Six of the eight Australian states and territories set the minimum age of criminal responsibility at 10 — the Northern Territory reversed its rise to 12 in October 2024, and a national process running since 2018 has produced no binding agreement to lift the age to 14

What is happening?

Australia has no national minimum age of criminal responsibility. Each of the eight states and territories sets the age in its own law, and the Commonwealth does not set one for the state and territory offences that make up the large majority of these matters.

As at August 2026, six of the eight set it at 10: New South Wales, Queensland, Western Australia, South Australia, Tasmania and the Northern Territory. Five of those six have never changed it. The Northern Territory is the exception in a different direction — it raised the age to 12 in 2023 and returned it to 10 on 2024-10-17, through the Criminal Code Amendment Bill 2024 passed in the October 2024 sitting, two years after the rise.

Victoria sits at 12 and stopped there. In April 2023 the state announced a two-stage plan, to 12 by the end of 2024 and to 14 by 2027 with a set of exceptions. On 2024-08-13 it abandoned the second stage. The rise to 12 commenced in September 2025, so half of the announced plan is law and the other half was withdrawn.

The Australian Capital Territory is the only jurisdiction at 14. It moved in two steps, to 12 on 2023-11-22 and to 14 on 2025-07-01, with exceptions for a small number of very serious offences.

Between the minimum age and 14 the presumption of doli incapax applies. For a child aged 10 to 13 the prosecution cannot establish criminal responsibility unless it proves beyond reasonable doubt that the child knew the act was seriously wrong. That is a rebuttable presumption, and it is tested in a courtroom after the arrest and the charge have already happened.

A national process has been running since 2018. A working group under the Council of Attorneys-General, later the Standing Council of Attorneys-General, examined a single national age. Its 2020 report recommended raising the age to 14 across the country, and the report was not released until December 2022 because the jurisdictions did not reach agreement. The group was reconvened in August 2022. Whether any binding conclusion followed was not established in this round.

Whose problem is this?

No single body can change this number, because the number is written in eight separate statutes and each of the eight can move without the others.

Where does this problem end?

The boundary here is the threshold itself and not what happens above it. A jurisdiction can leave the age at 10 and still detain almost nobody, and a jurisdiction can raise the age and make the sentences above it harsher. This document measures where the line sits, who moved it, and in which direction.

What is the state now, and what should it be?

Now

What it should be has two answers in this record, and only one of them was set by the governments that decide. That one is Closing the Gap target 11, which asks for the Indigenous detention rate for ages 10 to 17 to fall at least 30 percent below the 2018-19 baseline of 28.3 per 10,000 by 2031. It carries a date, it is measured on a published dashboard, and the assessment recorded for 2024-25 is no change from the baseline at low confidence. It names no minimum age.

The figure of 14 comes from the other answer, which is a set of recommendations. The attorneys-general working group recommended a single national age of 14 in its 2020 report, Disability Royal Commission recommendation 8.22 asked every government to reach 14 with no exceptions in 2023, and the United Nations committee on 2026-05-05 measured the current ages against an international standard of the same order. Three separate bodies name the same figure and none of them can set it, which the responses to 8.22 record directly: two governments did not accept it and three referred it for further consideration. This document therefore carries 14 as the value those bodies recommend rather than as a threshold any jurisdiction outside the Australian Capital Territory has undertaken to reach. The gap is not in the target. It is between the target and any instrument that would make it arrive.

How big is it?

Six of the eight jurisdictions set the age at 10, one sets it at 12, and one sets it at 14. That distribution is the size of the problem stated in the unit the problem is actually written in, because the age is a number in a statute and there are eight statutes.

The count that this document cannot give is the number of children who live under a minimum age of 10. That number is the whole of the exposure and it needs a single year of age population for the band from 10 to 13, split across the six jurisdictions. No source opened in this round carries it. The impact population is therefore recorded as not-derivable, and the missing term is stated with it.

The figures that do exist describe the sharp end rather than the exposure. In 2024-25 the detention rate for ages 10 to 13 was 3.6 per 10,000 against 47.3 per 10,000 for ages 14 to 17, so detention in the younger band is roughly an order of magnitude rarer than in the older one. That comparison is the strongest quantitative statement available here, and it cuts in two directions at once: it shows that the minimum age of 10 produces few detentions, and it shows that the arrests, charges and court appearances it does authorise fall on a very small group.

Who that group is has a clearer answer than how large it is. Aboriginal and Torres Strait Islander children were about 65 percent of the average daily detention population in 2023-24 while making up 6.5 percent of the population aged 10 to 17. The Closing the Gap series for the Indigenous detention rate for ages 10 to 17 ran 28.3 per 10,000 in 2018-19, fell to 20.5 in 2020-21 and stood at 25.7 in 2024-25, which is better than the baseline and worse than the best year in the series, and which the dashboard assesses as no change from the baseline at low confidence against a target of at least a 30 percent reduction by 2031.

Under what conditions does it arise?

