I’m seeing a lot of claims that Labor lost because Clive Palmer spent lots of money to get votes from Labor which he then handed to the LNP as preferences. I do not think this is a good summary of what actually happened.

First I should make it clear that I am not denying that Palmer may have had a significant impact on the election. I am just saying that the claim that he delivered votes to the LNP as preferences is overblown.

Parties Don’t Control Preferences

To some extent, a lot of people’s concerns regarding preferences arise from the old Group Voting Tickets (GVT) in the Senate. This was the system that was used from 1984 until 2016 whereby a vote above the line had preferences determined by the party according to a GVT lodged with the electoral commission. This meant that parties did control votes, and sometimes these preferences were determined by deals which may have seen voters’ preferences going to candidates they might not have chosen themselves.
GVTs were abolished in 2016. The system no longer works like this and also it never did in the House of Representatives (HoR). All preferences are determined by what the voter writes on the ballot. An above the line vote in the Senate means the order within a party list is determined by the party (i.e. it is what is written below the line), but there are no preferences between parties, which is the sort of thing which is relevant here.
The confusion over this is understandable, in particular I felt that a lot of reporting on preference deals would have suggested to people that parties did still control preferences.

How to Vote Cards

So what is the point of preference deals if parties do not control preferences? They can still have an influence using how to vote (HTV) cards. I think an important point is that generally parties are not handing out these cards to influence your preferences, they just want your number 1, but they also want you to make a formal vote, so they show preferences. Perhaps Palmer did have a significant interest in influencing preferences though.

For him to effectively steal Labor votes, which are essentially the accusations, then you need a scenario where someone would vote Labor over Liberal, but decided instead to vote 1 for Palmer, take a how to vote card, and then fill in the preferences on it putting Liberal ahead of Labor, even though that would not have been their own preference. I’m sure this scenario could happen, but is it really plausible that Palmer HTV cards would cause people to change preferences between the major parties en masse?

Another scenario is a person with no preference between Labor and Liberal who does choose UAP and then follows their how to vote. Given that this hypothetical voter has no preference otherwise then we can’t say for sure this has changed their vote from Labor though, if they were to arbitrarily choose then we can only give them a 50% chance of putting Labor first if they didn’t get a HTV.

While the full preference flows for this election are not yet published, the evidence from previous elections seems to be that HTVs do not have a large effect in this way when it comes to minor parties (on the other hand, there can be an effect in how many 1 votes you get depending on whether you have them or not). A big reason for this is that minor parties just don’t hand out anywhere near as many how to votes, and without doing so they can have no influence. Also I would suggest that while major party voters might not have much in the way of their own preferences below their number 1, minor party voters are much more likely to have an opinion of their own as their remaining preferences include both major parties and most of the reporting and the advertising during a campaign is about them.

Even when the preference flows are published this won’t tell us for sure who was influenced by a HTV. After all, it is not reasonable to assume that every UAP voter who puts Liberal above Labor did it because of a how to vote. Even by comparing their full set of preferences including all other candidates, a vote matching those does not necessarily mean that they put Liberals ahead due to the HTV, after all a UAP voter who already knew that they wanted Liberal before Labor could look at the HTV, see that it agreed with their preferences of the majors and then decide to use it to determine all of their preferences.
So the number of UAP votes which go Liberal before Labor will give an upper bound but I think the actual number influenced is much lower and pretty much impossible to determine for sure.

The UAP vote wasn’t big enough to have an effect anyway

I don’t think we really need data on how well the HTVs were followed, because even if we assume that every UAP voter who preferenced the Liberals did it because of a HTV, it is still not enough to change the election result. Note that the precise numbers in this section may be a bit different by the time you read this since counting is still underway, but enough has been counted that the basic points will stand.

My best guess is that even 20% of the UAP vote would be a fairly generous estimate of the impact of the HTVs.

Consider the seats that Labor lost in Queensland.

Herbert: The UAP vote was 5.76% but the 2 party preferred result has Labor on 41.79%. So even if you move back every hypothetical UAP voter who followed the how to vote card, which will be well less than the 5.76 (because this includes some who preferenced Labor), then the result is still a loss for Labor.

