Saturday, July 30, 2016

Eric Abetz And The 2016 Tasmanian Liberal Result

Advance Summary

1. The Tasmanian Liberal result at the 2016 federal election was very poor.

2. Not only was the Liberal primary low, but in the Senate race the Liberals failed to obtain preferences even from "right-wing" parties.

3. Liberal ticket leader Eric Abetz has offered a number of reasons for the party's poor results, some of which are clearly false.

4. In fact the Liberal ticket would have had more chance of winning five seats had a below-the-line campaign for Richard Colbeck been more successful, not less.

5. Although Senator Abetz polarises opinion and was the candidate most frequently placed last on Senate ballots, no conclusions about his popularity can be drawn directly from this result.

6. Historically, there is evidence that the Liberal Party performs worse in Tasmania (relative to the nation) when Senator Abetz is on top of the Senate ballot, but the difference is not quite statistically significant and could be caused by other factors.

7. There is no evidence in historic results that having Abetz on top of the ticket is an asset to the party's fortunes.

8. Tasmania is a historically Labor-leaning state, but this does not alone explain the poor result at this election.


Tasmania's results for the 2016 federal election are finalised. The outcome is a disaster for the local Liberal Party, but who is to blame for it? Was the Tasmanian Liberal campaign an innocent victim of national factors beyond its control, or was the failure partly caused by local factors?  Senator Eric Abetz, the man the left here loves to hate, has blamed so many things for the result, but could it be he is himself to blame? Even a little bit?  He doesn't seem to think so.

The Liberal Party has lost all three House of Representatives seats that it won from Labor in 2013, with a two-party preferred vote that looks likely to finish below 43% (the party's fourth or fifth worst in Tasmania's history).  The party has failed to increase its Senate holdings, in the process losing the seat of its only federal Minister.  It may seem that at least retaining its existing Senate seats is not too bad, but its stocks in the Senate were low because of a bad result in 2010 and bad luck in 2013, and any half-decent showing this time around should have won the party five Senate seats.  The party has furthermore seen the return of two Greens Senators and Jacqui Lambie, meaning that on health, education and other basic service issues there will be an 8-4 block of votes against it from Tasmania.

In the Senate, the party's primary vote was ordinary enough (down five points on 2013) but what really marked this as a shocker was the party's failure on preferences.  Through the last few weeks some analysts expected the party to win five seats, having looked at the "right-wing" list of micro-parties and assumed they would strongly preference the Liberals.  In scrutineering I'd been seeing a different story.

In fact Labor received more #2 above-the-line preferences than the Liberals from 13 of the 18 micro-parties running, and more top-6 preferences from 14 of them.  Parties whose voters preferred Labor to the Liberals included One Nation, Family First, Shooters Fishers and Farmers, Palmer United and Derryn Hinch Justice Party, all of which were assumed to be "right-wing".  (We'll need to wait for the button presses in other states to see if this is a specifically Tasmanian trend, or if the Coalition struggles to get Senate preferences from micro-parties elsewhere.)

Senator Richard Colbeck was overtaken in the count because the Liberals had been out-performed on above-the-line preferences even by a One Nation ticket that polled 2.5% of the vote and was barely seen in campaigning.  With a better preference flow, Senator David Bushby would have crossed quota earlier in the count and Colbeck would have probably won.

Abetz's Shopping List

The following is a not necessarily complete list of things Abetz has blamed or reportedly blamed for the Liberal Party's poor showings, mostly at some stage or other with reference to Tasmania, though some were just general federal comments:

