November 9, 2023

Labour lost their special votes advantage

The final NZ election results are out, and my idle spreadsheet-fuelled predictions are proven partly but notably not entirely correct. Most significantly, Labour almost entirely lost their special votes advantage. Since 2005, Labour had always increased their vote when the final results including specials were unveiled, and actually they’d been getting better and better on specials, peaking in 2017. This year Labour did almost exactly the same on specials as on preliminary votes. So this was the worst they’d done on specials since 2002.

Here is how every party did, and then a comparison to other MMP elections…

The other noteworthy factor was that National reversed their downward trajectory on special votes, getting their best performance on specials since 2008. However, they still dropped, as they always have. Even in a good year for them on specials, they still did badly enough to lose two seats from what they had been projected on preliminary results. National losing multiple seats on specials will be increasingly normal if the proportion of votes that are specials remains high or continues to grow.

Other parties did basically similar to how they’ve done before: TPM, Green, TOP, Legalise Cannabis, and conspiracist parties doing well on specials (though some not as well as before), ACT and NZF doing badly.

So, in summary, this led to the following changes in seat numbers from the party vote (with self-indulgent comments about how this compares to the predictions in my last blog):

Parliament will have 122 seats, with one more to come after the Port Waikato by-election. National + ACT will not be able to form a government on their own, as they will only have 59 seats (to grow to 60 after the by-election).
In theory there are multiple possible governments, the first three of which avoid New Zealand First:
- Teal Deal: National + Green, 63 seats (to grow to 64)
- Grand Coalition: National + Labour, 82 seats (to be 83)
- Getting the Old Gang Back Together: National + ACT + TPM, 65 seats (to be 66)
- Winston Kingmakes the Left Again: Labour + Green + NZF + TPM, 63 seats.
- Three rightwing parties that dislike each other: National + ACT + NZ First, 67 seats (to be 68). In practice this is the government we will get, because all other combinations involve parties that have (explicitly or implicitly) ruled each other out.
Labour did not pick up any seats on specials.
- This is contrary to most of my predictions.
National lost 2 seats on specials.
- This is consistent with some of my predictions (though other predictions were that they would lose 3, which would have been nice but aah well).
The Greens picked up one seat.
- This was predicted in most of my predictions. On some of my predictions they stayed on 14 and on others they went up to 16 but most said they’d go to 15.
TPM picked up one more seat on party vote entitlement (however, since they won more electorates than party vote entitlement, this actually doesn’t affect their own number of seats in Parliament, just the overall size of Parliament, as discussed in probably excessive detail in the last blog).
- This was predicted in most of my predictions.
No other party changed their seats from the preliminary predictions. ACT and NZF went down on specials as usual, but not enough to lose seats.
- This was predicted in basically all my predictions.
- There was one super hypothetical scenario on my predictions where ACT actually gained, but this coming to pass would require ACT to magically become good at specials again (and since they became a more-than-one-person party again in 2020, they’ve been bad at specials, just like the other medium-large right-wing parties are). NZF have been bad at specials for ages (retirees don’t do special votes). Realistically, ACT and NZF were definitely going to drop their vote on specials, not increase it.

How close were we to these seat numbers being different?

Changes in seats (or no change in seats) doesn’t necessarily clearly translate to change in votes, so generally it’s better to look at percentage changes in votes. Sometimes quite a small change in votes could push a party over the boundary to get one more or one fewer seats.

I’ve played around a bit, tweaking the the final numbers on the seats generator to see what kind of vote changes could have led to different seat numbers. Here are some ways that votes could have been different, leading to different seat numbers, in order of likelihood:

If Labour had taken 0.2% of National’s total votes and 0.1% of the Greens’ votes, Labour would have won the extra seat instead of the Greens.
If NZF had dropped 0.2% of their vote, to 5.88%, they would have gone down to 7 seats. That seat probably would have gone to National; could have been Labour, but only if almost all of these 0.2% votes went to Labour instead). But NZF still would have held the balance of power (National and ACT would have only had 61 out of 123 seats even after they get a bonus seat with the Port Waikato by-election).
ACT had gone down from 12 to 11 in the last stages of the election day count, so they would have had a fair way further to descend for this to drop to 10. But even if it did, the votes & the seat would have realistically gone to National anyway, so wouldn’t have changed the overall size of the right bloc.
While some scenarios had TPM staying on only 3 seats from party vote, realistically they would always have increased this to 4 as they always do so well on specials. Even if TPM had dropped 0.3% to 2.78% with (for example) 0.1% each going to Green, Labour, and National the seat numbers don’t change.

