KIEFEL CJ, BELL, KEANE, NETTLE, GORDON AND EDELMAN JJ. Since 1992, the Commonwealth Electoral Act 1918 (Cth) ("the Electoral Act") has required the scrutiny of votes in an election for each Division of the House of Representatives to include an indicative two‑candidate preferred count ("the Indicative TCP Count"). The Indicative TCP Count takes place, in a Division, after counting of first preference votes. It is a "count of preference votes (other than first preference votes) on the ballot papers that, in the opinion of the Australian Electoral Officer, will best provide an indication of the candidate most likely to be elected for the Division".
The plaintiffs were endorsed and nominated by the United Australia Party as candidates in a Division of the House of Representatives or for the Senate in the recent federal election, held on 18 May 2019. Prior to that election, the plaintiffs filed an application for a constitutional or other writ in this Court seeking to challenge the practice of the first defendant, the Australian Electoral Commission ("the Commission"), in making public, while polls remained open in some parts of Australia, one or both of the identity of the candidates selected by the Commission for the purpose of the Indicative TCP Count in a Division ("the TCP Candidates") and the progressive results of any of those indicative counts (collectively, "the TCP Information").
The plaintiffs' complaint was about when the Commission made the TCP Information public. The plaintiffs did not allege that the Commission could not make the TCP Information public after the polls had closed in all States and Territories. The plaintiffs put their case in two ways. First, they submitted that publishing the TCP Information before the polls closed in all parts of the nation was not authorised by the Electoral Act. In particular, they submitted that by publishing that information, the Commission would not be impartial or avoid the appearance of favouring one or more of the candidates. Second, they submitted that by publishing the TCP Information while the polls remained open in any part of the nation, the Commission "would impermissibly distort the voting system in a manner that would compromise the representative nature of a future Parliament" contrary to the constitutional mandate for direct and popular choice in ss 7 and 24 of the Constitution.
The plaintiffs' application for a constitutional or other writ, which was referred to a Full Court for hearing, proceeded on a statement of agreed facts. The defendants − the Commission, the Electoral Commissioner and the Australian Electoral Officers for each State, the Northern Territory and the Australian Capital Territory − filed submitting appearances. The Attorney‑General of the Commonwealth intervened and made written and oral submissions.
At the conclusion of the hearing before the Full Court on 7 May 2019, the plaintiffs' application was dismissed with reasons to be published at a later date. These are our reasons for joining in that order.
Publication of the TCP Information relating to a Division after the polls in that Division had closed, but before the polls had closed elsewhere in the nation, is authorised by s 7(3) of the Electoral Act. The factual foundation for the plaintiffs' statutory challenge was not established - for example, it was not shown that publication suggested "imprimatur" or appeared to favour one or more of the candidates. Moreover, there was no factual foundation for the contention that the publication of the TCP Information after the polls in a Division had closed, but before the polls had closed elsewhere in the nation, had any effect on the requirement for direct and popular choice in ss 7 and 24 of the Constitution.
There were three central difficulties with the plaintiffs' case: first, there were no facts showing that publication of the TCP Information had any effect on the electoral choices of voters in Divisions where the polls had not closed; second, the plaintiffs did not say that the Commission could not publish the results of the counting of the first preference votes after the polls had closed, even if other polls across the nation remained open; and, third, the Indicative TCP Count was based on votes cast and was a prediction of the candidate most likely to be elected - not an expression of any opinion by the Commission about whether that prediction pointed to a desirable or undesirable outcome.
The Constitution and the Electoral Act
Subject to the express and implied limitations with respect to federal elections, it is the Constitution that provides Parliament with the power, and responsibility, for establishing an electoral system which balances "the competing considerations relevant to the making of a free, informed, peaceful, efficient and prompt choice by the people".
It is the Electoral Act that gives effect to the electoral system chosen by the Parliament, in the exercise of that legislative power, with respect to federal elections. The Electoral Act establishes the Commission, the stated functions of which, relevantly, are:
"(1) ...
(a) to perform functions that are permitted or required to be performed by or under [the Electoral Act], not being functions that:
(i) a specified person or body, or the holder of a specified office, is expressly permitted or required to perform; ... and
...
(c) to promote public awareness of election and ballot matters, and Parliamentary matters, by means of the conduct of education and information programs and by other means; and
...
(f) to publish material on matters that relate to its functions".
Section 7(3) provides that "[t]he Commission may do all things necessary or convenient to be done for or in connection with the performance of its functions".
