This analysis was first published exclusively in the Sydney Morning Herald and The Age on April 22 2016. Full results were published on this site on May 14 2016. An updated result will be released in the coming week.
The dynamics of the upcoming Senate election are unprecedented, with the combination of the newly introduced Senate voting system and the calling of a double-dissolution election. MetaPoll uses a unique method to calculate the likely outcome in the Senate this unprecedented set of circumstances. Our data sources for Senate modeling including the MetaPoll House of Representatives (HoR) primary vote aggregate (which is derived from a variety of sources listed elsewhere), State-level HoR polling from NewsPoll & Ipsos from the most recent quarter, Senate & HoR primary votes from previous elections, and NSW upper house preference flows from the 2015 election. The model translates HoR voting intention into Senate voting intention at a State level, and then distributes preferences based on the newly introduced Senate voting system.
The inaugural Senate MetaPoll result shows both major parties failing to win an outright majority of seats. The Coalition would have lost two seats, from 33 seats to 31 seats. Labor would have gained three seats, from 25 seats to 28 seats.
On the cross-bench, the Greens would have gained one additional seat, bringing their total to 11 seats. The Nick Xenophon team would have quadrupled their current representation, picking up a total of 4 seats (one third of the SA seats). MetaPoll predicts that Jacqui Lambie (TAS) and Glenn Lazarus (QLD) will be the only other crossbenchers to retain their seats (see methodology section below).
The chart below shows the breakdown of results by State. The Coalition performs best in Western Australia with six seats, while Labor performs best in NSW, Victoria and Tasmania with five seats each. The Greens pick up two seats in each State with the exception of SA, where they and the major parties suffer at the hand of the Nick Xenophon Team. Unsurprisingly, the ACT and NT remain unchanged from the previous election with one seat each to the two major parties in both cases.
Unlike previous federal Senate elections, voting reform has meant that votes will be allowed to exhaust (and this change was upheld in a recent High Court challenge). While voters will be instructed to number at least 6 preferences if voting above-the-line, it remains to be seen how closely this will be followed. Given that voters are familiar with only numbering one box above the line, combined with the fact that there is nothing preventing parties from issuing ‘vote 1 only’ how-to-vote cards, we expect there to be a solid proportion of votes that are exhausted. In the NSW upper house, there are no instructions to nominate 6 preferences and around 80% of people choose to vote 1 only. For our model, we assume that after first preferences, 40% of votes then exhaust (around half of that of NSW). It is worth noting that ABC election analyst Antony Green has said that he expects the exhaustion rate to be closer to 20%. We have conducted sensitivity analyses around exhaustion rate and a lower exhaustion rate of 20% would have no impact on our final seat result for this particular release.
Preference data from previous federal elections is now redundant, as past preference flows are more reflective of backroom party deals than true voter intention. In our model, we use NSW upper house election data as our primary source for preference flows, as the voting system for that house is the most similar to our new federal Senate system. The one exception to this is in South Australia, where we have adjusted the flow of preferences to reflect the strong presence of the Nick Xenophon Team.
The core model is focused on what are expected to be the four main Senate parties: Coalition, Labor, Greens and Xenophon. For other minor parties and independents, we treat each on a case-by-case basis based on a variety of factors. Based on their 2013 primary vote and their strong personal brands, MetaPoll believes that Jacqui Lambie (TAS) and Glenn Lazarus (QLD) are the only candidates likely to win seats outside of the four main parties. The Liberal Democrats in NSW had a strong primary vote in 2013 off the back of a favourable ticket position and voter confusion over the similarity of the party name to the Liberal Party. Due to the combination of that ticket position being unlikely to be repeated, combined with the inclusion of party logos on tickets as part of Senate reform, we do not believe that their primary vote will be sufficient in this election to retain a seat.
Our method of extrapolating from lower house voting intentions given past Senate voting trends is the best available method in the absense of Senate-specific polling. Since this analysis was conducted, some Senate polling have been conducted by other firms. Additionally, MetaPoll plans to conduct its own Senate polls. The inclusion of these will improve the reliability of future releases.