Australian Federal Election- July 2, 2016

[jimrtex]

What sort of strategies were the parties advocating for marking the senate ballot?

Under the old system, presumably the advice would be: "It's insane to vote below the line, vote for us above the line", along with a reminder which column number they were in the state.

A voter who was inclined to vote for a single-issue party might try to analyze their preference order, but would quickly become confused or bored, and simply pick their issue. Any tactics would be left to the agents who would try to figure out other groups whose 77th preference to them, in exchange for a 63rd preference that they knew probably not reciprocate.

Under the new system, it might be risky having voters voting above the line, since they might continue the practice of marking only one party, resulting in an informal ballot. Parties with a full slate of twelve might be content with voters ranking their candidates below the ballot. There is no real harm in voters switching the order around, and probably is helpful.

For example with ALP at 4.17 quotas in NSW, the 5th candidate would be in 8th place for the final 4 seats after the initial 8 seats were given out (4 to the Coalition and 4 for ALP) under the old system. The Greens at 0.96 would get the 9th seat sometime, but that puts the ALP candidate 7th place for the final 3 seats.

A large party that encouraged voting above the line might also have to endorse other parties, and they really don't want to give the people the idea that it is OK to wander.

But spreading the first preferences out would likely keep the 4th and 5th candidates continuing, with more opportunity to win a 5th seat.

The single-issue parties could instruct their supporters who to vote for above the line, but would voters listen? And putting out a specific order, might be seen as conniving.

So what were the parties actually suggesting? What about the press? Are there any good government groups? If you voted, what did you do?

[jimrtex]

In this table:

First preferences by vote type

What are "Ticket Votes" and "Unapportioned"

Are "Ticket Votes" an Above-The-Line first preference for the group; and "Unapportioned" a Below-The-Line first preference for one of the candidates of the group?

Ticket Votes are generally 20%-25% of the votes for the group.

A few exceptions are:

Smaller Parties with full slates of 12 candidates (Greens and Christian Democrats) had Ticket Votes in the teens. But ALP and Coalition were in the typical range.

Science Party/Cyclist Party was in the teens - perhaps voters could make no sense of the joint group.

One Nation was over 30%.

[jimrtex]

Booth results in SEQ (primary vote plurality)

Obviously, these are still preliminary, since recounts continue.

As always, bigger version in gallery.

Could you explain the demographics behind the voting patterns?

There is a strong ALP vote along an east-west axis. Is that in Brisbane? Then to the north of that it begins trending Coalition. But north of that along the coast is an area that is more mixed, with a small ALP advantage. But the coastal area north of NSW is Coalition. Has Queensland had an influx like in Florida and California where the population was attracted by the climate (once air conditioning was available).

Why is there the concentration of support for One Nation to the west? Banjo Belt? To the west of that there is a small concentration - some sort of regional center?

And then the rural areas are Coalition.

[jimrtex]

So, have you guys seen the Senate forecasts on the website blog.geeklections.com.au ? The author seems very bullish on preferences flowing to the Nick Xenophon Team -- he's very confident of Luke Bolton (the top WA NXT candidate) overtaking Rachel Siewert on preferences, and seems to think Aidan Dalgliesh (the NSW candidate) has a very real chance of defeating David Leyonhjelm for the last seat.

He also thinks a Shooters Senator is likely to be elected in Tasmania (God willing), but that forecast might be off because the very high levels of BTL voting aren't being taken into account, and only 1 Green Senator being elected off of 1.46 quotas seems like an unlikely result to me.

How likely are Bolton/Dalgliesh/Allen to actually win?

He linked to a couple of websites which would help you prepare your ballot. On one you could select parties that you might be inclined to vote for, and then rearrange parties and candidates, and it would produce a ballot-like image.

Are you permitted to take these with you to the voting booth?  Do parties also prepare these?

I doubt that many voters understand how their ballot is actually distributed. If someone votes a lot of minor party candidates, it won't transfer 37, 38, 39, 40, etc.  It will transfer 37, and then skip over a bunch of excluded candidates to 43, etc.