The Electoral Act, among other things, provides for Divisions for the House of Representatives; the roll of electors for each State and Territory and each Division; the qualifications and disqualifications for enrolment and voting; the registration of political parties; the procedures by which candidates are nominated and required qualifications; the regulation of polling and compulsory voting; the appointment and role of scrutineers appointed by candidates; and the scrutiny of votes in an election.
This matter is concerned with the scrutiny of votes in an election for a Division of the House of Representatives. The scrutiny is conducted in the manner set out in s 274.
First preference votes are counted and then, after that count is recorded, that information is transmitted in an expeditious manner to the Divisional Returning Officer for the Division. It was an agreed fact not only that the progressive results of the first preference counts are released by the Commission and the scrutineers appointed by the candidates, but also that the publication of the progressive first preference counts has the capacity to affect electoral choices. The plaintiffs did not suggest that the release or publication of the first preference count was not authorised by the Electoral Act or that it infringed the Constitution.
If, in an election for the House of Representatives, there are more than two candidates for a Division, an Indicative TCP Count is conducted. That count is addressed in s 274(2A)‑(2C) of the Electoral Act. The process starts well before polling day. Given the limited nature of the plaintiffs' complaint, it is not necessary to describe the process of the Indicative TCP Count in elaborate detail. The ultimate purpose of the process is described in s 274(2A) as requiring returning officers "to conduct a count of preference votes (other than first preference votes) ... that, in the opinion of the Australian Electoral Officer, will best provide an indication of the candidate most likely to be elected for the Division" (emphasis added).
Section 274(2A)-(2C) does not prescribe the procedure by which the Australian Electoral Officer is to reach their opinion as to the identity of the TCP Candidates or how the Indicative TCP Count is to be conducted. The Commission's established practice for preparing for, and carrying out, the Indicative TCP Count in each Division comprised part of the agreed facts and is summarised below.
Commission's established practice
After nominations close, and before the polling day, the Commission's Election Management System selects a default set of TCP Candidates, based on the results of the previous election. The Australian Electoral Officer for each State and Territory reviews the default TCP Candidates for each Division in their State or Territory and either endorses the default selection or advises of any changes that should be made. Once the Australian Electoral Officer has identified the TCP Candidates, they are submitted to the National Election Manager for endorsement, who informs the Deputy Electoral Commissioner and the Electoral Commissioner as required. The Commission's identification of the TCP Candidates is finalised in the second week before polling day to enable the information to be entered into the Commission's Election Management System. Further changes cannot be made prior to election night.
Once the TCP Candidates have been entered into the Election Management System, a written direction with the names of the TCP Candidates in a Division is produced to the Assistant Returning Officer in that Division and placed in a sealed envelope. In the lead up to election night, Australian Electoral Officers actively monitor potential issues to be alert to potential deviations from the identified TCP Candidates.
At each polling place in every Division, after the close of polling (at 6.00 pm on polling day in the applicable time zone) and in view of the scrutineers, the Assistant Returning Officer opens the sealed envelope containing the names of the TCP Candidates and announces those names to everyone present (including the scrutineers).
Then, after the first preference count, the ballot papers for the TCP Candidates for that Division are removed to a secure area and the ballot papers for the remaining candidates are notionally allocated to the TCP Candidate for whom a higher preference has been expressed on the ballot paper. The result of the allocation of ballot papers to the TCP Candidates - the Indicative TCP Count - for each polling place is telephoned to the Divisional Returning Officer for the Division. The Divisional Returning Officer progressively enters the Indicative TCP Count for each polling place into the Commission's Election Management System. The Election Management System updates a webpage on the Commission's website, "The Tally Room".
From the close of polls in the principal time zone of a Division, or shortly thereafter, the Tally Room displays the identity of the TCP Candidates in respect of whom the Indicative TCP Count will be undertaken and, once information about the Indicative TCP Count for each polling place begins to be received into the Election Management System, the Commission uses a "matched polling place method" to calculate a "matched polling place projection" in each Division.
The matched polling place projection is a prediction of the Indicative TCP Count in the Division, expressed as a "swing" in the Division since the last election. It is calculated by comparing the Indicative TCP Count in a polling place with the votes received at the same polling place in the previous election. The Commission explained, in a publication regarding the electoral process, that the "matched polling place method relies on the empirical fact that swings to or from political parties or candidates tend not to vary greatly within electoral divisions" - that is, "swings within divisions tend to be consistent across polling places in a particular electoral division". The matched polling place projection is progressively updated as information from the Indicative TCP Count becomes available from each polling place. From approximately 2.00 am AEST the following day, the results of the Indicative TCP Count are displayed on the Tally Room.
