That 5% threshold and recent polling

Posted: August 18, 2017 in elections, New Zealand
Tags: , , ,

There has been some discussion on Twitter recently about how to describe what happens if the Greens don’t clear the 5% threshold, which is beginning to look like a real possibility and not just right-wing trolling. To illustrate this, I’m going to use the results of the recent Colmar Brunton poll, but I don’t actually think it’s definitive yet that the Greens are really polling below 5%1.

You will know if you’ve followed my blog since last election, or have seen my commentary from previous blogs or elsewhere online from even earlier, that I am very opposed to New Zealand’s current threshold, and believe it should be lowered to somewhere around 0.8%. (the amount to win a single list seat outright) These facts bear obligatory mention because we shouldn’t even need to have this discussion, as the Greens have enough support that they shouldn’t risk going under the threshold.

Back to the topic at hand, the first and most simple way to describe what happens if the Greens don’t clear the threshold is that essentially every Green vote doesn’t count, so is ignored for determining who wins List seats. It’s still worth voting even if you don’t want to give your Party vote to anyone else, as 4.3% (+/-1.25%) is well within the margin of error for getting over the threshold, and I think it’s likely that any polls that surveyed later than this one will probably show the beginnings of a rebound. The recent low Green polling seems to be largely (but not entirely) attributable to the Jacinda effect. A month, as this campaign has shown, is an incredibly long time in an election campaign.

The second description is regarding what happens to the seats. Let’s, for a moment, pretend that there is no threshold in New Zealand. (because that’s roughly what Parliament should look like at every election. You can also call this the “Green Party wins Nelson” scenario, if you like) This is what the recent Colmar Brunton poll would indicate, assuming Dunne loses in Ōhāriu, and Seymour wins in Epsom, and either Marama Fox or Te Ururoa Flavell wins an electorate:

CBpollaugAct: 1 seat
National: 55 seats
New Zealand First: 12 seats
Māori Party: 2 seats
Labour: 46 seats
Green: 5 seats
Total: 121 seats, 61 seats needed to govern

I’ll get to how this is calculated in a minute, but the upshot of this is that the Greens would win the 11th, 34th, 58th, 80th, and 102nd seats in Parliament. (Yes, each one is calculated separately)

If we add the threshold back in, we get the following results:

ACT: 1
National: 56
New Zealand First: 13
Māori Party: 3
Labour: 48
(Total/government threshold both as above)

Note that National, the Māori Party, and NZF each get a seat that “should” belong to the Greens, making things even easier for them in coalition negotiations. Labour gets 2. It is a little bizarre that National should ever get any seats that “belong” to the Greens, but by ignoring under-threshold votes, that means the seats are re-allocated somewhat proportionally.

Things are even more complicated, of course, than where the seats go, as you may have noticed that one of the Greens’ seats went to the Māori Party when we implemented the threshold rule, when they really didn’t have enough of the vote to warrant it. This is because the system we use in New Zealand isn’t actually governed by percentage of the vote.


This is the third way to describe what happens if a party falls under the Threshold- to get into the weeds on what the Sainte-Laguë method. (a mathematical system for allocating items using divisors) actually means. The image is a snapshot of my spreadsheet running the numbers for the seat calculation above (it’s a little primitive, I haven’t automatically coded it to cut off under-threshold parties yet, instead I just manually guarantee certain parties likely to win their electorates one seat in formulas, but it suffices with a little bit of checking-over for ties, (note there are two 116th seats here, those work fine as it will skip rank 117 in the formulas, but if there’s a tie for 120th it doesn’t note that, as there would be a tie-breaker then and it’s important to know who might lose a seat)

Essentially, we start off with the raw votes (it’s currently showing percentages multiplied by 1000, because it’s showing polling numbers, but in reality the first National Party number might by 1,000,000 or so) for each party, then we divide them by 3, by 5, by 7, and so on, until we’re satisfied we have a long enough list of numbers. (that’s usually by the time we get to dividing by 125 or so, which allows for 63 MPs for the largest party. If you were genuinely expecting a landslide you might go to) We then go through that list, and pick the largest number that hasn’t got an MP yet, and give that party the first MP, then the next largest for the second MP, all the way until we’ve got 1202. Because it’s sequential, it’s a little different to divvying up those seats according to the percentage of the total valid votes for over-threshold parties, as it doesn’t always exactly “round” the same way you would if you were looking at raw percentages of 120 seats.

If a party falls below 5%, we essentially just don’t bother to calculate their numbers for the list, and they get no seats allocated even though if we plugged them into the calculation they would likely be entitled to a fair number.

This divisor method is very friendly to small parties who aren’t discluded, but pretty proportional for parties that are polling in double digits. This is why I generally advocate for a 0.8% threshold- removing the threshold altogether would likely have awarded a seat to satire parties if people voted like they currently do, which seems a reasonable cutoff to me. There’s also something inherently fair about making it hard to win that first seat, but easier to grow larger from there.

So, in conclusion: don’t panic yet, and even if this poll result is real, there is plenty of reason to keep voting for the Greens, it just means there’s more campaigning to do, not the least is that if they don’t secure that 5%, my vote and plenty of other people’s votes won’t be counted towards changing the government, as I’m not Party voting Labour this election under any circumstances. And I won’t even start in the main article about the whole “I’m voting Labour so they’ll be large enough to form a Government” mess3.

Before your Regularly Scheduled Annotations, I will make a brief apology for not talking about other recent events in a timely fashion. I’ve had stuff going on personally that put me off writing about politics online for a while, but it’s now sorted.

1 Firstly, because the margin of error has plenty of room above 5%,
Secondly because this is only the first poll showing that and you need at least two polls to confirm this sort of thing, and
Thirdly because the Colmar Brunton is one of the least friendly polls to the Left in general, with Reid Research’s one coming in a close second, ie. TV news polls slant a bit to the right on average. That’s not to say they’re hugely likely to be out by more than the margin of error, but it is to say that they are more right-wing than the overall polling trend.

That said, Green supporters who want them in Parliament, and Labour supporters who don’t want Green votes wasted should be acting strategically right now as if this result is exactly as disastrous as it looks because you can’t afford to waste time when you might dip below the threshold.

There are likely several things going on here- I suspect the biggest one is that Ardern’s popularity is letting Labour eat a lot of the left and centre vote. This poll also started in the period where Metiria’s resignation was fresh, and a lot of the newer and lower-information Green supporters may have gotten the impression from bad news coverage that the Party had pushed her into resigning, rather than that she did so because she could no longer bear the disproportionately severe media attention on her private life. Add to that circular coverage regarding poll results, and it’s no surprise that this result was even worse than the previous one; I had expected polls to go down before coming back up, and I am genuinely hopeful that this will be the lowest poll the Greens will get in the campaign.

2 Or less if any independent MPs are elected. For instance if Raf Manji defeats Gerry Brownlee in Ilam, only 119 list seats will be allocated, as independents reduce the proportional size of Parliament instead of creating overhangs like electorate parties do. I’ll probably talk about this some more in an upcoming post, as I think it’s actually a bit of a silly rule.

3 I will of course concede that there are genuine reasons one might vote for Labour. Thinking they need to be the largest party in order to form the government is not one of them. It’s actually far more important that you give your vote to the party you want to have the strongest voice in government or opposition, rather than to the one you think is most likely to be the biggest party in the type of government you would like.

  1. […] Cross posted from […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s