Today we shall have a look at the recent Victorian state voting intention polling which has been produced in the past few weeks. From the state-wide polls we shall make another of our state projections.
There have been two state-wide polls released by Newspoll and Morgan which have both shown a swing away from the Liberal-National Coalition since the last state election. A Newspoll of NSW state voting intention with 1288 respondents was published in The Australian on Tuesday. The polling was taken over the past two months. Roy Morgan also released a Victorian state voting intention poll taken over July and August of 479 Respondents.
Below is a one off poll average using both polls weighted by sample size. The table of the averages is below and a second table indicating the swings since the last election on primary vote and TPP.
Estimated preference flows from the 2010 state election (Green: 76-24 ALP, Others: 53-47 Coalition) are used to calculate the state wide TPP vote
As you can see there has been a minute swing away from the Coalition since the last election. However given the relatively close TPP results at the last election and the bare Coalition parliamentary majority of 2 that the previous election returned, any swing may blow the Baillieu’s government’s majority over. Therefore all eyes turn towards the seat projection.
Given that Victoria uses compulsory preferential voting, the seat projection is more like its federal counterpart with a simple application of TPP swing. A probability is calculated for each seat in the same manner of the Australia poll average using a standard deviation of 3%. The sum of the probabilities for each party is the number of seats each party wins in the seat projection.
In Victoria whilst there are no crossbenchers we are confronted with the 4 inner city seats of Melbourne, Richmond, Brunswick and Northcote where the Greens finish in the TCP count. We shall treat these seats as crossbench seats and will not award them against the incumbents unless there is specific polling evidence in the form of seat polling to confirm a change of MP. These 4 seats are outside the pool of Labor/Coalition seats where the respective probabilities are summed. It should be noted that given Green preference flows between the majors, the Liberals would not register any probability to win these seats given current state-wide party support levels.
We set out the seat projection below.
The result is a slightly increased Liberal-National Coalition majority of 4 despite the slight TPP swing away from the blue team of 0.6%. Why is this so?
As noted above the seat projection is calculated by estimating the probabilities of each of the major parties to win each parliamentary seat. The probabilities are summed and rounded to the nearest whole number to provide a number of seats won by each party. The probability is calculated by using a cumulative probability function with a standard deviation of 3%. The practical effect of this method is seats which are close to 50% after the addition of the state wide swing, will not have a probability of 1 of 0 but of a fraction in between.
Therefore if after the application of the state wide swing there more seats close to 50% on one side or the other this will tip the projection by a seat or two to the party with the MPs on the less vulnerable margins.
The current distribution of seats in Victoria is a classic example. On the current pendulum, Labor has 9 seats under a TPP margin of 2%. The Coalition only 3 seats in a similar situation. If there is a negligible swing away from the Coalition it is still possible that given the variability of the swing that the Coalition can pick up a Labor seat or two.
The reason for this complexity is it is designed to reflect that fact that a uniform swing whilst a simple and valuable theoretical tool does not strictly exist in practice. Swings are generally normally distributed with a tight standard deviation of 3%, sometimes larger in smaller contests or in elections with large swings (i.e. NSW 2011).
A final note about sophomore surges
What about sophomore surges? As noted in our last post we have not included this factor in the seat projection simply because only just before the next election will we know which MPs are contesting and which MPs are going to call it a day.
Nevertheless given the tight contest we estimate how the sophomore surge may affect the result. We shall assume for this exercise that all incumbent MP’s will recontest their seats.
Sophomores are first term MP’s that recontest their seats. They are said to be contesting elections for a second time and thus this is where the name sophomore comes from as the American term for a second year high school/college student. Currently there are 25 seats which have first term MPs, 15 Coalition, 10 Labor.
Most Coalition first term MPs won their seat off Labor at the last election, 12 of the 15 Coalition first term MP’s are on margins of 8% or less. On the Labor side, 5 of their first term MPs are on margins of 8% or less versus the Coalition.
Therefore the Coalition gets a slight advantage under the sophomore surge. A table is set out below of the projected seat numbers under different sized sophomore surges.
As you can see when the sophomore surge hits three percent the Coalition gains an extra seat of Labor under the projection.
All in all the seat projection is indicating a tight election with an edge to the Coalition if one was held today. However there is no election today. The next election is in 2 years and will be held on redistributed boundaries which will shift the projection parameters. However the above is our best estimate of the current state of play.