* "destablising" below-the-line campaigning for Richard Colbeck
* superannuation changes
* failure to listen to backbenchers
failure of national issues to resonate
national issues swamping the state campaign
failure to attack Labor on negative gearing
GetUp! attacks on Andrew Nikolic
failure to tackle GetUp!
The Recreational Fishers Party (this link has one of Abetz's more comprehensive interviews)
failure to address cost of living pressures
* failure to discuss the carbon tax
* failure to run on union corruption
* focus on innovation, science and technology
* failure to embrace all sides of the party (code for putting Tony Abbott on the frontbench)
* the structure of the campaign
* centrism and ignoring disgruntled Abbott-supporting voters after the leadership change
failure to run on national security
* failure to campaign on free speech for racists and other bigots
* failure to highlight a legal dispute about safe spaces on a university campus

It is of course untrue and unreasonable to say that Abetz has blamed everything but himself.  He has yet to blame the Hodgman government, the floods in June, Devil Facial Tumour Disease, Pokemon Go, or me.

(Note added 13 August: More of the same in a very detailed interview by Matt Smith from the Mercury.)

At the Declaration of the Poll, Abetz made further comments about negative portrayals of him in the media (which he suggested were disproved by him being elected first), and people who apparently blame him for "flat tyres and cats on roofs", though it is not clear he was relating those charges to the election result.

Did The Colbeck Campaign Cost The Liberals Their Chance Of Winning Five?

Most of the above reasons are hard to objectively assess.  But one we can knock over is the scapegoating of the below-the-line campaign for Colbeck.  Abetz alleges that had the same effort put into campaigning for Colbeck been put into campaigning for an above-the-line Liberal vote, the Liberals would have won five. I have no doubt whatsoever this is false.

Had the Liberals recorded the same vote and preference shares, but with the vote more concentrated in the top of the ticket, they wouldn't have been remotely in the hunt for a fifth seat.  Given their modest primary and feeble performance on preferences, their only path to victory was to beat Nick McKim for the final seat by means of the Ginninderra Effect (warning: uber-wonky link).

The only way they could have done it was by having their last two candidates as evenly split as possible at certain points, which would have involved Richard Colbeck having a higher personal vote as a share of the Liberal ticket than he actually got.  Had Colbeck's primary been around the range 5.1-5.5%, with the overall Liberal ticket vote the same, Colbeck and Bushby would have both got over the One Nation tipping point and almost certainly have both gone on to beat McKim.

Also, had Colbeck and his supporters rolled over and taken Colbeck's demotion, some voters who voted for Colbeck would not have voted Liberal at all, as can be seen from the large leakage rate from Colbeck's preferences (26% leaked, with around half of these going to Lisa Singh).   So without the Colbeck BTL campaign it is arguable whether the party would have even had 4.23 quotas rather than, say, 4.1.

Of course it can be argued that the Colbeck campaign damaged both the Liberal primary and the Liberal preference share in the Senate.  Yet the poor Liberal primary and Liberal preference share were completely consistent with the party's lousy Reps result (I crunched a lot of numbers on this as a way of checking that what I was seeing in preferencing - extraordinary as it seemed - was actually believable).  So the claim then must be that the Colbeck campaign trashed the Liberal vote in both the Reps and the Senate.

But the circumstances that applied to the Liberal Party with Richard Colbeck were very much the same as those that applied to Labor with Lisa Singh - indeed arguably Labor was a more extreme case.  Yet Labor thrived in the Reps, and while their Senate primary vote was boosted by votes taken from the Greens by Singh, they also excelled on Senate party preferences.  So the whole case just collapses.

Both parties would have been four-quotas-and-out without their below-the-line rebels. The party with the more successful below-the-line rebellion was rewarded for its own preselection dumbness and won five seats.  With a higher Colbeck vote, they would have both been thus rewarded.

Recreational Fishers

This is another one I can dispose of.  The Recreational Fishers Party ran a vigorous campaign including TV advertisements and robocalls and polled very well in the House of Reps seats it contested.  Senator Abetz has described it as a Labor front, given that it involved a well-known union figure and that the party's policies were basically ALP plus fishing rod and minus supertrawler. The party polled an average 5.6% in the three Reps seats that it contested.

