Average below 80 per cent

	House	Coalition Odds ($)	Labor Odds ($)	Coalition Win Probability (%)
2016-03-23	Betfair	1.24	4.52	78.47
2016-03-23	CrownBet	1.14	5.50	82.83
2016-03-23	Ladbrokes	1.18	4.75	80.10
2016-03-23	Luxbet	1.16	5.00	81.17
2016-03-23	Sportsbet	1.20	4.50	78.95
2016-03-23	TABtouch	1.21	4.30	78.04
2016-03-23	William Hill	1.20	4.50	78.95

	We 16-Mar	Th 17-Mar	Fr 18-Mar	Sa 19-Mar	Su 20-Mar	Mo 21-Mar	Tu 22-Mar	We 23-Mar
Betfair	80.00	80.00	79.21	79.07	79.21	80.27	78.94	78.47
CrownBet	84.27	84.27	84.27	84.27	84.27	84.27	84.27	82.83
Ladbrokes	80.10	80.10	80.10	80.10	80.10	80.10	80.10	80.10
Luxbet	81.76	81.76	83.33	83.33	83.33	83.33	81.17	81.17
Sportsbet	81.04	81.04	81.04	81.04	81.04	81.04	81.04	78.95
TABtouch	80.27	80.27	80.27	80.27	80.27	80.27	-	78.04
William Hill	81.04	81.04	81.04	81.04	81.04	81.04	78.95	78.95

Betting markets

On the day after the Turnbull maneuver, let's have a quick look at the betting markets (noting that TABtouch is down for maintenance).

	House	Coalition Odds ($)	Labor Odds ($)	Coalition Win Probability (%)
2016-03-22	Betfair	1.23	4.61	78.94
2016-03-22	CrownBet	1.12	6.00	84.27
2016-03-22	Ladbrokes	1.18	4.75	80.10
2016-03-22	Luxbet	1.16	5.00	81.17
2016-03-22	Sportsbet	1.17	5.00	81.04
2016-03-22	William Hill	1.20	4.50	78.95

%	Tu 15-Mar	We 16-Mar	Th 17-Mar	Fr 18-Mar	Sa 19-Mar	Su 20-Mar	Mo 21-Mar	Tu 22-Mar
Betfair	80.00	80.00	80.00	79.21	79.07	79.21	80.27	78.94
CrownBet	84.27	84.27	84.27	84.27	84.27	84.27	84.27	84.27
Ladbrokes	80.10	80.10	80.10	80.10	80.10	80.10	80.10	80.10
Luxbet	81.76	81.76	81.76	83.33	83.33	83.33	83.33	81.17
Sportsbet	81.04	81.04	81.04	81.04	81.04	81.04	81.04	81.04
TABtouch	80.27	80.27	80.27	80.27	80.27	80.27	80.27	NaN
William Hill	81.04	81.04	81.04	81.04	81.04	81.04	81.04	78.95

Aggregation update

With the 51-49 Newspoll result favouring the Coalition in today's Australian, it is time to update the polling aggregation. The result is unchanged on last week: 52.3 per cent for the Coalition to 47.7 per cent for Labor.

Turning to the primary votes:

On the attitudinal polling, Turnbull is in net negative territory for the first time (according to Newspoll).

In aggregate, the attitudinals look like:

Betting Market Update

It has been a relatively quiet week on the markets ...

	House	Coalition Odds ($)	Labor Odds ($)	Coalition Win Probability (%)
2016-03-18	Betfair	1.21	4.61	79.21
2016-03-18	CrownBet	1.12	6.00	84.27
2016-03-18	Ladbrokes	1.18	4.75	80.10
2016-03-18	Luxbet	1.14	5.70	83.33
2016-03-18	Sportsbet	1.17	5.00	81.04
2016-03-18	TABtouch	1.18	4.80	80.27
2016-03-18	William Hill	1.17	5.00	81.04

%	Fr 11-Mar	Sa 12-Mar	Su 13-Mar	Mo 14-Mar	Tu 15-Mar	We 16-Mar	Th 17-Mar	Fr 18-Mar
Betfair	78.88	78.61	79.02	79.02	80.00	80.00	80.00	79.21
CrownBet	84.27	84.27	84.27	84.27	84.27	84.27	84.27	84.27
Ladbrokes	82.83	80.10	80.10	80.10	80.10	80.10	80.10	80.10
Luxbet	81.76	81.76	81.76	81.76	81.76	81.76	81.76	83.33
Sportsbet	81.04	81.04	81.04	81.04	81.04	81.04	81.04	81.04
TABtouch	80.27	80.27	80.27	80.27	80.27	80.27	80.27	80.27
William Hill	81.04	81.04	81.04	81.04	81.04	81.04	81.04	81.04

DD Redux

The Government has two broad options with the possible double dissolution, each with its own challenges:

• Go early (ie. a DD-election on or before 1 July),

• Pros: no need for a long election campaign, and if the DD is held early enough (say in April), the Government can leave the problem of getting a supply budget for 2016-17 until the last week of June and a new parliament.

