WeRise · Policy platform
Criminal Justice
The human story
Why Criminal Justice matters
A safer justice system gives victims choice, demands repair and reduces the churn into prison.
Australia’s criminal justice system is becoming more expensive, more unequal and less capable of delivering the outcomes victims reasonably expect: safety, answers, acknowledgement and a genuine reduction in future harm. At 30 June 2025, 46,998 adults were in Australian prisons – an annual increase of 2,595 people, or 6%.1 This is not evidence of a system that has solved crime. It is evidence of a system increasingly reliant on custody as its central response.
The harm is not distributed equally. Aboriginal and Torres Strait Islander people made up 17,432 prisoners in 2025.2 In the March quarter of 2026, the average daily Aboriginal and Torres Strait Islander imprisonment rate was 2,701 per 100,000 adult First Nations people.3 For children, the injustice is even starker: in 2023–24, First Nations young people aged 10–17 were detained at 27 per 10,000, compared with 1.0 per 10,000 non-Indigenous young people – about 27 times the rate.4
The alternative is not “restorative justice or public safety.” It is a system that combines firm legal accountability with stronger pathways to repair, rehabilitation and prevention. Australia spent $5.43 billion on prisons alone in 2024–25, while the average recurrent cost per prisoner was $440.60 per day when capital costs are included.9 A restorative justice framework will not eliminate those costs overnight. But it can help redirect a small, disciplined share of justice spending toward outcomes that prison often fails to produce: recognised harm, completed repair and fewer future victims.
The practical case
Criminal Justice: the findings to notice
People and everyday life
Who Criminal Justice affects
Victims, young people, First Nations communities and the public all pay when accountability stops at punishment.
The first people affected are victims and survivors. Many want conventional prosecution and sentencing; that option must remain available. Others want information, acknowledgement, practical repair, a voice in what safety requires or the chance to ask questions a court process cannot answer. Restorative justice expands choice – it does not impose reconciliation. New Zealand’s 2016 victim survey found 84% of surveyed restorative-justice participants were satisfied with their conference and 81% would recommend it to someone in a similar situation.15 Those findings are encouraging, but Australian implementation must measure survivor wellbeing and adverse outcomes rather than assume overseas results transfer automatically.
Case study
A tenant in regional Queensland has their home damaged by a 17-year-old neighbour during an impulsive break-in. The court can order a penalty, but it may not answer why the young person targeted the home, whether they understand the fear caused or how the damage will be repaired. In a voluntary, professionally facilitated conference, the tenant can bring a support person, ask questions, describe the impact and negotiate a realistic repair plan. The young person must acknowledge responsibility, complete repair and connect with practical supports. If either party does not consent or risk assessment finds the process unsafe, the case proceeds through ordinary pathways.
Children and young people are also directly affected. In 2023–24, 499 First Nations young people were in detention on an average day, and 460 First Nations children aged 10–17 accounted for 65% of all detained children in that age group.16 First Nations children aged 10–17 were detained at 27 per 10,000, compared with 1.0 per 10,000 non-Indigenous children.17 These figures do not mean that First Nations children are inherently more harmful; they show a justice system interacting with entrenched structural inequality. Community-led restorative pathways can reduce unnecessary escalation, but they must be governed and delivered with First Nations communities rather than simply targeted at them.
The direction of change
What better Criminal Justice could look like
Commonwealth standards and funding; state justice powers; First Nations governance; victim-led safeguards.
Australia should establish a National Restorative Justice Framework through a National Cabinet agreement, supported by tied Commonwealth funding and delivered by states and territories. Criminal law, policing, courts, prisons and most youth justice functions are state and territory responsibilities. The Commonwealth’s practical role is to fund capability, set nationally consistent minimum standards, support research and data infrastructure, and use Closing the Gap mechanisms to embed shared decision-making with Aboriginal and Torres Strait Islander peoples.
