Should How-To-Votes be Banned at Australian Elections?

Australian is unusual among western democracies in permitting active campaigning outside polling places on election day.

Despite sharing much electoral heritage with Australia, New Zealand sits at the opposite end of the election day campaigning spectrum. Not only are all forms of election day campaigning banned, but all signs erected in the campaign must be removed before election day.

On regulating election day campaigning, most countries sit nearer New Zealand than Australia. Election day is viewed as a time for considered contemplation by voters, not as an opportunity for boisterous last day campaigning.

Australia also has some of the world’s most complex methods for completing ballot papers. Almost all countries use a single cross to vote, some a second cross, and some have a limited form of preferential voting. Some countries, notably the USA, complicate simplicity by holding multiple elections on the same day.

Only Australia requires voters to complete a sequence of numbers for every square on the ballot paper under rules with no allowance for error.

And no other country compels voters to engage with such a complex voting system on pain of being fined.

Thanks to compulsory voting, polling places are the last chance candidates have to engage with undecided and disinterested voters who in other countries probably wouldn’t turn up to vote.

And thanks to full preferential voting, candidates and parties that have attracted a primary vote have enormous interest in ensuring that voters correctly number all other squares to complete a formal vote.

In the Northern Territory, so vigorous had polling day campaigning become that the term “running the gauntlet” was coined for the challenge voters faced to enter a polling place.  The Territory is also where polling place signage reached its apogee, fences around schools used as polling places being wrapped in continuous roll campaign paraphernalia.

The tactic has since been copied down south, as the example below of Trent Zimmerman’s campaign posters at the 2015 North Sydney by-election shows.

One of the problems with such blanket signage is that it generates disputes. Workers for other candidates and parties turn up later and get into arguments about having the right to display their own signs.

There is clearly a problem when party workers engage in over-vigorous polling place campaigning. Voters should never feel intimated when turning up to do their civic duty and legal obligation.

Some states have moved to limit the size and quantity of signage outside polling places, and even limit what activist can wear. Other jurisdictions require registration of how to vote material. Voters interviewed on why they use pre-poll voting mention the intensity of campaign activity on polling day as one of their reasons.

Tasmania bans any polling day campaigning and regulates the pre-election distribution of how-to-votes and preference recommendations. The ACT bans campaigning within 100 metres of a polling place. There are special reasons why Tasmania and the ACT do this, which I will return to at the end of this post.

Experience with “the gauntlet” has pushed the Northern Territory to move further down the path of regulating polling place campaigning. The NT Electoral Act permits how-to-votes to be distributed from a designated area within the regulated zone of a polling place, with the guiding principle that voters are free to approach campaign workers, but not the other way around.

The NT’s move to restricting how-to-vote distribution in 2016 coincided with the introduction of optional preferential voting. Optional preferences delivered a record low informal rate in 2016. It will be interesting to see how restrictions on how-to-votes and the return of full preferential voting interact in 2020.

Regulating conduct outside polling places is a good idea, but I don’t agree with blanket bans on how-to-votes. With compulsory voting and full preferential voting, banning how-to-votes can only increase the level of informal voting.

Many who call for bans use ending preference “deals” as justification, without understanding that influencing preference is not the main purpose of how-to-votes.

One of the original purposes of how-to-votes was to let voters know the party affiliation of candidates, something that became less relevant after 1984 when party names were first printed on Commonwealth ballot papers.

Yet how-to-votes remain important as a final appeal by candidates for a first preference vote, and as a guide on completing a formal vote.

The example below shows the daunting prospect that once faced voters on giant Senate ballot papers with no party names and enforced full preferential voting.  (ALP how-to-vote, NSW Senate election, 1974)

How-to-vote cards are a consequence of compulsory voting and full preferential voting. Only Australian jurisdictions with optional or limited preferential voting have banned or restricted the distribution of how-to-vote material.

Like it or not, if voters are required to number every square for a formal vote, then candidates and parties are going to offer a guide on casting a formal vote.

But Always Remember, it is Voters that Determine Preferences, not Candidates and Parties

A constant mistake made in referring to preferences on how-to-votes is to say a candidate or party is “directing” preferences.

A candidate or party may “recommend” preferences on a how to vote, but they cannot direct preferences. Parties and candidates cannot override the preferences that a voter writes on their ballot paper.