• Cons: the Senate term will be reckoned from 1 July 2015 and will result in a Senate with around two years before a half Senate election is needed (a full year before the next House of Representatives election is needed); or

• Go late (ie. a DD-election on 2 July or thereafter), 

• Pros: the Senate term will be reckoned from 1 July 2016, and will effectively align the next half Senate election with the next election for the House of Representatives.

• Cons: because the last day for calling a DD is 11 May, it will be a long election period;  it will be impossible if the Government cannot secure a supply budget for 2016-17 through what looks like an increasingly hostile Senate; and if the Senate count goes beyond 11 August, ministers who are senators will need to resign their ministerial positions. 

Yesterday, the Senate has indicated that it will not be recalled a week early to consider a supply budget. And I suspect there is no guarantee that it would would pass supply bills on 10 May when it next sits. In which case, I doubt there can be a 2 July (or later) double dissolution. It would be reckless to ask for and for the Governor General to grant a double dissolution without a supply budget for 2016-17.

This then brings back on to the table the question of an early double dissolution in say late April. My suspicion is that an early DD (with all the pain of a shortened subsequent Senate term) can only be worthwhile if the government can bring on and have rejected (or at least have the Senate refuse to consider) its legislation for an Australian Building and Construction Commission (ABCC). If this does not happen before Thursday this week, I suspect the early DD option is also off the table.

Without a double dissolution, the normal election timing would see a Federal election on a Saturday on or after 6 August, in the second half of 2016.

Links

I previously wrote about double dissolution options for the current Parliament on:

• 2 January 2016
• 21 February 2016

Ipsos 53-47 (and a rant on polling methodology)

The Fairfax-Ipsos poll for March has been released. It estimated the national two-party preferred vote share for the Coalition at 53 per cent (a one percentage point improvement on the February poll).

Adding this poll to my aggregation sees no change. The Coalition's national two-party preferred vote share remains estimated at 52.3 per cent.

I saw quite a lot of twitter outrage on this poll as I went to bed last night. The outraged were asking, how could Morgan and Ipsos have the Coalition at 53 per cent, when Newspoll-2 and Essential have it at 50 per cent? The outraged were disinclined to believe the latest Ipsos poll.

I must admit my biases run the other way. Newspoll-2 and Essential do not appear to behave in a manner that is entirely consistent with statistical theory (with the caveat that Newspoll-2 is relatively new and the number of observations in this series is small). Both appear a little under-dispersed. There is less variation in the national population estimates from poll to poll than you would expect for the reported sample sizes, assuming the sampling frame is based on a random selection from the entire population. I am disquieted by this under-dispersion. It suggests there may be issues with the sampling methodology.

The magic of making statistical inferences for the entire population based on a small sample only works when that sample has been randomly selected from the entire population. The maxim I apply is when the poll results are too consistent from one poll to the next (for the sample sizes), then the poll results may be too good to be true.

No polling house achieves the gold standard of a randomly selected sample.

The use of telephone polling axiomatically excludes those voters without a phone. It may also exclude those whose phone numbers are unlisted. It is likely (for example) that voters in nursing homes will be under-represented. Excluding mobile phones will reduce representation from those aged under 35 years (especially those living in share houses without a land-line). But including mobile phones might see those young people living with their parents over represented (as they can be accessed by their parents land-line and their mobile phone). Not withstanding these problems, with good sample design, telephone polling gets closer to the gold standard than other approaches.

More challenging is the two-stage sampling frame when internet polling a panel that has been drawn from the whole population. The process of identifying and selecting such a panel is likely to result in biases within the panel. The selection bias here is that people interested in politics and people with good internet access are more likely to become panel members. This selection bias will then be magnified into the final poll results (especially if some panel members are regularly re-polled over time). While a pollster can normalise the internet panel results to the entire population (using the results of a general election or by running separate telephone polls), there is no guarantee that when there is a shift in voter sentiment, this shift will be equally reflected in the non-randomly selected internet panel.

The attraction of internet polling is cost. An internet panel poll is much cheaper to run than a phone poll (especially a human interfaced phone poll). But the downside for the cost equation may be an unquantifiable sampling error (which may move randomly over time as the internet panel is refreshed).

Why does this matter? In the United Kingdom last year they had a complete failure of opinion polling in the lead up to the general election. Pollsters in that case did not have representative samples. They did not include sufficient older voters in their samples, and their samples included too many politically active younger people (The preliminary findings of the inquiry are here, which have been reported in the Guardian here and the Telegraph here). The bifurcation we are seeing in the polls at the moment may reflect similar issues in Australian polling.

[For more information, see the links at the bottom of this page.]