The framework should begin with four linked streams. First, every jurisdiction should provide an early-intervention youth conferencing pathway for eligible children and young people, with referral available before deep court involvement where lawful and safe. Second, adult victim–offender conferencing should be available pre-sentence, post-sentence and post-release for eligible matters where responsibility is accepted.
Third, victims should have access to restorative communication options even when a direct conference is unsuitable – for example, shuttle dialogue, written questions and answers, indirect accountability processes or victim-impact dialogue. Fourth, specialist domestic, family and sexual violence pathways should be limited to independently assessed cases, operate under survivor-led consent and include specialist trauma, coercive-control and cultural-safety expertise.
Costs and choices
The trade-offs behind Criminal Justice
The fiscal case is not that restorative justice is free – it is that custody is extraordinarily expensive and often poorly targeted.
Restorative justice requires real investment. A credible national framework needs trained and independently accredited facilitators, victim advocates, legal advice, culturally safe First Nations-led services, interpreters, therapeutic pathways, data systems, evaluation and secure case management. It would be irresponsible to claim a national net saving before jurisdictions have costed a common model and run implementation pilots. All proposed program-cost figures below are therefore projections, not established budget estimates.
Read the complete costs and trade-offsFollow the evidence
Read further on Criminal Justice
- ABS Prisoners in Australia, 2025. 30 June 2025 snapshot; release December 2025 with April 2026 update. Use reference date rather than generic 2026 label.
- AIHW Youth justice in Australia 2023–24, First Nations young people in detention, Table S75a. Ages 10–17, average day, 2023–24; 27 versus 1.0 per 10,000. Not all youth justice supervision.
- ACT observational evaluation 2018, executive summary PDF p5 and results PDF p14. Observational adjusted association, not randomised causal proof or guaranteed Australian implementation effect.
- Productivity Commission, Report on Government Services 2026, Corrective services. Tables 8A.1 and figure 8.11: 2024–25 prison expenditure and average daily costs including capital. Average costs are not marginal savings.
Criminal Justice sits within a wider set of WeRise policies.
- Foundations of Dignity: more WeRise policies in this pillar
- WeRise: all 29 policies across seven pillars
Key findings
- 46,998 adult prisonersAustralia’s adult prison population reached 46,998 people on 30 June 2025, up 6% in one year. Source: Australian Bureau of Statistics · Prisoners in Australia, 2025 · 2026
- 17,432 First Nations prisonersAboriginal and Torres Strait Islander people made up 37% of Australia’s adult prison population in the 2025 prison census. Source: Australian Bureau of Statistics · Prisoners in Australia, 2025 · 2026
- 27×First Nations young people aged 10–17 were about 27 times as likely as non-Indigenous young people to be in detention on an average day in 2023–24. Source: Australian Institute of Health and Welfare · Youth Justice in Australia 2023–24 · 2025
- 23% fewer offencesIn the ACT’s long-run observational evaluation, restorative-conference participants had a 23% lower frequency of offending after statistical adjustment than comparable young people processed conventionally. Source: Broadhurst et al. / ACT Justice and Community Safety Directorate · Restorative Justice: An Observational Outcome Evaluation of the ACT Program · 2018
Part 01
The Case for Change
Why Criminal Justice matters
Punishment alone is not accountability
A safer justice system gives victims choice, demands repair and reduces the churn into prison.
Australia’s criminal justice system is becoming more expensive, more unequal and less capable of delivering the outcomes victims reasonably expect: safety, answers, acknowledgement and a genuine reduction in future harm. At 30 June 2025, 46,998 adults were in Australian prisons – an annual increase of 2,595 people, or 6%.1 This is not evidence of a system that has solved crime. It is evidence of a system increasingly reliant on custody as its central response.