Longman: This is much closer. Labor are on 46.7, the UAP vote is 3.31. If you move all of the UAP to Labor as preferences then that would put Labor just ahead. But this is not a reasonable thing to do. For a start, there is no way that UAP preferences broke 100% to the Liberals, this never happens. So some of that 3.31% is already counted in the 46.7% two party preferred for Labor. I am confident it is more that 0.2% so that means Labor falls short, but of course we are still also assuming that none of these people chose themselves to preference UAP after the Liberals, which is completely unreasonable. So the amount potentially swayed by the HTV is less still.

Looking at seats retained by the Liberals, we only need consider those with a 2PP for the Liberals of 55 or less.
The following can immediately be ruled out:
Reid: UAP on 1.9, Lib 2PP 53.4.
Robertson: UAP 2.7, Lib 2PP 54.6
Wentworth: UAP 0.7 Lib 2PP 51.8
Dickson: UAP 2.2 Lib 2PP 54.6
Leichardt: UAP 4 Lib 2PP 54
Boothby: UAP 1.9 Lib 2PP 51.9
Casey: UAP 2.6 Lib 2PP 54.9
Deakin: UAP 2.0 Lib 2PP 55.0
Higgins: UAP 1.2 Lib 2PP 54.6
La Trone: UAP 2.5 Lib 2PP 54.5
Swan: UAP 1.8 Lib 2PP 52.8

This leaves the following closer ones:
Chisolm: UAP 1.6 Lib 2PP 50.8
Braddon: UAP 3.7 Lib 2PP 53.2
Bass: UAP 4.8 Lib 2PP 50.4
and Macquarie which is 50/50 with a UAP vote of 4.0

Note that Macquarie may still be won by Labor and Chisolm is going to be subject to an appeal.

I would rule out Braddon and Chisolm as not being close enough to be affected by UAP. You need to assume that up to 50% of the UAP vote is people who would have voted Labor but didn’t because of UAP how to votes and I just don’t think that is reasonable. If evidence to the contrary comes out of analysis of the preference flows I’d reconsider.
In fact none of these seats are surprising losses for Labor, Chisolm was held by the Libs, the two Tasmanian ones change hands regularly and Macquarie was subject to a redistribution which made it marginal.

I think this leaves just Bass or Macquarie as close enough with a big enough UAP vote for HTVs to have possibly had an influence. This is not enough for the Liberals to lose the election.

What about Palmer’s ineligible candidates?

There has been a lot of talk on twitter about accusations that some of Palmer’s lower house candidates would be disqualified under Section 44 of the constitution. This has no effect on anything because they are nowhere near winning.
Now in principle, a losing candidate can still effect the outcome as the voting system does not have the independence of irrelevant alternatives property. In fact, essentially no voting system does unless you allow even worse features. This isn’t a problem here because it requires particular circumstances where the order of elimination of candidates affects the outcome. This just isn’t happening in any of these lower house seats, they are all easily competitions between two specific candidates and the order of elimination of the UAP candidate has no effect.

So basically, to challenge based on the ineligibility of a candidate you would need to show that it affected the outcome, but it wouldn’t have. You take all of the UAP 1 votes, and just reallocate them to their next preference, and because in every case they were eliminated early anyway then you get exactly the same outcome.

Some are suggesting that these ballots should be removed entirely, but this would actually have the effect of making elections more manipulable. At the moment, you get no advantage from running ineligible candidates, this is because of full preferential voting. On the other hand, if those votes were removed then you could rig an election by running a candidate you know to be ineligible with similar policies to your opponent to try and pull away some of their votes, and then get someone to challenge to have those votes removed effectively taking out a bunch of votes that would otherwise have gone to your opponent. With preferencing this does not work because those votes only move on to where they would have gone had you not run at all.

Also it is worth noting that the only remotely close seat where questions have been raised regarding the UAP candidate is Chisolm, but my prediction is that once the break of preferences is known, there probably wouldn’t be enough votes in it to sway it even if you moved all UAP votes with Liberal prefs to Labor. A court challenge would be unlikely to achieve anything anyway because I suspect the court would just take people’s preferences as an indication of their intent had the disqualified candidate been absent, and not make guesses as to whether people changed their preference based on a how to vote. (Just to clarify, this is completely different to the actual challenge which is being mounted over Liberal Party advertising)