The 2019 Senate election was the second conducted under changes introduced in 2016. The changes continued to use proportional representation by single transferable vote, and retained the divided ballot paper in use since 1984, and . A thick horizontal line continues to divide the ballot paper into two voting options, ‘above the line’ (ATL) for parties and groups, or ‘below the line’ (BTL) for candidates.
The changes abandoned full preferential voting in favour of partial preferential voting, and ended party control over between-party preferences.
Before the changes, voters could only mark a single square when voting ATL, the ballot paper imputed to have the chosen party’s full list of preferences as registered with the Electoral Commission.
The new system abolished the tickets and allowed ATL voters to give second and further preferences, ballot paper instructions suggesting at least six preferences. Above the line votes continued to give parties and groups control over preferences between their own candidates, but ended party control over preferences to other parties and candidates.
Previously a BTL vote required a voter to mark preferences for all candidates on the ballot paper. Under the new system, ballot paper instructions stated that BTL voters should mark at least 12 preferences.
In an earlier post I went into the political impact of these changes and how the system performed at its second test, its first at a half-Senate election. (See How the new Senate Electoral System Performed at its first Half-Senate Election test.)
In this post I’m going to look at how voters reacted to the new electoral system and whether they voted above or below the line. For each option, I look at how many preferences voters completed.
This will be the first of several posts over the next fortnight going into detail of how the Senate count unfolded in each state, how preferences flowed, and what impact parties and their how-to-votes had on preference flows.
About the Data Source
All tables have been drawn from ballot paper data published by the Australian Electoral Commission (AEC). The data contain a record of marks on all formal ballot papers admitted to the count. It includes all valid numbers on each ballot paper, both above and below the line. The file has had to be cleansed to resolve ballot papers marked both above and below the line, and to remove additional preferences beyond breaks or duplicates in preference sequences. There are no informal votes in the data set.
What is a Formal Vote?
The ballot paper instructions to voters are to mark at least six preferences for an ATL vote, or at least 12 preferences for a BTL vote. The Electoral Act also includes directions to counting staff on how to deal with ballot papers not completed according to the instructions. These are the so-called savings provisions and are –
- Any ballot paper marked above the line is formal as long as it had at least a valid first preference. This includes a single tick or cross imputed to be a 1. By this rule, any ballot paper with fewer than six ATL preferences is formal.
- Any ballot paper marked below the line with at least six preferences is formal.
- A ballot paper is formal once the minimum number of preferences is reached. If a subsequent preference on a formal ballot paper misses or duplicates a preference, the ballot paper is formal up to the break or duplicate.
- Any ballot paper marked both above and below the line counts as a BTL vote if the BTL preference sequence is formal.
- Any ballot paper marked both above and below that was informal below the line defaults to use the ATL preference sequence if it is formal.
For these reasons, the tables below include totals for votes that were incomplete according to the ballot paper instructions, but remained formal thanks to the savings provisions.
Summary of Ballot Papers
Table 1 below provides percentages of formal ballot papers by category for each state, territory and nationally. The categories are –
- ballot papers with a single ‘1’ above the line
- ballot papers with 2-5 ATL preferences
- ballot papers with the recommended 6 ATL preferences
- ballot papers with 7-12 ATL preferences, 12 being the recommended number of preferences below the line
- ballot papers with more than 12 ATL preferences
- BTL ballot papers
Table 2 contains the same analysis for the 2016 election. The number of groups on the ballot paper is shown next to the state code.
You can peruse the tables in full below, but the key highlights are –
- There wasn’t a lot of change between how voters completed their ballot papers in 2016 and 2019.
- At both elections, NSW saw the highest proportion of votes counted that had only an ATL first preferences, 4.3% in 2016 and 5.5% in 2019. Both figures relate to the use of optional preferential voting at NSW elections, including on the state’s upper house ballot paper. A NSW state election was held two months before the Federal poll. Informal voting is always higher in NSW at House elections, and usually due to a high rate of ‘1’ only voting.
- The proportion of ballots with fewer than 6 ATL preferences was 7% overall, ranging from 3.6% in the ACT to 9.6% in NSW. Without the savings provision, these ballot papers would have been informal.