How special votes affected electorates

Māori electorates: I said that Te Tai Tokerau and, likely, Tāmaki Makaurau were too close to call based on past special vote performances. 2017 and 2020 results ever-so-slightly favoured Labour’s Kelvin Davis to narrowly hold onto the seat, but with lots of uncertainty. (I lacked the data to do a projection for Tāmaki Makaurau.) It turned out Te Pāti Māori won both seats. In Te Tai Tokerau, TPM’s Mariameno Kapa-Kingi actually turned a 500 vote preliminary deficit into a 500 vote win by winning 3000 extra votes from specials to Labour’s Kelvin Davis’ 2000… big numbers considering Kapa-Kingi won with only 10,428 votes total. In Tāmaki Makaurau, the margin of victory for the TPM candidate is only four votes… yes, four single votes… so this will go to a recount and is essentially a tie. But TPM’s performance especially in TTT was likely helped by TPM delivering their normal excellent special votes performance while Labour failed on specials. Here are tables with the Te Tai Tokerau results this year and compared to 2017 and 2020.

General electorates: I said in my last blog as a vague hunch that one or two seats would probably flip from National to Labour on the final count. In fact two did (Nelson and Te Atatū). I also added a more specific guess on the morning before final votes were announced when I saw a list of four seats identified as close (the Labour candidate just ahead in Mt Albert on preliminary votes and the National candidates just ahead in Nelson, Te Atatū and Banks Peninsula). My last minute guess was that Labour would win all four of these. I was right about three out of four but in Banks Peninsula, the National candidate actually extended their lead on the final count and won the seat by a relatively comfortable margin of 396 (after being 83 ahead on the prelim count).

Why the changes this time?

This is all entirely speculative.

I have three possible theories about why Labour did no better on specials than ordinary votes this time, and the first two of them would also be reasons why National did less bad on specials than they had been doing. There are not really any other significant changes to explain about other parties.

A theory that some pundits were predicting was that overseas voters would abandon Labour because they were annoyed about the border closures and compulsory user-pays quarantine for people wanting to come home. This may explain some of it, even though overseas voters are only 78,030 or 12.9% of special votes. I will admit I have been sceptical about this theory, because I was an overseas voter who suffered a government that literally wanted to let the elderly die (and to a large extent succeeded in that), looking in at NZ’s best-in-world COVID policy and feeling more warm towards NZ Labour than I probably ever had before. Also, I had no intention to come back while MIQ was in place. However, I am willing to grant that the picture would have looked quite different to an overseas voter in Australia whose own government didn’t do too bad (relatively), who has always expected it would be relatively easy to pop back home, and who really wanted to come home in 2021 when a lot of other countries were giving up on COVID protections but Aotearoa wasn’t. So maybe there is truth to this pundits’ theory.
Labour seemed to narrow the gap with National near the end of the electoral campaign when Hipkins finally had a good debate and National were going off the rails saying, basically, that the government with ACT and NZ First that they were (and are) set to form would be bad and that NZ First are racist but that they will work with them anyway, etc. Labour did somewhat better and National did somewhat worse on votes cast on the day than on advance votes (compare: advance ordinary votes with preliminary votes). So IF special votes are more likely to be advance votes, then this could explain why special votes were more National-leaning than Labour-leaning. BUT I don’t know if special votes are more likely to be advance votes or not. I’ve asked the Electoral Commission if this info is available (yes. I am a nerd.).
A lot of special votes are students or people who move around a lot, who are staying outside the electorate in which they’re enrolled when the vote. A lot of others are first-time voters who enrol and vote on the same day. (I’m not sure if we have the data on how many of each group there are.) These groups are generally left-leaning and may have taken issue with Labour watering down climate commitments, not delivering on Ardern promising to prioritise it as “her generation’s nuclear free moment,” again ruling out any kind of wealth or capital gains tax, and failing to do anything about housing unaffordability. Ardern was a cautious centrist Labour leader, more timid than Helen Clark even after getting an unprecedented majority vote in 2020, and Hitchens pitched himself further right than that, offering no hope of any action on the major issues affecting the world and the future. Some of these voters may have gone back to the Greens or TPM (or voted Green or TPM in their first election instead of voting Labour). Perhaps similar to the demographics who (on both ordinary and special votes) abandoned Labour for Green or TPM in electorates like Rongotai, Wellington Central, and the Māori electorates.

October 29, 2023

National could drop 3 seats on the final count y’know, or: Blog in Which Spreadsheets Drag Me Back Into NZ Political Blogging after Three Years

Warning/promise: lovely spreadsheets and infodumps follow. Since I last blogged I’ve learned I’m autistic, and I’m leaning into it.

We’re most likely staring down the barrel of Winston Peters, noted liar and stoker of conspiracism and surely Aotearoa’s most powerful politician of recent decades, again entering the government. If he keeps his word (big if) and if Labour, Green and Te Pāti Māori do, he will not this time be playing the left and right blocs against each other. Instead Winston First (H/T my friend Peter for that name) will “kingmake” Christopher Luxon as PM, with ACT also playing a role in the government. This may mean NZ First and ACT cancel each other out on the severity of neoliberal capitalism, but grasp hands and go forth in unity on opposing Te Tiriti. It will mean a lot of clashing egos.

However, NZF holding the balance of power wasn’t immediately clear when the election results came pouring in, as early results had National and ACT with an outright majority on their own. In fact, even at the end of election night, provisional votes showed Nat+ACT with the slimmest possible majority (61/121 seats) if final vote %s match provisional vote %s.

