Prison Program Structure Models: Cohorts, Tiers, Mentorship Pipelines, and Outcomes from Leading U.S. Correctional Programs

Georgia's prison system has run past its design capacity for fifty years, and the physical plant — locks, water, cooling, fire systems, and kitchens — has degraded faster than the state has documented or repaired it. GDC's own PREA audits, the Department of Justice's October 2024 findings, DPH food-safety scores, and the December 2024 Guidehouse assessment converge on a facility estate that requires "immediate intervention," while the state's $600 million repair program and a $436.7 million new-build near Washington State Prison move forward without a published facility-by-facility capacity or maintenance inventory.

Georgia runs a four-rung segregation ladder — Tier I, Tier II, the Tier III Special Management Unit, and the STEP step-down programmes — and publishes a population count for exactly one rung. The only Tier population GDC publishes is 154 people in the Special Management Unit as of 2026-09-01, a side effect of the SMU being carried as a standalone facility in the monthly Inmate Statistical Profile; the Tier I and Tier II units inside ordinary prisons are absorbed invisibly into those prisons' totals and are counted nowhere. The best per-facility segregation dataset GDC publishes — the PREA audit corpus — puts system-wide segregation capacity at 3,401 cells across 30 facilities, a figure that is a cell count, a floor, and a mix of three audit vintages, not a population.

Georgia pays incarcerated people nothing for regular prison labor — kitchen, laundry, janitorial, groundskeeping, construction, and facility maintenance — making it one of roughly seven states with a true zero-wage system. That unpaid labor generates an estimated $100+ million annually in value for Georgia counties alone, while the families of incarcerated people are charged commissary prices that run 67% to 1,150% above retail, producing $18.76 million in documented state commissary profit in 2024. This page synthesizes GPS's labor, commissary, family-cost, and convict-leasing collections to show that the same extraction logic — confine, compel, charge — has run continuously from 1866 to the present.

Georgia measures reentry success with a single narrow number — felony reconviction within three years — which has ranged from roughly 24 to 31 percent over the past decade and excludes misdemeanor arrests, parole and probation violations, reoffending after year three, and everyone who dies before the window closes. Behind that number, the state releases roughly 12,000 to 16,000 people a year — more than half at sentence expiration with no parole supervision — into a reentry system with 2,344 transitional beds, no dedicated reentry line item in the GDC budget, and a first-two-weeks death risk 12.7 times the general population's. The interventions that reduce recidivism are among the best-evidenced in corrections; Georgia's spending on them is a rounding error.

Georgia has repeatedly designed, piloted, and evaluated evidence-based rehabilitation models — cognitive-skills curricula, earned-time incentives, intensive supervision, reentry housing, addiction treatment — and then starved them of scale and funding. The national evidence base is unambiguous: cognitive-behavioral programs reduce recidivism 20–30%, correctional education cuts recidivism odds by 43% and returns $4–$5 per dollar spent, and vocational completers in Georgia recidivate at roughly half the general rate. Georgia's FY2027 budget nonetheless cut education line items while adding $22.1 million in new surveillance spending, and the state still publishes no outcome data that would let the public judge whether any of its programs work.

Georgia runs a three-tier segregation system whose written limits run in the opposite direction of its population: Tier I is capped at 30 days, Tier III at 24 months with a statewide retention panel, and Tier II — the programme that holds the most people — has no outer limit in policy at all. The only Tier population figure GDC publishes is a bookkeeping side effect of the Special Management Unit being carried as a facility, and a systematic review of 110 GDC reports and all 34 distributions of its monthly statistical series found no published count of the Tier or restrictive-housing population anywhere. What is documented instead is 39 percent diagnosed mental illness in the Special Management Unit, a cell count of at least 3,401 segregation cells across 30 facilities, and named deaths inside isolation.