1. The power is in eight places and the recommendation is in one. Every review that has examined this question has addressed itself to all governments at once, and no government needs any other one to act. A body that can only recommend to eight separate decision makers has to win eight times, and it fails if it loses once. The 2020 working group report is the clearest instance — it recommended a single national age of 14 and could not even be published for two years, because publication itself required an agreement that did not exist.

2. The national table is a table of agreement, not of decision. The Standing Council of Attorneys-General is where this question goes, and what it produces is a position that each jurisdiction then chooses whether to legislate. There is no mechanism in the record here by which a majority at that table binds a minority, and no Commonwealth power over the state offences at issue that would substitute for one.

3. A rise can be reversed by the next parliament. The Northern Territory raised the age to 12 in 2023 and a subsequent government returned it to 10 in October 2024. Nothing in the arrangement makes a rise harder to undo than to enact, so a jurisdiction that moves first carries the risk of moving alone and being moved back, while a jurisdiction that waits carries no risk at all.

4. The threshold and the presumption do their work at different moments. Doli incapax filters heavily, and in New South Wales the proportion of matters for children aged 10 to 13 that ended in a proven outcome fell from 76 percent in 2015-16 to 16 percent in 2022-23, with more than half of matters withdrawn by the prosecution in 2022-23. But that filter operates in court. The arrest, the charge and the time between them happen first, and they happen to every child the minimum age exposes, including the large majority whose matters do not survive the presumption.

5. The direction of travel in some jurisdictions is toward severity above the threshold rather than toward the threshold itself. Queensland assented to an adult sentencing expansion in December 2024 and widened it in 2025 to 47 offences, applying adult maximum, minimum and mandatory penalties to children from age 10 and removing detention as a last resort. The minimum age was untouched, which means the threshold can stay where it is while what sits above it gets heavier.

6. The children affected are concentrated in the group with the least weight in the decision. About 65 percent of the average daily detention population in 2023-24 was Aboriginal and Torres Strait Islander children, from 6.5 percent of the age group. A cost that falls on a small and specific population, and a benefit that is claimed for the general public, do not enter an electoral calculation on the same terms.

What has been tried?

Three kinds of instrument have been tried and none of them binds. Reviews recommend, treaty bodies observe, and individual parliaments legislate for their own territory only. The one jurisdiction that reached the recommended age did so alone, the one that reversed did so alone, and the national process that exists to make those decisions the same decision has been running for eight years without producing a common answer.

What was found?

Why is it still unsolved?

Coordination failure — eight jurisdictions each hold the whole of the decision for their own territory, and the national process that exists to align them can recommend but cannot bind.

A recommendation that every jurisdiction must accept before it takes effect is a recommendation that any one of them can defeat by doing nothing. That is the shape of the record here. The working group recommended a single national age of 14 and the report sat unreleased for two years because the jurisdictions did not agree, and when it was released it changed the law in none of them.

The second reason is that the incentives run against moving first. A jurisdiction that raises the age acts alone, carries the political cost alone, and can be reversed by its own next parliament — which is not hypothetical, because the Northern Territory raised the age in 2023 and returned it to 10 in October 2024. A jurisdiction that waits carries no cost at all and keeps the option. Victoria shows the intermediate case: it announced two steps and took one, so the state that moved furthest short of the Australian Capital Territory still ended up below the figure every review names. When moving is reversible and waiting is free, the equilibrium is the status quo in six places and a scatter in the other two.

The third reason is that the presumption of doli incapax makes the threshold look less consequential than it is, and therefore easier to leave alone. In New South Wales the proportion of matters for children aged 10 to 13 ending in a proven outcome fell to 16 percent in 2022-23 and the prosecution withdrew more than half of them. Read one way that is a system already declining to convict very young children, which weakens the case for urgency. Read the other way it is a system that arrests and charges children in order to discover, months later in a courtroom, that it should not have. The filter sits after the part that the minimum age controls, so the argument that the filter is working is also a description of how many children pass through the part that is not filtered.

The fourth reason is that the pattern is stable because nothing about it produces an event. A jurisdiction that keeps its age at 10 announces nothing, publishes nothing and misses no deadline. Recommendations arrive from royal commissions, working groups and treaty bodies and none of them has a date attached that anybody must meet. The one measure with a date, Closing the Gap target 11, is assessed as no change from the baseline at low confidence, and that assessment carries no consequence either. A position that requires no defence is cheaper to hold than any position that requires a decision.

What observation would mean it is solved?

Candidates — (a) all eight jurisdictions set the minimum age at 14 and none reverses within a full parliamentary term (b) arrests, charges and detentions of children aged 10 to 13 fall and stay near zero (c) the Closing the Gap indicator for Indigenous detention in the 10 to 17 band falls toward the 2031 target.