But the final vote % never do match the provisional vote %s. The release of the final count with all the specials always leads to different results to those projected from preliminary results. National will almost certainly “lose” seats (i.e. end up with fewer seats than the 50 they are projected to get on prelim votes) – unless things go very differently this election than previous ones. This much is clear to anyone who’s ever been nerdy enough to play with election results in a spreadsheet.

This blog represents the results of me doing said spreadsheeting (updating my analysis from six years ago: 1, 2). If you are interested in looking at my spreadsheet for yourself, it’s here.

You may also enjoy the work of Graeme Edgeler or the NZ Herald. Full disclosure: Edgeler seems to have different numbers to me from the 2020 election (eg he says National did 22% worse on specials than preliminary vote, whereas I make it 26%). I can’t figure out whether we’re using a different calculation or slightly different category of votes or whether one of us has made a mistake. The Herald seems to use the same numbers as me for past elections. But they get a slightly different projection of seats when using the 2020 election as a predictor (not when using the 2017 one). I think this must be due to slightly different rounding used (and rounding is unavoidable to squish one election’s patterns into another election, as explained below).

Anyway, here’s the TL;DR: On almost all special vote calculations I do, National are predicted to win 47 or 48 seats (as compared to 50 projected from preliminary results) and ACT are predicted to win 11 (the same as predicted in preliminary results). Add in National’s expected extra overhang seat after the Port Waikato by-election and National and ACT are looking at either 59 or 60, and Parliament will have at least 121 MPs after the by-election, theoretically as many as 124. So National will need NZF as well as ACT (unless they can make it work with another party. Which they can’t).

Labour, Green, and/or TPM are set to gain seats from the preliminary projection. Between them, on all these scenarios, these parties gain the 2-3 seats that National lose on party vote entitlement. TPM’s situation is complicated. They would not actually get an extra seat if they gain one more on party vote entitlement, because they currently have one more electorate than party vote entitlement. They could, however, gain up to 2 more electorate seats where they are nipping at Labour’s heels on preliminary votes.

Now for a couple of skippable sections of info before moving onto the good stuff (i.e. the numbers and the lovely screenshots of spreadsheets).

How we can predict special vote results from available info (because sadly this isn’t as simple as you might hope)

We have the preliminary votes and turnout from this year (2,254,441 preliminary votes) and the expected total number of special votes according to the Election Commission (570,000). And we have the counts of both preliminary and final votes for all the previous MMP elections (except the first one, 1996, where I have never found the preliminary vote count online). By deducting the preliminary votes from the final count, we basically have the number of special votes cast for each party and candidate.

(Technically the difference between preliminary and final account is not the exact number of special votes. This is because when counting special votes, they also re-count all non-special votes and sometimes discover mistakes. Eg in 2020, HeartlandNZ actually went down from 1,293 on the preliminary count to 914 on the final count; this obviously doesn’t mean that they got a negative number of special votes. I don’t know how common these mistakes are, but I just have to hope that basically the difference between the counts is roughly equal to the number of specials, or else this whole exercise is even more of a waste of time than it already was.)

Anyway, with this info we can predict how many special votes each party will get by predicting whether they’ll do better or worse on specials than they did on preliminary votes. Some parties typically do better on specials (most left wing parties, especially small ones) and some typically do better on prelim votes (most right wing parties). If we predict that a party will do twice as well on specials as on preliminary votes, and they got 10% of preliminary votes, then we predict they’ll get 20% of specials. So then we predict 20% of the 567,000 expected special votes, add that to 10% of the preliminary votes and voila, we can predict their final votes and percentage. (That example party will obviously hope there are lots of specials.)

How do we predict how parties will do on specials compared to preliminary votes? By past performance of them or (more tentatively) similar parties. Maybe (also more tentatively) by speculating about whether specific conditions of this election or this party at this time will make them do better or worse on specials than normal. (For example, some people are guessing that overseas voters, who are some of the special votes, will either love Labour for doing better at saving lives from Covid than any other government, or hate them for stopping them from coming home for too long in the course of that saving of lives.) Then you can punch in your predicted final party vote %s into the MMP seat allocation calculator to calculate numbers of seats from those predicted numbers of votes.

Here is each party’s/category of parties’ previous special votes performance, by which I mean their % in specials compared to their % in prelims. (E.g. if a party gets 10% of preliminary votes and 9% of special votes, their special performance/% in specials compared to % in prelims will be 90%).

Screenshot from spreadsheet: Table showing Special vote performance, i.e. Special votes % compared to prelim votes %, for parties/categories of parties since 1999, plus best/worst party votes since 1996.

BTW, I have just realised that these are in the wrong order: Labour should be ahead of National because they are meant to be sorted in order of best ever party vote. Oh well, I'll fix that for the next time I do this.