- The proportion of votes counted with the suggested minimum six preferences was 80.0%, slightly down on 81.6% in 2016. The lowest rates were 60.3% in Tasmania and 65.4% in the ACT, where many voters used the BTL option thanks to experience with the Hare-Clark system at local elections. The low rate of 6-preference ATL votes in the NT (66.5%) was not due to BTL voting but due to a much higher rate of voters going beyond 6 preferences on a 9-column ballot paper. The ACT with 7-columns also saw a higher rate of voters going beyond 6 preferences.
- The proportion of voters going beyond 12 preferences rose as the size of the ballot paper shrank, going from 0.8% in NSW with 35 columns, to 4.0% in Tasmania with 16. There were not enough columns to go beyond 12 in the ACT or NT.
- Tasmanian (27.7%) and the ACT (22.2%) recorded by far the highest rate of below the line voting, again presumably due to experience with the Hare-Clark electoral system at local elections. The Tasmanian figure was down slightly from 28.1% in 2016, the ACT figure up on 15.2% in 2016. The rate of BTL voting rose in NSW, largely due to the activities of Jim Molan, who was fourth on the Coalition ticket but campaigned for below the line votes and recorded 137,325 votes or 2.92% of the state vote. Molan recorded 42.6% of the BTL vote in NSW.
- Nationally there was a slight rise in below the line voting from 6.5% to 7.3%.
- The number of voters who completed all above the line squares by state were NSW 16,854 (0.4%), Victoria 8,845 (0.2%), Queensland 11,579 (0.4%), Western Australia 18,733 (1.3%), South Australia 29,372 (2.7%), Tasmania 6,038 (1.7%), ACT 24,074 (8.9%) and Northern Territory 8549 (8.1%). Many more voters may have tried to complete all numbers but made a mistake. Percentages are of all formal ballot papers.
Table 1 – Above and Below the Line Voting – 2019 Half-Senate Election
| Percentage of Ballot Papers | ||||||
|---|---|---|---|---|---|---|
| Counted as Above the Line – Number of Preferences | Below | |||||
| State (Groups) | 1 | 2-5 | 6 | 7-12 | >12 | Line |
| NSW (35) | 5.5 | 4.1 | 78.6 | 4.2 | 0.8 | 6.9 |
| VIC (31) | 3.1 | 3.6 | 83.0 | 3.8 | 0.7 | 5.8 |
| QLD (26) | 1.7 | 3.1 | 82.5 | 4.6 | 1.2 | 6.9 |
| WA (23) | 2.7 | 3.1 | 80.8 | 5.9 | 2.0 | 5.5 |
| SA (16) | 3.0 | 2.7 | 79.6 | 3.8 | 3.3 | 7.6 |
| TAS (16) | 1.3 | 2.4 | 60.3 | 4.9 | 4.0 | 27.1 |
| ACT (7) | 1.6 | 2.0 | 65.4 | 8.9 | .. | 22.2 |
| NT (9) | 2.6 | 3.0 | 66.5 | 19.5 | .. | 8.3 |
| Australia | 3.5 | 3.5 | 80.0 | 4.5 | 1.2 | 7.3 |
Source: Calculations by the author derived from ballot paper data files published by the AEC. (See this link).
Table 2 – Above and Below the Line Voting – 2016 Double Dissolution Election
| Percentage of Ballot Papers | ||||||
|---|---|---|---|---|---|---|
| Counted as Above the Line – Number of Preferences | Below | |||||
| State (Groups) | 1 | 2-5 | 6 | 7-12 | >12 | Line |
| NSW (41) | 4.3 | 4.1 | 81.3 | 4.3 | 0.6 | 5.4 |
| VIC (38) | 1.9 | 3.5 | 83.9 | 4.5 | 0.8 | 5.3 |
| QLD (38) | 1.6 | 3.3 | 83.6 | 4.6 | 0.8 | 6.1 |
| WA (28) | 1.9 | 3.4 | 83.7 | 4.2 | 1.2 | 5.5 |
| SA (23) | 2.0 | 2.9 | 79.5 | 5.3 | 1.8 | 8.5 |
| TAS (21) | 0.9 | 2.1 | 61.3 | 5.0 | 2.7 | 28.1 |
| ACT (10) | 1.2 | 1.7 | 70.7 | 11.1 | .. | 15.2 |
| NT (7) | 2.1 | 2.4 | 51.3 | 35.6 | .. | 8.6 |
| Australia | 2.6 | 3.5 | 81.6 | 4.9 | 0.8 | 6.5 |
Source: Calculations by the author from data released by AEC and used to calculate Tables 2 and 3 in ‘Senate ballot paper study 2016 – Number of preferences’.