However, the swings in two of the seats the Recreational Fishers ran in were actually much lower than the swings in Denison and Franklin, where they weren't running at all.  Moreover, while over 11,000 voters voted for the Recreational Fishers in the Reps, the number of Tasmanian voters who followed their how-to-vote card in the Senate was exactly 44.  The Recreational Fishers were also very much less successful in the Senate, where they polled 0.7%.  So most likely the votes gathered by the Recreational Fishers were "none of the above" votes from people who rejected both majors and the Greens but found almost nobody else to vote for.  They preferenced Labor (to the extent they did) not because Kevin Harkins told them to, but because they were going to anyway.

It may well be the Liberals would have lost both Lyons and Braddon just on Green preferences even if the other party preferences split 50:50, but we'll have to wait for the full distributions to see.

Personal Votes For (And Against) Abetz

Since Abetz has not himself explored any role he might have had in the outcome, it falls to me to do some statistical soul-searching for him.  What does the result say about whether Abetz is popular or unpopular?

Unfortunately, not all that much, because it only shows the extremes.

To start with Abetz being elected 1, all this shows is that he was top of the ticket for the Liberal Party, which, while polling just behind Labor, had a lower below-the-line vote for non-leading candidates.  As a sign of popularity it's meaningless.  Had the campaign gone well for the party he would have been #1 by at least half a quota and they would not have had to rely on candidate splits or on right-wing preferences.

Moving on to Abetz's personal vote, he obtained the fourth-highest below-the-line vote in the state (behind Singh, Colbeck and Lambie and a whisker ahead of Whish-Wilson).  But that was still only 2.57%.  Indeed Abetz's personal share of his party's vote was lower than that for every other ticket leader in Tasmania except for Labor's Anne Urquhart.

Many people have tried lining up Abetz's vote against Colbeck's and saying that Colbeck must be more popular than Abetz.  It doesn't fly.  A person whose favourite candidate was Colbeck had to vote 1 Colbeck, but someone whose favourite candidate was Abetz may well have been happy voting 1 Liberal above the line, if they didn't wish to reorder candidates in the Liberals or other parties.

Conversely, a person who voted 1 Abetz below the line may not have been an Abetz fan as such.  If they put the Liberal ticket in strict order, they may have been wanting to reorder their preferences among candidates for other parties (like one I saw that went 1-6 down the Liberal ticket then 7-12 up Labor's).

All this said, 46% of Abetz's #1 below-the-lines were not #2s for Stephen Parry.  From this we can conclude that Abetz was the favourite candidate of at least 1.2% of voters.  The figure is probably several times higher, but we have no way of knowing.

At the other end of the scale, data-scraping on #58s (votes from people who went all the way, often specifically in order to put someone last) has shown that Abetz was by far the most detested candidate.  40.3% of voters who used the number 58 put it next to Abetz's name, and by the time you correct for those who missed some numbers I am told this rises to 41.4% - in any case more than seven times the next highest.  However, very few voters go through all the way (only about 2.2%).  (These numbers also ignore the number of voters who stopped a few places from the end, many of whom may have left Abetz's box blank.)

Abetz was thus certainly the least liked candidate for just under 1% of voters, but there were doubtless many more who would have put him last if forced to, but who couldn't be bothered voting 1 to 58.  As for those who did, Club Eric#58 has 3202 members, 100 of whom voted 1 for Richard Colbeck.

Abetz's Track Record

It's difficult to tell what effect a Senator has on the party's result because there are so many other things that influence it.  Opponents of the Senator have been quick to point out there was an even worse Liberal result the last time he was top of the ticket as well, in 2010.  In that year the Liberals lost Guy Barnett's Senate seat and were thrashed in every House of Reps seat in the state.

Supposedly, the same Tasmanian voters who were said to be not interested in technology and innovation this time around were so interested in it six years ago that they ran screaming from the Liberal ticket over broadband policy.  Also, the same voters who were supposedly unimpressed with a Turnbull-style campaign this year thought even less of an Abbott-style one in 2010.