So when doing this kind of prediction, people often say “assuming all parties do the same on specials as they did in 2020/2017/etc.” Unfortunately, this isn’t actually possible. This is because the original prelim vote numbers are different. So if you assume everyone gets the same specials:prelim ratio as another election, you will end up predicting the wrong number of total votes. Eg: last election, parties who do well on specials won a majority, but this election, parties who do badly on specials won a majority. So if everyone did exactly as well as last time on specials compared to prelim votes, there won’t be enough parties doing well on specials to make up 100% of the votes: to be precise, it would only account for 96.9% of votes. So you have to adjust the predictions either up or down to make it match up to 100% of votes. (If you don’t do this, and just punch in your not-equalling-100% predictions into the MMP seat calculator, the calculator will do this adjustment automatically.) These adjustments mean that you’re not actually saying everyone did the same as the comparator election. What you are saying is that the ratio of everyone’s special performance to everyone else’s is the same. So that’s what I’m doing in the “adjusted to make 100%” columns in the tables below – evenly adjusting all the numbers in the column to the left either up or down to make the scenario actually make sense.

On overhangs (also not simple)

Sometimes a party wins more electorates than the number of seats their party vote entitled them to. When this happens, they don’t have to give back any electorates, and they don’t get to take anyone else’s party vote seats. Instead they add extra seats to Parliament beyond the regular 120. This is called an overhang and we can say that a party effectively has “overhang seats.”

On preliminary votes Te Pāti Māori is winning four electorates but only has enough party votes for three seats, so there is an overhang of one. TPM is expected to improve both its party vote and its electorate vote on specials, but we don’t know by how much. Their party vote could go up so that they qualify for four seats on party vote, thus no overhang if they win the same four electorates they’re currently winning. Or they could win one or two other electorates where they aren’t too far behind Labour on prelim votes. So they could have anywhere between 4 and 6 total seats and create an overhang of anywhere between 0 and 3.

There is also a by-election in late November in the Port Waikato seat due to the sad death of a candidate during advance voting. The rules say when this happens, the electorate is not decided at the general election, but instead in a by-election a few weeks later. And the seat becomes an overhang seat, even though it most likely wouldn’t have been one if it was decided in the general election. National is almost certain to win this seat; Labour aren’t even running. So National will also gain an extra overhang seat in a few weeks and have one more seat than their party vote in the general election qualifies them for.

Alright, now for the figures of my predictions. I have used several different methods to predict party vote results and also done a quick calculation to give an indication about the Māori seats.

Predicting using past election results

As explained above, we have to adjust up or down to make the numbers work, so predicting using past election results is not a reproduction of actual special vote performances from past elections, but a reproduction of the ratio of the parties’ special vote performances to one another.

Anyway, if we do this with the 2020 or 2017 party votes, both quite different elections, we get the same results when translated to number of seats: National gets only 47 seats, 3 less than predicted on the preliminary count. Labour, Green, and TPM each gain one seat on party vote entitlements. TPM’s seat doesn’t actually give them any more seats, it just catches their party vote entitlement up to the four electorates they’ve won, thus removing the overhang (though as mentioned above, they could also win 1-2 electorates, thus an overhang returns.)

Screenshot from spreadsheet: Table showing 2023 guess using 2020 special performance

Screenshot from spreadsheet: Table showing 2023 guess using 2017 special performance

If we do this with the 2014 results, National only loses two seats, which go to Labour and Green. TPM doesn’t gain any more seats in their party vote entitlement, so there is an overhang of one seat from their four electorate seats (with potentially 1-2 more overhang seats).

Screenshot from spreadsheet: Table showing 2023 guess using 2014 special performance

Predicting using a trend line

I figured out a linear trend line for each party’s results from 1999 or from when the party first contested the party vote up to 2020, and looked at where the line points for 2023.

I’m not a statistician so I don’t know if this is a good thing to do. I don’t know whether the votes follow linear patterns, or how to tell. Or whether another form of trend line would have been better. I do note, though, that these numbers produce predicted numbers of votes that come closest to 100% of any of my calculations, and thus they require the least adjustment.

On these calculations National again loses 3 seats, but this time two of them go to Labour and none to Green (probably due to Green having their worst ever specials performance of 122% last election, but Labour generally trending better, making their trend line predictions almost the same, which is probably not accurate). The other party vote seat lost by National goes to TPM (i.e. it reduces the overhang).

Screenshot from spreadsheet: Table showing 2023 guess using a linear trend line to predict special performance

Predicting ‘best ever’ or ‘worst ever’ scenarios

Here I have tried plugging in each party’s best ever and worst ever previous special vote performance going back to 1999 or whenever each party started. This involves some deep dives e.g. back to 1999 when Labour actually did worse than National on specials. I only really did this one for curiosity. Obviously, not everyone can do their best or their worst on the same night in a zero sum game like a set of votes. So these numbers need huge amounts of adjustment (from 118.56% and 88% down/up to 100%), meaning no party actually ends up doing that amazingly well or badly after all on the adjusted numbers.

