote by groups for their lower preferences. There was some variation by region, though voters in the metropolitan regions showed a slight tendency to vote more by groups than country voters did. Again there were greater variations by party, ranging from 65.07% for Group E in Northern Victoria (headed by Stefano de Pieri) to 80.65% for the Liberal Party. It is difficult to detect a meaningful pattern in the scores for each group. What is significant are the differences between these scores and rates at which voters stayed within a group for their first preferences. Broadly, the differences were smaller for the major parties. For example, 84.57% of Liberal voters kept within the Liberal group and 80.65% voted by groups for all their preferences, a difference of 3.92 percentage points, while the corresponding figures for the Country Alliance were 90.29% and 79.08%, a difference of 11.21 percentage points. This shows that although supporters of major parties were more inclined to stray from that party, the major party voters who stayed within the party group nearly all voted by groups when completing their lower preferences.
Ballot paper order
Another subject of analysis was the degree to which voters numbered their ballot paper in the same order as on the ballot paper, both for their favoured group and for other groups.
In Victoria as a whole, 67.89% of below-the-line voters voted in ballot paper order for their preferred group. This figure disguises a marked variation by party, ranging from 53.79% for Labor voters to 92.63% for Country Alliance voters and 97.82% for supporters of Group H in Northern Victoria. Voters for the major parties, particularly Labor and Liberal, were much less inclined than voters for the smaller parties to keep to ballot paper order. It may be that one of the main reasons why major party supporters would have voted below the line would have been a disagreement with the order of candidates on the party ticket (as in the case of Elaine Carbines in Western Victoria Region), leading to a high proportion of them voting in a different order. In contrast, voters for the smaller parties, who tended to vote for these parties as a protest, probably had no objection to the order of their party’s two or three candidates on the ballot paper.
In geographical terms, the proportion of voters keeping to ballot paper order for their preferred group ranged from 59.12% in Western Metropolitan Region to 80.34% in Northern Victoria. Country voters were more inclined than metropolitan voters to vote in ballot paper order, except in Western Victoria Region, which was affected by the Carbines campaign.
Turning to the number of voters who voted in ballot paper order for all the groups on which they expressed an opinion, much the same pattern is evident, except at a lower level. Statewide, the figure was 57.61%, about 10 percentage points lower than the score for the preferred group. Voters for the Labor and Liberal parties were the least inclined to keep to ballot paper order, and metropolitan voters recorded lower scores than country ones (except for Western Victoria Region). One noticeable feature was that the score for Labor and Liberal voters showed a comparatively slight decline compared to their score for their preferred group. For example, the figure for Labor voters was 47.87%, 5.92 percentage points lower than the proportion of Labor voters who kept to ballot paper order in the Labor group. By contrast, the Democrats’ score for all groups was 17.51 percentage points lower than their score for the Democrat group alone. Although Labor and Liberal voters were more disposed to vote in a different order for their own party’s candidates, not many of them then went on to vote in a different order for other groups.
Number of preferences
Below-the-line voters had the option of either voting “1” to “5”, or continuing to number as many squares as they wished. Instructions on the ballot paper and the VEC’s information campaign stressed that below-the-line voters should number the squares from “1” to at least “5”. Voters who did not number all the squares may not have realised that their votes would exhaust at their final preference, reducing the power of their vote to decide the result. The actual number of exhausted votes depended on the flow of preferences in each region. In the event, 29,445 votes, or about one fifth of the total, were exhausted during the count.
Overall, 57.15% of below-the-line voters only voted “1” to “5’. The second largest group were those who numbered all the squares on their ballot paper, comprising 21.86% of the total. Small numbers of voters stopped at each intermediate point on the ballot paper, without any pattern being apparent.
In party terms, the proportion of voters contenting themselves with the minimum number of squares ranged from 42.3% for the Australian Democrats to 71.65% for Group D in Northern Metropolitan Region (headed by Joseph Kaliniy). Supporters of the major parties were much more inclined to stop numbering at 5 (Labor 67.03%, Liberal 64.77%, National 65.26%); it seems that these voters, having made up their own minds about the order of the candidates for their own party, then decided to rest on their laurels. In contrast, voters for the smaller parties tended to have lower proportions of “1 to 5” voters, particularly the Democrats and the Greens (43.41%). Voters for the non-party groups varied, but most voted “1” to “5”.
The picture for voters who completed all the squares on the ballot paper was the converse of the above. Proportions ranged from 10.34% for Group H in Northern Victoria Region to 30.53% for the Democrats, with low scores for the large parties and high scores for most of the smaller parties. It may be that voters for the smaller parties, who tended to vote for them as a protest, were determined to make their own decisions about all the candidates on offer.
The number of squares completed varied considerably across Victoria. The region with the lowest proportion of “1 to 5” votes was Southern Metropolitan (48.88%), and the region with the highest proportion was Western Metropolitan (68.15%). The Eastern, Northern and Southern Metropolitan Regions had comparatively small numbers of “1 to 5” voters, while the other metropolitan and most of the country regions were higher than average. These variations partly reflected the differing strengths of the parties in the regions. However, it is noticeable that the three regions with lower “1 to 5” scores (and higher proportions of completed ballot papers) shared higher than average proportions of professionals and internet users. Perhaps the social characteristics of the residents of these regions influenced the way they voted below the line.
Conclusions
This study of the below-the-line votes at the 2006 State election leads to the following conclusions.