Electoral projections Archive

NSW headed for colossal landslide

By
Ben Raue
 – February 17, 2011

Yesterday’s Nielsen poll has Labor still on track for a catastrophic defeat. Along with the January Galaxy poll, the Nielsen poll has the ALP polling only 34% of the two-party-preferred vote, compared to the Coalition on 66%. On primary votes, the Coalition has hit a new high with 53%, compared to 51% in the Galaxy poll, with the ALP on 22% (compared to 20% in Galaxy) and the Greens on 13% (compared to 15% in Galaxy).

There is a general consensus that such a result would produce a massive defeat for the ALP, way out of proportion to the voting figures. Antony Green’s swing calculator has the ALP winning only 14 seats compared to 73 Coalition seats on a uniform two-party-preferred swing. In the next election, however, many contests will not be between Labor and the Coalition, with Greens and independents coming in the top two in many seats in Sydney and the country. In addition, two-party-preferred figures in polls are based on preference flows from minor parties remaining consistent. Yet it appears that the Greens preference flow to Labor will be greatly diminished. Both of these factors suggest that the impact could be worse than Antony Green’s calculator predicts.

I developed my own calculator which instead calculates swings on primary votes, based on proportional swing, which means a party’s vote will swing more heavily in areas where their vote is higher. This reflects the expectation that the swing against Labor will be more heavily concentrated in its heartland and marginals, rather than in Coalition seats, and that the Greens vote appears to have the most potential to grow in areas where it is already strong.

Before laying out what my calculator produced, it’s worth noting that many flaws still remain. Like a simple pendulum calculator, it relies on the concept of a uniform result. It doesn’t take into account the abilities and appeal of individual candidates, either at the current election or at the last. As an example, Macquarie Fields appears much more marginal than its neighbouring seats of Campbelltown and Liverpool largely because of the 2005 by-election and 2007 election which saw a particularly strong Liberal candidate and a local ALP hit by repeated scandals. It is unlikely to experience such a strong swing as other seats that have not previously swung so hard.

Neither calculator can factor in the strong Liberal candidates in Keira and Cabramatta, both of which are some of Labor’s safest seats in the state on paper. It is a particular problem when it comes to independents. The calculator assumes that defeated independent MPs will run again in Pittwater and Manly, while ignoring the independent candidate in Wollongong who many are tipping to win the seat.

In addition, the calculator allows the user to make changes to the estimated preference flows in a contest between any two of Labor, Coalition, Greens and Independents. These assumptions could be wrong. For example, I assume that 30% of Greens preferences will go to Labor, and 15% to the Coalition, which is substantially less than in 2007.

Having said that, the result the calculator produces is:

  • Liberal – 56
  • National – 19
  • Labor – 10
  • Greens – 3
  • Independents – 4

After the break, I break down these figures, map them out on a map of NSW, and give you a link to where you can download the calculator yourself.

Read the rest of this entry »

Tags: ,

Margins in strong Greens seats

By
Ben Raue
 – February 18, 2009

There are a handful of seats in the House of Representatives where the Greens are within spitting distance of winning the seat. In nearly all of them, however, the true margin is hidden by the AEC’s methods of measuring marginality. Margins are calculated by referring to the two-candidate-preferred vote, but this doesn’t take into account the possibility that a candidate in third place may overtake the second-place candidate and be in a winnable position.

Greens marginal seats run into what psephologists call “non-monotonicity”, whereby one group’s preferences are not reciprocated. In monotonic system, a candidate cannot be harmed by receiving an extra vote. In recent federal elections, Liberal preferences in strong Greens seats have been directed to the Greens over Labor, putting the seat in a position where a Green in third place might win against Labor in a 2CP count, on Liberal preferences, but the preferences are not reciprocated, with Greens voters preferencing Labor over Liberal. Thus a seat may be registered as a 19.5% margin for Labor, as in the case of Sydney, even though the Labor vs Greens margin would be much smaller. Even if the result isn’t reversed by a change in the order of elmination, it’s possible that a seat can shift from “Very Safe” to “Marginal”.

At the 2007 election, only one Greens candidate managed to come second after preferences, in the seat of Melbourne, where the Greens overtook the Liberals on preferences from minor candidates and registered over 45% of the two-party-preferred vote, making the seat marginal. But there are a number of seats that fit the criteria of Greens marginals but don’t appear as such on the pendulum. These criteria are:

  • A reasonably high Greens vote (above 15%).
  • A large gap between the two major parties, with the weaker of the two (usually the Liberal) very low (usually below 30%).

The Greens managed to attract 82.5% of Liberal preferences in Melbourne at the 2007 election. In order to determine the next five most marginal Greens seats, I translated this proportion of Liberal votes to the Greens, to determine a two-party-preferred vote between Labor and Greens. I excluded all seats where the margin between Liberal and Greens is greater than the Labor-Greens margin. For example, in the case of Grayndler, I calculated a 60-40 margin, and the Liberals beat the Greens by 2%. I assume that if the Greens were to gain 10% of the 2PP vote they would manage to overtake the Liberals.

Anyway, here is my calculations.

  1. Melbourne – 54.7-45.3
  2. Sydney – 55-45
  3. Grayndler – 60-40
  4. Cunningham – 61-39
  5. Denison – 61-39
  6. Batman – 62-38

Comments, anyone?

Tags: , ,

UK election animation

By
Ben Raue
 – February 17, 2009

ukmap

I’ve come across an interesting election animation, showing the electoral map of every UK election since the Reform Act of 1832, showing the geographical distribution of Conservative, Labour and Liberal/Liberal Democrat seats since that time as a cartogram.

You can clearly see the shift over time, with the appearance of Labour in the late 1800s and the decline of the Liberals from the time that the universal franchise was introduced in the early 1920s.

Tags: ,