I myself admit to the sin of occasionally saying “directed” when I mean “recommended”.

Recommendations are meaningless unless a voter is given a candidate’s how-to-vote. The more how-to-votes a candidate distributes, the more voters might choose to follow it. Voters that don’t receive a how-to-vote can only match a candidate’s preference sequence by deduction or chance. The fewer the candidates on a ballot paper, and the clearer the ideological choices, the greater the chance that a voter can guess a candidate’s recommended sequence without a how-to-vote.

For all the talk of what proportion of preferences flow from one party to another, there is only one thing a voter needs to know about preferences. 100% of the preferences for a voter’s ballot paper will flow where the voter sends them. When preferences are being distributed, a candidate cannot override what the voter has written on their ballot paper.

So if preference “deals” are not the point of how-to-votes, what are they for? Let me set out the reasons in descending order of importance.

Reason 1 – Vote for Me

How-to-votes are distributed to attract first preference votes. The vast majority of how-to-votes distributed are about first preferences, not further preferences.

Candidates want you to vote for them, to give them your first preference vote. Even candidates who don’t expect to win, who are more interested in influencing your further preferences, still want your first preference.

Thanks to compulsory voting, a small proportion of voters turn up to vote with no idea who the candidate are, though even the most disinterested usually know the party names.

Party names did not appear on national ballot papers until 1984. Before then, how-to-votes were crucial in letting voters know which party a candidate represented.

Giving voters a how-to-vote highlights the candidate’s name, their party, their party leader, and maybe a policy or two. It is the best way of attracting the attention of an undecided or disinterested voter.

For major party candidates, most likely to finish in the final two competing for victory, the overwhelming reason to distribute how-to-votes is to attract first preference votes.

But candidates also want to make sure their voters number all the other boxes on the ballot paper correctly. (This is less important under optional preferential voting.) Which leads  to the second major reason for distributing how-to-votes.

Reason 2 – Make Your Vote Count

If a candidate or party attracts a voter’s first preference vote, it is then hugely in the candidate’s interest that the voter’s ballot paper is numbered correctly.

Not because those preferences will ever be counted, but rather because if the voter doesn’t complete the sequence of preferences correctly, the vote is informal. There is no point attracting the first preference of a voter if the voter then fails to complete a formal vote by numbering all the squares.

To meet the aims of Reason 2, parties like to publish how-to-votes with obvious sequences. Straight line sequences are preferred to sequences that jumble the numbers.  We don’t know whether voters transcribe numbers from a how-to-vote top to bottom or whether they search out a sequence and number the boxes in order. But the view is that if voters are given an ordered sequence on a how-to-vote, they are more likely to successfully transcribe those numbers on to a ballot paper.

(See Reason 4 below as to why preference sequences on how-to-votes have become more complex in recent years.)

Reasons 1 and 2 are overwhelmingly why major parties distribute how-to-votes.

Reasons 1 and 2 are also important for minor parties. But for candidates that expect to be excluded in the distribution of preferences, their recommended ordering of further preferences can be the main point of campaigning.

Reason 3 – Follow my Preferences

For a candidate likely to be excluded in the distribution of preferences, their influence on the contest is determined not by their first preference vote, but by how their preferences flow on exclusion.

The importance of preferences is often over-stated. As many as two-thirds of Australian electorates require preferences to be distributed to determine the winning candidate, but fewer than one in ten see the leading candidate on primary votes defeated after the distribution of preferences.

Research by the Victorian and South Australian Electoral Commissions has shown that up to 45% of Labor, Liberal and National Party first preference voters follow how-to-votes, while the rates for minor party voters is much lower. As major parties hand out more how-to-votes than minor parties, the above figures will reflect how many voters received how-to-votes as well as what proportion of voters wish to follow them.

Politically, the destination of a preference is more important than the route by which the destination is reached. Green voters have low rates of following how-to-votes, but 80% of Green preferences reach Labor ahead of the Coalition, even if they a follow different path. For instance, the proportion of  Green voters following a how-to-vote diminishes whenever a candidate other than the Labor candidate is given second preference, without effecting the overall preference flow to Labor.

The first minor party that drew influence from preferences was the Democratic Labor Party (DLP). It split from the Labor Party in the mid-1950s over the issue of Communism in the unions and Labor Party. Until 1972 the DLP was important for using its preferences re-elect the anti-Communist Coalition.

