In Moree, on the Newell Highway roughly seven hours' drive north-west of Sydney, shopkeepers have boarded up the same windows more than once, and the servo on the edge of town fitted steel roller doors it had never needed before. In Bourke, further west again, one of the town's few remaining businesses has grown used to reporting stolen cars that turn up burnt out on a back road days later, sometimes linked to a name police already know. Neither town is unfamiliar with hardship — both carry decades of drought, closed industries and chronic under-investment — but what changed after 2022 was the visibility of a specific pattern: a small number of children and teenagers, many already known to police and the Children's Court, cycling through arrest, bail and further alleged offending faster than the system around them could respond. That pattern, repeated across a handful of regional local government areas, is what pushed the Minns government to do something no NSW Labor government had done in a generation: make it harder, even if only temporarily, for a repeat young offender to walk back out on bail.
What the government did
Under the Bail Act 1978, a young person charged with an offence is generally entitled to a presumption in favour of release, reflecting the long-standing legal principle that detention before trial, especially for a child, should be a last resort rather than a default. In March 2024, facing sustained pressure from regional mayors, police and residents after a run of high-profile alleged offences involving stolen cars and break-ins in towns including Moree, Bourke, Coffs Harbour and Dubbo, NSW Parliament passed an amendment to the Bail Act creating a new, additional bail concern for 14 to 18-year-olds charged with serious break-and-enter, robbery or motor vehicle theft offences while already on bail for a similar alleged offence. In practice, courts and police bail authorities were now required to treat a further alleged offence committed while on bail as a serious factor weighing against release, tightening a threshold the government argued had become too easy to clear.
The law was paired with a package of regional measures — additional NSW Police resourcing for the north-west and Mid North Coast, funding for youth support, diversion and after-hours services, and community-safety grants for the towns most affected (approx.; amounts vary by budget line and program — see NSW Budget Papers, Department of Communities and Justice, Youth Justice NSW). Ministers were explicit that bail alone was not meant to be the whole answer. What made the change unusual, and politically significant, was that it was not made permanent: it carried a 12-month sunset clause, due to lapse automatically unless Parliament actively voted to renew it, a compromise struck to secure crossbench support and to test the change against evidence rather than lock it in as settled policy. As at the date of this analysis, the government has renewed the amendment at least once beyond its original 2025 sunset date, citing continuing regional pressure, while an independent statutory review of its effect continues (approx.; see NSW Parliament Hansard for the precise renewal date and terms).
The strongest case
A small number of repeat offenders were doing outsized harm, and the towns most affected said so directly
The government's case rests on a real and specific problem, not a generalised panic about young people. BOCSAR's own recorded crime data shows offences such as break-and-enter and motor vehicle theft are heavily concentrated in a handful of regional command areas — Moree Plains, Bourke, Walgett, Coonamble and Brewarrina among them — at rates that have, at times, run several times higher than the state average (approx.; BOCSAR, NSW Recorded Crime Statistics). Local police, business owners and community leaders, including Aboriginal elders in some of the towns worst affected, described a familiar and repeating cast of names moving through arrest, bail and further arrest, with victims — often other local families, many of them Aboriginal themselves — losing patience with a process that appeared, in the worst cases, to release the same young person within days of an alleged serious offence. BOCSAR and comparable interstate research have both found that a relatively small share of young offenders account for a disproportionate share of recorded youth offending, so targeting bail settings specifically at repeat, already-on-bail offenders was a more precisely calibrated response than a blanket toughening of bail for every young person charged with any offence. Add to that the practical toll a wave of break-ins and car theft takes on a town too small to easily absorb repeated victimisation — rising insurance costs, businesses closing early, families talking about leaving — and the government's argument that it owed these communities a visible, urgent response is hard to dismiss. Attaching a 12-month sunset clause, rather than legislating the change permanently, was itself a mark of institutional honesty: an acknowledgement that the evidence for what it would achieve was not yet proven, and a commitment to test it rather than assume it.
The evidence
Two years on, BOCSAR's own data is more equivocal than either side of the debate tends to admit. Recorded rates of break-and-enter and motor vehicle theft in the worst-affected regional command areas have moved within the normal range of year-to-year volatility that regional crime statistics show, rather than settling into a clear, sustained fall that can be attributed to the bail change specifically (approx.; BOCSAR, NSW Recorded Crime Statistics, 2023–2026). Isolating the effect of one bail provision from a package that also included extra policing, funding and, in some towns, more intensive youth operations is difficult even for BOCSAR's own researchers, and the Bureau has been publicly cautious about attributing any local change to the law itself rather than to the surrounding response.