So could the prominence of Senator Abetz have been a factor?  The obvious counterexample is 2004, in which the Abetz-led Liberal ticket easily won three seats and the party gained two House of Reps seats on the back of Mark Latham's ill-advised forestry policy.

Abetz has now been top dog on the Liberal ticket four times, so I thought I'd have a look at the last eight Senate elections to see what degree of possible impact might be detected.  Here there are two possible hypotheses:

1. When Abetz is on the ticket the party's Senate vote suffers in Tasmania, but its House vote doesn't.
2. When Abetz is on the ticket both the party's Senate vote and House vote suffer in Tasmania.

Rather than looking at raw votes in Tasmania (which will usually be higher when the Coalition does well federally and weaker when it doesn't) I thought I'd look at how the Tasmanian vote differs from the national vote, and also at how the Tasmanian Reps vote and Senate votes differ from each other.  There are many different ways these things can be measured.

The following table is the result:

This takes quite a bit of unpacking:

House RPR = the difference between the House of Reps primary for the Coalition in Tasmania and the Coalition's national primary.

House ARPR = the same as House RPR but adjusted for the Coalition's expected advantage or disadvantage once sitting member effects are considered in Tasmania.

House R2PP = the difference between the House of Reps 2PP for the Coalition in Tasmania and the Coalition's national 2PP

House AR2PP = the same as House R2PP but again adjusted for sitting member effects

SEN Rdiff = the difference between the Senate primary for the Coalition in Tasmania and nationally

SEN-H diff = the difference between the Senate primary and the Reps primary for the Coalition in Tasmania

Sen-H addiff = the same as above adjusted for sitting member effects in the Reps.

SD = standard deviation (for the Abetz and non-Abetz years)

Diff = the average difference between Abetz years and non-Abetz years

Some discussion of this table and what it means:

1. There is no evidence that having Abetz on top of the Senate ticket causes the party to do badly in just the Senate.  Most likely either he causes the party to do badly in both houses, or he doesn't have any effect.

2. On average, the party has done 1-2.5 points worse in Tasmania (depending on indicator) relative to the rest of the country when Abetz has been top of the Senate ticket compared to when he hasn't. Because of the small sample size, none of the differences are statistically significant at the 95% confidence level (a few fall just short of being so).  Even if a difference was statistically significant that would not necessarily prove Abetz was the cause.

3. The average differences between Abetz and non-Abetz years exist because of the Abetz years 2010 and 2016, in which the party did much worse in the state (in relative terms) than in the non-Abetz years 2007 and 2013.  Prior to that Abetz years were not worse than non-Abetz years.  So some might say Abetz wasn't always a liability but has become one in recent years.

4. If having Abetz on the top of the ticket was an asset then this table would be likely to reveal it, and with the possible exception of 2004, it doesn't.

So, did Abetz just have the sheer bad luck to be top of the ticket when the party flopped abysmally for other reasons in both 2010 and 2016?  Or did his own prominence in these campaigns have much to do with it? By statistical standards, the case against Abetz is far from proven.  By political standards though, he sure looks guilty.

Problems With The Tasmanian Liberal Campaign

The Tasmanian Liberal campaign was widely criticised before the election, and not just over the demotion of Colbeck.  The campaign failed to preselect a woman or indeed any kind of "diversity candidate" for any of the eight winnable seats.  The lone female candidate (out of 11) who did run for a safe Labor seat was a religious-morality activist and ex-Christian-Democrat whose campaign was a trainwreck.  The party's approach to preselection was slammed by Hobart Lord Mayor Sue Hickey, who referred to the party preselectors as being on average very elderly, obsessed with religious morality issues and treating the Senate as a rewards club that they didn't want to waste star candidates on.