There are two situations where the Commission adopts a different approach. The first is where one of the identified TCP Candidates for the Division did not contest the previous election or is endorsed by a party which did not endorse a candidate in that Division in the previous election. In that situation, the Tally Room displays the results of the Indicative TCP Count on a progressive basis, rather than a matched polling place projection.
The second is where, on election night, it becomes apparent that one or both of the TCP Candidates in a Division should no longer be included in the Indicative TCP Count - for example, because a candidate polling first or second on first preferences was not one of the TCP Candidates identified by the Commission. In that situation, the matched polling place projection and the results of the Indicative TCP Count in that Division are masked from public view on the Tally Room ("TCP Exception") to avoid incorrect consideration of the eventual winner. A TCP Exception does not stop the counting or entry of the results of the Indicative TCP Count. However, because changes to TCP Candidates can only occur from the Sunday after polling day, the TCP Exception remains in place until the new TCP Candidates are recorded in the Election Management System.
After the polls close, the Commission also maintains a "real‑time" media feed that provides to media organisations and interested third parties the identity of the TCP Candidates in each Division as well as the Indicative TCP Count for each Division, updated by the Commission every 90 seconds on election night.
Thus, after the polls close in each Division, the TCP Information is made available in different ways to different people and organisations. The TCP Candidates are identified to scrutineers in a Division, to the public via the Tally Room and to media organisations and interested third parties via the media feed. The progressive results of the Indicative TCP Count in a Division are provided to the scrutineers, and to media organisations and interested third parties via the media feed. The matched polling place projection for a Division or the progressive results of the Indicative TCP Count in a Division are published to the public via the Tally Room.
Before saying something more about the plaintiffs' complaints, it is necessary to record why s 274 of the Electoral Act was amended to include the Indicative TCP Count.
Legislative history
In the 1990 election, there was a delay of some days in determining election results despite the fact that the government was returned with a majority of eight seats. As a result, the Joint Standing Committee on Electoral Matters ("the JSCEM") in 1990 recommended the addition of "a new step to the House of Representatives scrutiny process to guarantee that scrutineers would have the opportunity to readily observe a 'two-candidate preferred vote' in each polling place on election night".
After the amendment to give effect to that recommendation was introduced in the Senate in 1992, the JSCEM in 1992 expressed two concerns about the Commission's original method of identifying the TCP Candidates, namely, the possibility of the Commission "getting it wrong and jeopardising an early result on election night" and "the effect on the electoral system of two candidates appearing to be the 'two most likely' in the judgment" of the Commission. The JSCEM in 1992 identified two measures to address those concerns. First, to maximise the chances of identifying the correct TCP Candidates, the Commission would take into account all relevant objective data including, but not limited to, historical performance. Second, the Commission would keep confidential the identity of the TCP Candidates until the close of polls. It was not in dispute that both of these recommendations are given effect in the Commission's established practice for selecting the TCP Candidates and conducting the Indicative TCP Count.
When s 274(2A) was introduced into the Electoral Act, the new procedure was intended to "provide the public on [election] night with an early indication of the two-candidate preferred result in most electorates". That reflected the recommendations of the JSCEM in 1992 that it was "highly desirable that the public and candidates know the result of the count as it becomes available" and, thus, that not only should the results of the first preference count be transmitted "immediately" but "[t]he result of the provisional two-candidate preferred distribution should be transmitted as soon as possible from each polling place, and transmitted in at least three batches from the Divisional Office to the National Tally Room". Again, it was not in dispute that both of these recommendations are given effect in the Commission's established practice for conducting the count of the first preference votes and the Indicative TCP Count.
Lack of factual foundation
The plaintiffs' contentions about the Indicative TCP Count process, which underpinned both their statutory and constitutional challenges, lacked a factual foundation. It is convenient to address each in turn.
No demonstrated effect on electoral choices
It was an agreed fact that there was no practicable means, in the time available before the 2019 election, to quantify the extent or likelihood of the effect, if any, on the electoral choices of voters who become aware of the identity of the TCP Candidates in any Division, or the results of the matched polling place projection or the Indicative TCP Count in any Division, if the voters had not already voted and when polls were still open in the Division in which they were electors.