The harm is not distributed equally. Aboriginal and Torres Strait Islander people made up 17,432 prisoners in 2025.2 In the March quarter of 2026, the average daily Aboriginal and Torres Strait Islander imprisonment rate was 2,701 per 100,000 adult First Nations people.3 For children, the injustice is even starker: in 2023–24, First Nations young people aged 10–17 were detained at 27 per 10,000, compared with 1.0 per 10,000 non-Indigenous young people – about 27 times the rate.4
Restorative justice does not mean excusing crime, forcing survivors to forgive or replacing courts in every case. It means creating a rigorously assessed option through which a victim can choose to ask questions, describe impacts, set conditions for repair and receive acknowledgement; meanwhile, the person responsible must accept responsibility, engage with the consequences of their conduct and complete tangible reparative actions. The process can sit beside prosecution and sentencing rather than displace them.5
There is Australian evidence that this can improve outcomes. The ACT’s long-run evaluation compared 1,143 restorative conferencing participants with 4,668 young people dealt with through conventional processes. The authors found that, after adjusting for differences between groups, restorative participants committed offences 23% less frequently over the follow-up period; matched analysis estimated a 30% reduction.6 At 12 months, the matched estimate of re-apprehension was 20% for restorative participants and 29% for comparable non-participants.7 These figures do not prove that every conference prevents crime. They show that carefully designed conferencing can be associated with substantial reductions in the frequency and timing of future offending.
The victim case is equally important. In the ACT’s earlier program evaluation, 93% of participants reported being pleased with the conference outcome, while 97–99% reported being treated with respect, able to speak and believing the process was fair.8 Such outcomes matter because procedural dignity is not a public-relations extra; it is part of what justice owes people harmed by crime.
The alternative is not “restorative justice or public safety.” It is a system that combines firm legal accountability with stronger pathways to repair, rehabilitation and prevention. Australia spent $5.43 billion on prisons alone in 2024–25, while the average recurrent cost per prisoner was $440.60 per day when capital costs are included.9 A restorative justice framework will not eliminate those costs overnight. But it can help redirect a small, disciplined share of justice spending toward outcomes that prison often fails to produce: recognised harm, completed repair and fewer future victims.
Part 02
Evidence and Experience
Criminal Justice: what other countries show
What rigorous international evidence shows
New Zealand · United Kingdom · United States · Australian Capital Territory comparison
International evidence supports restorative justice as a targeted, voluntary and professionally facilitated complement to conventional justice – not as a universal substitute. The strongest findings are generally in victim experience, procedural fairness, restitution completion and modest reductions in repeat offending. Effects vary by program type, offence, participant readiness and study design.
The United States Office of Juvenile Justice and Delinquency Prevention cautions that more rigorous evaluations often find smaller effects than weaker studies, underscoring the need for independent Australian trials and transparent reporting.10 New Zealand’s national experience demonstrates that restorative processes can be embedded within a mainstream justice system, while UK randomised trials provide rare cost-effectiveness evidence. The ACT is included as a directly relevant domestic comparison system because it demonstrates what statutory, specialist implementation can look like in an Australian jurisdiction.
New Zealand · 2008–2013 cases; outcome analysis published 2016
Police- or court-referred adult restorative justice conferences delivered through approved providers; no universal participant payment. Agreements can include apology, reparation, treatment engagement or other negotiated repair. Matched national analysis of restorative justice cases from 2008–2013; the analytic sample size has not yet been independently checked for this discussion draft.
- 34% of restorative justice participants reoffended within 12 months, compared with 39% of matched non-participants.
- This represents a 5-percentage-point reduction in 12-month reoffending.
- Restorative participants committed 26% fewer offences per offender in the following 12 months than matched non-participants.
- The reoffending reduction was reported as 15% in the following 12 months and 7.5% over three years.
- In the 2016 victim survey, 84% of surveyed victims were satisfied with their conference and 81% would recommend restorative justice.
New Zealand shows that restorative conferencing can be delivered at a national-system scale while retaining victim choice and independent facilitators. Australia should not copy the model mechanically: it should embed Aboriginal and Torres Strait Islander governance, local legal safeguards and consistent national evaluation standards.
England and Wales · 2001–2008 trials; synthesis published 2013 and 2015
Restorative justice conferences offered alongside prosecution and sentencing for eligible offences; no cash payment to participants. Delivery costs were compared with estimated crime costs prevented. Seven UK tests in the restorative justice conferencing cost-effectiveness analysis; part of a broader program of 12 randomised trials across Australia and the United Kingdom.