Abetz was widely seen (including by Colbeck) as having a hand in Colbeck's demotion, just as he is believed to have had fellow religious-righter Guy Barnett dumped to #3 in 2010 for trying to bump him off the top position (this caused Barnett to lose his seat).  The Senator maintains that he could not control the Liberal preselectors if he wanted to.  I doubt that anyone believes that, but perhaps he doesn't have to. It may be that the party is so far to the right that that level of active involvement is only attractive to those who don't care if you lose the battle so long as you are fighting the culture war.  It doesn't seem to be a problem at state level though - the Hodgman government is overall much less right-wing than the federal lineup.

The Liberal federal campaign in Tasmania seemed quite active.  But as Karen Middleton's Saturday Paper article suggested, the vast predominance of Liberal sign and TV advertising was no substitute for connecting effectively with voters.

It seems that the Liberal campaign got the worst of both worlds.  Some potential Liberal voters (or preferencers) might have been receptive to an Abetz-style campaign but were still scared off by Liberal federal positioning on health. Labor's scare campaign on Medicare was way over the top but as absurd as the core charge was, it was code for things that are actually happening on a smaller scale as health care becomes more user-pays and many clinics cease bulk-billing.  On the other hand, some more moderate and financially secure prospective Liberal voters were open to supporting Malcolm Turnbull but looked at who they had to vote for locally to do so (especially Abetz) and were aghast.

Abetz's Pre-Election Output

What was Abetz talking about in the months leading up to the election?  Well, he talked about himself, at quite some length.  He spent time telling voters Jacqui Lambie was a Green.  He talked about how he might keep voting against same-sex marriage even if a plebiscite supported it.  He talked about the Greens some more and how you shouldn't spray your vote around.  He talked some more about same-sex marriage and all those poor souls who would suffer depression if the State stopped discriminating against people on the basis of their sexuality.  He talked about biblical principles and rights and obligations that cannot be infringed by government (except if they relate to gay people of course) and even about beer. Did I mention that he talked about the Greens?   (Oh and other Liberal candidates talked about him too, to no effect.)

Of course these are issues he was quoted on in the press, and they may not be representative.  But if we check out his press releases from the leadup to the election, they consist largely of announcements relating to goodies for electorates (aka porkbarrelling), and attacks on Labor and the Greens.  These issues that Abetz thinks the party should have talked some more about - where were they?  Mostly what he talked about (in trying to deter people from voting Labor or Green) was costings.

Going back to the period following the removal of Tony Abbott it was obvious the Senator was politically "damaged goods" and verging on delusional following his own dumping from the ministry.  He was not a team player and seemed more interested in flying the flag for Abbott than in the good of the party (which is amusing given his claims about destabilisation and stability).  In the leadup to the removal of Abbott, Abetz had had to take leave for family reasons and it often appeared that he was under strain.  At this time I suggested it would be in the interests of the party to demote him to number two (where he would still be reelected) to take the focus off him and reduce the chance of him being a liability in the campaign.  Richard Colbeck seems to have paid the ultimate price for reaching the same conclusion that the party might act in its own electoral interests rather than as a perpetual Eric Abetz vanity machine.

What Liberal voters are like

Abetz appears to think that while the party's national campaign was just a bad fit for Tasmania because of its economic profile and susceptibility to Mediscare, the party should also have done more to appeal to a right-wing base on issues like border protection, union corruption, the carbon tax and free-speech culture wars.

The funny thing is that the rump who still voted 1 Liberal in the Senate were not as right-wing as people think.  This election has revealed some very interesting things about the pluralistic voting preferences of voters for all parties.  Only 61% of Liberal above-the-line voters preferenced a party that could be described as right-wing by any stretch of the imagination (and here I include even Lambie, who is only right-wing on selected issues). 18% preferenced Labor, part of a long-running Tasmanian tradition of supporting both major parties in order to curb the power of the Greens.  Many people who vote Liberal do so not because the party matches their opinions, but in spite of it not doing so, because they come from a social background in which everyone votes Liberal.  The same is true to some degree of any party.

