← The record
The record · Evidence review

Diversion & Alternatives

Do the alternatives to jail work? The research says sometimes — and it turns on two things: whether treatment is chosen or forced, and whether the “alternative” becomes its own trap.

Record layer. A neutral review of national research — strengths and limits alike. We are not pro-jail or pro-diversion; we report what the evidence shows.

The honest answer: it depends

“End the jail and use the alternatives” is a real position with real evidence behind it — and real limits. Diversion, drug courts, community treatment, supervised release, and reentry programs are neither a cure nor a failure in the research. Their effects vary by design and population. Two variables decide whether an alternative actually works, or just moves the revolving door somewhere else.

Where alternatives work

Well-designed diversion can beat jail on cost and recidivism

<10% vs ~68%
Recidivism in some nonprofit reentry/entrepreneurship programs vs. the national rate
Wolff et al., 2024; Nazarko & Suvorova, 2025
$21M saved
Net savings, ~47% return per dollar, for CA parole + community services vs. prison
Lindquist-Grantz et al., 2021
Lower re-arrest
13-district federal ATI study — less re-arrest & incarceration, no loss to community safety
Petre & Tomita, 2022
~$8–12k/yr
Program cost per person vs. ~$43k/yr for prison
Wolff et al., 2024
Where it is mixed or fails

Not all alternatives work — and some cost more

No clear effect
A meta-analysis of juvenile diversion found no overall recidivism effect — only family-based treatment and some restorative-justice programs significantly reduced it
Mathew, 2019; Wright, 2022
Sometimes costlier
Poorly designed community corrections produced high new-crime costs; incarceration was sometimes cheaper once those were counted
Bernard et al., 2020
“Limited but promising”
Substance-use diversion shows reduced recidivism, but very few cost-effectiveness studies and mixed economics so far
Surprenant, 2019
Depends on access
Cost-effectiveness hinges on real access to quality treatment after diversion — without it, the model underperforms
Haynes & Larsen, 1984
How to read these numbers

A lot of the “success” is who stays, not what works

Before trusting any single figure, one caution runs through this whole literature: much of the apparent gap between alternatives and jail is who enters, refuses, and drops out — not the program itself (Olver et al., 2011; Mitchell et al., 2021).

People who complete a program are different at the start — lower prior justice involvement, more motivation, more support. Comparing completers to everyone else overstates the effect: in one community-diversion study, completers had 3.7% re-arrest vs. 31.2% among those not retained (Ozturk et al., 2022), and non-completers were roughly six times more likely to re-offend within a year (Zarling et al., 2021). Dropout can even be actively harmful — quitters sometimes did worse than matched controls (Anderson & Medendorp, 2024; Rizzo et al., 2025). Stronger designs (intent-to-treat, matching) shrink these effects but can’t fully remove them (Halstead & Poynton, 2016; Beaudry et al., 2021).

The honest question isn’t “did the program work?” It’s “for whom, at what level of retention and dosage?” It’s also how a reported “success rate” gets inflated — by counting graduates only instead of everyone referred.

The two variables that decide it

1 — Chosen vs. forced

The coercion question

You often can’t force treatment — and forcing it may not help

In Washington, treatment for mental illness or addiction generally cannot be compelled unless a person is a danger to self or others or gravely disabled — about 85% of states require dangerousness tied to mental illness for involuntary commitment (Werth, 2001). Experts recommend voluntary care, with civil commitment as a last resort (Hadland et al., 2021).

And the evidence that involuntary confinement for addiction works is mixed and generally weak — several systematic reviews find equal or worse outcomes than voluntary care (Bahji et al., 2023; Neven et al., 2024; Pasareanu et al., 2016; Pilarinos et al., 2019; Jain et al., 2018). “Just force them into rehab” is neither legally simple nor clearly effective.

But it is not as simple as “voluntary good, forced bad.” Legal pressure is different from legal confinement: mandated participants often stay in treatment longer because they can’t simply walk away, and retention drives outcomes — entering treatment as an alternative to detention raised completion odds 4.67× (Haviv et al., 2024), and Germany’s compulsory addiction treatment beat prison (63.4% vs. 80.3% recidivism at seven years; Schalast et al., 2025). The catch: some of what looks like “voluntary works better” is just healthier people choosing to stay.

2 — Help vs. a new trap

The dependency question

An “alternative” can become its own revolving door

Alternatives can create dependency whether they are private or government-run. Probation, diversion, monitoring, and treatment often come with fees, fines, and legal financial obligations that trap mostly poor people in debt and repeat court contact — “cash-register justice” and a “treatment-industrial complex” (Kleiman, 2011; Appleman, 2022; Link, 2021).

Agencies can grow dependent on fine-and-fee revenue — some municipalities draw 10–25%+ of revenue from it — turning courts, police, and probation into revenue collectors (Sobol, 2017; Wamsley, 2019). Higher monetary penalties are associated with greater recidivism (Piquero & Jennings, 2017). The same fiscal lever runs the other way, too: funding alternatives instead of prisons, or rewarding counties for handling people locally, measurably shifts systems away from incarceration (Wodahl et al., 2024; Harding et al., 2022; Holzer, 2022).

The bottom line

Well-designed diversion — voluntary, with real treatment access, and not loaded with fees — can cut recidivism and cost less than jail. Poorly designed — coerced, underfunded, or fee-driven — can match or worsen the revolving door it was meant to replace. The useful question is not “jail or alternatives.” It is which design, and who is being forced into it.

Sources

Peer-reviewed research, surfaced via Consensus. Effectiveness & cost: Petre & Tomita (2022); Wolff et al. (2024); Nazarko & Suvorova (2025); Lindquist-Grantz et al. (2021); Haynes & Larsen (1984); Bernard et al. (2020); Mathew (2019); Wright (2022); Surprenant (2019). Coercion & commitment: Werth (2001); Hadland et al. (2021); Bahji et al. (2023); Neven et al. (2024); Pasareanu et al. (2016); Pilarinos et al. (2019); Jain et al. (2018). Financial dependency & incentives: Kleiman (2011); Appleman (2022); Link (2021); Sobol (2017); Wamsley (2019); Piquero & Jennings (2017); Wodahl et al. (2024); Harding et al. (2022); Holzer (2022). Selection, dropout & retention: Olver et al. (2011); Mitchell et al. (2021); Ozturk et al. (2022); Zarling et al. (2021); Anderson & Medendorp (2024); Rizzo et al. (2025); Halstead & Poynton (2016); Beaudry et al. (2021); Haviv et al. (2024); Schalast et al. (2025). Full citations and DOIs on file.