Sentencing — punishment, recidivism, capacity

The system's response — sentences, recidivism, capacity

Since 2017 the Swedish parliament has passed a series of sentence increases (gang crime, weapons, sexual offences). At the same time, Kriminalvården's occupancy exceeds 100% at several facilities. Here are the numbers — and why research on 'what works' is divided.

~41 %

Share of people who reoffend with a new conviction within 3 years (Brå, latest final statistics). The level is higher for prison sentences (~50%) and lower for lighter penalties. Recidivism isn't the same as new crime in the same category.

Two questions often conflated

1. What is Sweden doing? — sentence increases, more pre-trial detention, more places. Measurable. 2. What works? — Deterrence, incapacitation and rehabilitation are three different mechanisms research weighs differently. We do not claim that tougher sentences work as fact — that's not what the data unambiguously shows.

Recidivism by sentence type — 3-year reoffending

Share reoffending with a new conviction within three years of the index event, by primary sentence. Clearest gap between light penalties and prison — but the link is not one-way: prison cohorts also carry heavier criminal histories.

Source: Brå — Recidivism statistics (final).

Timeline — sentence increases 2017→

• 2017 — Heavier sentences for aggravated weapons crime.
• 2018 — Consent law (negligent rape).
• 2020 — Heavier sentences for gang-related crime.
• 2022 — New expulsion law; easier to deport on grave crime.
• 2023–2024 — Search zones, anonymous witnesses, doubled sentences for gang crime.
• 2024 → Safety zones, covert measures against children under 15.

What works? Three perspectives that research weighs differently

Deterrence

Theory: tougher sentences raise the 'price' of crime. Empirically, certainty of detection matters more than length of sentence for deterrence (Brå, US NIJ).

Incapacitation

A person in prison commits fewer crimes outside by definition. Long-run effects are contested — prison reinforces criminal networks and can raise long-run recidivism.

Rehabilitation

CBT-based programmes, work and addiction treatment show measurable effects on recidivism in several meta-analyses — but the effect size is small to medium and varies.

Capacity — the prison service

Kriminalvården's forecast projects a sharp rise in places needed — from around 9,000 prison places in 2023 to over 20,000 in the mid-2030s. Occupancy already exceeds 100% in several facilities. That is a capacity fact independent of whether harsher sentences 'work' in policy terms.

What the data does NOT say

— Recidivism is measured as new convictions within 3 years. Unsolved crimes don't appear; a drop can mask poorer clearance.
— Time series after reforms are short — 3–5 years are needed before recidivism rates can be evaluated.
— Research on 'what works' is not unified. General claims about sentence effects — in either direction — are interpretation.
— Crime statistics and sentencing statistics are different sources and not always directly comparable.