- Across the UK experiments, the estimated benefit–cost ratio was £8 in crime costs prevented for every £1 spent delivering restorative conferences.
- The seven UK tests were assessed as cost-effective over a two-year follow-up.
- Victims assigned to conferences were less fearful of repeat attack, more satisfied with case handling and less likely to desire violent revenge than control victims.
- The program’s findings are strongest for victim outcomes and reductions in repeat offending for some, not all, participants.
UK evidence suggests that a restorative service can produce benefits even when it supplements rather than replaces prosecution and sentencing. Australia should independently cost its own model; UK benefit–cost figures cannot be transferred directly because prison, policing, court and victim-service costs differ.
United States – four victim–offender mediation sites · Evaluation published 1993; follow-up generally one year
Victim–offender mediation associated with juvenile courts in Albuquerque, Minneapolis, Oakland and Austin; no participant payment. Agreements could include restitution and other reparative commitments. Four program sites; the study used multiple datasets, comparison groups and participant surveys.
- Offenders who met victims completed restitution at a rate of 81%, compared with 58% for similar people in court-administered restitution programs without mediation.
- Reported reoffending was 18% for mediation participants and 27% for comparable non-participants.
- Subsequent offences by mediation participants tended to be less serious.
- Victim–offender mediation produced high levels of client satisfaction and perceived fairness.
- The cross-site evaluation cautioned that lower overall recidivism was not statistically significant across all analysed comparisons.
Restitution and concrete repair should be central measures of accountability, not decorative additions. The mixed recidivism evidence also reinforces the case for strong evaluation rather than treating restorative justice as automatically effective.
Australian Capital Territory · Participants followed to September 2016; evaluation published 2018
Statutory restorative justice conferencing for young people; no participant payment. Facilitated conferences can create repair agreements and operate as a diversion or alongside formal justice pathways. 1,143 restorative conferencing participants compared with 4,668 young people processed conventionally.
- The multivariate analysis estimated a 23% lower frequency of offending for restorative participants over the average follow-up period.
- Matched analysis estimated a 30% lower frequency of offending.
- At 12 months, estimated re-apprehension was 20% for matched restorative participants and 29% for the matched comparison group.
- At 10 years, estimated re-apprehension was 54% for restorative participants and 64% for matched non-participants.
- The evaluation warns that unadjusted group comparisons can mislead because restorative participants had different baseline characteristics and risks.
The ACT provides the most immediately transferable evidence for national reform: specialist facilitation, legislated safeguards and long-term administrative-data evaluation. Its results justify expansion, but not complacency: each jurisdiction should publish comparable outcomes, including cases that do not proceed and reasons for non-participation.
Part 03
Costs and Trade-offs
The cost of Criminal Justice
Spend less on churn, more on repair
The fiscal case is not that restorative justice is free – it is that custody is extraordinarily expensive and often poorly targeted.
Restorative justice requires real investment. A credible national framework needs trained and independently accredited facilitators, victim advocates, legal advice, culturally safe First Nations-led services, interpreters, therapeutic pathways, data systems, evaluation and secure case management. It would be irresponsible to claim a national net saving before jurisdictions have costed a common model and run implementation pilots. All proposed program-cost figures below are therefore projections, not established budget estimates.