Now that I have got that off my chest, let's turn to the other headline, which is the decline in the attitudinal polling for Turnbull (which has been more consistent).

Which can be aggregated as follows:

Links

• Good piece on Newpoll-2 methodology by John Stirton in the SMH of 9 May 2015. 
• A more detailed explanation on opinion poll sampling by Data Analysis Australia. 
• Kevin Bonham's excellent Field Guide To Australian Opinion Pollsters (from 2013).
• John Curtice's discussion about why the polls got it so wrong in the British general election. 
• Australian press council guidelines on reporting opinion polls.

Palmer on 0.3

Seeing Clive on Insiders this morning reminded me that I have not looked at his vote share for some time. So here it is.

From its peak at 6.7 per cent of the national vote share, 0.3 per cent is a substantial decline.

Notes on the Beta model I use for tracking Palmer United can be found at the bottom of this page.

BMU

Friday is my usual betting market update (BMU). Betfair has fallen below an 80% implied probability of a Coalition win at the next election.

	House	Coalition Odds ($)	Labor Odds ($)	Coalition Win Probability (%)
2016-03-11	Betfair	1.21	4.52	78.88
2016-03-11	CrownBet	1.12	6.00	84.27
2016-03-11	Ladbrokes	1.14	5.50	82.83
2016-03-11	Luxbet	1.16	5.20	81.76
2016-03-11	Sportsbet	1.17	5.00	81.04
2016-03-11	TABtouch	1.18	4.80	80.27
2016-03-11	William Hill	1.17	5.00	81.04

%	Fr 4-Mar	Sa 5-Mar	Su 6-Mar	Mo 7-Mar	Tu 8-Mar	We 9-Mar	Th 10-Mar	Fr 11-Mar
Betfair	81.83	82.97	83.33	83.33	83.33	82.49	81.15	78.88
CrownBet	84.27	84.27	84.27	84.27	84.27	84.27	84.27	84.27
Ladbrokes	82.83	82.83	82.83	82.83	82.83	82.83	82.83	82.83
Luxbet	82.83	82.83	82.83	82.83	82.03	82.03	81.76	81.76
Sportsbet	82.83	82.83	82.83	82.83	82.71	81.04	81.04	81.04
TABtouch	81.17	81.17	81.17	81.17	81.17	81.17	80.27	80.27
William Hill	82.71	82.71	82.71	82.71	82.71	82.71	81.04	81.04

Some betting market movements

There is some movement to report on the betting market in the past 48 hours.

In the next table we have the implied probabilities (expressed as percentages) for a Coalition win from each of the houses I track for each day over the past week.

	Th 3-Mar	Fr 4-Mar	Sa 5-Mar	Su 6-Mar	Mo 7-Mar	Tu 8-Mar	We 9-Mar	Th 10-Mar
Betfair	83.09	81.83	82.97	83.33	83.33	83.33	82.49	81.15
CrownBet	84.27	84.27	84.27	84.27	84.27	84.27	84.27	84.27
Ladbrokes	82.83	82.83	82.83	82.83	82.83	82.83	82.83	82.83
Luxbet	83.70	82.83	82.83	82.83	82.83	82.03	82.03	81.76
Sportsbet	82.83	82.83	82.83	82.83	82.83	82.71	81.04	81.04
TABtouch	81.17	81.17	81.17	81.17	81.17	81.17	81.17	80.27
William Hill	82.71	82.71	82.71	82.71	82.71	82.71	82.71	81.04

Poll Update No. 2

The release of Newspoll in today's Australian caught me a little by surprise, arriving just after I had completed yesterday's aggregation. I was expecting Newspoll to be released next week to coincide with the next Parliamentary sitting week (a typical Newspoll practice). Anyway it's out and it is unchanged on a fortnight ago (50/50). Dropping the result into the aggregation yields a national two-party preferred voting estimate of 52.3 per cent for the Coalition and 47.7 per cent for Labor.

It is worth noting that we are tracking the new Newspoll as being 1.3 percentage points biased towards Labor (in comparison with the other opinion polls, and assuming the house effects of all polling houses sum to zero). Much of that house-effect bias has been evident in the poll results since the elevation of Malcolm Turnbull (where Newspoll-2 has been consistently towards the bottom of the pile).

Turning to the primary voting intention model, again assuming that house effects sum to zero, we can see the following.

The other thing of note in today's Newspoll is the continuing decline in the attitudinal polling for Malcolm Turnbull. Let's start with the raw data before moving the attitudinal aggregation. Of note in the negative PM polling, whereas the former Newspoll tracked closely to the other pollsters, the new Newspoll is significantly more negative than its competitors.

Moving to the comparative results for both the Prime Minister and the Leader of the Opposition, we can see a small improvement in the net-happiness position of the Leader of the Opposition, a more significant decline for the Prime Minister, in the context where the Prime Minister is still well ahead overall.

Technical information on the polling aggregation process can be found here.