The Territory built the response and never built the scoreboard.
Between 2016 and 2026 the Australian Capital Territory built most of the components of a serious domestic and family violence system, paid for partly by a levy on every household. The architecture is far stronger than it was. What was never built is the other half — the part that measures whether any of it works.
This is the public the findings in fifteen minutes, with every figure drawn from a document anyone can obtain.
Four numbers, each checkable in an afternoon
No performance monitoring strategy or evaluation framework for the Safer Families Levy — ACT Auditor-General, Report 10 of 2024.8 of 13 deceased victims had been wrongly recorded by police or another agency as the person causing the violence, while alive — ACT death review, 2025.0 of 17 requirements of AS ISO 31000:2018 — the risk standard the ACT Government mandates for everything it does — are fully met by its 2022 family violence risk framework.2019–20 was the last year the Magistrates Court published protection-order application statistics.Underneath sits a single technical failure. The ACT requires risk to be managed to a recognised standard across all of its functions. Procurement is assessed that way. Public events are assessed that way. Staff travel is assessed that way. The process that can remove a person from their home without notice is not — and so the Territory cannot tell when it gets that process wrong.
It is also an assessment of the two bills now before the the Family, Personal and Sexual Violence Legislation Amendment Bill 2025 and the Crimes (Coercive Control) Amendment Bill 2026. Three best and three worst features of each, set out plainly. Both are well-intentioned. Both respond to real failures. Both attach serious consequences to a decision the Territory has never assessed against written criteria, never recorded in a form a court can read, and never counted.
And it is constructive. Seven things that could start this week. Five reforms, with the legislation drafted and the cost estimated. Four checks any reader can run in an afternoon without trusting a word of it.
The verdict is "not proven" — not poor, and not good either. Several programs show promising results and there is no evidence of waste in the ordinary sense. There is also no evidence to support a finding of value for money, because the data needed to answer the question is not collected in a form that permits it.
Julian Talbot is a risk-management practitioner, author of a dozen books including Control Effectiveness Criteria, Evidence, and a Scale Two Assessors Can Share, lead author of the Security Risk Management Body of Knowledge, a Fellow of the Risk Management Institution of Australasia and of the Institute of Strategic Risk Management. This edition contains no argument that is not made at length, with sources, in the companion technical volume.