The DLP was also the first minor party to benefit from the introduction of Senate proportional representation in 1949. Contesting lower house seats and handing out how-to-votes had the double goal of influencing House preferences and attracting votes in the Senate.

The Australian Democrats were the most important minor party in the last quarter of the twentieth century, but mainly through their place in the Senate. At lower house elections the Democrats issued split-ticket or double-sided how-to-votes, voters making their own choice on preferences. At the height of the Democrat’s influence, there were fewer seats where preferences changed the first preference winner. Unlike the DLP or Greens, the Democrats were a centrist party and the party’s supporters displayed less ideological delineation in preferences flows.

In 1990, the Labor Party used television ads to directly appeal to Democrat voters for preferences, resulting in strong flows of Democrat preferences to Labor.  Many of these may have come from ex-Labor voters, shifting their first preference to the Democrats but not ready to cross over and vote for the Coalition. Democrat preferences leaned towards Labor more strongly after 1990 than before.

Since the 1990s, the Greens have had a significant impact on election outcomes, delivering victory to as many as a dozen trailing Labor candidates  at some elections. The overall proportion of Green voters who preference Labor often exceeds 80%. Generally, the stronger the Green first preference vote, the stronger the flow of preferences to Labor.

Labor preferences flow even more strongly to the Greens in the rare cases where distributed, and strong flows of Labor and Green preferences have played a part in electing Independents in safe Liberal and National Party seats.

Until the 2019 Federal election, Pauline Hanson’s One Nation rarely produced strong flows of preferences. The highest vote for One Nation often occurred at elections where there was a significant decline in Coalition support. Tactics such as recommending preferences against sitting members tended to hurt the Coalition more than Labor.

That trend reversed in 2019 with One Nation preferences flowing 65% to the Coalition, much higher than at previous elections. United Australia preferences also flowed around 65% to the Coalition, much higher than the even split for the Palmer United Party in 2013.

In my opinion, the higher One Nation and United Australia preferences in 2019 may stem from the nature of the electoral contest. In the past One Nation raged against both parties, where in 2019 Bill Shorten and Labor were more clearly the enemy, a perception added to by Clive Palmer’s enormous advertising spend attacking Labor. That both parties polled more strongly in Queensland where the anti-Labor mood was strongest suggests a strong ideological and issues specific source to the preference flows. It is hard to know whether how-to-votes played a part.

Many minor parties have no influence on preferences as they don’t distribute how-to-votes, resulting in random or evenly divided preferences. Some very small parties, such as the Australian Christians and Christian Democrats, deliver strong flows of preferences to the Coalition despite the lack of how-to-votes. Sometimes it  is the ideology of a party and its supporters that determines preferences flows rather than suggestions on how-to-votes.

At the Northern Territory’s Johnston by-election in February 2020, the Greens chose to send a message to the Gunner Labor government on fracking policy by recommending preferences to the Territory Alliance ahead of Labor. 56.9% of Green voters still gave preferences to Labor, but this was well down on the normal flows.

Contrast this with the Liberal Party’s swap from recommending preference for Labor over the Greens since the 2013 Federal election. Where previously Liberal preferences flowed two-thirds to the Greens, afterwards they flowed two-thirds to Labor. Liberal voters seemed more amenable to following suggested how to votes than smaller and ideologically focused parties like the Greens. Note that in only a small number of seats are Liberal preferences distributed.

For all the talk of preference “deals”, the preferences of most minor parties are recommended and flow in obvious directions. How-to-votes almost always recommend preferences to parties that voters expect to receive the recommendation. Parties that try to recommend preferences tactically rather than ideologically, such as Green preferences in Johnston, will only weaken the flows rather than reverse them.

It is optional preferential voting, as used at NSW elections, that has been the bigger threat to Green preference flows to Labor, a fact that played a part in the Queensland Labor government’s decision to bring back full preferential voting in 2016.

And despite all the focus on minor party preferences, the key point to repeat is that most how-to-votes are distributed by major parties whose preferences aren’t distributed. And those parties’ major interests are Reasons One and Two listed earlier, not about influencing preferences flows.

Reason 4 – Virtue Signalling, Who We Hate Most

When ballot papers had at most only only four or five candidate, organising a how-to-vote was easy. Put your candidate first, your major opponent last, and then come up with a sensible sequence for everyone in between.

