Statistics New Zealand chief executive Liz MacPherson has announced her resignation, after the release of a damning independent review into the array of problems surrounding last year’s census.
“As leader of this organisation, I take full responsibility for the shortfalls identified in the report,” MacPherson said in a statement.
State Services Commissioner Peter Hughes asked MacPherson to stay on until Christmas, but said she “has done the right thing” by opting to resign.
“She has taken full responsibility for what went wrong and she is doing an excellent job of fixing it.”
The review found that “weaknesses in overall governance and strategic leadership at the programme level” were responsible for the bungled census and recommended that the general model be kept on for 2023. It found that two of Stats NZ’s four objectives for the 2018 census were not met, and a third “is likely to be very challenging”.
These goals were to meet statutory requirements, at least maintain the quality of information that the 2013 census managed, improve the timeliness of information released after the census by 20 percent, and to reduce the average cost by 5 percent over two census cycles.
The census was plagued by low response rates, a blown budget, and IT problems as the national statistics agency tried to roll out New Zealand’s first ever digital-first census.
Census coverage sees considerable decline
One in ten people didn’t take part in the census at all, and another one in nine only partially participated. This means around 83.3 percent of the national population responded fully – below not only the 94 percent target but also the bottom tolerance line.
The review shows that these numbers grew worse for traditionally under-represented sub-groups. Just 68.2 percent of Māori responded fully, down 20 percentage points from the 2013 census.
Asians saw a 10 percent drop to 81.7 percent, Pasifika a 23 percent drop to 65.1 percent, and youth a 13.5 percent drop to 75 percent.
The 2018 goal for each of these groups was 92 percent, already a step down from 2013’s 94 percent objective.
Flaws in decision-making highlighted
The review highlighted 10 key decisions that the census team took which “had a significant impact on the ultimate outcome of the 2018 programme”. Chief among these was the work that went into planning a risk management framework.
“The primary and almost exclusive focus was placed on the risk associated with potential failure of the online collection system.
“Other catastrophic risks that were identified, such as lower-than-expected response rates and difficulties in recruitment of field staff, did not appear to have appropriate risk mitigation measures in place to adapt quickly when/if they materialised,” the report said.
It also found that Stats NZ waited until 10 days after census day to follow up with those who didn’t respond. “We believe management was slow to react,” the report concluded.
Problems relating to the supply of paper forms were not adequately handled. When Stats NZ’s field operations team said it would need 1.8 million English paper packs, the agency’s vendor “did not have sufficient capacity to print the required volumes”. Stats NZ did not contract another printer and received half a million fewer English packs than requested.
In addition to targeting the agency’s failures, the review has made 16 recommendations for future censuses and broader reforms to Stats NZ’s governance structures. It insisted that the next census go ahead in 2023, instead of heeding the National Party’s call for a start date as early as 2021.
The review also asked for more funding for Stats NZ so that the organisation can better prepare for contingencies like those seen in 2018.