Refine By

Our Pick

### BERNARD KEANE

Encryption opponents rush to exploit the victims of Paris
Refine By
Refine By
Our Pick

### MYRIAM ROBIN

Junket watch: Sharri Markson pinged for non-disclosure
Where culture lives

## A Few Charts.

Just thought I’d whack up a few charts of things I’ve been looking at over two-party preferred margins lately.

First up, Newspoll TPP margins for every Newspoll going back to Novemember 1985. Newpsoll had a few periods where they didn’t calculate TPP results in their polling, so I’ve constructed a TPP series using the primary vote estimates and distributing preferences the further away from an election you get, the less the results get distributed according to that previous election and the more they get distributed on the basis of the results of the approaching election (click on these to blow them up).

It’s interesting how election results interrupt the longer running trends – sometimes it’s the dreaded narrowing, sometimes its the opposite.

Next up, if we measure the TPP vote of the winning party for every election back to 1949, as well as the  percentage of seats that the winning party achieved from that vote and chart them with a scatter plot and a trend line (quadratic trend) we get:

This gives us an alternative to the pendulum for projecting how many seats a given TPP vote would be expected to deliver to the victorious party. If we transform those percentage of seats won to actual seats in a 150 seat parliament (which we have today) we get:

Which is really just a bit of fun, although the cluster of results where a party won government by getting less than 50% of the vote makes things a bit more complicated here.

Finally, a little chart on the way that One Nation permanently reduced the Coalition primary vote. It’s again using Newspolls but I’ve also added a quadratic time trend through the results with a regression that has a One Nation dummy variable in it (which starts at the first Newspoll after the One Nation Party was created).

I hope you find it as visually interesting as I do in terms of the pronounced impact of One Nation on the Coalition vote.

UPDATE:

The US Election Intrade data has suffered its weekly update, showing Obama getting his second week of probability flight in a row.

• 1
steve_e
Posted August 11, 2008 at 1:06 pm | Permalink

There is a delicious irony of the “support” the Libs provided to Pauline and her ilk and its outcome over time.

This impact and the previously identified gap in the younger age group demographic support for the Coalition parties supports the view that to lead the Coalition from its current position back to Government will be significant tasks.

They will require concerted efforts that are beyond the capacity of a diffident ex Treasurer.

• 2
David Richards
Posted August 11, 2008 at 3:49 pm | Permalink

Given both the One Neuron and demographic effects.. merely maintaining the current position is a challenge that may be beyond any of the contenders. Short of a near 180 degree turn around in coalition policies, it’s hard to see any way they have a realistic future.

• 3
caf
Posted August 11, 2008 at 4:13 pm | Permalink

The One Nation effect won’t be permanent. It should already be washing out of the system – something like a classic damped shock response.

• 4
Topher
Posted August 11, 2008 at 4:40 pm | Permalink

So where did all the One Nation voters go after One Nation imploded?

I guess it’s possible that as the Liberals were picking up One Nation voters, they were losing small l liberal voters at a similar rate.

It’s an indictment of the democrats that they couldn’t pick up support from them. The natural home for those voters would have been the democrats.

• 5
Labor Outsider
Posted August 11, 2008 at 5:13 pm | Permalink

Possum

Interesting graphs.

I would second Topher’s question – the primary votes have to go somewhere – as he says, one possibility is that they permanently lost part of their small-l liberal base – a possibility supported anecdotally through conversations with young professional friends of mine who once regarded themselves as liberal supporters but no longer do. Tampa and Iraq also contributed.

That would then feed back to your earlier posts claiming that the fact that the coalition’s base is dying faster than they are being replaced endangers the coalition in the long run.

So, the real question is what would the coalition need to do to restore its primary vote to a level high enough to have a workable majority?

Let me take a shot. I think the coalition is going to go through a very difficult few years. The current party is dominated by a conservative wing with views that don’t reflect the centre ground of Australian politics. That wing will be very difficult to dislodge – at least until they receive at least one, and maybe two more federal election thumpings. At that point, they will be finally forced to confront the political reality that they either shift back to to the centre, or be prepared for a new party to step into the vacuum.

The best analogy I can think of is the Tories in the UK. After being routed by Blair in three elections, the tories finally realised that they needed a leader that spoke to middle Britain, not just retired euroskeptics in the south.

• 6
Possum Comitatus
Posted August 11, 2008 at 6:27 pm | Permalink

LO and Topher – this actually deserves a thread on it’s own. Tomorrow we’ll do the usual Newspoll Tuesday affair and another article on the One Nation effect. The regression results against the TPP and primary votes of each party suggest a few things in terms of where those Coalition voters went – in a pretty consistent and rational fashion actually, so I’ll post those regression results as well as visuals tomorrow and see if we can get to the bottom of it.

Is there anything else anyone thinks might be worth exploring with this?

• 7
charles
Posted August 11, 2008 at 6:46 pm | Permalink

The interesting question is: Did the Liberal vote drop because one nation took it, or because the liberals were so discussed with the Liberal party response they vowed never to vote Liberal again. I know that set includes at least one.

• 8
Posted August 11, 2008 at 6:50 pm | Permalink

Possum,

In relation to your question (6) what happened to the ALP primary vote during the formation the Coalition’s ‘ramp’ (pre and post One Nation)?