His model seems reasonable enough, given the lack of experience with the new system. Voters might behave like they do in NSW. But maybe they won't. There was likely more news coverage of something that was part of a national election.

In NSW there was a strong tendency to vote the minimum required below the line. So supporters for  the larger parties are likely to vote for their party and quit. If someone votes 1-12 for ALP, will they go ahead and vote 13-24 Green? Or will they start hopping around, so that their Green Party preference s will either have been elected or excluded?

If you decide not to vote for ALP or LNP or Greens or anything like a conventional party, are you going to find 20 or 30 candidates to vote for and then place ALP (etc.) at 31-42?  I am dubious.

So I think there will be a hemorrhage of exhausted ballot (exhausted will probably be the true 10th candidate elected, with 11, 12, and 13 not reaching quota but being declared winners. I also believe that someone who votes bunches of candidates from many parties below the line is not going to vote too consistently, and may be somewhat attracted to personality, or celebrity, etc. Some voters probably see the groups as 20 different variants of the OMRLP.

His initial estimates for each state were based on guessed support. I don't think NXT performed as well as expected outside of SA. His Live Senate predictions are based on the actual reported first preference, along with his predicted preference flows.

It is pretty easy to forecast about 11 seats in each State, so his forecast may actually only be of the last seat:

NSW: LNP 5, ALP 4, Gre 1, ON 1, (other 1)
Clement says the last will be NXT or Liberal Democrat

VIC: LNP 4(?), ALP 4, Gre 1(?), Hinch 1, (2 with a reasonable chance that one or both will be LNP and/or Gre)
Clement is pretty sure it will be LNP 5th, and Gre 2nd

QLD: LNP 5, ALP 4, ON 1, Gre 1, (other 1)
Clement is almost certain on Liberal Democrat. I think this a bit bullish for starting at 0.35 quotas

WA: LNP 5, ALP 4, Gre 1(?), ON 1, (other 1, with possible Gre)
Clement is very strong on NXT, and somewhat skeptical on ON

SA: LNP 4, ALP 4,  NXT 3, Gr 1.
Clement agrees

TAS: ALP 5, LNP 4, Gr 1(?), Lambie (1), (other 1, with possible Gre).
Clement is very strong on Shooters, Fishers and Farmers. But starting at 0.17 quotas requires a lot of favorable growth

The personality groups (NXT, Hinch, Lambie) seem to have done particularly well, which makes the other predictions easier since their support is likely coming from preferences that would be scattered.

[jimrtex]

Booth results in SEQ (primary vote plurality)

Obviously, these are still preliminary, since recounts continue.

As always, bigger version in gallery.

Could you explain the demographics behind the voting patterns?

There is a strong ALP vote along an east-west axis. Is that in Brisbane? Then to the north of that it begins trending Coalition. But north of that along the coast is an area that is more mixed, with a small ALP advantage. But the coastal area north of NSW is Coalition. Has Queensland had an influx like in Florida and California where the population was attracted by the climate (once air conditioning was available).

Why is there the concentration of support for One Nation to the west? Banjo Belt? To the west of that there is a small concentration - some sort of regional center?

And then the rural areas are Coalition.

I'll discuss the voting patterns, as you've asked, but you'll probably want to refer to Google Maps regarding place names. One thing to note before we get started, remember that these are primary vote plurality - in some areas near the downtown, the left vote is split between the Greens and Labor. In some of the rural areas in Wright and possibly Blair, the right vote is split between the LNP and Hanson. Some of the voting patterns may shift once I post 2CP maps (I've done a bit of work on them, but they haven't finished these counts in some of the electorates where it was unclear who would finish second vs third).

primary vote plurality - "primary" refers to first preference? Does a candidate other than the first or second on 1st preferences, ever win? Do they even finish second?