That has become more difficult in recent years as the number of candidates has increased, and as how-to-votes have been increasingly examined by opposing parties and the media for the signals they send on who a candidate or party find most objectionable.

Decades ago all parties used to put the Communists last, but the fewer candidates on ballot papers in those days prevented how-to-votes getting too complex.

The 1998 election was the modern turning point for how-to-votes as virtue signals. Labor and the Coalition both went out of the way to put Pauline Hanson’s One Nation last. Coalition resolve on that unity ticket has waned in recent years for a variety of reasons.

Trying to decide who to put last can seriously complicate the sequence of preferences displayed on a how-to-vote. And complex sequences are in conflict with Reason 2 listed earlier, insuring votes are formal.

In 2013, Federal Labor MP Dr Andrew Leigh (Fenner, ACT) was criticised for the design of his how-to-vote. Dr Leigh is one of the few people who has done research on how-to-votes, and certain in the knowledge his preferences would never be distributed, he produce the simplest possible how-to-vote sequence that ended up listing an anti-Islamic candidate higher than the Greens.

In a fit of high dudgeon the Canberra Times attacked Dr Leigh. It is instructive to read his defence.

In a letter to The Canberra Times on Tuesday, Mr Leigh said the major parties’ preferences in Fraser were “irrelevant”.

Mr Leigh said Labor preferences in Fraser had never been distributed since the seat’s establishment in 1974, and his how-to-vote card was aimed at making it easier for Labor supporters to lodge a valid ballot paper.

His own preference would be to “put the Greens above the Liberals, above any party with a racist agenda”, but said instead he opted for the “simplest” numbering on his how-to-vote cards to reduce confusion.

His instructions are to number the candidates in the order they appear on the ballot paper, beginning with a “1” next to himself, then continuing down the page, before returning to the top of the list to number the remaining candidates.

“The priority for us was formality, to minimise informality rather than to worry about the extremely small chance that Labor preferences would be distributed for the first time in the seat’s history,” Mr Leigh said. “This isn’t the Senate, where there is a good chance that preferences get distributed. This is a house seat, where the race is always between Labor and the Liberals. And Labor and Liberal preferences have never been distributed.

“My concern is that Labor supporters put a one next to my name and then number the other boxes any way they want to, and they can do so in the knowledge that those preferences are extremely unlikely to be distributed.”

Mr Leigh said his record showed that he supported multiculturalism in Australia. He said it would be “disappointing” and “inaccurate” for voters or other candidates to associate his views with those of the Rise Up Australia party.

When asked how he would feel should his preferences come to be distributed to the Rise Up Australia candidate – however small the chance – Mr Leigh said he wasn’t even entertaining the thought.

“I simply think it is near impossible that the Labor preferences would be distributed in Fraser,” he said.

(Canberra Times, 4 September 2013)

Dr Leigh was entirely right, but he met resistance for having placed Reason 2 (ensuring formality) above the evolving Reason 4 (virtue signalling.)

And sometimes the “My preferences will never be distributed” defence fails. When the Queensland Liberal and National Parties decided to preference One Nation ahead of Labor at the 1998 state election, National Senator Bill O’Chee dismissed criticism by arguing Liberal and National preferences would never be distributed.

They were distributed in 22 of the 89 seats, helping to elect 11 One Nation members. A few months later at the Federal election, the Coalition put One Nation last everywhere.

Reason 5 – Just Vote 1 / Number all the Squares / Put party X last

Party how-to-votes have always been about influencing electors to vote for a party (Reason 1), how to cast a formal vote (Reason 2), or how to direct preferences (Reason 3).

What none of these do is send a message to voters who aren’t voting for the party. What can a party do with a message and a how-to-vote to influence voters for other parties in a manner that helps your party. How can a how-to-vote diminish the chances of another party by influencing the choices of voters for that party?

This method has evolved in NSW and Queensland under optional preferential voting. It is a tactic designed to influence voters on how few or how many preferences to complete.

It was perfected by Peter Beattie and the Queensland Labor Party at the 2001 state election. With the non-Labor vote split between the Liberal Party, National Party, One Nation and the (ex-One Nation) City-Country Alliance, the Labor Party switched from recommending full preferences on how-to-votes to advocating “Just Vote 1”. It was a slogan used everywhere, on how-to-votes and in big letters on signs,and  sometimes on signs that were dressed up to look like electoral commission signs.