Instead, the plaintiffs referred to and relied on three published articles, each of which concerned overseas jurisdictions - the United States of America, France and Denmark. The plaintiffs submitted that the studies recorded in these articles supported a proposition that voters who were yet to cast their ballot may be influenced by the release of election results elsewhere or by the publication of opinion polls or exit polls while voting booths remain open, sometimes referred to as the "bandwagon effect". Taken as a whole, the articles do not assist the plaintiffs.
The studies recorded in two of those articles related to electoral systems markedly different from the system of compulsory preferential voting, and parliamentary elections, prescribed by the Electoral Act in Australia, with the systems under consideration including such features as non‑compulsory voting, national presidential elections, "sequential" voting over a period of weeks and months, and voting on weekdays. The third article considered the effect of election polls on party support in the abstract.
The studies did not address the effect, if any, of an action similar to that of the publication of the TCP Information. The studies did not address a situation where the TCP Information relates to candidates in a given Division, and is released after the close of polls in that Division, but the claimed effect on voting is said to occur in other Divisions, where the polls remain open.
Thus, the studies do not establish that the action of the Commission in making public the result of a predictive exercise intended to assist with counting, or even a count based on that exercise, had an effect similar to, for example, opinion or exit polls on voters who had not yet voted. Even then, the views expressed by the authors of these studies do not all point towards conclusions of the kind the plaintiffs put as the foundation for their arguments. In particular, it was said in one study that no significant effect on turnout was identifiable in parliamentary elections, as opposed to presidential elections.
Thus, there was nothing to support any finding that the publication of the TCP Information by the Commission prior to the polls closing across the nation distorted the voting system in any relevant way.
Not sufficient that voting "may be affected"
More loosely, the plaintiffs contended that it was sufficient that voting "may be affected" by release of the TCP Information before the close of polls nationally to render the publication invalid. Not only was there no factual foundation for such a contention, but the parties agreed that other sources of information have the capacity to affect the electoral choices of voters, including the progressive first preference counts for candidates released by the Commission and scrutineers; the results of opinion polls published in close proximity to the election; other media reporting on the election, including reporting of the results of exit polls; the presence of party officials and/or candidates near polling booths; the offer of, and prevalence of, "how to vote" cards at polling booths; and the number of parties and candidates listed on the ballot for Senate elections in that State or Territory, and the order in which those parties and candidates are listed on the ballot paper. How, if at all, publication of the TCP Information about one Division might interact with these matters, which it was agreed may affect choices of voters in another Division, was not explained.
Selection of TCP Candidates not inaccurate or misleading
The plaintiffs also sought to rely on the idea that the selection of candidates for the Indicative TCP Count was inaccurate or misleading. Again, there was no factual foundation for that contention.
The process adopted by the Commission has been addressed. It is a predictive exercise and the identification of the TCP Candidates is generally accurate. Indeed, the matter proceeded on agreed facts that in relation to the Commission's identification of the TCP Candidates in the two preceding federal elections, only two Divisions in the 2013 election, and none in the 2016 election, did not include the eventual winner. Moreover, it must be recalled that any imprecision in the identification of the TCP Candidates is addressed by masking the matched polling place projection and the results of the Indicative TCP Count from public view on the Tally Room when it appears that one or both of the TCP Candidates in a Division should no longer be included. And after the polls have closed in that Division and the TCP Candidates are announced, that masking can occur at any time.
No imprimatur or partiality in publication of TCP Information
Contrary to the plaintiffs' submission, publication of the TCP Information does not constitute the Commission giving any imprimatur to any particular candidate or outcome. As a matter of statutory construction, the Indicative TCP Count is a prediction after the close of polls. The TCP Candidates remain secret in those Divisions where polls have not yet closed. It is only after the close of polls in a Division that the prediction is made public. And the progressive results that are published by the Commission are based on the votes cast and counted. Publication does not constitute any expression of opinion by the Commission about the desirability of the results that are published. It is not an expression by the Commission of any opinion favouring one candidate over another and, thus, is not a form of partiality.
Indeed, the Commission is alert to the need to avoid any appearance of partiality: it explained as early as 1996 that the identification of the TCP Candidates is not made public prior to polling day "so as to ensure that the [Commission] is not seen to be giving any public endorsement to the perceived popularity of any candidates contesting the election".
