At least 225,000 Nats said “No” to asset sales

Rob Salmond  is currently running a business selling data based political advice. But he also runs the Polity blog and has given us permission to syndicate his posts when we find something interesting to repeat. Long time readers will remember him from the brilliant posts on 08wire.org during bthe 2008 campaign, so expect us to often find something of interest. Like this post looking at who voted No on the referendum. That number of previously National voters must be scaring the National party at present.  

The results are in on asset sales, with an overwhelming rejection of National’s programme. I estimate at least 225,000 of the No votes came from National voters.

Among the very strong rejection of National’s asset sales programme is another problem for the government – about a third of the votes against them likely came from their own 2011 supporters. In this post I outline an evidence-based estimate about what kind of people voted in the referendum, and how their asset sales position lines up against their 2011 ballots.

Which people voted?

Turnout in the referendum was 44%, which is very good for a postal ballot just before Christmas in conjunction with no other election. But who are they? You have to be something of a political animal to vote in a referendum like this one. A good working assumption is that these were the 44% of people who are the most interested in politics.1

Who did those people support in 2011?

The New Zealand Election Study can give us a good idea who the likely referendum voters supported in 2011. 44% of the population represents all the people in the NZES who said they were “very interested” in politics, along with a little under two thirds of the people who said they were “fairly interested.” That group of the population has a partisan preference broadly the same as the rest of the population, as you can see in the Table below.

Party in 2011 Estimated Support across… Estimated referendum voters
Likely referendum voters New Zealand
National 48.5% 47.3% 646,000
Labour 26.0% 27.5% 347,000
Greens 13.4% 11.1% 178,000
NZ First 7.5% 6.6% 100,000
ACT 1.3% 1.1% 17,000
United 0.5% 0.6% 7,000
Maori Pty 1.7% 1.4% 23,000
Mana 1.1% 1.1% 15,000

How did each party’s former supporters vote?

We know for sure that 895,000 of these people voted “No” to asset sales, while 433,000 voted “Yes.” So what is the minimum number of National supported who must have voted no? Let’s start by assuming that absolutely all the 2011 Labour, Greens, NZ First, United, Maori Party, and Mana voters voted “No.” That is probably an overstatement. Then assume all the ACT voters voted yes.

That adds up to 17,000 Yes votes and 670,000 No votes, leaving only National voters unassigned. As the Table below shows2, at least 225,000 National voters – over a third of all the National supporters who likely voted, said No.

Party in 2011 Likely referendum votes
Yes No Total
Labour 0 347,000 347,000
Greens 0 178,000 178,000
NZ First 0 100,000 100,000
ACT 17,000 0 17,000
United 0 7,000 7,000
Maori Pty 0 23,000 23,000
Mana 0 15,000 15,000
National 416,000 227,000 646,000

This has to be enormously concerning for National. The people who voted for National then rejected National’s flagship policy are very much at risk come 2014. Many are likely centrists, part of the “squeezed middle”who National hasn’t delivered for so far this term. They’ve now already rejected National at the ballot box once.

If National loses these same people to the left again in 2014, it is curtains for John Key.

UPDATE: There are also good clues from the detailed results that the turnout was not just leftie strongholds lining up to say no. Cases in point: National strongholds Clutha-Southland, Taranaki-King Country and Wairarapa had above average turnout, while Labour fortresses Mangere, manukau East, and Manurewa all came in with less than 30% turnout.

1. I know Labour’s turnout targeting was focused on people who we thought had high political interest, so Labour’s campaign didn’t expand the turnout universe into low interest people

2. Some numbers in the table below do not sum properly, due to rounding

Powered by WPtouch Mobile Suite for WordPress