The Progress Report

I’ve updated my Election Forecast with the results of the final Colmar Brunton and Reid Research polls. Either National or Labour could form government, but they would both need the help of NZ First, Labour would also need to add the Māori Party to the mix.

My forecast has the Greens creeping closer to, but still below, the 5% threshold. I feel that my weighting against them has been too aggressive, and I would be surprised if they don’t make it back to Parliament.

While National have gained ground in the last two weeks, it will still be a very, very close result. Labour will make massive gains, and with a good ground game on Saturday could still take the prize.

It’s certainly been the most interesting election I remember in years.

My genuine thanks go out to all the candidates, staff and especially volunteers. No matter what party you stand for, the work you’ve done over many months ensures that New Zealand has a healthy democracy. Get some sleep on Sunday!

Electorate level polls in New Zealand are problematic.

They almost always use tiny samples, and pollsters have significant problems knowing the people they’re calling are actually enrolled in the seat they’re asking about.

And taking all that into account, it’s never done regularly so you can’t even look for a simple trend.

So, for those reasons, predicting electorate level results is difficult at best, and it’s why I haven’t built any change in electorates into my Election Forecast.

However, I promised a few people I’d write a list of which seats I think will change hands next Saturday. Here they are:

Auckland Central – Labour gain from National

Ōhāriu – Labour gain from United Future

Maungakiekie – Labour gain from National

Christchurch Central – Labour gain from National

Waimakariri – Labour gain from National

Whanganui – Labour gain from National

Waiariki – Labour gain from Māori Party

 

So there you have it, seven seats I think Labour will pick up on the day. And I no longer think any of Labour’s seats (eg Hutt South, Te Tai Hauauru, Napier) are at risk.

If I’d written this list two months ago it would have been very different!

I’ve updated my Election Forecast with the results of tonight’s TVNZ/Colmar Brunton poll. The update results in Labour gaining two seats, one from National and one from the Māori Party. The Greens are still forecast to be under the 5% threshold.

This weekend I’m going to go through my data and make my predictions for which electorates are going to change hands. I’ll be very interested to see what others guess as well.

Nine days to go.

I’ve updated my Election Forecast with the results of tonight’s Colmar Brunton poll. While the gap has significantly closed between National and Labour, the Nats are still slightly ahead in my forecast. Both major parties could form a government with the support of NZ First.

I’ve also done a breakdown of who would actually make it into Parliament on these numbers. With Labour polling almost 20 points higher than they did in 2014, it’s a fairly safe bet that they’d pick up additional electorate seats. For the purposes of this I’ve made a guess that they’d pick up Auckland Central, Ōhāriu, Maungakiekie, Christchurch Central and Waimakariri – which were all less than 3,000 votes out of reach in 2014.

If you disagree on which seats would change hands, you can use my National and Labour calculators to see who would make it in at any level of support.

Here’s what Parliament would look like:

Continue reading “Election Forecast updated – and what Parliament would look like”

I’ve updated my Election Forecast with the results of two recently released polls – one from UMR which ended three days ago, and one from Roy Morgan which ended on the seven days ago.

Neither of these polls taken individually offer much interesting new information – they both show a closing of the gap between Labour and National as swing voters start moving towards Jacinda Ardern’s Labour Party.

However, what is interesting is that adding these two new data points has tipped my election model substantially.

It’s designed so that individual polls can’t have considerable impact, but the sheer weight of support now behind Labour has meant that with these two new polls my forecast for Labour seats has sky rocketed from 36 to 44 – with them coming from National, NZ First and the Greens.

On these numbers, Jacinda Ardern could form a government with the backing of NZ First and the Greens, and Bill English could form a five-headed hydra of a government with ACT, the Maori Party, United Future AND NZ First.

I would be very worried if I was a National back bencher.

 

 

I’ve updated my Election Forecast with the results of today’s Colmar Brunton poll. It’s good news for Labour, and bad news for National and the Greens.

You may have noticed that Colmar Brunton have also done an electorate-level poll in Ohariu which puts United Future’s Peter Dunne 14 points behind Labour’s Greg O’Connor. It’s only one poll with a small sample, but it does not surprise me. If there is any further polling in that seat showing similar results, I’m going to update my forecast to give the seat to Labour, meaning United Future and their over-hang seat will be gone from Parliament.

I’ve made two changes to my Election Forecast.

Firstly, I’ve managed to import all the UMR polling data that has been released into the public domain, going back to January this year (though polls older than 90 days will have no impact on the forecast). Given they are polling at least fortnightly I’m really hoping their results continue to be released.

Note that UMR polls don’t release numbers for ACT, Maori Party, United Future or any other minor parties.

I’ve also fixed a bug that was impacting the weightings. The impact of this will be that the forecast will be more heavily impacted by newer, bigger polls. Older and smaller polls won’t have as much of a factor in the result. While it wasn’t my intention, this means the forecast will start to show the impacts of the Jacinda Ardern affect sooner.

Bring on more polls!