Commentary: Predictions for the 2011 general election

Here are Horizon's predictions for the 2011 General Election.

We've waited until the last minute to post them, so as to take into account the latest polls by Horizon, Reid Research, Colmar Brunton, DigiPoll and other polling companies.

Polling the electorate during an intense election campaign is dynamic and changes have occurred in the past week so hopefully we're close to the mark on election day.

We predict a higher turn-out compared with 2008.

In our most recent poll, 81.4% said they will definitely vote and another 10% said they will probably vote. This was 6.3% up on the previous poll. Given the 1.7% margin of error for the survey, the rise in numbers voting could be 4.6%.

Intentions are always more definite than behaviour ("I'll definitely go to the gym more often this year!") but we think "definitely vote" is a good proxy for actually voting.

We think the turnout will be about 81% of the adult population. By the time you take into account population change, we think there will be over 2.5million votes cast for the first time in New Zealand.

In terms of MMP, we're predicting 56% OF THOSE THAT VOTE will vote to retain MMP, and 44% will vote to change. However, we also think at least 10%will not vote on that question. Probably more.

(Just over 20% of our panel said they were undecided on the issue only days out from the election, so things could change.)

Let's deal with the minor parties first, as whether they get across the 5% threshold or win an electorate seat affects the other parties.

Conservative: We think on the basis of our own polling, the Conservative Party will most likely not be in the next parliament. We have the Conservatives at under 5%, as a share of the definite vote expressed as a percentage of the total 18+ population, as do other polls we have seen.

While a mid-September poll by Research First put Conservative party leader Colin Craig ahead in the Rodney seat, we feel there is not enough recent information to say that he's going to win there. We think the party will get about 4% of the vote and miss out on Parliament.

Act: The most recent NBR survey put Act candidate John Banks behind in Epsom. Our polling as well as others’ shows Act below 5%, so we believe Act will not be in the next parliament. We think they will get around 2% of the final adult population vote.

Greens: We are showing the Greens at well over the 5% threshold, which is important as it seems unlikely they will win an electorate seat. We have them at about 12% of the "definitely will vote" so if that translates to election day, that is quite a lot. They do seem to be riding high at the moment, so maybe this will be a game-changing election for them.

New Zealand First: This is one of the main areas where our polls differ from telephone ones. We have them on between 7.9% and 11.3% - huge, compared to other polls, one of which has them at just over 5% (before applying that poll’s 3.6% margin of error).

We think 11.3% this may be too high, but that's what we're showing at November 24. We think there may be a group of centre-right voters who, while they look favourably on John Key and National, want assets to be retained. Our research shows older people in particular are opposed to assets sales, and this may be driving that. (48% of the population oppose National's asset-sale plan, whereas this percentage goes up gradually once people are above 65, until by age 75 and over, it's 66%).

NZF has one of the most loyal supporter-base, with a 79% retention rate from the previous election. (Labour and National are in the high 60s). They also appear to be picking up 6% of previous-National-supporters, which translates to an additional 3% on their 4% last election.

Labour: We have Labour on 29%, which is in the ball park of other polling. We think this is a reasonable estimate of its final election result.

United Future:  Peter Dunne's retention of his electorate seat is not certain, but we're predicting he will keep it. We don't think he will bring any other MPs into Parliament with him, as we have UF hovering around 1%.

Maori Party: We think it is likely that the Maori Party will win an electorate seat, but we have them at only 1% or so, so it looks like they will only get the electorates they win outright. Other polls indicate they’re likely to win at least three electorates.

Mana Party: We have them polling at under 5% (close to 3%) so it depends on whether they get an electorate seat or not. We think they will -Hone Harawira in Te Tai Tokerau - but that's it. However, this means he will bring another couple of MPs into Parliament house with him.

National: We have perennially shown National to be lower than the telephone polls show them. We have them in the thirties, at last survey most probably attracting 33.9% of the total adult population saying it they will definitely vote.

For the record, about 38% of the panel responding in our final pre-election poll said they voted National at the last election and this is slightly higher than National actually got.

(Remember that over 20% of the adult population didn't vote, and this is why that number looks lower that the figure people are used to.) So the panel is certainly not skewed left. If anything, it may be skewed slightly right, with results being corrected by our weighting system to provide a representative sample of the adult population.

As far as we know we’re the only published poll weighting responses by party vote 2008.

(At Horizon we monitor the percentages of people in various categories- like age, for example - and we recruit via the purchase of lists people who fit the criteria we want.)

So, what percentage of the vote do we think National that will get on election night? Given the higher turnout, we think they will get about 40% of the votes cast (in 2008 this was 45%) but will gain a few more points from the parties that don't make it to the threshold. (They will gain the most from these as the largest party on election night.)

Our polling shows that about 6% of people say they will make up their mind in the polling booth. And if that's stated behaviour - that people will admit to something socially frowned-on - then we think the real figure is higher. While those who didn't vote last time are showing signs of favouring Labour and the minor parties, we also think the "times are tough - don't change horses in mid-race" message by National will sway uncommitted voters to not risk another party. This may help National in the booth itself.

Our polling finds the factor exerting most influence on people’s party vote choice (about 90 in every 100 voters) is “confidence in a party’s ability to manage the economy”. Which might explain why National has placed such great emphasis on saying it is best economic manner in volatile times.

Once we take out the few extra per cent of people who vote for the very small parties we can apportion the seats in Parliament. We think there may be a very slight overhang as the Maori Party will win again more a higher percentage of electorate seats than the party vote. We're suggesting a 122-seat Parliament, the same as in 2008.

Based upon this, we are predicting a minority National-led government which will have arrangements with (mainly) New Zealand First and the Greens (to a lesser degree) to pass legislation.