The current cost baseline is clear. Australia spent $5.43 billion in real net operating expenditure on prisons in 2024–25, and the Productivity Commission reported recurrent expenditure of $440.60 per prisoner per day including capital costs.11 This equates to approximately $160,819 per prisoner-year if the daily rate is annualised, although it should not be treated as a marginal saving from every avoided prison day: prison infrastructure, staffing and contracts create fixed costs. []
The economic case is strongest where restorative practice diverts suitable matters from deeper system contact, improves agreement compliance, reduces offending frequency or helps prevent a return to custody. The United Kingdom’s experimental evidence estimated £8 in crime costs prevented for every £1 spent on restorative conferencing, but this cannot be converted directly into an Australian budget forecast.12 Australia should publish gross costs and realised offsets separately, conduct an independent cost-benefit evaluation after three years and make continuation funding contingent on safety, victim experience, equity and outcome thresholds.
| Item | Amount | Note |
|---|---|---|
| National restorative justice capability design, standards, accreditation and data framework | Not yet costed nationally | Commonwealth-led design work, national minimum standards and interoperable outcome reporting; a formal costing should precede implementation. |
| State and territory delivery – facilitators, victim advocates, legal and therapeutic supports | Not yet costed nationally | Recurring program cost depends on eligibility, referral volume, rural delivery and specialist practice requirements. |
| Average recurrent prison expenditure | −$440.60 per prisoner per day | National recurrent cost per prisoner per day, including capital costs, in 2024–25; this is a system-cost benchmark, not automatically cash-releasable savings. |
| Annualised recurrent prison-cost benchmark | −$160,819 per prisoner-year | $440.60 times 365 ; derived annual comparison only. |
| Australian prison operating expenditure | −$5.43 billion in 2024–25 | Net operating expenditure on prisons nationally; excludes some associated costs such as prisoner health and transport/escorts under reporting definitions. |
| International restorative-conferencing benefit–cost result | +£8 in crime costs prevented per £1 spent | UK experimental estimate; informative but not transferable as an Australian fiscal forecast. |
| Net cost estimate row | Undetermined pending Australian pilot costings and independent evaluation | WeRise should reject invented savings claims; publish gross expenditure, avoided service use and net fiscal effects after pilot evaluation. |
The strongest objections to Criminal Justice
This is the strongest intuitive objection: wrongdoing deserves punishment, and a conference, apology or agreement may appear inadequate – especially to victims of serious violence.
The response is not to deny the value of proportionate punishment or public denunciation. Restorative justice should not be compulsory, should not replace prosecution where safety or public interest requires prosecution, and should not be used to discount serious offending automatically. Its distinct contribution is more demanding personal accountability: a person responsible must confront the impact of their conduct, answer questions where appropriate and complete repair. For many offences, that may be more meaningful than a passive sentence. Evidence supports modest reductions in repeat offending in some settings, but not a blanket deterrence claim.
This risk is real.
Victims may feel implicit pressure from family, police, prosecutors, service providers or a desire to appear cooperative. The risk is particularly acute in domestic, family and sexual violence, where coercive control may continue after an offence. The answer is a survivor-led model: independent advocacy, extensive preparation, free legal advice, separate meetings, withdrawal at any time, no adverse consequences for refusal and specialist risk assessment. The ACT’s Phase Three evaluation found meaningful victim-survivor benefits but also barriers, limited referrals and implementation constraints. That argues for cautious specialist practice, not indiscriminate expansion.
Correct.
Restorative programs vary substantially, and people who participate may differ from people who do not. The ACT evaluation itself found that raw comparisons could imply worse long-term re-apprehension for participants, while adjusted and matched analyses found reduced offending frequency. US evidence reviews likewise find that stronger research designs tend to show smaller positive effects. WeRise should not promise that restorative justice “stops crime.” The defensible claim is that it has credible evidence of victim and procedural benefits, with promising but variable evidence on reoffending, warranting independently evaluated scale-up.
Correct again.
Restorative justice cannot substitute for housing, health care, disability support, income security, school inclusion, bail reform, legal assistance or community-controlled prevention. It can reduce system contact in selected cases and shift power toward community-led responses, but it is one element of a decarceration strategy. Presenting it as a complete answer would obscure structural causes identified by the ALRC and Closing the Gap data.
A centrally designed program could reproduce exactly the problem it seeks to address: standardisation without self-determination.
The remedy is to distinguish national rights and safeguards from uniform delivery. National standards should guarantee consent, independence, data transparency and funding; local models should be co-designed and governed with First Nations communities, victim-survivor services and regional organisations. Funding should flow to Aboriginal and Torres Strait Islander-controlled services where communities choose that approach.
