Each possible cause is an institution, policy or condition, never a person. Each is put to the same four tests and ranked by one
rule, closest to proven first. How a cause is tested and ranked.
Open1 of 4 support, 0 against; partly supported needs one more.
What would settle it: Renewal-to-depreciation ratios by sector over time, set against each sector's service series, holding workforce and demand constant. Councils report the ratio through the Auditor-General; central government agencies are not required to report it.
Open1 of 4 support, 1 against; partly supported needs one more.
What would settle it: Model quarterly citizen departures to Australia on the real wage ratio and the unemployment gap. If the wage ratio adds nothing once the gap is included, the productivity gap sets only the baseline outflow, not the surge.
Open0 of 4 support, 0 against; partly supported needs two more.
What would settle it: A register of every Crown project over $50m with each decision date and the spend at that point. Compare write-offs within 12 months of every election with other months, across every term, and against cost-estimate growth, then against Australian states.
Cite Policy churn: programmes reversed at and between changes of government
macrohard.nz, “Policy churn: programmes reversed at and between changes of government”, The Hollowing Index, tested 2026-09-28. https://macrohard.nz/causes/electoral-cycle-churn
Open0 of 4 support, 1 against; partly supported needs two more.
What would settle it: Link departures to prior tenure and region in the Stats NZ Integrated Data Infrastructure, and test whether renters and residents of high price-to-income regions leave at higher rates, controlling for age, occupation and the NZ-Australia unemployment gap.
Open0 of 4 support, 0 against; partly supported needs two more.
What would settle it: An agency panel from 2018: changes in contractor share, unplanned turnover and tenure set against each agency's own service measures, with a lag of one to three years. The hypothesis predicts that agencies losing capability first degrade first.
Cite Erosion of public-service capability and institutional memory
macrohard.nz, “Erosion of public-service capability and institutional memory”, The Hollowing Index, tested 2026-09-30. https://macrohard.nz/causes/public-service-capability
Open0 of 4 support, 1 against; partly supported needs two more.
What would settle it: A district panel from 2014: elective performance and hospital demand against the 65+ share and its growth, and a peer panel of ageing against public investment and waiting times. Older, faster-ageing places should degrade most.
Cite Demographic ageing and fiscal pressure crowding out public goods
macrohard.nz, “Demographic ageing and fiscal pressure crowding out public goods”, The Hollowing Index, tested 2026-09-30. https://macrohard.nz/causes/demographic-ageing
Open0 of 4 support, 0 against; partly supported needs two more.
What would settle it: Match every disclosed donation to the donor's sector and to later decisions benefiting that sector, and compare decision rates for donors with matched non-donors. A null result refutes the access pathway; a large gap supports it.
Open0 of 4 support, 0 against; partly supported needs two more.
What would settle it: Look for control, not perception: an official finding of improper influence over a spending or regulatory decision, or beneficiaries winning decisions at rates above matched non-beneficiaries. Continued absence as symptoms deepen would refute it as the main cause.