The East-West Labor vote is typically just to the South of Brisbane, starting at about Beenleigh, in the East (maybe South-East), through to Ipswich in the West. The strong Labor area near the centre of this band is Inala. The strong Labor area in the East is Woodridge/Slacks Creek and probably down to Marsden. These areas have all been traditional Labor-voting areas. Ipswich itself is in Blair (the N-S aligned seat with many Labor booths in the South, although it used to be in Oxley (the seat that looks a little like "^" just to the East of Ipswich. The Eastern half of Oxley is within the Brisbane city limits, the Western half is in Ipswich. The next seat to the East is Rankin, which is based predominantly on Logan city. Ipswich and Logan can really be considered part of "Greater Brisbane" - they are linked to Brisbane by freeways and commuter railway lines. The electorate to the North of Oxley and Rankin is Moreton. The Labor vote in the middle is around Rocklea and Moorooka, and up through Yeronga and Yeerongpilly.

There appear to be a lot of industrial areas. was this the result of a deliberate development policy? South Brisbane is more competive, so this would be more white collar or service workers,and not the wealthy who headed to the higher ground. Does Brisbane flood?

Demographically, as I noted, the East-West band has been traditional Labor, Moorooka and Yeerongpilly and up to Dutton Park (this is where we start heading up into Griffith) had a sizeable community of Sudanese refugees settle there during the Howard era. Western Yeronga, however (the area in the bend of the river), is relatively "leafy" - tree lined streets, large blocks with renovated houses. The riverfront end obviously has river views. As you head further North, into Dutton Park and West End and Highgate Hill, these were traditionally inner-city Labor areas, which have now become more popular with uni students. The Liberal-voting areas to the West across the river are traditional Liberal areas, but also have many uni students (University of Queensland is there). You'll note the particularly strong LNP areas in the Western suburbs of Brisbane. This area is hilly, tree-filled, with large blocks and is quite affluent.

Downtown is on the peninsula with the Botanic Gardens on the tip? Or is it Fortitude Valley? Are some of the bridges to the south side transit or pedestrian bridges, they seem narrow.

At the Eastern end of the Brisbane electorate, you can see Ascot/Hamilton/Clayfield. This area is again quite affluent, the horse race tracks are there, the houses are older Queenslanders that have been renovated inside (Google image search that housing type - or look on realestate.com). Again, quite a wealthy area. Back in 1996, this area was in Lilley, and West End was in Brisbane - which is why in Howard's landslide, Lilley and Griffith both went Liberal, but Brisbane stayed Labor, whereas now Brisbane is a Liberal seat and Griffith and Lilley are both in Labor hands. That's not a gerrymander allegation, just an observation. The current boundaries are probably more natural.

I had thought you wanted me to research "renovated inside". It appears that "inside" is used as a noun, "It has a nice renovated inside", rather than a preposition, "It is renovated inside". The houses with renovated insides appear to favor an open plan. Is this modern, and requires a tearing down of interior walls, and adding structural support; or is it characteristics of Queenslanders? The obvious characteristic of this type of house is the deep verandas (sleeping porches?), and the sometimes elevated foundation, both presumably to help circulate air before AC. But it would also help to have few interiors walls. The verandas would provide a visual expansion of space. What does "raising" refer to?

You can see the area around Bulimba (across the river from Hamilton), which has gentrified over the last... fifteen years, I'd guess, give or take a few years. The Labor areas in Lilley and across the river in Bonner are both industrial areas (as you might be able to gather from the lack of polling booths in the area, which are a defacto indicator of population density), and traditionally Labor. The mid-sized island at the mouth of the river is the Port of Brisbane. The strong Liberal area in the centre of Bonner is probably the "bible-belt" of Brisbane, in the somewhat-affluent, suburban manner, (as opposed to the Appalachian version).

I don't see an obvious way to get across the river from Bulimba, so was an area that was originally more affordable because of lack of access, or were these more for port workers, whose jobs have moved further east or been automated. I'm not sure I understand where the Liberal area in the center of Bonner is: Carseldine, Bridgeman Downs, Aspley?

Explain Belmont? It appears to be agriculture, but the lot sizes (farms) are too small, but they are too large for an ordinary suburb.