Statutory challenge
As the factual foundation for the plaintiffs' statutory challenge has not been established, it remains necessary only to address the contention that the release of the TCP Information before close of polls nationally is not authorised by the Electoral Act.
The Indicative TCP Count is a predictive statutory exercise that will best provide an indication of the candidate most likely to be elected for a Division in the House of Representatives. The Attorney-General of the Commonwealth submitted that publication of the TCP Information was authorised under s 274(2A). It may be accepted that that sub-section, with its reference to "best provide an indication", recognises that the TCP Information can be, even should be, published. But, be that as it may, the functions given by s 7 of the Electoral Act to the Commission to "promote public awareness of election and ballot matters" and to "publish material on matters that relate to its functions" are functions that include publishing election results which, as s 274(2A) provides, include both the first preference count and the TCP Information.
Section 7(3) then gives the Commission power to do all things necessary or convenient to be done for or in connection with the performance of those and its other functions. The language of s 7(3) is broad and it is for the Commission, in its discretion, to determine how that power is exercised.
The plaintiffs did not say that the Commission could not publish the results of the progressive first preference count in a Division after the polls in that Division closed even if other polls were still open. That is unsurprising: the Commission's power to publish the first preference count under s 7(3) existed before s 274(2A)-(2C) was enacted.
The plaintiffs did not say that the Commission could not publish the TCP Information. The plaintiffs did not dispute that it was necessary or convenient for the Commission to publish the TCP Information. The plaintiffs' complaint was limited to the timing and mode of publication of the TCP Information in circumstances where there were no facts showing that publication of the TCP Information had any effect on voters in other Divisions where the polls had not closed.
Having regard to the statutory framework and the legislative history, s 7(3) empowered the Commission to publish the TCP Information as soon as the polls closed in a Division. It was open to the Commission to decide that publication of that information was both necessary and convenient for or in connection with the performance of the Commission's functions to "promote public awareness of election and ballot matters" and to "publish material on matters that relate to its functions".
That conclusion is consistent with both the 1990 and 1992 reports of the JSCEM and the extrinsic materials which record that the purpose of the Indicative TCP Count was to provide the public with an early indication of who was most likely to be the elected candidate. There would be no point to the Indicative TCP Count if the count was not made public until the actual count was completed.
In any democratic vote, electors would expect to have information about the vote made publicly available, in the interests of transparency, in order to have confidence in the maintenance of the electoral system chosen by the Parliament in the exercise of its legislative power with respect to federal elections, and to achieve "promptitude, certainty and finality in the declaration of the poll".
Thus, the Commission has power under the Electoral Act to publish the TCP Information. Once the Commission has the power, the issue of how it is preferable for that power to be exercised is not a matter for the Court. It is possible that questions as to the limits to that power might arise, but they do not in this case.
As the plaintiffs have not established their factual contention that publication of the TCP Information conveys "imprimatur", it is unnecessary to address the plaintiffs' written submission that the Commission's conduct contravened an implied statutory limitation against partiality. A deviation from a statutorily prescribed course might say something about the validity of the exercise of a statutory power, but that does not arise here.
Constitutional challenge
What has been said earlier about the factual bases on which the case proceeded is reason enough to reject the plaintiffs' arguments that publication of the TCP Information in relation to a Division, after the polls in that Division had closed but before the polls had closed throughout the nation, is unconstitutional.
There is no factual foundation for the plaintiffs' contention that the publication of the TCP Information, while the polls remained open in any part of the nation, "would impermissibly distort the voting system in a manner that would compromise the representative nature of a future Parliament", contrary to the constitutional mandate for direct and popular choice in ss 7 and 24 of the Constitution.
In particular, the agreed fact that there was no practicable means, in the time available before the 2019 election, to quantify the extent or likelihood of the effect, if any, on the electoral choices of voters who became aware of the identity of the TCP Candidates in any Division or the results of the matched polling place projection or the Indicative TCP Count in any Division - if they had not already voted and when polls were still open in the Division in which they were electors - means that there is no basis for the plaintiffs' complaint.
Moreover, the Indicative TCP Count and the publication of the TCP Information prior to the polls closing across the nation do not imply the Commission's support or imprimatur for the particular TCP Candidates that are selected. As a result, the question of whether it would be constitutionally problematic for a government agency to endorse or support particular candidates for election does not arise in this case.
Costs
The plaintiffs should pay the defendants' costs, being the costs that the defendants incurred up to and including the date that they filed their submitting appearance.