Government looks for a better way to count Māori

After repeated, failed attempts at capturing Māori communities in census surveys, Statistics NZ is taking a new approach to co-designing its mandatory five-yearly data gathering exercise.

And as the Government relaunches National’s social investment approach, the need for high-quality data on Māori, and other marginalised communities, is paramount.

Earlier this week Stats NZ began its public consultation on the future of the census, which would include a specific focus on working with hard-to-reach communities from the outset.

“The data landscape is changing,” government statistician Mark Sowden said. 

The traditional approach of people going out into the field and collecting data on a large scale was no longer viable. Surveying was becoming more expensive, for diminishing returns.

This was even more true for hard-to-reach, or diverse, communities, such as people with disabilities, Pasifika and Māori. 

The state was bad at counting Māori, and while this was not a new problem, it persisted.

In 2023, Stats NZ fell just shy of its 90 percent response rate target for the general population, coming in at 90 percent (administrative data was used to boost the figure to 99.1 percent coverage). 

The agency also set a 90 percent target for Māori, Pasifika, Asian, and 15 to 29-year-olds, but they were, respectively, 76.7 percent, 79.9 percent, 91.5 percent, and 84.8 percent.

Following the 2018 Census, independent reviewers and an external data quality panel recommended deepening and strengthening the relationship with Māori, through “governance arrangements, early engagement and co-design”.

“2023 Census, whilst a significant step up from prior censuses, did not achieve the level of partnership and co-design expected by Māori,” a recently released census review said. “This will likely require increased capability within Stats NZ and refreshed governance arrangements with Māori.”

Independent reviewers, Murray Jack and Geoff Bowlby, said there was a strong “relationship focus” at senior levels of Stats NZ, but within the census programme “the relationship defaults to a transactional rather than a partnership one”.

There were initiatives and relationships with some iwi, but they began too late – the design of the field collections had been made by the time iwi became involved.

In the places where Stats NZ fostered direct relationships with iwi organisations, the response rate significantly lifted: Toitū Tairāwhiti, from 70.6 percent to 80.1 percent; Te Whānau-ā-Apanui, from 63.5 percent to 79 percent; and Ōhua, from 69.1 percent to 69.6 percent.

But those results came at significant expense. The nationwide cost was $13 a head – largely thanks to high digital uptake, in these areas the cost was $40 a head. This factored in the use of food vouchers and Warriors tickets as a form of koha, which was a form of acknowledgment or aroha for the time and information these communities relinquished.

One successful initiative, taken when it was realised Māori returns were lagging, was the ad hoc involvement of the Whānau Ora Commissioning Agency, which collected responses from 10,000 Auckland households that hadn’t responded. This returned responses from 29,426 people.

“If a future census is to build on the good work done in 2023 and the pilot experience, then engagement with Māori on the collection-approach process design and support needs must start now,” the review said.

And that’s exactly what Stats NZ was doing.

Waipareira Trust chief executive John Tamihere (Ngāti Porou ki Hauraki, Whakatōhea) said by the time Stats NZ came knocking on his door, the 2023 Census was just 90 days away.

He hoped this time around the Government would partner with them sooner, to design a census that would get a higher response rate across the board.

“It’s an extraordinarily important exercise,” Tamihere said.

“You can’t run a First World country, with a First World economy, with First World policy, without the data.”

Without counting Māori, and particularly poor Māori, the Government would end up tailoring policy to a “well-off, well-covered group”.

Being able to understand things like the number of people who were at higher risk of respiratory or cardiac issues, in order to properly design and fund preventative and early intervention policies, was only possible with the right data. Without that data, these conditions would be caught later, costing the state more money, and costing more people their lives.

“They identified the wrong people to ask the right questions.”

Waipareira Trust chief executive John Tamihere

There was historical mistrust in the state, paired with a growing mistrust in democratic institutions, which was magnified by Covid-19, Tamihere said.