The result was a surge in non-Labor voters who just voted one. At an election that became a Labor landslide, the decline in non-Labor voters giving preferences delivered extra seats to the Labor Party, winning several seats from unusually low levels of first preference support.

The tactic may have been a success for Labor in 2001 and 2004, but by the end of the decade it had become a problem as support for the Greens rose. Now it was Labor advocating preferences, and the new LNP (partly formed to by-pass optional preferential voting) advocating “Just Vote 1”. The collapse in the flow of Greens preferences to Labor made the Bligh government’s defeat in 2012 all the greater.

By the 2015 election, new Premier Campbell Newman had generated considerable voter antipathy. Labor and allies perfected a different message to third party voters. The LNP again ran with “Just Vote 1”, but there were signs everywhere saying “Number all the squares” (often in a very friendly Australian Electoral Commission purple) and “Put the LNP Last”.

The proportion of third party preferences exhausting nosedived, and the proportion flowing to Labor rose dramatically. Labor won nine seats from second place on preferences in 2015, a very high number for optional preferential voting, and preferences from other conservative minor parties flowed to Labor.

Labor’s “Put the LNP last” message was so strong that Labor preferences almost elected Pauline Hanson in Lockyer despite Labor’s anti-Hanson how-to-vote.

The Queensland tactics have been repeated in NSW, the Coalition putting out signs saying “Just Vote 1”, Labor advocating “Number all the squares”, and “put Labor last ” or “put the Coalition last” becoming more common.

And the “Put Labor/Liberal/National Last” has begun to appear more commonly at elections conducted using full preferential voting. If the Coalition appeals to third party voters to “put Labor last”, it of course means any preference on the ballot paper would reach the Coalition candidate first. It is away of capturing preferences without appealing for them.

“Number all the squares” signs in Chinese and coloured in AEC purple were prominent at the 2017 Bennelong by-election. They were authorised by the Labor Party, though in this case more about minimising the informal vote than trying to influence the vote. It is more an example of Reason 2, maximising the chance of your voters casting a formal vote.

Back to Tasmania and the ACT

As noted earlier, Tasmania bans any campaigning on election day and the ACT limits the distribution of campaign material within 100 metres of a polling centre.

Both states use the multi-member Hare-Clark electoral system. Both states use limited preferential voting, with ballot paper instructions saying voters should show five preferences, the number of vacancies to be filled.

Recommendations of preferences is not banned, but they are restricted by candidates having to agree to being listed in an order on a how-to-vote. The practicality of this is that how-to-votes do not show between party preferences, and very few show within party preferences.

The Labor and Liberal Parties do not recommend within party preferences. An attempt by the Tasmanian Labor Party to impose a how-to-vote in 1980 led to the introduction of “Robson rotation”, the randomisation of candidates from ballot paper to ballot paper. At the first ACT election using Robson rotation, the ACT Labor Party tried to enforce a how-to-vote order, but it failed due the rotating order. Distribution of how-to-votes near polling places was then banned.

The Greens usually issue a preferred ordering of candidates in campaign literature and some advertising. Green candidates are usually designated as leading, second and so on.

As a party that only tends to elect a single member in multi-member electorates, the Greens use this ordering to concentrate vote in their lead candidate. The Labor and Liberal parties, polling more than one quota, are happy to spread their vote across several candidates.

In proportional preferential systems like Hare-Clark, the way to maximise the return of members is to only stand as many candidate as you think you can elect, and to try and split your party’s vote evenly across candidates. That is how Irish elections are conducted under a similar electoral systems.

The rule on minimum preferences under Hare-Clark obstructs this tactic. It forces larger parties stand as many candidates as there are vacancies to fill, and also means voters get a greater choice of candidate. Parties also over-nominate because casual vacancies are filled by countback.

So the ban on how-to-votes is not about the sort of between party preference flows that occur in single member electorates. It is about choice of party candidate.

That’s why raising Tasmanian and ACT bans on how-to-votes as justification for banning  how-to-votes at Federal election misses the relevant context.

The ACT and Tasmania do not use full preferential voting and so how-to-votes have less relevance as a solution to informal voting. The ban under Hare-Clark is about preserving a voters right to choose between a party’s candidate. At a Federal election, a party only stands one candidate, so advocating a vote for the party is also advocating a vote for the candidate.