Three new polls reinforce an unchanging overall political landscape and underscore a recent trend that is bad for Labour, good for the Greens, and bad for the left generally

Three new polls came out over the weekend. Collectively, they show National gaining further ground, increasing its support to level last seen in late 2009. All the smaller parties are slowly losing ground.

Labour lost about as much support as the Greens gained in these polls, confirming and evening strengthening a broad trend I identified last week.

Using some basic regression analysis on our poll-of-poll estimates, the chart above shows that the proportion of ex-Labour voters who are moving over to the Greens has been rising during the year. In the first half of the year, a 1% drop in the Labour vote was associated with about a 0.2% increase in the Green vote on average, while in the last three months a similar drop in Labour support now leads to around a 0.45% increase in Green support. This trend has three implications:

  • Those who think the drop in Labour support translates directly into a decreased probability of a change of government are being overly pessimistic;
  • Those who think the growth in Green support translates directly into an increased probability of a change of government are being overly optimistic.
  • An increasing number of left-leaning supporters prefer the Greens brand of people and policies to Labour’s brand of people and policies. But, as we will see below, there is not an increasing number of left-leaning supporters.

As the second chart shows, total left-leaning support has been in slow decline since around June after being more or less stable through the first half of the year.

This period of broadly-left decline is the same period in which the Greens have experienced their big rally. Taken together, these trends show that Labour is bleeding support from both edges, losing some left-leaning voters to the Greens and also losing some moderate supporters to either National or one of the small centrist parties.

In turn, this analysis may suggest that Labour’s is less a problem of ideological positioning than it is of communications and execution.

In addition, there was a Marae Investigates poll of Maori voters, showing an interesting race developing in Te Tai Tokerau, a seat crucial to the Mana party’s chances of getting back into parliament. Someone at TV3 didn’t appear to fully understand the difference between this poll (of one non-random subset of voters) and the other polls of the entire electorate.

I hope it was a short-staffed weekend and they had the cleaners doing the political reporting, because that kind of inferential error is pretty basic. For obvious reasons, the Marae poll is not in our poll of polls.

Comments (8)

by alexb on October 03, 2011
alexb

I disagree that a Green surge would not help change the govt, after all, the chances of the Greens going with National are very slim, and National's potential coalition parties are looking very shaky. A strong Green result would compensate for any awful showing by Labour, and could just sneak a broad coalition of left-ish parties over the line.

by Matthew Percival on October 03, 2011
Matthew Percival

alexb the votes from National's coalition partners are going to National, so effectively the bloc has the same percentage as before. So provided United Future, ACT and the Maori Party can each win a seat, a proposition which seems highly likely, National will lead a strong majority post election.

by Tim Watkin on October 03, 2011
Tim Watkin

It depends what you mean by a surge Alex. It would have to involve taking votes off National, something that simply hasn't happened to any significant extent so far this term.

Your broad left coalition is missing at least 5%, even if they can swing the Maori Party on board, which seems unlikely as long as Turia is in charge.

Rob, I'm curious about the NZF factor and the fact that it's so much bigger in the Horizon polls, which unlike the mainstream polls push the undecideds. Perhaps you could do a post sometime about what that all means.

by william blake on October 03, 2011
william blake

Traditionally the AB's choke in the quarters; this year they will have Daniel's groin injury as an explanation, however, J.K. smiling and waving with the victorious 'Boks will loose him that crucial 10%, forcing a Blue / Green coalition.

J.K. will wear hemp, Brownlee will mine himself and create a trade surplus, and in the distant future Dans groin will hold pride of place next to Phar Lap in that Australian museum.

by Rob Salmond on October 04, 2011
Rob Salmond

@alexb: What Matthew and Tim said.

@Tim: I don't put much faith in the Horizon polls, to be honest. They are so far out of whack with everyone else, and many of the existing pollsters have fairly good track records behind then, whereas Horizon's is short and spotty. Here's more from me on that:

http://pundit.co.nz/content/a-polling-scandal-and-a-saviour-on-the-horizon

http://podcast.radionz.co.nz/mwatch/mwatch-20110731-1008-mediawatch_for_31_july_2011-048.mp3  (the last 10 minutes of so)

So I'll wait until some of the traditional pollsters start showing NZ First pushing upwards before I sit up and take notice.

by Tim Watkin on October 04, 2011
Tim Watkin

I remember you writing that. (Although I don't listen to Mediawtach, as I'm working!). It was more the point about undecideds and specifically NZF that I was curious about.

Thanks for the NZF answer - perhaps something about how pollsters deal with undecideds at some point...

by alexb on October 04, 2011
alexb

Re, the left coalition being 5% short - It isn't inconcievable that Act will turn into 3% of wasted right wing vote if Banks fails in Epsom, which is very possible. As well as that, the Maori Party and United are both tied up in electorates that face strong challenges from Labour, meaning those seats could go. If NZ1st slip over 5%, thats another 6 seats, and if the 2002 effect (major party polling very highly losing out to minor parties on election day) hits National, with support going to the quite centrist looking Greens, then they could be looking at 15 seats. Add up 15 for Green, about 35-40 for Labour, Hone, and however many Winston drags in, and you are getting to about 60-61. Even if National gets up to 58-60 seats, they might find their partners have all vanished. Its requires a few ducks to fall into line, and would be a practically unworkable coalition, but I wouldn't bet too soon on a National coalition getting the numbers, given the failings of their support partners.

by Tim Watkin on October 05, 2011
Tim Watkin

That's a lot of very big maybes having to come in, Alex.

You're needing Labour and Green support to both grow, National to fall, ACT and United Future to fail despite National deals, and Winston to rise again. That goes against all the polling trends.

But you never know...

Post new comment

You must be logged in to post a comment.