National has put in another commanding poll performance, yet the short-term prospects and long-term ambitions of the Maori Party could yet have a signficant impact on this year's election race

Another poll, another list of government parties showing National and its coalition partners looking good for another term – ACT is there with Epsom; United Future has Ohariu; and with more than both of those put together is the under-appreciated Maori Party. Yet nothing it does this year should be taken for granted, least of all by National.

The One News-Colmar Brunton poll was more good news for National, giving it a majority that would allow it to govern alone. It's an election year launching pad few even in National would have dared hoped for so deep into its second term and is reflected in a striking 2014 upturn in the left-right gap in Pundit's Poll of Polls. No-one, however, expects such a lead to survive an election campaign. So those minor parties are vital, and while Richard Prebble gets headlines talking of an eight-seat ACT renaissance, the Maori Party's existing strength is barely mentioned.

Why? In part because the party is in a state of transition as its mastermind, Tariana Turia, prepares for retirement and its other founder, Dr Pita Sharples, looks to be in trouble in his seat, Tamaki Makaurau. Waiariki MP and recently minted co-leader Te Ururoa Flavell looks the only safe bet come election day.

Yet one or two seats could yet be crucial come coalition negotiations and every pollster, in accordance with tradition and the past two terms, bundles the government's current coalition neatly together into a potential post-2014 government.

It looks good for National, but the reality is foggier. A combination of hints from the Maori Party itself and political logic suggests that National would be wrong to be presumptuous.

The Maori Party's decision in 2008 to ally itself with the fresh and popular new Key administration was a political masterstroke. It gave the new Prime Minister a statement about mana-enhancing Treaty partnership and unity that sent a message to his party's base (at a time of immense high spirits) that his was a government that wouldn't play the race card; in return the Maori Party got concrete commitments to Whanau Ora and two seats at the table. Perhaps more than else, it made the Maori Party relevant and told the left that after several generations the Maori vote could no longer be taken for granted.

The six years since however have not been as mana-enhancing as the party's co-leaders might have hoped. John Key – however solid his personal relationship with these coalition partners – has not seized the opportunity to make race relations part of his legacy and the commitment to Whanau Ora has tended to shrivel come budget time. Worst of all, its supporters have drifted away, unconvinced that the deliverables were worth the deal.

At Ratana Pa this year, the party was dropping hints left, right and centre about its future, clearly signalling – to the right this time – that it still shouldn't be taken for granted.

Consider this from a Fairfax story, January 24:

When asked whether it was time for the Maori Party to move away from National, Sharples conceded "it probably is".

"I think we've already started to be honest," he said.

This did not necessarily mean moving closer to Labour "but we have certain things in train and I think it's time we hoed into those and finished those off and some of those are with the support of National Government and some of them are not."

Progress on issues such as adequate housing and addressing poverty had not been fast enough, he said.

That message of impending separation from National was reinforced by his successor. This from TVNZ:

Maori Party co leader Te Ururoa Flavell says he thinks it's fair to say that the the party has got the message that "our people" are not comfortable with the relationship with National.

"And so it's important then that we open that discussion up again with our people, to take our time on it and to get it right," he told reporters at Ratana.

"Three years is a long time, and with the possibilities of a different coalition it's important that we take it seriously and we don't just follow the same pattern in the past. So we'll be doing that."

The language is unambiguous – "we don't just follow the same pattern". Put simply, the Maori Party can't afford to. Given that the majority of those who voted for Turia, Sharples and Flavell in their electorates gave their party vote to Labour, the Maori Party's tilt to the right could never last forever. The Key deal came with a time limit and my guess is that Flavell, who leans a little more to the left anyway, is calling last orders.

Turia's retirement makes such a tactical change easier. For her the move away from Labour was in part personal – she felt betrayed over the Foreshore and Seabed bill. With her exit, hostilities between the Maori Party and Labour can be put to bed.

Most of all the political maths is compelling. The Maori Party's poll trajectory has been gently downward since 2008 and has stayed below two percent for months now in our Poll of Polls. If the party goes with National for a third term, it is putting its very existence in jeopardy.

As for its oft-repeated ability to deliver gains for Maori, well, there's no sign of that improving if the existing coalition were to get another term; even less so if that coalition has to make room for Colin Craig's Conservatives. If the government's numbers slipped as you'd expect in a third term – and especially if Key decided to walk away – the temptation to dig out that old race card may prove too much for National to resist.

So the Maori Party will be looking at its short-term prospects and most of all its ability to survive this year's election with more relevance than, say, United Future. In other words, it needs electorate wins and that means giving past voters a new reason to turn out.