Below the Line Voting
Table 3 below replicates the ATL analysis in the above tables for 2019 below the line votes. With below the line votes, any ballot paper with fewer than 6 BTL preferences was informal. The savings provision where a vote with only six of the 12 suggested preferences completed insures that any voter completing a BTL vote having read the ATL instructions, would end up completing a formal vote.
Table 3 shows percentages for the following categories of BTL vote –
- ballot papers with the minimum 6-preference allowed by savings provisions
- ballot papers with 7 to 11 preferences
- ballot papers with the suggested 12 BTL preferences
- ballot papers with between 13 preferences and one fewer than the number of candidates (Max). This upper limit varied from state to state.
- ballot papers where a voter successfully completed a sequence of preferences for every candidate on the ballot paper (Max).
- the last columns shows an average of the number of valid preferences on formal ballot papers. In all states the median value was 12, the skewed distribution meaning the average was higher than the median.
The number of candidates is shown on the left next to the state code.
I don’t have comparable 2016 data available, but the following comments can be made on the 2019 data.
- The highest rate of 6-only preferences was 7.2% in NSW and Victoria, both states also totalling 10% with fewer than 12 preferences. The NT had the highest rate with fewer than 12 preferences (10.8%) and the ACT the lowest (3.0%). Nationally 5.2% of ballot papers had only 6 formal preferences, and in total 8.0% had fewer than 12 preferences. Without the savings provision, these ballot papers would have been informal.
- As noted with ATL votes, the number of preferences completed increased as the number of candidates declined. A voter forced to hunt around the giant NSW ballot paper (35 groups + ungrouped, 105 candidates) was perhaps more inclined to give up after locating 12 candidates, while voters in SA (42 candidates), Tasmania (43), ACT (17) and NT (18) had smaller ballot papers that could be read without folding up the side of the voting screen.
- The smaller number of candidates in the ACT and NT produced much higher rates of voters numbering all the squares.
- Of the states, NSW had the lowest average preferences per BTL sequence (13.9) and WA the highest (18.7). The figures were lower in the ACT and NT due to their being fewer candidates.
- The number of voters who completed all below the line squares by state were NSW 3,450 (0.07%), Victoria 10,147 (0.27%), Queensland 7,569 (0.26%), Western Australia 6,975 (0.48%), South Australia 12,333 (1.16%), Tasmania 11,938 (3.39%), ACT 20,055 (7.42%) and Northern Territory 2,224 (2.12%). Many more voters may have tried to complete all numbers but made a mistake. Note that the above percentages are of all formal votes while the Max column below shows percentages of below the line votes.
Table 3 – Below the Line Voting – 2019 Half-Senate Election
| Percentage of Below the Line Ballot Papers | ||||||
|---|---|---|---|---|---|---|
| State (Cands) | 6 | 7-11 | 12 | 13 to
< Max |
Max | Avge Prefs |
| NSW (105) | 7.2 | 2.8 | 72.6 | 16.3 | 1.1 | 13.9 |
| VIC (82) | 7.2 | 2.8 | 55.7 | 29.6 | 4.7 | 17.5 |
| QLD (83) | 3.8 | 3.0 | 70.2 | 19.3 | 3.8 | 16.2 |
| WA (67) | 4.2 | 3.0 | 61.0 | 23.1 | 8.8 | 18.7 |
| SA (42) | 2.7 | 2.3 | 61.9 | 18.3 | 14.9 | 17.3 |
| TAS (44) | 2.4 | 2.7 | 62.2 | 20.1 | 12.5 | 17.0 |
| ACT (17) | 1.0 | 2.0 | 58.8 | 4.8 | 33.4 | 13.7 |
| NT (18) | 5.5 | 5.3 | 57.8 | 6.0 | 25.5 | 13.2 |
| Australia | 5.2 | 2.8 | 65.2 | 19.8 | 7.0 | 15.9 |
Calculations by author from ballot paper data files published by the AEC. (At this link.)
Excellent resource Antony. Hope you’ll be posting regularly.
Can you explain how the senate ATL voting operates as far as counting the votes? That would be appreciated by a group of friends
COMMENT: An above the line vote is imputed to be for the candidates of that group as listed below the line. So if you vote 1-6 above the line, it is treated as being a vote for the candidates of your first choice party in order, then the candidates of your second choice party in order, then the candidates of your third choice party, and so on.