Julian

• 9
charles
Posted August 11, 2008 at 6:51 pm | Permalink

And reading back through the other posts it would seem we are all asking the same question.

• 10
JP
Posted August 11, 2008 at 7:31 pm | Permalink

Where did the primaries go when One Nation imploded?

Well the first place I’d look is at the formation at around the same time of the Australian Greens. Adding the Greens primaries to that chart may be interesting, as they went from effectively zero in 1996 when they weren’t even a unified party, to the 10% odd they get now, leaving only 90% to be divvied up by the majors.

Secondly, the second leg of the Coalition Primary chart looks to me like a recovery until the 2004 election, when the Coalition got their Senate majority and started introducing all the odious legislation that they’d been saving up for years, like WorkChoices, which previously would have had zero chance of getting through the Senate. But that’s history now, and I’d expect their primary to start its upwards drift again (as oppositions do over time), not the downward trend suggested by the chart. What would that Quadratic Time Trend look like with a “Senate Majority Effect” dummy variable, I wonder?

• 11
Possum Comitatus
Posted August 11, 2008 at 8:04 pm | Permalink

Charles – consensus!

JP – I’ll throw that in tomorrow as well.

• 12
Michael
Posted August 12, 2008 at 11:18 pm | Permalink

Your seats-votes graph is fun, but I’d make the same comments here as I made to Professor Jackman on 30 September last year:

“…it’s worth asking whether it’s valid to use every election from 1949 to 2004 to estimate an underlying bias figure, since that seems to rest on the assumption that the underlying relationship has been the same throughout the period. But a lot has changed in the structure of elections. Seats now have more equal enrolments, such inequality as exists appears on a casual glance to be less correlated with the party holding the seat than was previously the case, redistributions are more frequent, and voting no longer has so strong a class basis. I wonder how much the seats/votes outcome in 1949 can really contribute to our knowledge of what the seats/votes relationship looks like in 2007? (Or to put that another way, does the risk of working with a possibly misspecified model outweigh the benefit of including additional sample points?)

By the way, what was the fitted equation?

Another question. In the early part of the period 1949-2004, such bias as existed looked quite “structural”, a feature of unequal enrolments, and differential concentration of majorities, with lots of wasted votes in safe seats. My impression is that these days, that’s less important, and “reversal” results such as that in 1998 will arise predominantly when a party does extremely well in the marginals. One theory would be that governments, or incumbents, now have an intrinsic advantage in marginals, but maybe the jury’s still out on that; the other hypothesis would be that they just benefit from a lucky outcome of a predominantly random process. But if that’s the case, then maybe a result like 1998 should be treated as an outlier. (Or maybe it should be assumed that the variance of the error term is increasing over time, which again raises the question of the model, and estimation procedures.)

What do you think? I wonder how your curve would look if it was estimated using just post 1983-reforms data, and with some sort of robust estimation process to reduce sensitivity to outliers?”

• 13
Possum Comitatus
Posted August 13, 2008 at 1:59 pm | Permalink

Michael – you’re completely right, it’s not particularly useful because of the changing nature of the electorate.

I didn’t even measure the fitted equation, I didn’t think it would have been big enough to warrant! With those results where the winner gets less than 50% of the vote, I think it requires each election where that happens to be looked at on their own merits. In 1998 One Nation was the cause, in 1990 it was the geographical distribution of the vote which had a lot to do with it. Each time it happens lately it seems to be because of a set of relatively unique circumstances.

If we take just those elections from 1983 onward and do it again (and transposing expected seats into 2008 Parliament terms) we get:

http://possumcomitatus.files.wordpress.com/2008/08/michealschart.jpg

Unfortunately, with only ten observations there’s not really enough data points for any regression type measurements. It looks good to they eye – but that could be entirely due to chance!

• 14
Michael
Posted August 13, 2008 at 10:35 pm | Permalink

Re the seats-votes relationship, there’s another more subtle issue to do with model specification. In a two-party relationship, when the vote for one party tends towards 0% or 100%, the variance in the number of seats won tends to zero, so the underlying relationship is heteroskedastic, and should strictly speaking be fitted using a generalised least squares model – though ordinary least squares will generally be pretty much OK, at least around the 45-55% domain in which the overall two-party vote typically falls.

This sort of curve fitting goes back quite a long way, see, for example, E R Tufte, “The Relationship between Seats and Votes in Two-Party Systems”, American Political Science Review, vol. 67, 1973, pp. 540-54; or for even more fun, R E Quandt, “A Stochastic Model of Elections in Two-Party Systems:, Journal of the American Statistical Association, vol. 69, 1974, pp. 315-24.

And of course there’s the good old cube law, and its variants.

• 15
David Richards
Posted August 16, 2008 at 12:28 am | Permalink

If it is because the small ls deserted the Libs and went elsewhere as a result of the extreme right wing policies of Howard’s mob… they aren’t exactly doing the right thing to woo them back. They are dead ducks unless they try to recapture the small ls, and rope in the fresh young brumbies to replace the old swayback carthorses that are heading to the glue factory faster than a racehorse with a J2 up its jakzi.

• 16
Posted August 31, 2008 at 10:23 am | Permalink

I’d agree with the other posters that the One Nation chart is fascinating. It really pinpoints where the Libs began to lose their way. I suspect they will be in the wilderness for some time to come. I feel sorry for the likes of Petro Georgio and Judy Moylan.