Could Turia and co swing one last big win out of National before the election to display the benefits of being inside the tent? If not, surely the prospect is that Labour will want the Maori Party more – and be willing to promise it more policy wins as part of a centre-left coalition. It's easy to see that more could be gained from a coalition including the Greens and Mana than it will from one with the Conservatives and ACT.

In that case, the question is what the Maori Party and Labour both want most and what they might be open to trading.

Perhaps Labour can be taken at face value and will go for broke in the Maori seats, attempting to drive the Maori Party out of parliament. But that's a gamble too.

Whichever way it plays out, those three red dots reliably placed in National's column at each poll are by no means seats John Key can bank on. And that's just one more reason that even with his commanding lead in the party vote and preferred Prime Minister polls, John Key is a long way from being able to rest easy ahead of election day.

Comments (13)

by stuart munro on February 25, 2014
stuart munro

For the Maori party it is as it was for the Arabs and the British:

It is better to be an enemy than a friend to the British; if you are their enemy they will buy you, but if you are their friend they will sell you.

Rather like NZ and John Key really.

by Richard Aston on February 25, 2014
Richard Aston

Stuart , chilling but well put .

I think the maori party better pull its thumb out and openly declare its leaving National and aligning with Labour , it might just stand a chance of reversing its decline.

I question on MMP maths ...Thinking of Labour and Maori as one group what would produce the most seats ( electorate and list)

1 Labour ditchs Maori and gets most of the Maori electorate

2 Labour and Maori align closely and Maori get all the electorate seats plus maybe 5% party vote



by Ian MacKay on February 25, 2014
Ian MacKay

That much discussed poll though. 11% Undecided. That many either for or against the current Government should be significant don't you think?

by Richard Aston on February 25, 2014
Richard Aston

Good point Ian , the non voters are a large group these days and from what I have read do not tend towards National.

Also this was a landline phone survey , which would cut out a large group , young people and lower social demographics.



by Andrew Robertson on February 25, 2014
Andrew Robertson

Hi Richard.

The 2013 Census tells us that 85.5% of NZ households have a landline telephone. This means that landline RDD (random digit dialing) sample frame non-coverage is up 6.1 percentage points, from 8.4% of households in 2006 to 14.5% of households in 2013.

This isn’t the percentage of individuals not covered by landline sampling frames. It’s the percentage of households.

Between 1% and 2% of households will not have access to either a cell phone or landline, so this means the proportion of cell phone only households has increased from somewhere around 6% in 2006 to somewhere around 12-13% in 2013.

Other than door-to-door surveys (which can take a couple of months to carry out and process), there’s still no other single sample frame that provides the coverage of a landline sample frame.

  • Simply adding cells to the sample frame won’t increase coverage by much, because you’re after people who have a cell but not a landline (all the others are already covered by your landline sample frame).
  • The Electoral Roll can be a useful sample frame, except the process of matching names/addresses to phone numbers reduces coverage down to around the 40% mark. Also, the phone records used for tele-matching can be based on home and vehicle ownership databases, so they introduce a socio-economic bias to your sample.

Landline surveys will generate a substantial sample of young people. Survey companies will also generally weight their data to adjust for biases. Non-coverage is certainly an increasining issue, I agree, but in my view non-response is a bigger issue. Cells vs landline coverage is just the more obvious one.

I'd be careful about assuming that landline surveys cut out lower soc-ec groups too much. Cell phone calls are still fairly expensive in NZ once you've used up free minutes. Landline calls are free and telco often bundle landlines with other products.

by Richard Aston on February 25, 2014
Richard Aston

Thanks Andrew great info.
I work with low income families and we increasing see cell phone only , prepaid , cheaper that a land line. Not very scientific though I agree.

Tell us more about the non- response rate . If I understand that right its not included in the margin of error and has a social statistical value in its own right eg would it indicate voter apathy ?


by stuart munro on February 25, 2014
stuart munro

While I personally would prefer the Maori party to align with Labour, I think Flavell should draw a few lines in the sand for any coalition partner.

He knows his constituency better than I do, but an outright ban on pokies might be one of the firm conditions he could make that would definitely benefit the communities he serves. This would also be consistent with his own long held views. The Greens at least would not find such a stance problematic.

It would be more palatable to voters than a taste for ministerial perks.

by Tim Watkin on February 25, 2014
Tim Watkin

Yeah thanks Andrew, good to know. While you say the landline frame is still a large one and statistically sound, I have one idea I'd like your thoughts on: it's a particular type of household that is mobile-only and so those 12-13% of households could be almost 100% of a certain type of New Zealander who is not being polled - does that make sense? Or are we all too many different types of types for that to be a problem, if you get my drift?

