Offenders serving community sentences are more likely to keep jobs, earn more – new research

When should offenders be sent to prison and when is it better to keep them in the community under close supervision?

New Zealand confronted that choice in 2007 when it introduced home detention, community detention and intensive supervision^[1] as alternatives to short prison terms.

At the time of the reform, New Zealand relied heavily on prison for criminal justice. The prison population stood at 189 per 100,000 people, compared to an OECD average of 136^[2], and prisons were operating above capacity.

Two decades later, these non-custodial sentences are now a substantial part of the justice system.

In the 2024–25 financial year, the courts convicted and sentenced 50,800 people, but only about 15% received imprisonment. About one in five were sentenced to home detention, community detention or intensive supervision for offences including traffic offending, acts intended to cause injury, theft and burglary.

We conducted two studies to examine what followed when sentencing shifted away from short prison terms towards community-based sanctions, focusing on work and reoffending, respectively.

We found offenders are more likely to retain work and earn more if they stay in the community, without raising the risk of new substantive reoffending.

The budget case for community-based sentences is straightforward. A day in prison costs NZ$552^[3], compared with $116 for home detention. But sentencing policy should not be judged purely on fiscal arithmetic.

Keeping offenders in the community may help them hold on to work and family ties, but it also raises concerns about deterrence and public safety. Understanding the trade-offs between short prison terms and community-based sanctions matters in New Zealand because most offenders now remain in the community.

Keeping jobs, earning more

Our research regarding work^[4] found the 2007 reform improved offenders’ labour market outcomes.

Comparing first-time offenders sentenced in the year before and after the reform, and adjusting for broader changes over time, monthly earnings over three years were about $107 higher under the post-reform regime. This is about 6% more than they otherwise would have earned, or roughly $3,850 per offender in total.

Part of this gain is pragmatic. Offenders sentenced before the reform were more likely to spend time in prison immediately after sentencing, while those sentenced after the reform remained in the community and able to work. But the earnings effect goes further than this.

Even beyond the first 24 months, when any short prison sentence under the old regime would have ended, earnings remain higher. Across ten years after sentencing, this adds up to a cumulative gain of roughly $7,800 per offender.

The strongest earnings gains appear among offenders who already had a foothold in the labour market before sentencing. This suggests community-based sentences helped offenders hold on to better and more stable jobs, which would have otherwise been disrupted by a short prison term.

Any reoffending stems from breach of conditions

Our study focused on reoffending^[5] revealed a more nuanced picture. Recorded recidivism rises under the post-reform regime, by about 8.7% after one year, 9.5% after two years and 9.6% after five years.

However, this increase does not appear to reflect more substantive offending; the increase is driven by convictions for breaching sentence conditions.

In New Zealand, serious breaches are classified as offences against justice, and it is those that are driving the higher recorded recidivism. This suggests the reform increased breaches of sentence conditions, rather than new substantive offending.

New Zealand’s experience is not unique. Internationally, community-based sanctions and electronic monitoring are used both as a substitute for prison^[6] and a form of supervision following early release^[7].

These sentences also replace different amounts of custody, ranging from short prison terms to the final months of longer sentences. The institutional settings differ, with some countries offering stronger reintegration support and more active labour market programmes than others.

Even across that variation, the evidence is broadly consistent. Studies from Europe^[8], the UK, Argentina^[9] and Australia^[10] find no or lower reoffending, while Danish^[11] and recent Swedish research^[12] also points to better employment and earnings outcomes.

Our findings add to this by showing that the labour market gains seen in Denmark and Sweden can also arise in New Zealand’s leaner welfare-state setting, while the reoffending rates fit the broader international pattern that prison alternatives at least do not lead to clear increases in substantive reoffending.

Taken together, the evidence makes a favourable case for community-based sentencing instead of short periods of imprisonment. Such sentencing delivers fiscal savings alongside stronger labour market outcomes and higher future tax-paying capacity for offenders, without increasing substantive reoffending.

In the New Zealand context, important questions remain about effects on offenders’ families, victims and the public’s sense of safety. Evidence from Danish research^[13] nevertheless provides an encouraging lead, suggesting that community-based sentences also benefit offenders’ family members.

References

1. ^{^} introduced home detention, community detention and intensive supervision (www.beehive.govt.nz)
2. ^{^} compared to an OECD average of 136 (www.oecd.org)
3. ^{^} day in prison costs NZ$552 (www.corrections.govt.nz)
4. ^{^} research regarding work (www.dropbox.com)
5. ^{^} study focused on reoffending (doi.org)
6. ^{^} substitute for prison (www.journals.uchicago.edu)
7. ^{^} form of supervision following early release (www.journals.uchicago.edu)
8. ^{^} Europe (journals.sagepub.com)
9. ^{^} Argentina (www.journals.uchicago.edu)
10. ^{^} Australia (direct.mit.edu)
11. ^{^} Danish (www.scup.com)
12. ^{^} Swedish research (www.sciencedirect.com)
13. ^{^} Evidence from Danish research (academic.oup.com)