Does Brisbane have plans for another airport? One runway seems to be small for an isolated capital city, with international flights.


<Snipped Lots of Interesting Stuff, Because message over 11000 characters>

I was looking up some population data, and found that while Queensland and Western Australia had grown faster than NSW and Victoria, that this had been reversed in recent years. Is this a hiccup, or start of a new trend?

Queensland and Western Australia had doubled their population from about 1975-2010, while the other areas had required from around 1955-2010 (2.0% vs 1.3%)

I also had thought Brisbane was much more dominant population-wise. But even considering the metropolitan area it isn't. Gold Coast and Sunshine Coast are remarkably populous - around 500,000 persons each. Gold Coast is substantial enough to be hosting the Commonwealth Games.

Your map doesn't appear to include Rockhampton, Mackay, Townsville, and Cairns. Are these all LNP areas?

You mentioned Sir Joh. Since his Country Party was able to dominate both the Liberals and Labor, why have the Liberals and Nationals formally merged in Queensland. I would think that the opposite might be true since the Country Party had been more successfully. At the federal level the two parties aren't really partners.

[jimrtex]

Declaration of House of Representatives results

What is interesting is that there is an option to have the web page read aloud.

The state abbreviations of NSW, Vic, Qld, WA, SA, Tas, ACT, and NT are read as:

New South Wales
Victoria
Queensland
Waugh or Roy
Essay
Tasmania
Australia Capital Territory
Entee

Further down:

WA is pronounced as Washington

51 Cattley St
Burnie

is "51 Cattley, Saint Burnie"

Suite 14W-16W is "Sweet 14 DoubleU to 16 Watt"

5/1-3 Cornpark Circuit is "May 1 to 3"

16 July 2016 is read as July 16, 2016

Times are listed as 10.30am and read as 10 point 30 A M.  Is this conventional in Australia?

Are Level 4, Level 7, Level 13, etc. school buildings?

[jimrtex]

Button pressed in Tasmania. 5 Labor (notably including Lisa Singh, elected ahead of Catryna Bilyk), 4 Liberal (not including Richard Colbeck, who didn't make it), 2 Greens, and 1 Jacqui Lambie. Last seat was between second Green and first One Nation candidate. Nothing too surprised, but at least movement is finally happening in the Senate.

EDIT: As an aside, the race for the last seat between Nick McKim (Greens) and Kate McCulloch (One Nation) was very close -- the margin was just 21247-21106, or 141 votes. (McKim was elected under quota, as the quota was 26,090). On an amusing note, Singh actually reached the quota on preferences from Colbeck. Some voters voted for the ideologically-opposite dissidents from both major parties. Lol.

Some observations.

Slates were limited to six candidates. Is this due to a rule peculiar to Tasmania? "You can't fool us. You were trying to get us to take our shoes off!"


The summary of first preferences showed:

339,159 formal votes 96.52%
12,221 informal votes 3.48%
351,330 total 96.04%

339159 / 351330 is indeed 96.52%

But what is the "total" 96.04% of? Is that turnout? Are there any consequences for the 3.96% who didn't vote.


I had thought that there was almost a reversal of above the line and below the line votes (from preliminary results), but 73.2% of ALP votes were "ticket" votes, and 75.7% of Coalition voters were "ticket" votes. These were the only groups with 6 candidates.

Were ballots that ranked the candidates 1,2,3,4,5,6 below the line reclassified as ticket votes?


The new voting system appears to work as intended. With 12 to be elected, you might expect more outsiders to be elected.

On first preferences:

ALP 4.3670 quotas.
Coalition 4.2284
Greens 1.4504
Lambie 1.0788

Total 11.1246
Others 1.8876

So you would kind of expect other parties to be able to scramble together enough for a 12th seat.

But after the initial 8 candidates were elected, and their surpluses distributed:

ALP (3) + 1.3584
Coalition (3) + 1.2087
Greens (1) + 0.4525
Lambie (1) + 0.0692

Total (8) + 3.0888
Others 1.9112

Rather than immediately electing their 4th Senator and leaving a smaller remnant, both ALP and the Coalition split between two candidates each with too much support to prevent early elimination, but with plenty of room to accept transfers. Meanwhile transfers among the others could not be organized, and they ended up being dissipated or exhausted.