Richard, it's an interesting theory about the MP committing to Labour. I've heard some talk about that. It may give them a boost, but it also would deny them the NZF strategy of being able to say they'll work with anyone and so being able to win voters from both sides of the spectrum. I guess the argument in favour of declaring a Labour preference is that sooo many Maori voters are on one side of the spectrum, perhaps keeping the door open to the other lot is less of a problem.

I was reminded yesterday that Dunne at one point ruled out working with Labour after the election, but stayed in coalition with them in the meantime... Was that 2008? Anyway, it's interesting to speculate on what the MP could be thinking...

by Andrew Robertson on February 25, 2014
Andrew Robertson

Hi Tim

Yeah I get what you mean. What you're describing is a situation where non-coverage would have a really big impact on a poll (assuming all of that 'almost 100%' vote on election day).

I don't think the situation in New Zealand is that extreme though.The 2012 Stats NZ Household Use of Information and Communication Technology is helpful for illustrating this. If you look at the Excel tables (Table 1b), you can see the proportion of households who have a landline by annual household income. The data show there is definately a skew toward lower incomes/away from higher incomes for 'no landline' households (no landline being the best proxy we've got for cell only households). However a large majority in each income band still have a landline. Also, a fair proportion of higher income housholds also also do not have landlines. So I don't think we can say the 'cell only households' are all similar in some way.

The big questions for me are:

  1. Can polling companies adjust for these skews through weighting or by setting quotas for hard to reach groups (eg, are the 'young cell only people' dramatically different from the 'young non-cell only people')? If the cell only households are only 'a bit different' and not 'dramatically different', then a careful weighting scheme or quota sampling approach might be enough to compensate.
  2. How important is 15% non-coverage compared to, say, a 40% refusal rate? If cell only households are only 'a bit different', then I'd argue that the refusal rate is a bigger issue, and that pollsters should focus on maximising fieldwork practices and maximising response rates.

Another thought - when it comes time to change, I'm not convinced that simply adding a cell phone sample frame is necessarily the best way to go (it could be, but I'm not sure). There may be other ways and methodologies to increase coverage (which I'm not too keen to post about :) )

Don't take me the wrong way. I definately think non-coverage is becoming a bigger problem, but the potential sources of error in a poll are endless (and non-coverage is just one). In my view a good pollster is one who tries to understand all of these sources, and tries to reduce them, adjust for them, or cancel them out. They won't always get it right - but I think that's the nature of measurment in a context where there are so many variables.

by Andrew Robertson on February 25, 2014
Andrew Robertson

Hi Richard

That's right. The margin of error relates only to the precision you lose by taking a sample rather than a census (I think it's better to describe it as the margin of sampling error). There are many other potential sources of error.

Yes, assuming consistently good fieldwork practices a chance in response rate can give an indication of a change in public interest in the survey topic.

If you're interested there is more on polling and surveys on my blog -

by Tim Watkin on February 26, 2014
Tim Watkin

Thanks Andrew. Yes, my question was whether non-cell homes were somehow similar, but I take your point that there don't seem to be any distinct characteristics.

As for refusals, I don't see how you can do anything about it or weight for that. Presumably those who refuse aren't distinct either – they could be any income level, gender or ethnicity – but even if they were distinct if they hang up quickly you can't be sure who refused and if they're in anyway similar to the previous refuser (apart from the fact that they don't like pollsters!). Tricky.

by Andrew Robertson on February 27, 2014
Andrew Robertson

Hey Tim

You're right. Surveys assume that non-respondents are the similar to respondents. That's a big assumption, and my view it's an incorrect one. I think this is a much bginger deal than the cell phone vs landline issue. It's not helped by the increase in sales calls and 'sugging' in recent years.

You can do something about it though, although you can't eliminate it. You can put significant effort into driving as high a response rate as possible and training your interviewers well. This is the reason I'm not a big fan of 'quota surveys', because they get very low response rates. Having said this, there are so many variables at play that the final pre-election accuracy of a poll is determined by many factors other than just the sampling approach.


by Richard Aston on February 27, 2014
Richard Aston

With Mat McCarten as Labour's chief of staff and given his background with MP and Mana will we see a more cohesive partnership between labour , the MP and mana? Implying Labour doing a maori electorate deal?

Post new comment

You must be logged in to post a comment.