RODNEY HIDE: Specials

The specials could sink Luxon's majority. They probably won't.

On election night the centre-right won. National took 42 seats, ACT 13 and NZ First 12. That's 67. Labour, the Greens, Opportunity and Te Pāti Māori have 64 between them.

The House is 131 seats, so a majority is 66. The centre-right has one seat to spare.

One seat is not much. And there are still 399,365 votes to count.

These are the special votes. They come from people who voted overseas, voted by phone dictation, voted outside their own electorate or weren't on the printed roll. 

The Electoral Commission estimates there are 399,365 of them, including 96,947 overseas and dictation votes, and it warns the total may change. They go into the official result on Friday 27 November.

RNZ ran a piece headed "How 399,000 votes could cost Christopher Luxon his majority". True, they could. I wanted to know how likely it is. So I ran the numbers.

Why the specials matter

Specials lean left. In 2023 there were 603,257 of them, and four electorates changed hands once they were counted: Nelson, Te Atatū, Te Tai Tokerau and Tāmaki Makaurau.

This year there are a third fewer. The rules changed. Enrolment now closes 13 days before election day, and same-day enrolment is gone. A lot of 2023's leftward lean came from people enrolling at the last minute. There will be fewer of them this time.

How we worked it out

We took the ordinary count as final. For every electorate, we estimated how many specials are coming, based on how many it had in 2023. Then we estimated how those specials would vote. 

We started from how the electorate voted on election night and shifted it by how far that electorate's specials strayed from its ordinary vote in 2023.

Because of the enrolment change, we didn't assume the specials would lean as hard as in 2023. We set the lean somewhere between half and the full 2023 strength. Then we added a dose of random luck, nationally and seat by seat.

Then we shared out the seats, as the Electoral Commission will. And we did that 40,000 times.

Think of it as rolling the dice 40,000 times and counting what comes up. It doesn't tell you what will happen. It tells you the odds.

Did it work? We tested the method by using the 2020 specials to predict 2023. It picked three of the four seats that changed hands. Not bad.

The answer

The centre-right keeps its majority on its own in 70 per cent of runs.

If the specials lean as hard as they did in 2023, that drops to 62 per cent. If they lean half as hard, it rises to 78 per cent.

What about the other 30 per cent? That's where Opportunity decides.

Most of it, 26 per cent, is a dead heat. Picture the centre-right on 65 seats in a House of 130. Opportunity could stop a centre-right government but couldn't put the left in.

In only 3.7 per cent of runs do Labour, the Greens, Te Pāti Māori and Opportunity have a majority together.

And Labour, the Greens and Te Pāti Māori on their own? Zero. Not one run in 40,000.

The overhang twist

National won 7 more electorates than its party vote entitles it to. That overhang is why the House is 131 and not 120.

So when National loses an electorate on specials, Labour gains nothing. Labour already has its full share of seats through the list. National loses a seat, the House shrinks by one, and the bar for a majority drops a little.

That's why 65 seats can mean three different things. In a House of 129, it's a majority. In 130, it's a dead heat. In 131, it's one short.

Our middle estimate of the final party vote has Labour on 29.2 per cent and National on 28.7. On our numbers, Labour finishes ahead on the party vote and still sits in opposition. National's party vote is buying it next to nothing. The seats were won in the electorates.

The seats to watch

Wairarapa is the big one. National leads there by 139 votes with about 6,500 specials to come. It's a coin toss. We give Labour a 51 per cent chance of taking it.

If National hangs on in Wairarapa, the centre-right majority is safe in 94 per cent of runs. If it loses, that falls to 47 per cent.

It's also our shakiest call, because the boundaries changed. When we used the old Wairarapa electorate's own 2023 specials instead, the chance of it flipping dropped to 34 per cent.

Next is NZ First's 12th seat. It's the last list seat in the share-out, and NZ First is fighting Labour and Opportunity for it. We give NZ First a 40 per cent chance of keeping it. If it does, the majority is near certain, 99 per cent. If it doesn't, it's a coin toss.

Mt Roskill is the outside chance. Labour leads by 262, and National takes it in 16 per cent of runs.

Wellington Bays will very likely go from Labour to the Greens. That changes nothing. The Greens' extra electorate just replaces a list seat.

My call

RNZ is right that the specials could cost Luxon his majority. More likely than not, they won't. Seven times in ten, the centre-right governs on its own.

But three in ten is not nothing. It's in play.

So come 27 November, watch Wairarapa and NZ First's last seat. If National holds the one or NZ First keeps the other, the centre-right is almost certainly home. If both go, start ringing Opportunity.

Tomorrow we will analyse what it means for Opportunity.  Spoiler: it is not good -- either way.

Rodney Hide is a former Minister and ACT Party leader