Stats NZ says planned cost cutting unlikely to produce successful census in future
Monday, 24 December 2018
Statistics New Zealand is unlikely to meet cost cutting expectations for future censuses, as Census 2018 looks to cost an extra $6 million.
The department has repeatedly pushed-out an initial October 2018 deadline for releasing the latest census data, while scrambling to fill gaps in a result which missed 1 in 10 New Zealanders. The data is now expected to be released as late as August 2019.
The stalled and flawed census data has attracted concern and criticism from government departments and economists, with some fearing court action may result.
And the political blame game over funding of the first truly digital census has already begun.
**READ MORE:
* Jonathan Milne: Statistics NZ boss should be nervous
* Election 2020 at risk of being compromised due to census delays
* Statistics NZ pushes Census 2018 deadline further as issues continue
* The Electoral Commission wants Parliament to make voting changes**
Stats NZ was expected to produce a modernised census while reducing the average cost by five per cent within two census cycles.
The total five-year budget for the census, granted by the previous National government, was $120m. It's difficult to compare census budgets, which rise and fall with each cycle, but the previous census in 2013 cost about $90m.
Stats NZ continues to deny requests for an interview, but a spokesman provided said in answer to written questions that 'an overall reduction in cost across two census cycles seemed an appropriate target at the time'. .
It was now 'unlikely' there would be cost savings at the 2023 Census in keeping with the goal, the spokesman said.
The debate surrounding the funding of Census 2018 was 'robust', with 'strong expectations that Stats NZ has considered all options, including the frequency of the census, and that the department has looked for efficiencies and savings'.
Current spending forecasts had Census 2018 costing about $6m more than the $120m budget, and 'Stats NZ expects to fund a substantial proportion of this via its own baseline'.
Independent economist Brian Easton said he and six other statisticians have suggested to Government ministers that Stats NZ abandon Census 2018 and run another census in 2021, instead of 2023.
'Our concern is that we'll end up in court wasting a hell of a lot of time. To give a particular example, if a District Health Board ends up having a bad deal, they might take the Government to court.'
Stats NZ ruled this out, saying there was not enough time to prepare another census for 2021.
Prime Minister Jacinda Ardern earlier this month blamed the previous National government for poorly funding the census.
'There was an expectation on Statistics NZ to reduce costs by the last Government by 5 per cent over two full census cycles. That was at the same time that Stats NZ did some risk analysis and advised that they would need an increase in support over a five-year period in order to do their jobs properly.'
But National's State Services spokesman Nick Smith pushed back on the description of a funding pullback, saying the 2018 census was 'generously funded' with a budget 20 per cent larger than the previous one.
A NEED FOR DATA
If National were to dispute the results it could have a serious impact on the integrity of the electoral system, as electorate boundaries are drawn based on census results and can have serious electoral consequences.
Stats NZ has acknowledged internally a 'high risk' it may not meet its statutory obligation of providing census population data for the redrawing of electoral boundaries.
Commission chair Sir Hugh Williams QC told a parliamentary select committee earlier in December that for the necessary redraw to be done for Election 2020, Stats NZ would have to release census data no later than September 2019.
Education Minister Chris Hipkins has told Stuff said he would recommend Cabinet replace the broken decile school funding about four years earlier than planned, because Statistics NZ won't be able to tell schools how many children are from impoverished neighbourhoods – not reliably enough, not soon enough.