However, it’s interesting to note that even though this is a completely artificial set of numbers, the outcomes aren’t too different to the above (or to each other). In both these scenarios, National loses two seats which go to Green and TPM (i.e. one more MP to Green and reducing the overhang by one).

Screenshot from spreadsheet: Table showing 2023 guess using best ever special performance

Screenshot from spreadsheet: Table showing 2023 guess using worst ever special performance

In the next two analyses I compared “best ever” for one side of the political spectrum and “worst ever” for the other side. Of course, my categorisations of “left” and “right” might be contestable. Also, this shouldn’t necessarily be seen as “best possible special votes for the right” etc, because obviously the best outcome for National and ACT would be for their fellow right-wing party NZ First to fall below 5%.

These scenarios don’t require that much adjustment, at least compared to the last set, because it’s built in that there is a balance of good and bad special performances.

These (adjusted) numbers are at the more extreme ends of realistic possibility in one direction or another, as you might expect.

The ‘best ever for right/worst ever for left’ has National only losing one seat and ACT actually gaining one, meaning essentially no change from preliminary results. These are the kinds of results National and ACT will need to be able to survive without Winston First – but as the first table with all special performances shows, neither party has had these kinds of results for many years – especially not with ACT as a more-than-one-person party.

The ‘best ever for left/worst ever for right’, on the other hand, has National losing four seats (two going to Green, one to Labour, and one to TPM i.e. cancelling out that overhang).

I don’t expect either of these scenarios to eventuate but they would still be (sort of) within past precedent.

Screenshot from spreadsheet: Table showing 2023 guess using best ever for the right/worst ever for the left

Screenshot from spreadsheet: Table showing 2023 guess using best ever for the left/worst ever for the right

Two Māori seats: Too close to call

As I keep alluding to, Te Pāti Māori could win some extra electorate seats on specials: Te Tai Tokerau and Tāmaki Makaurau. In these two Māori electorates, the TPM candidate is only about 500 votes behind the Labour candidate, and since TPM do better than Labour on specials, they could catch up.

I have done a quick analysis of the Te Tai Tokerau to see how likely this is. I didn’t do Tāmaki Makaurau because I couldn’t find the preliminary results from previous elections anywhere. (Edgeler uses the final count of advance + on the day votes, but that is not actually the same as preliminary votes.) I found the Te Tai Tokerau preliminary results from previous elections on the Northern Advocate website so here we go.

This is a lot less reliable than the nationwide party vote analysis because the numbers are a lot lower and the parties don’t all run every time. Also, TPM are a different party now than they were a few years ago, so their voter patterns might have changed. I haven’t gone back any further than 2017, largely because of the changing landscape of kaupapa Māori parties.

Screenshot from spreadsheet: Table showing Te Tai Tokerau - Special vote performance, i.e. Special votes % compared to prelim votes %, for 2017-2020, including all parties/independents who have contested the seat in those elections.

But even going back to 2017 reflects influence of this changing landscape: as you will see from the above table, actually in 2017 TPM didn’t run a candidate in Te Tai Tokerau; they implicitly endorsed Hone Harawira of Mana. Also the Greens didn’t run in 2020. So I’ve had to use the Greens’ 2017 result on the 2020 analysis. I also did two versions of the 2017 analysis: one where I give TPM their candidate’s 2020 result, and one where I give TPM Harawira’s 2017 result.

Oh, and I’ve also guesstimated that 1.29% of specials will be cast in Te Tai Tokerau (i.e. a 1.29% estimate of 567,000, which is itself also only an estimate), as this was the case in 2020. So there’s a whole lot of guesswork and approximations and proxies going on.

Screenshot from spreadsheet: Table showing How many specials do we expect in Te Tai Tokerau? 6,517 (27.09% of all TTT votes) if, like in 2020, 1.29% of all special votes are cast in TTT

Anyway, FWIW, here are what the numbers tell us: If 2020 patterns hold, Labour’s Kelvin Davis will just hold onto the seat but the margin will narrow to about 100 votes. If 2017 patterns hold (but using TPM’s 2020 result in lieu of a 2017 result), ditto. But if 2017 patterns hold (but this time using Mana’s 2017 result in lieu of a TPM 2017 result), TPM’s Mariameno Kapa-Kingi will win the seat by the narrowest of margins.

Screenshot from spreadsheet: Table showing Te Tai Tokerau - 2023 guess using 2020 special performance

Screenshot from spreadsheet: Table showing Te Tai Tokerau Te Tai Tokerau - 2023 guess using 2017 special performance version 1 (using TPM 2020 performance in lieu of a TPM 2017 result)

Screenshot from spreadsheet: Table showing Te Tai Tokerau - 2023 guess using 2017 special performance version 2 (using Mana 2017 performance in lieu of a TPM 2017 result)

Basically: it’s too close to call. And I expect the same is true in Tāmaki Makaurau. TPM could win one eat, both, or neither on specials, and none of those outcomes would be surprising. The only real surprise would be if either seat is not close.

Other electorates

The other 5 Māori seats are not close. The results won’t change.