As the minor groups were being excluded, around 40% of their votes were transferring to ALP, the Coalition, Greens, or Lambie.

Most of the transfers to ALP and the Coalition appear to have been ticket transfers.

After the initial distribution of the surpluses until Colbeck was excluded, Bushby the 4th Coalition candidate, received 86% of the transfers to the Coalition. The only notable exception was when the 6th Coalition candidate was excluded, and Colbeck received about 2/3 of the transfers. These were likely voters who voted 6,5,4,3,2,1.

Between the two final ALP candidates, Bilyk received about 81% of the transfers to he and Singh.

Singh was elected because of her large share of 1st Preferences. That she was put over the top by Colbeck transfers was not too remarkable, given that all his votes were below the line. His transfers were:

Bushby (Liberal) 74%
Singh (ALP) 13%
Exhausted 6%
McCulloch (One Nation) 3%
McKim (Greens) 2%
Bilyk (ALP) 2%.

With only 5 of 58 candidates remaining, most of the votes other than to Bushby may have been fairly deep on the preference lists. 3/4 of Colbeck's support would be from Liberal voters who didn't like the party's order. It would not matter if they ranked Bushby 2nd or 6th, he was the only Liberal remaining. There might have been a few who deliberately skipped Bushby if they somehow felt that would help Colbeck.

If you are ranking below the line, there is nothing that would be personally attractive about a 4th ALP candidate, a 2nd Green candidate or a One Nation candidate. There would be some sympathy for Singh since she was in the same place as Colbeck.


Both ALP and the Coalition could do better if they could develop a strategy for voters to vote below the line. Something along the line of "You Choose" along with biographies that would attract votes on a non-ideological basis.

[jimrtex]

The button has also been pressed in Western Australia; the result is 5 Liberal, 4 Labor, 2 Greens, and 1 One Nation, with all those elected having made it over the quota fairly comfortably and only one unsuccessful candidate mustering up even half a quota (Kado Muir of the Nationals). Amusingly the final Labor Senator made it over the hump on preferences from Shooters candidate Andrew Skerritt (and One Nation and the Greens both made it off the elimination of the Nationals).

There seem to be questions about the eligibility of the elected One Nation Senator Rod Culleton connected to a larceny incident; if he were deemed ineligible, the party would be able to recommend a replacement. #2 on the WA ballot, and the likely replacement, is a certain Peter Georgiou, the brother-in-law of the elected Senator.

In terms of term length, either of the two discussed methods seems to give 3 Liberal, 2 Labor, and 1 Green longer, six-year terms while 2 Liberal, 2 Labor, 1 Green, and One Nation get three-year terms.

Unlike in Tasmania, there were no secondary candidates. The Liberals had 5.004 quotas, and after distributing of surpluses had 0.987 quotas, almost all behind their 5th candidate.

Almost all the secondary candidates were eliminated before moving onto eliminating the first candidate of any group. The 5th Liberal candidate was soon pushed over the top, but with only an eleven vote surplus. After the distribution of the initial surpluses ALP and Greens together had about one quota, of the four outstanding, with three more to be elected. Because the transfers of the smaller parties were not directed, both a 4th ALP candidate and 2nd Green was elected, along with a One Nation candidate.

With five quotas of first preferences, you should be able to elect six candidates. The Liberals would have been better off with the Nationals in a group. It would actually be desirable for voters to vote below the line for a National Candidate, as long as they followed through with the Liberal candidates. Even if they stuck a few odd other candidates in their list, they would eventually flow back to the Liberals.

[jimrtex]

Unlike in Tasmania, there were no secondary candidates. The Liberals had 5.004 quotas, and after distributing of surpluses had 0.987 quotas, almost all behind their 5th candidate.