This has created a further barrier to data collection.

The number of “hard refusals” in 2023 rose to 20,000, from 6000 in 2018. And a “very large number” of the people who refused to fill in the census survey cited general anti-government sentiment.

“They identified the wrong people to ask the right questions,” Tamihere said.

But iwi organisations weren’t the state, or a regulator, and they had established trust and connections among their community.

It was crucial they were involved in future censuses and data gathering, if the Government wanted to get enough, high-quality data about Māori.

Stats NZ’s Sowden said the organisation was aware of the need to begin that co-design process early.

The agency had identified Māori, as well as Pasifika, disability communities, and the rainbow community as some of the groups that needed particular attention.

And it wasn’t just about lifting coverage rates, there were also nuances and sensitivities in how questions were phrased and how any surveys were delivered, Sowden said.

On Wednesday, Stats NZ began its six-week public consultation on the future face of the census.

Following the six-week feedback period, Stats NZ would make recommendations to the minister, who would get sign off from Cabinet, then full steam ahead. This would give Stats NZ a further 18 months to work alongside these groups to co-design surveys and data-gathering, before the census would begin in earnest.

Sowden had not been shy in advocating for the wholesale overhaul of data-gathering in Aotearoa.

In 2018 the census cost $123 million and in 2023 it cost $326m. This rapid increase in costs reflected Stats NZ’s need to double the number of field workers, and did about five times as much community engagement.

Sowden said without a change, costs would continue to balloon, but return on investment would drop.

Sowden told Newsroom the future was a census without surveys, but that was unlikely to happen during his tenure.

Stats NZ would soon be going to the minister with a recommendation on the shape of the 2028 census. Sowden would recommend a census that had a smaller, more focussed survey component, and a greater reliance on admin data. The exact details of this, including how large the surveying would be was not yet clear.

“Social investment is going to be, I think, a game-changer for the way that government uses data.”

Government Statistician Mark Sowden

In 2018, New Zealand was forced to use admin data to plug the gaps, following a low response rate. But this put New Zealand ahead of the pack in the global shift away from expensive field surveying.

Administrative tax data and things like Stats NZ’s Household Economic Survey gave good quality data, for a relatively low price, Sowden said.

But there were limitations.

Administrative data was not good at the “descriptive piece” – things like answering: ‘How many cigarettes do you smoke a day?’ or ‘How much volunteer work do you do?’ And not all admin data was created equal.

When people filled out forms in the emergency room, they weren’t thinking about the quality of data, they were focused on getting medical treatment, Sowden said. So, that had to be accounted for when analysing health data.

Likewise, data from justice, police and corrections likely had negative framing, Sowden said.

Māori were over-represented in the justice system, so Stats NZ was likely to gather a disproportionate amount of “deficit data” from Māori, so that needed to be worked through when Stats NZ considered the future of the census and data-gathering, generally.

There had long been discussion about groups that weren’t captured by data, particularly admin data. This concern was expressed by experts in Stats NZ’s recent decision to cancel the Living in Aotearoa Survey – how would they count vulnerable, hard-to-reach groups without this longitudinal study?

But Sowden said groups with the highest need generally had a lot of interactions with the state, through things like welfare support. That meant there was a lot of admin data on these groups.

But as this Government rolled out its social investment approach the need for specific data would grow.

“Social investment is going to be, I think, a game-changer for the way that government uses data.”

On Thursday, Finance Minister Nicola Willis officially launched her standalone Social Investment Agency – as tipped by Newsroom in March. 

The new agency would be tasked with setting up a Social Investment Fund, setting cross-agency standards for social investment practice, data and evidence infrastructure and leading an ongoing review of social sector spending to measure outcomes. 

Sowden said not only would this approach need more population data, but the Government would also need to collect data on the effectiveness of its policies – something that could take decades to be borne out.

That evaluation aspect would be a “big step-change” from what the Government was currently doing with data.