If anyone is concerned that moving to somewhere near the centre on social issues, and hence being even a possible voting option for most young people, will lose the party votes to right-wing micros, then that is groundless.  The party lost those votes at this election anyway.

Why wasn't this predicted?

While psephologists predicted this election rather well Australia-wide, we did quite badly for Tasmania.  I don't think any of us forecast that the Liberals would completely wipe out here.  It was on the cards they could, but it seemed more likely (even at the end) that they'd lose just one or two Reps seats. Some analysts expected Bass to fall (because of its history of dumping incumbents and because of economic and health issues), but nothing like as heavily as it did.

This was even though it was obvious that the Liberals were "too far to the right" (as Sue Hickey says) in the state, running a ticket full of very conservative Abbott supporters who would be useless in a Turnbull government.  It was also obvious that the federal campaign was a risky fit for the state and that Tasmanian support had been placed at risk by the state-income-tax thought-bubble.

The major reason many of us got it wrong was that we believed the local polling.  Tasmanian local polling at the 2013 federal election was skewed to the Liberals, but that was part of a general national skew that at this election greatly abated.  Except that in Tasmania, it didn't: the polls were almost as far out as they were last time, no matter who did them or who they were commissioned by.

Indeed the only poll to get it close to right (on average) in its Tasmanian sampling was Morgan, which I'd previously sent to the naughty corner for having Labor much too high last time around.  While the last published Morgan reading for Labor in Tasmania (63%) was silly, the average of their last five (54.9) was only 2.5% off.

It seems that based on the historic table above, that even after adjusting for the personal votes of new sitting members, we should have been more sceptical of all polls suggesting the Liberal Party was getting anything above, say, 47% two-party-preferred statewide (that would be -3.5 in the House RPP column, for instance).  Some polls ran contrary to the strong pattern of recent decades that Tasmania is usually a Labor-leaning state. It is also not the case that the state swung late, since pre-poll and postal voting also showed large swings against the Liberals in Bass, for example.

Any debate about Abetz and the Tasmanian campaign failure then needs to be seen through the prism that Labor was always going to win the 2PP vote and get some sort of swing back in the state.  I think the argument that the 2016 campaign was a bad fit for Tasmania only goes so far, because the historic patterns suggest that demographically any Coalition campaign is a bad fit for Tasmania.  But for the second time in two Abetz cycles, the party has done even worse than normal after adjusting for this.

The strong 2013 result can be seen as a combination of a strong federal vote and a backlash against the then state government.  It probably did not have much to do with who was on the federal ticket.  But the poor results in both 2010 and 2016 leave plenty of room for an argument to be run that Abetz's prominence in the campaign is quite a problem.


  1. Eric Abetz is a dinosaur and believes Tasmania is the same as in was in the 1970s. At the state government level there was a clean out of far right liberals after the 1996 rout and whilst i am not a fan of "Teflon Will", he has a good reading on the state voters. Whilst Abetz continues to control the Liberal part they will drift away from the voters and getting rid of Colbeck and anyone who doesn't abide by the far right agenda, the liberals will struggle in a state where the number of voters required to win senate and house seats is much less than anywhere else in the country.

  2. Thanks - Very incisive article. Nice to see Abetz's handwaving actually countered by hard evidence.
    Any thoughts on whether Tasmania is becoming more Labor-leaning (or at least more progressive), with most of the population growth centered on (greater) Hobart?
    Is there any chance of the ALP winning Denison back from Wilkie?

  3. I think Wilkie would have to do something really dumb, or nothing at all for a very long time, to lose the seat back. At this election he became even more entrenched.

    I haven't looked closely at the demographics though I suspect that would make just a minor difference. Tasmanians voted Liberal in huge numbers at the 2014 state election but I don't know if we will see anything like that again for a while.

  4. Hello Kevin,
    Just want to thank you for all the work you put into the recent Federal election. You kept us up to date and informed. I really appreciated your insights, predictions and vast knowledge. Great job!