Almost all the secondary candidates were eliminated before moving onto eliminating the first candidate of any group. The 5th Liberal candidate was soon pushed over the top, but with only an eleven vote surplus. After the distribution of the initial surpluses ALP and Greens together had about one quota, of the four outstanding, with three more to be elected. Because the transfers of the smaller parties were not directed, both a 4th ALP candidate and 2nd Green was elected, along with a One Nation candidate.

With five quotas of first preferences, you should be able to elect six candidates. The Liberals would have been better off with the Nationals in a group. It would actually be desirable for voters to vote below the line for a National Candidate, as long as they followed through with the Liberal candidates. Even if they stuck a few odd other candidates in their list, they would eventually flow back to the Liberals.

Isn't this at least partly to do with the ticket system not balancing the candidates as Irish parties try to do?  After 9 counts the Liberals in WA were left with four candidates elected after reaching quota (which was 105,091), one candidate on 101,888 looking certain to be elected, and their other two candidates way back on 1,202 and 647 and so heading for elimination before there was much chance to pick up transfers.  If they'd tried to get six candidates on around 5/6 of a quota they'd probably all have got elected.

What is surprising is that there was no attempt to do so.

In Tasmania, it appears that there were a couple of dissident senators who had been listed low, and made a personal appeal for votes. While it represented dissent within the party, it was still overall good for the party.

In South Australia, the Liberals had 4.236 quotas on first preferences. After the distribution of surpluses they had four elected and 0.227 quotas left over. This placed them 6th overall with 3 remaining to be elected, and Xenophon and Greens with very healthy bases, so they were 4th overall for the final seat.

Had they split the final two: 0.618 and 0.618 or even 0.718 and 0.518 that would have put them in 3rd and 4th place with 4 remaining to be elected. One Nation and Family First would have been forced out. While they might not have ranked the Liberals high, they wouldn't have materially ranked ALP or the Greens higher.

With 6 remaining it was:

Greens 0.926 quotas (+.147 gained since distribution of initial surpluses)
Xenophon 0.907 (+.142)
ALP 0.619 (+.121)
Family First 0.550 (+.161)
One Nation 0.539 (+.146)
Liberals 0.326 (+0.099)

But split the final two for Liberals and you squeeze out Family First and One Nation which probably puts you clear of ALP.

And even if it doesn't, the eliminated candidate will elect the other candidate on transfers. With the candidates of each party lined up in columns, it is not likely that there would be much failure to not rank all of them high. Even if they stick an ALP candidate up high for some personal reason (he lives across the road, etc.) it is unlikely to be _the_ ALP candidate who remains.

[jimrtex]

Isn't this at least partly to do with the ticket system not balancing the candidates as Irish parties try to do?  After 9 counts the Liberals in WA were left with four candidates elected after reaching quota (which was 105,091), one candidate on 101,888 looking certain to be elected, and their other two candidates way back on 1,202 and 647 and so heading for elimination before there was much chance to pick up transfers.  If they'd tried to get six candidates on around 5/6 of a quota they'd probably all have got elected.

Victoria is the first state where a party has had a full slate of 12 candidates, such that a voter could vote below the line entirely within a party column and have the vote considered to be formal. The Greens had 12 candidates, while ALP had eight, and the Coalition seven.

There was surprising amount of leakage among Green voters (around 2/3) but that may just be Green voters being Green. I suspect that both ALP and Coalition voters would be a little more loyal, since they are seeking to be a governing party.

The Coalition has 0.293 quotas after the distribution of the 8 initial surpluses. They could have been in trouble for their 5th seat, except that the party that did the best job of gaining transfers, Family First was in 10th place of those seeking the final 4 seats at 0.150 quotas.

By the time the field had been reduced to 9 candidates seeking the final three seats, ballots were going deeper on preferences, either being exhausted, or voters were tired of ranking minor parties and were just filling out the last parties on their ballot.

So while the Coalition did get 5 seats on 4.3 quotas, they made it harder than they had to.

Victoria ended up with 1.118 exhausted ballots, and the trash can was the #1 choice for the final three exclusions.