Part 04
The Australian Design
How Criminal Justice would work in Australia
A national guarantee, delivered locally
Commonwealth standards and funding; state justice powers; First Nations governance; victim-led safeguards.
Australia should establish a National Restorative Justice Framework through a National Cabinet agreement, supported by tied Commonwealth funding and delivered by states and territories. Criminal law, policing, courts, prisons and most youth justice functions are state and territory responsibilities. The Commonwealth’s practical role is to fund capability, set nationally consistent minimum standards, support research and data infrastructure, and use Closing the Gap mechanisms to embed shared decision-making with Aboriginal and Torres Strait Islander peoples.
The framework should begin with four linked streams. First, every jurisdiction should provide an early-intervention youth conferencing pathway for eligible children and young people, with referral available before deep court involvement where lawful and safe. Second, adult victim–offender conferencing should be available pre-sentence, post-sentence and post-release for eligible matters where responsibility is accepted.
Third, victims should have access to restorative communication options even when a direct conference is unsuitable – for example, shuttle dialogue, written questions and answers, indirect accountability processes or victim-impact dialogue. Fourth, specialist domestic, family and sexual violence pathways should be limited to independently assessed cases, operate under survivor-led consent and include specialist trauma, coercive-control and cultural-safety expertise.
The ACT offers important implementation lessons. Its scheme has statutory foundations, a dedicated unit and specialist expansion into domestic, family and sexual violence. The Australian Institute of Criminology found Phase Three was working effectively overall, while identifying limited referrals, stakeholder reluctance and resourcing delays as practical obstacles.13 In 2023–24, the ACT reported 27 agreements and a 96% compliance rate among agreements that had reached completion or were being monitored at the reporting point.14 This is promising but not a scale substitute for a national framework.
Minimum protections should include: voluntary and revocable participation; no adverse inference from declining; independent legal advice; informed consent; trained victim advocates; culturally safe and accessible delivery; clear confidentiality rules; evidence-based risk assessment; separate preparation meetings; and a prohibition on using restorative processes to bargain away safety or reduce accountability. A person responsible for harm must acknowledge responsibility before any direct process proceeds. Restorative participation must not become a shortcut around prosecution where public protection requires formal action.
Implementation should occur in stages. Year one should establish standards, governance, workforce mapping and data definitions. Years two and three should fund pilots in every jurisdiction, prioritising regional access and First Nations-controlled delivery. At the end of year three, an independent evaluation should report victim wellbeing, safety incidents, agreement completion, reoffending frequency, equity, costs and reasons people were excluded or declined. Years four and five should scale proven models. This is a disciplined reform program: national ambition, local delivery and public evidence.
Part 05
People and Transition
Who Criminal Justice affects
Justice should repair the harm it sees
Victims, young people, First Nations communities and the public all pay when accountability stops at punishment.
The first people affected are victims and survivors. Many want conventional prosecution and sentencing; that option must remain available. Others want information, acknowledgement, practical repair, a voice in what safety requires or the chance to ask questions a court process cannot answer. Restorative justice expands choice – it does not impose reconciliation. New Zealand’s 2016 victim survey found 84% of surveyed restorative-justice participants were satisfied with their conference and 81% would recommend it to someone in a similar situation.15 Those findings are encouraging, but Australian implementation must measure survivor wellbeing and adverse outcomes rather than assume overseas results transfer automatically.
Case study
A tenant in regional Queensland has their home damaged by a 17-year-old neighbour during an impulsive break-in. The court can order a penalty, but it may not answer why the young person targeted the home, whether they understand the fear caused or how the damage will be repaired. In a voluntary, professionally facilitated conference, the tenant can bring a support person, ask questions, describe the impact and negotiate a realistic repair plan. The young person must acknowledge responsibility, complete repair and connect with practical supports. If either party does not consent or risk assessment finds the process unsafe, the case proceeds through ordinary pathways.