(a) is the direct test and it is not sufficient on its own. The Australian Capital Territory reached 14 with exceptions for a small number of very serious offences, and Victoria announced a rise to 14 that also carried exceptions before abandoning it altogether. An age of 14 with a wide enough carve-out and an age of 10 can produce the same arrests. Durability is the other half of it, which is why a reversal window has to be part of the observation — the Northern Territory satisfied a version of this test in 2023 and failed it in 2024.

(b) is closer to what the threshold is actually for and it is the hardest to read. The detention rate for ages 10 to 13 was 3.6 per 10,000 in 2024-25, which is already low, and no source opened here publishes a national count of arrests or charges in that band at all. A fall in the count could mean fewer children are being brought into the system or it could mean the same children are being handled through some other route that produces no count. Detention is also the last stage of a sequence the minimum age governs from the first, so it measures the residue rather than the exposure.

(c) counts the thing that matters most and attributes it least. The Indigenous detention rate for ages 10 to 17 covers eight years of age, of which the minimum age governs four, and it moves with policing, bail law, sentencing and court capacity in eight jurisdictions at once. It fell from 28.3 in 2018-19 to 20.5 in 2020-21 and rose again to 25.7 in 2024-25 without the minimum age changing anywhere except two territories. A movement in this series cannot be assigned to the threshold without a decomposition that no opened source provides. The three have to be read together, and (a) has to be read with a reversal window attached.

What is it connected to?

Fills with researchchild protection and out-of-home care as the pathway into police contact, policing and bail practice for the 10 to 13 band, school exclusion, conditions inside youth detention, deaths in custody and the prosecutions that follow them, and the minimum age set in comparable jurisdictions outside Australia. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How many children are exposed. The Australian Bureau of Statistics distributes single year of age population as spreadsheets, and those were not opened in this round, so no count exists here for the 10 to 13 band nationally or by jurisdiction. The only national figure obtained is a total resident population of 27,400,013 at December 2024, which cannot be turned into an age band without a term that no opened source supplies.
• How many children aged 10 to 13 are arrested, charged or detained in a year. The national youth detention report returned HTTP 403 on both its page and its document, so a nightly detention average of 884 for the June quarter of 2025 and a share of 5 percent aged 10 to 13 reach this document only through search summaries. Those two figures and the published rate of 3.6 per 10,000 do not sit on a common base that any opened source states, so neither this document nor its impact population estimate uses them.
• Whether the national process is still running. The 2020 working group report text was not read and the record after the August 2022 reconvening was not examined. Whether the process formally concluded, lapsed or continues is unresolved, and that distinction is the difference between a failed negotiation and an open one.
• What each government actually said about recommendation 8.22. The territory government page and the federal publication of the recommendation both refused automated requests, so the wording of the responses reaches this document through a summary published by a state guardian office. Where a government qualified its answer, this document has no way to detect the qualification.
• What the New South Wales review recommended in full, and what will be done. Seven recommendations including codification were reported, and the text of them was not read. The government answer as at 2025-10-25 was that it would consider the review and set out a reform path, and no decision has been published since.
• What the Senate inquiry will find. The report is due 2026-08-31, which is after this status check date, so nothing here anticipates it.
• How many children aged 10 to 13 are in Northern Territory detention. The reporting gives a comparison, the highest rate in the country and close to double the national average, and no figure.
• The basis of the overrepresentation ratio in the advocacy explainer opened here. That document dates from 2022-09-08 and names no primary source for its headline ratio, so the ratio is not used anywhere in this document and the overrepresentation figures above come from a different source with their own as-of.

See the evidence

No Australian primary government record of the minimum age itself was read directly. The spine of the jurisdictional picture is an encyclopedia table, corroborated jurisdiction by jurisdiction where an opened source happened to touch it — the Northern Territory position by an opened news report, the Victorian position by an opened premier release and an opened commissioner statement, the Queensland direction by an opened state government page. Where those corroborations exist the figures agree. Where they do not exist, and that includes the Australian Capital Territory step to 14 and the exact date of the Northern Territory reversal, this document rests on the table alone and the findings above are graded accordingly. Twelve of the nineteen rows carry a URL. Seven are blank with the reason recorded rather than filled with a substitute link, and among those seven are four documents that would be the primary record for claims this document makes: the federal publication of recommendation 8.22, the parliamentary inquiry page, the working group report and the territory news release. The strongest evidence here is quantitative and comes from two sources that opened cleanly — a government dashboard for the detention rates and the Closing the Gap series, and a departmental media release for the New South Wales doli incapax figures. The weakest is the national process, which reaches this document as a summary of a report that was never read, and that weakness is load-bearing, because the claim that a national push has stalled since 2018 rests on it. One row is present and deliberately unused: an advocacy explainer whose headline ratio names no primary source, kept in the table because it was opened in research and excluded from the text because it cannot be traced further. Two figures that would have sized the exposed population are similarly present and unused, the nightly detention average and the share aged 10 to 13, because the report carrying them refused every automated request and the numbers reached research only as search summaries that cannot be reconciled with the published rate. This is a Path A output, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range Not derivable

The reason and what is missing are listed under “What is missing” below