I haven’t bothered to look at the close general electorates to see how likely they are to flip.* My suspicion without data is that one or two will flip National to Labour. Changed results in general electorates might change which individuals are heading to Parliament, but it won’t change the overall numbers, as no party has a chance of winning an electorate that isn’t already guaranteed representation through other electorates and/or the 5% threshold. TOP’s Raf Manji, for example, is well behind National’s Hamish Campbell in Ilam. Based on these two parties’ past performance on specials, I expect Manji will narrow the gap, but nowhere near enough. Also, no party other than TPM is gonna get overhang seats (weirdly National is closest to getting overhang seats, because they’re winning 45 electorates on prelims, and they’ll only get 47 on 48 – at a push, 46 or 49 – on party vote entitlement. But the 45 number is not going to go up, and their party vote seat entitlement won’t go below 45).

* Edit on the morning before special votes are announced: According to the Spinoff’s daily Bulletin email, “Ones to watch include Nelson, where Labour’s Rachel Boyack currently trails National’s Blair Cameron by just 54 votes; Mount Albert, where Labour’s Helen White leads National’s Melissa Lee by 106 votes; Banks Peninsula, where National newcomer Vanessa Weenink is 83 votes ahead of Labour MP Tracey McLellan; New Lynn, and Te Atatū, where National newbie Angee Nicholas leads veteran Labour MP Phil Twyford by a stunningly tight 30 votes.” I’m gonna go ahead and guess Labour wins all four of those.

September 13, 2020

The difference between Green and Labour: a tale of two Finance Ministers

So the Greens co-leader James Shaw recently made a mistake. In his role as Associate Finance Minister approving funding for “shovel-ready” projects, he fought hard for a private “Green school” to get funding to expand their buildings and, therefore, their student capacity. There are many problems with what he did: forgetting to oppose private schools as per Green policy; supporting an approach to environmentalism based on individual education of wealthy elites’ children, rather than systemic change; finally showing some spine around the cabinet table and for this; being so out of touch with his party’s kaupapa and membership that he actually thought his actions would be seen as a “win” for the Greens. It was a big mistake and it doesn’t say anything good about Shaw’s political judgment.

But here’s what happened next: the Greens’ membership, supporters, and former MPs flipped out. They rightly criticised what Shaw did, instead of sycophantically defending his actions because he’s the party leader. And Shaw called an emergency meeting with members, admitted it was a mistake, apologised unreservedly, and tried to do whatever he could do to reverse his actions.

Meanwhile, Grant Robertson, Shaw’s fellow Wellington Central-based MP and carpool buddy*, was working on an announcement of his own. Robertson is also effectively the second most important leader of his party, and he’s also a Finance Minister… and not just an associate one, but the proper one.

Robertson’s announcement was that his party, Labour, are finally going to increase tax on the rich. First, they’ve brought in a new top income tax rate of 39% on income above $180,000. And second… there is no second. That’s it. In this country of notoriously low taxes on the rich, in the greatest economic crisis since the Great Depression, their revenue policy is: Bringing back the Helen Clark/Michael Cullen top income tax rate of 39c, but to qualify for that top tax rate you have to earn as much as someone working 183 hours and 10 minutes per week on the minimum wage. No wealth taxes, just a small income tax change that even right-wingers think is too low. Australia, the UK, and other OECD countries have higher income tax on the rich, lower taxes on the poor, and of course capital gains taxes, even before the COVID recession… but Labour’s pathetic approach is literally lower taxes on the rich than what Don Brash proposed as National leader in 2005.

So what happened next after Robertson’s announcement? Well, there was dismay and anger from the left and from the centre (this example from No Right Turn is characteristically concise and well-reasoned: If not now, when?). But did Labour’s membership and supporters revolt? Did Robertson, like Shaw, call an emergency members’ meeting, apologise profusely for his massive misjudgment, and do everything in his power to rectify his mistake? Nope. None of that. The leader of the country’s biggest union, the PSA, even welcomed the announcement. The policy stands, and will probably become government policy after the election, unless polls change and the Greens get some leverage. Yet another opportunity for Ardern’s and Robertson’s promised “transformational change” has been wasted.

And therein lies the difference between the Green and Labour parties.

* “carpool buddy”: My partner saw Robertson giving someone a ride in his Labour car during the 2017 election campaign, and she’s like 76% sure it was Shaw.

October 21, 2018

Quick blog on how National sees its Asian MPs

Bridges responds with “Which is what we’ve got at the moment, right?” before going on to insult some of his pākehā MPs. (Image from The Spinoff)

I recommend this Spinoff article about National leader Simon Bridges’s and former protege Jami-Lee Ross’s conversation about the relative merits of different Asian ethnicities on the National candidates’ list. (The full conversation, recently leaked by Ross, is here.)

I have often noticed that Labour and the Greens are bad at putting forward elected representatives of Asian, especially Chinese descent. National have put forward more Asian and Chinese MPs, and I thought this was to their credit. However, this conversation gives insight into the ugly motivations behind National’s ethnic inclusion.