What has moved unambiguously is the number of children held in custody before their case is even decided. BOCSAR's custody statistics show the large majority of children in NSW youth detention are unsentenced — held on remand, not yet found guilty of the offence that put them there — and that share has continued a rise that predates the 2024 amendment rather than reversing it (approx.; BOCSAR, NSW Custody Statistics). Aboriginal children, who make up a small fraction of the state's youth population, continue to account for roughly half the youth detention population, a gap the amendment was not designed to close and that community legal centres and Aboriginal legal services argued, during the 2024 debate, it risked widening (approx.; AIHW, Youth Justice in Australia).
The shift is not new — the proportion of NSW children in detention who are unsentenced has been rising for the better part of a decade, tracking a national pattern AIHW has documented across most states — but a law explicitly designed to make bail harder to get for a defined group of children could only be expected to add to that trend, not slow it, and the government has not published evidence separating the amendment's specific contribution from the rise it was already riding.
Deterrence versus diversion
The core empirical question behind the bail law is whether making detention before trial more likely for a repeat young offender actually reduces further offending, either by that individual or by deterring others. On the evidence BOCSAR itself has published, the answer is thin. The Bureau's broader research on bail and reoffending — built up over years of evaluating changes to the Bail Act for adults and children alike — has generally found that tightening bail increases the number of people held in custody without a matching, demonstrated fall in reoffending; custody itself, particularly short and disruptive stints on remand, is more often associated with interrupted schooling, lost housing and family disruption that can make later reoffending more, not less, likely. None of this proves the 2024 amendment made outcomes worse. It does mean the government legislated ahead of proof, which is why it wrote in the sunset clause — and the honest reading of the record since is that the review it promised has not yet produced public evidence that the law is achieving the deterrent effect it was sold on.
A law written with a use-by date is an admission that nobody yet knows if it works — and two years on, the government still cannot show that it does.
Set against that thin evidence base is a comparatively well-evidenced alternative that receives a fraction of the political attention. Bourke's Maranguka justice reinvestment project — an Aboriginal community-led initiative backed by Just Reinvest NSW, redirecting resources toward family support, mentoring and early intervention rather than policing and courts alone — was the subject of an independent KPMG impact assessment in 2018 that found falls in juvenile convictions and days spent in custody, alongside a positive economic return on the funding invested (approx.; KPMG / Just Reinvest NSW, Maranguka Justice Reinvestment Project, 2018). The NSW Youth Koori Court, which involves Aboriginal elders directly in sentencing and has since expanded beyond its original Parramatta registry, has similarly been the subject of BOCSAR evaluation pointing to reduced or delayed reoffending among participants compared with mainstream court processes (approx.; BOCSAR, Youth Koori Court evaluation). Both programs are older than the 2024 bail amendment, both have a published evidence base the amendment still lacks, and both remain comparatively small and precariously funded next to the recurring cost of holding a growing remand population in custody at roughly $2,000 or more per child, per day (approx.; Productivity Commission, Report on Government Services).
A law built to be temporary
The sunset clause was meant to force a reckoning: either the evidence would justify making the change permanent, or the law would lapse and the government would rely on its funding and diversion measures instead. What has happened in practice is closer to a third option — renewal without a public verdict, extending an emergency provision on the strength of continuing community concern rather than of demonstrated results. That is a defensible short-term political choice in towns still reporting high rates of break-ins and car theft; it is a harder position to defend as a lasting approach to criminal justice, because it converts a law built for scrutiny into one renewed by default, which is exactly the drift a sunset clause exists to prevent.
The towns that pushed hardest for the change have not been made whole by it either way. Moree, Bourke and their neighbours still record some of the highest youth offending rates in the state (approx.; BOCSAR, NSW Recorded Crime Statistics), the amendment's central promise — fewer repeat offences by children already before the courts — remains unproven in the government's own data, and the state's youth detention system is absorbing more unsentenced children than it was two years ago. None of that means the government was wrong to act; the underlying problem was real, and doing nothing carried its own cost to victims and to public confidence in a system that seemed unable to respond. It does mean the response chosen was the one with the least evidence behind it, running alongside two the state's own research body has evidence for, and that a temporary law has now outlasted the emergency framing that first justified it.
The risk to the future
Remand without rehabilitation entrenches the problem it was meant to fix
If the bail amendment continues to be renewed without published evidence that it reduces reoffending, New South Wales risks locking in a cycle where more children — disproportionately Aboriginal children from the same regional towns the law was meant to protect — spend longer in custody before trial, with the disrupted schooling, housing and family ties that make future offending more likely, while the underlying rate of youth crime in Moree, Bourke and similar towns stays largely where it started. Detention capacity built or expanded to hold a growing remand population becomes a recurring budget cost that is politically difficult to wind back even if the review finds it has not worked, crowding out funding for the justice reinvestment and diversion programs that do have a published evidence base behind them. The test the government set for itself, in writing a use-by date into its own law, was whether the evidence would justify keeping it. Two years on, the honest answer is that the evidence has not yet been shown publicly — and every renewal without it makes the eventual choice, keep it or let it lapse, harder to make on the facts rather than on the politics of the moment.