← All Policies

Justice / Law Enforcement: Adopt South Korea's Approach

Grade B β€” Moderate Evidence

Reduce justice / law enforcement spending to the cheapest high-performer floor. South Korea achieves Life Expectancy 83.57 at $6037/cap; United States gets 76.93 at $12848/cap.

Rank #7 of 22 policies

Welfare Score
+39
Causal Confidence
70%
Policy Impact
58%
BH Average
78%

πŸ“Š Bradford Hill Criteria Scores

Temporality100%
Plausibility100%
Coherence99%
Consistency93%
Analogy85%
Strength of Association75%
Biological Gradient66%
Specificity59%
Experiment25%

πŸ’₯ Impact Breakdown

Income Effect
+167%
Health Effect
+5%
Combined Welfare
+39

πŸ§ͺ Natural Experiments

Real-world before/after data from jurisdictions that implemented this policy.

Norway β€” Rehabilitative Prison System

Intervention year: 1998 Β· Focus on rehabilitation over punishment; maximum 21-year sentence; open prisons; education/vocational training

Recidivism Rate (2-year)
-34.2%
p=0.004
Incarceration Rate
+10.8%
p=1.000
Recidivism Rate (2-year)(% re-offend within 2 years)
1928371998 β€” Policy enacted199020191998
Incarceration Rate(per 100K population)
5263741998 β€” Policy enacted199020211998
Sources: Norwegian Correctional Service annual reports Β· Pratt (2008) Scandinavian Exceptionalism Β· Bhuller et al (2020) American Economic Review β€” RCT on Norwegian prisons

🌍 Criminal Justice by Country

How countries compare on this policy domain. The US row is highlighted.

CountryIncarceration/100KHomicide/100KRecidivismApproach
Japan380.24900%Strict discipline, regimented rehabilitation in prisons. Very high conviction rate (99%+). Strong community-based crime prevention (koban system). Social shame as deterrent.
Singapore1810.22400%Strict deterrence with rehabilitation. Caning and mandatory death penalty for drug trafficking. Community-based programs (Yellow Ribbon Project). Strong social safety net reduces recidivism drivers.
Norway560.52000%Rehabilitative / restorative: focus on reintegration, humane conditions (Halden model), maximum 21-year sentences (preventive detention possible). Education and job training in prison.
South Korea930.52300%Deterrence-based with rehabilitation focus. Electronic monitoring widely used. Restorative justice pilot programs. Low crime rate reflects cultural factors.
Netherlands590.64700%Pragmatic / rehabilitative. Closed several prisons due to declining crime and incarceration. Emphasis on electronic monitoring, community service, and post-release support.
Germany670.84600%Rehabilitative, dignity-centered. Constitutional principle: resocialization is the goal. Prison conditions modeled on life outside. Day-release and open prison common.
Australia1600.94500%Mixed rehabilitative/punitive. High Indigenous incarceration (disproportionate). Drug courts, diversionary programs. Victoria introduced a supervised injecting room.
United Kingdom1291.04500%Mixed punitive/rehabilitative. Overcrowded prisons. Recent emphasis on reducing short sentences and expanding community sentences. Probation service re-nationalized after private-sector failures.
New Zealand1651.15200%Increasingly rehabilitative. Halo/Kowhiritanga programs for Māori. Arms Amendment Act 2019 (post-Christchurch). High Māori over-representation in prisons.
France1051.25900%Mixed. High incarceration relative to Nordic peers, chronic overcrowding. Alternatives to incarceration expanding. Specialized courts for minors.
Finland511.63100%Rehabilitative with open prisons. ~30% of prisoners in open institutions. Focus on normality principle β€” life in prison should resemble life outside as much as possible.
Canada1042.04000%Mixed approach. Federal system emphasizes rehabilitation and safe reintegration. Restorative justice programs for Indigenous communities. Drug treatment courts available.
United States5316.37600%Predominantly punitive. Mass incarceration, mandatory minimums, three-strikes laws, private prisons (8% of federal inmates). Cash bail system. Growing reform movement: some states reducing sentences, diversion programs expanding.
Brazil38122.47000%Punitive in practice, with severe overcrowding (often 2x capacity). Militarized policing. High rates of pretrial detention. Gang control of prisons is a systemic challenge.
Mexico16825.25000%Mixed punitive. Transition from inquisitorial to adversarial justice system (2016). High impunity rate (~95% of crimes unreported or unsolved). Military involvement in drug enforcement.

πŸ“‹ Policy Details

Type
budget allocation
Category
justice law enforcement
Recommendation
reallocate
Current Status
United States spends $12848/cap, ranks 25/26. 2.1x overspend.
Recommended Target
South Korea model ($6037/cap floor). $2309B/yr savings β†’ Optimization Dividend.
Rationale

Cheapest-high-performer analysis: South Korea achieves Life Expectancy 83.57 at $6037/cap. United States at $12848/cap (2.1x overspend). Top 3: South Korea ($6037), Singapore ($7868), Australia ($8330). Savings: $2309B/yr β†’ $17,361/household/yr as Optimization Dividend.

Blocking Factors
political opposition

πŸ”¬ Evidence Assessment: Bradford Hill Criteria

The Bradford Hill criteria are nine principles used to establish evidence of a causal relationship between a policy intervention and its outcomes. Originally developed for epidemiology (1965), they provide a structured framework for evaluating whether an observed association is truly causal. Each criterion is scored from 0 to 1.

Strength of Association75%

How large is the association between the policy and the outcome? Larger effect sizes increase confidence in causation.

Consistency93%

Has the relationship been observed across different populations, settings, and times? Replication strengthens causal claims.

Temporality100%

Does the policy change precede the outcome change? Temporal ordering is a necessary condition for causation.

Biological Gradient66%

Is there a dose-response relationship? More of the policy leads to more of the effect? Gradients support causation.

Experiment25%

Is there evidence from randomized controlled trials or natural experiments? Experimental evidence is the gold standard.

Plausibility100%

Is there a plausible mechanism explaining how the policy causes the outcome? Mechanistic understanding increases confidence.

Coherence99%

Does the causal interpretation fit with existing knowledge? The relationship should not contradict established facts.

Analogy85%

Are there analogous policies that have produced similar effects? Similar interventions with known effects support the claim.

Specificity59%

Is the effect specific to this policy rather than a general phenomenon? Specific associations are more likely causal.

How is the Causal Confidence Score calculated?

The Causal Confidence Score (CCS) of 70% is a weighted average of the nine Bradford Hill criteria. Experiment and temporality receive higher weights since they provide the strongest evidence for causation. The CCS is then combined with the estimated effect magnitude to produce the Policy Impact Score (PIS) of 58%.

See the Optimal Policy Generator paper for full methodology.

← All Policies

Analysis: Β· Optimitron OPG

Optimitron β€” The Evidence-Based Earth Optimization Game