The starkest and most widely-quoted comment, “two Chinese would be more valuable than two Indians”, was made by Ross, who could be seen to have been deliberately goading Bridges to sound bad on a secret recording.

However, Bridges is far from off the hook—fact I’d say he’s more guilty than Ross of the cynical racist way of using rather than representing Asian communities. The topic of relative value of ethnicities was brought up in the conversation by Bridges, who said it “depends where we’re polling” as to whether they should have “Two Chinese” or “One Chinese [and] one Filipino”. The implication is that the only reason to have non-white people in the caucus is to win votes from those communities. Ross’s comment was a reflection and expansion of Bridges’ point. Also, Bridges agreed with Ross’s stark comment by offering a reflection and expansion of his own (rather than stepping in as a leader to correct Ross for expressing the issue in terms of relative value). Perhaps worst of all, Bridges also immediately thought of the two current Indian National MPs, saying “which is what we’ve got at the moment, right?”. This reveals that he thinks in these ways about his existing MPs, not just about hypothetical future MPs during a conversation about strategic nomination decisions to get ethnic communities on board.

September 16, 2018

How to complain about Heather du Plessis-Allan’s racist comments

Women's_Mau_leaders_and_committee_in_front_of_octagonal_Mau_office_ca_1930_-_AJ_Tattersall Some leaders of the Mau, a non-violent movement for Samoan liberation from harsh colonial rule by New Zealand. The New Zealand police killed eleven protestors at a Mau demonstration in 1929.

What were her racist comments?

Some items from Mediawatch (Radio NZ) about radio host Heather du Plessis-Allan’s racist comments:

Broadcaster stands by Pacific Islands “leeches” claim (16 Sep 2018)

Mediawatch Midweek: 12 September 2018 (12 Sep 2018)

Media go OTT on PM as RNZAF VIP (9 Sep 2018)

How to complain:

The official process to complain to the Broadcasting Standards Authority is here. You have to complain to Newstalk ZB first, and then they have to respond within 20 days. If you’re not happy with your response you can escalate it to the BSA.

You can complain to Newstalk ZB either on their website here, or by filling out this BSA form and posting it to Newstalk ZB.

Here are some tips from the BSA about effective complaints. I wish I’d read this before I submitted my complaint.

The details you’ll need:

1. Station name, programme name, date, and time.

It is important to get these details right. ZB are using imprecise details as an excuse not to respond to complaints.

The station name is Newstalk ZB. The programme name is Wellington Mornings with Heather du Plessis-Allan, and it’s on at 8:30am to 12 noon on weekdays.

My best guess for the date of the first programme is Monday 4 September (as Mediawatch says it was the same day Barbara Dreaver was detained). My best guess for the second programme where she doubled down on her comments was Tuesday 12 September, because Mediawatch talked today about “last Tuesday”, but if it was Tues 5 Sep they could have covered it last week. I have asked Mediawatch on Twitter for confirmation of these dates, or – even better – for the full audio.

2. Precise details of what was said.

I suggest getting this from listening to the excerpts on the Mediawatch episodes. If I get the full audio, I’ll post that here.

What your complaint needs to say:

Your complaint should explain why the broadcast breached at least one of the eleven standards listed in the Radio Code of Broadcasting Practice. There are some tips here about what each standard pertains to.

An example you can use (my complaint):

I just wrote this off the top of my head, and I would probably write it differently if I had read all the tips before writing it.

Nonetheless, it’s an example, so it may save you some time. (If anyone else has any examples, please let me know.)

Anyone is welcome to copy this as they wish for their own complaint. I don’t mind if you adapt it or not. No need to ask for permission or to cite me.

Which standards do you think were breached?

Standard 1 – Good taste and decency
Standard 3 – Children’s Interests
Standard 5 – Law and Order
Standard 6 – Discrimination and Denigration
Standard 8 – Balance
Standard 9 – Accuracy
Standard 11 – Fairness

Why do you think the programme breached those standards?

Please note that my complaint is both about the original broadcast from Heather du Plessis-Allan and also about the following Tuesday’s broadcast where du Plessis-Allan defends and stands by her comments. I have attempted to distinguish the two broadcasts in my below comments where practical.

Standard 1 – Good taste and decency. “Current norms of good taste and decency should be maintained, consistent with the context of the programme and the wider context of the broadcast.” Heather du Plessis-Allan said that the Pacific Islands (referring to independent countries and territories of Aotearoa New Zealand) “don’t matter”, asking rhetorically “what are we going to get out of them”, with the implied answer from her following comments being that we get nothing, as “they are nothing but leeches on us”. She also made other insulting comments about certain societies and people in them, as outlined in more detail below. This type of insult to entire societies breaches current norms of decency, as demonstrated by the widespread worldwide outrage earlier this year at US President Donald Trump’s comments referring to countries as “shithole countries”. du Plessis-Allan used a synonym for “shithole countries”, namely “hellhole” for Nauru. She also made similarly insulting and sweeping claims, such as referring to these societies as “nothing but leeches on [NZ]”, referring to “welfare sponging” in relation to some NZ citizens’ rights to superannuation in NZ territories, and suggesting in the following Tuesday’s broadcast Niue does not contribute anything to its own upkeep but that New Zealand aid is “funding all of Niue”.