Can you give examples of both above and below the line voting, using fictitious characters of course ie’ Mickey Mouse, Donald Duck etc.
If voting above the line, I assume you could list six candidates by name only, in whichever order you chose whereas if you vote by group above the line, so for example if there were 6 candidates standing for an individual party you could list all six in your order of preference? If for example you wanted to list 3 from a party and 3 from a different party, then you could list both groups of 3 in your order of preference. Similarly I assume you could list 3 from your first preference ie. 1,2,3 and then say 3 individuals from different parties 4,5 and 6 or 4 and 5 from a party and an individual 6?
COMMENT: If you vote above the line you are accepting each party’s list of candidates as printed on the ballot paper. If you want to change the order of a party’s candidate, or you want to pick and choose between candidates of different groups, then you have to vote below the line for candidates. You can’t vote for candidates above the line, only parties. If you want to vote for candidates you have to vote below the line.
Thanks Antony,
Question: To cast a formal vote on the Senate paper, ATL the pamphlet in the mail says 6 preferences should be marked whereas your blog implies up to 6. I have probably misinterpreted your blog.
Years ago I was a Returning officer and I was trained that the formality rules were a little broader than electors were told.
Any chance of a clarification to sort out a friendly local discussion?
COMMENT: The instructions are to mark a minimum six preferences above the line, or a minimum 12 preferences below the line.
What happens to your preference if you vote “1” above the line? How is it allocated?
COMMENT: If you vote 1 for a group above the line, it is imputed to be preferences for the candidates of your chosen party in the order they are listed on the ballot paper. As you only marked one square, your vote has no preferences beyond the chosen group. It is best if you mark further squares beyond 1 so that your vote has preferences for other parties. That’s why the instructions suggest a minimum 6 preferences.
I seem to recall a suggestion that voting for only 12 candidates BTL in 2019 was not enough to avoid exhausting your vote too soon. What is your advice for 2022 If you want to vote for your choice of candidates BTL but have no desire to fill in over 100 boxes?
COMMENT: Whether your vote exhausts depends on who you vote for rather than the number of preferences completed. If you give all your preferences to candidate of parties with little chance of election then your vote will exhaust. If you include preferences somewhere in your list for candidates of parties with a chance of electing members, then your vote won’t exhaust. Exhaustion is a function of how long candidates remain in the count and whether you preferenced them, not the number of preferences you complete.
Hi Antony, I’ve noticed that your 2016 table here don’t match the data in Table 2 of this report in the parliamentary library: https://www.aph.gov.au/About_Parliament/Parliamentary_Departments/Parliamentary_Library/pubs/rp/rp1718/SenateVotingSystem#_Toc504654990 (link should jump straight to table). Eg 87% of voters are said to have given 6 ATL preferences in that report, while here you give 81.6%.
Is that likely a difference in data definitions, or an error somewhere?
COMMENT: A combination of being different elections and different definitions. Their numbers are a percentage of ATL votes not all votes and they are for the 2016 election.
Hi Antony – I’m still a little confused as to how preference votes are allocated when using the above the line method. If you place a 1 next to party A (which has 5 candidates below the line), and a 2 next to party B (which has 4 candidates below the line), would that be the same as marking below the line as preference votes 1-5 for Party A and 6-9 for party B in the respective orders they were displayed? Or, has that action given first preference to the first person below the line listed in Party A, and second preference to the first person listed in party B below the line? Thanks for your help.
COMMENT: Your first answer, all the candidates of group then the second group. All votes stay with the first preferences until distributed on exclusion or election of the candidate.
Who and when decided to change from marking 1 above the line to marking 6 above the line.
AND your explanation is saying that a vote is valid even when repeating a number e.g.1,2,2,3,4,5,6,
COMMENT: ‘1’ only above the line voting applied from 1984 to 2014. It worked in conjunction with group voting tickets where all above the line votes had party preference tickets applied. There was so much abuse of group voting tickets at the 2013 election that the Joint Standing Committee on Electoral Matters recommended it be abolished. The Turnbull government adopted most of the recommendations in legislating for the new Senate system ahead of the 2016 Federal election.
The savings provision that any vote with fewer than six above-the-line preferences would be formal was adopted to avoid a rise in informal voting caused by voters continuing to use the pre-2016 rules.