Children and young people are also directly affected. In 2023–24, 499 First Nations young people were in detention on an average day, and 460 First Nations children aged 10–17 accounted for 65% of all detained children in that age group.16 First Nations children aged 10–17 were detained at 27 per 10,000, compared with 1.0 per 10,000 non-Indigenous children.17 These figures do not mean that First Nations children are inherently more harmful; they show a justice system interacting with entrenched structural inequality. Community-led restorative pathways can reduce unnecessary escalation, but they must be governed and delivered with First Nations communities rather than simply targeted at them.
Case study
A 15-year-old First Nations boy in an outer-Brisbane community is charged with property damage after repeated school exclusion, unstable housing and contact with police. A punitive pathway risks deepening distrust and disconnection. A community-controlled conference, if the victim agrees, can bring family, Elders, a youth worker and the affected small-business owner together to set a reparative plan: direct acknowledgement, safe practical restitution, school re-engagement and housing support. The goal is accountability connected to the conditions that make further harm more likely – not an excuse for the conduct.
The public benefits if restorative justice reduces future victimisation and increases compliance with repair agreements. In the ACT’s observational study, restorative participants had a statistically adjusted 23% lower frequency of offending over follow-up.18 The opposition will come from people who regard punishment as justice in itself; from agencies concerned about new workload and liability; and from victim advocates who fear coercive or poorly resourced practice. Their concerns must shape the safeguards. A model that pressures survivors, trivialises violence or displaces due process would betray restorative principles. A model that offers meaningful choice, high standards and transparent evaluation can strengthen public confidence.
Part 06
Related policies
Criminal Justice sits within a wider set of WeRise policies.
- Foundations of Dignity: more WeRise policies in this pillar
- WeRise: all 29 policies across seven pillars
Notes
Sources and Supporting Documents
- ABS Prisoners in Australia, 2025. 30 June 2025 snapshot; release December 2025 with April 2026 update. Use reference date rather than generic 2026 label.
- AIHW Youth justice in Australia 2023–24, First Nations young people in detention, Table S75a. Ages 10–17, average day, 2023–24; 27 versus 1.0 per 10,000. Not all youth justice supervision.
- ACT observational evaluation 2018, executive summary PDF p5 and results PDF p14. Observational adjusted association, not randomised causal proof or guaranteed Australian implementation effect.
- Productivity Commission, Report on Government Services 2026, Corrective services. Tables 8A.1 and figure 8.11: 2024–25 prison expenditure and average daily costs including capital. Average costs are not marginal savings.
Numbered citation mapping and the remaining international evaluation figures are still being checked for the full research document.
Record
Revision History
26 October 2026: Fred approved publication of the initial discussion draft. Australian primary references recovered. International comparisons and full citation mapping remain under review; proposed national delivery has not been independently costed.
Specific proposals
Making Criminal Justice practical
The direction set out in this policy is the starting point. This is where individual measures can be developed so people can examine what would change, who would benefit and how it could work.
As each measure develops, its evidence, delivery arrangements, costs and open questions will be set out alongside it. The areas below come from the existing policy; they are starting points for detailed proposals, not final costed commitments.
Commonwealth standards and funding; state justice powers; First Nations governance; victim-led safeguards.
Australia should establish a National Restorative Justice Framework through a National Cabinet agreement, supported by tied Commonwealth funding and delivered by states and territories. Criminal law, policing, courts, prisons and most youth justice functions are state and territory responsibilities. The Commonwealth’s practical role is to fund capability, set nationally consistent minimum standards, support research and data infrastructure, and use Closing the Gap mechanisms to embed shared decision-making with Aboriginal and Torres Strait Islander peoples.
The framework should begin with four linked streams. First, every jurisdiction should provide an early-intervention youth conferencing pathway for eligible children and young people, with referral available before deep court involvement where lawful and safe. Second, adult victim–offender conferencing should be available pre-sentence, post-sentence and post-release for eligible matters where responsibility is accepted.