Standard 3 – Children’s Interests. “Broadcasters should ensure children can be protected from broadcasts which might adversely affect them.” Whilst in the following Tuesday’s broadcast, du Plessis-Allan suggested that her comments were about the countries and their leaders, rather than the individual people in Aotearoa or in the islands, she also referred to individual people, such as people who live in New Zealand and then move back to Niue or other NZ territories with pension portability. Other comments in the original broadcast seemed also to refer to people, such as talking of “leeches” which is an insult that is typically applied to people rather than countries. (Other comments were more clearly about countries, such as calling Nauru a “hellhole”.) It cannot be reasonably claimed that no Pasifika children listening to the broadcast would take du Plessis-Allan’s comments as insulting to them as people, by reducing them to “nothing but leeches” who offer no benefit to New Zealand. It cannot be reasonably denied that this “might [have] adversely affect[ed] them”.

Standard 5 – Law and Order. “Programmes should not actively promote serious antisocial or illegal behaviour, including violence, suicide, serious crime and substance abuse.” Whilst not included in the list of examples, racism and/or xenophobia towards Pasifika peoples is serious antisocial behaviour. The programme actively promoted resentment towards Pacific Island nations as being nothing but “leeches” who “do not matter”, except, apparently, insofar as we should be upset at funding them.

Standard 6 – Discrimination and Denigration. “Broadcasters should not encourage discrimination against, or denigration of, any section of the community on account of sex, sexual orientation, race, age, disability, occupational status or as a consequence of legitimate expression of religion, culture or political belief.” “‘Discrimination’ is defined as encouraging the different treatment of the members of a particular section of the community, to their detriment. ‘Denigration’ is defined as devaluing the reputation of a particular section of the community.” du Plessis-Allan encouraged discrimination against a particular section of New Zealand citizens, namely Nieuan, Cook Island, and Tokelauan NZers who have moved from Aotearoa back to one of these three NZ territories. She opposes their right to receive NZ superannuation, whereas she does not oppose this right for other NZ citizens who also qualify for superannuation by living in New Zealand for the requisite number of years. du Plessis-Allan also encouraged, and indeed engaged in, denigration of a section of the NZ community (Pasifika people in NZ and its territories) and a section of the global community. She devalued the reputation of this section of the community by saying they do not matter, and that NZ does not get any benefit from them, because they are “nothing but leeches” on NZ.

Standard 8 – Balance. “When controversial issues of public importance are discussed in news, current affairs or factual programmes, broadcasters should make reasonable efforts, or give reasonable opportunities, to present significant points of view either in the same programme or in other programmes within the period of current interest.” du Plessis-Allan brought into the public conversation a controversial issue of public importance, namely whether the Pacific Islands have any value or whether they are “nothing but leeches on us” that “do not matter”. While I have not listened to every Newstalk ZB broadcast since then to see what other views were aired, I am not aware of significant effort from Newstalk ZB to present other significant points of view on this issue of public importance.

Standard 9 – Accuracy. “Broadcasters should make reasonable efforts to ensure that news, current affairs and factual programming:
• is accurate in relation to all material points of fact
• does not mislead.” “The requirement for accuracy does not apply to statements which are clearly distinguishable as analysis, comment or opinion, rather than statements of fact.” du Plessis-Allan made factually inaccurate comments, such as that “the Pacific Islands are nothing but leeches on us” that New Zealand does not benefit from these nations (despite many economic and other benefits to New Zealand, according to various official reports and statistics), that Niue is entirely funded by New Zealand, and that Nauru is a “hellhole”. She also gave a misleading impression of the rights of people in NZ territories to receive NZ superannuation if they qualify for it by having lived in NZ for the requisite amount of years, by portraying this right as “welfare sponging” and as an unfair imposition on New Zealand, without clarifying that this is the same right as that enjoyed by other NZ citizens who qualify for superannuation. These comments could be construed as statements of analysis, comment or opinion (and therefore exempt from this standard), but in later broadcasts she insisted that the “hellhole” description of Nauru was “factually correct” (she used the purported factuality as a defence of her right to make the comments).

Standard 11 – Fairness. “Broadcasters should deal fairly with any person or organisation taking part or referred to in any broadcast.” “If a person or organisation referred to or portrayed in a broadcast might be adversely affected, that person or organisation should usually be given a fair and reasonable opportunity to comment for the programme, before the broadcast.” du Plessis-Allan did not deal fairly with the Pacific Island nations (both nation-states and territories of New Zealand), their governments, and their people, when referring to them as “leeches” and saying they “do not matter”, saying that Niue is entirely funded by New Zealand, and saying Nauru is a “hellhole”. I am not aware of Newstalk ZB giving these people and organisations “a fair and reasonable opportunity to comment for the programme”.