Recidivism Reduction & Program Outcomes Modeler

Author

SectorPrompts.com

How to read this report. Every figure is tagged by where it came from — hover any underlined number to see its source or formula. A full breakdown of every value is in the appendix.

This report contains 44 figures you provided, 25 calculated figures (25 of them independently re-checked), and 27 stated assumptions.

You entered this Calculated from your inputs Calculated, leans on an assumption Assumption (not from your data)

Executive Summary

The FY2019-20 release cohort’s three-year conviction rate of 39.1% is the lowest recorded since departmental reporting began — an -8.5%-point decline from the FY2016-17 rate of 47.6% — yet this headline obscures a 46.8-point definitional spread: the same cohort’s three-year rates span 17.4% (return-to-prison) through 39.1% (conviction, primary measure) to 64.2% (arrest), meaning any rate cited without its definition is operationally uninformative for budget or policy purposes.

  • Declining trend is real, but the gain has slowed and the primary measure remains above 36%. Cohort-over-cohort reductions were -3%, -2.7%, and -2.8% percentage points respectively — consistent but not accelerating. Reaching the 36% target requires sustained structural intervention, not trend continuation alone.
  • The reconstruction audit confirms the published rate. Computed as (7,567 felony + 5,828 misdemeanor) ÷ 34,215 tracked individuals = 39.1%%, matching the published 39.1% within rounding. Prior cohorts reconstruct equivalently (FY2016-17: 47.6%%; FY2017-18: 44.6%%; FY2018-19: 41.9%%), confirming methodological consistency across the series. The two-denominator structure — 34,215 for conviction and arrest rates, 34,493 for return-to-prison — must be disclosed whenever rates are compared across definitions.
  • Program-achievement associations are promising but unadjusted. Career technical education (CTE) achievers show a 26%% three-year conviction rate versus 40%% for non-achievers (14%-point gap); academic education achievers show 27.7%% versus 39.9%% (12.2%-point gap). Both are correlational associations — participants self-select and differ from non-participants in ways the departmental reports do not adjust for; neither figure should be treated as a causal program effect.
  • Program throughput is the critical bottleneck. Of the 34,493 individuals released, approximately 12,954 enrolled and 2,042 completed — a completion rate of 15.8%% against an illustrative improvement target of 85%%. Only 1,511 CTE completers avoided conviction, yielding a modeled ROI of 2.59× based on a 2017 Legislative Analyst cost basis of $10,000/participant-year applied to FY2019-20 outcomes — a period-mixed estimate that should be updated with current cost data before budget reliance.
  • The Legislative Analyst’s 2017 conclusion — that the legislature could not assess which programs most effectively reduce recidivism — has not been formally resolved. That finding predates this cohort; the department’s own reporting still does not publish risk-adjusted program comparisons, wage-record employment outcomes, or 12-month breakdown rates. Until those data gaps are closed, per-program cost-effectiveness cannot be determined from published sources.

Outcomes Scorecard

Note: Three metrics are data-unavailable, not merely off-target. “Not Met” status on remaining metrics reflects the gap between current performance and illustrative improvement targets. CTE program rates are used where whole-cohort breakdowns are unavailable.

Recidivism Reduction Funnel

Note

Key insight: Of 34,493 individuals released, the funnel narrows sharply at completion — only 15.8%% of those enrolled finish the CTE sub-program, and 1,511 individuals reach the no-reconviction endpoint. Throughput loss between enrollment and completion is the largest single attrition stage.

12-Month Outcome Glide Path

Note

Key insight: This chart shows the path required to reach the 36%% target conviction rate — not improvement already achieved. The current CTE-completer rate stands at 26.0%%, leaving a 13.1%-point gap to close. The glide path is a linear interpolation; no departmental 12-month breakdown is published, and no monthly milestone data exist to plot against it.

Methodological disclosure: The glide path interpolates linearly from 26.0%% (CTE completer rate, three-year follow-up) to 36%% over 12 months. No departmental monthly recidivism breakdowns are published; no actual month-by-month observations are available to plot as achieved progress. The required path is illustrative only. Completion rate is held flat at 15.8%% because no improvement target has been formally stated by the department.

Definition Spread: Three Rates, One Cohort

The 46.8%-point spread between the lowest and highest published rates for the same FY2019-20 cohort makes definitional precision a prerequisite for any budget or policy comparison:

Definition Rate Denominator Scope
Return to Prison 17.4%% 34,493 released State custody only; excludes county jail
Conviction (Primary) 39.1%% 34,215 tracked Felony + misdemeanor; excludes 278 lacking RAP sheet
Arrest 64.2%% 34,215 tracked Any arrest; no conviction required

The return-to-prison rate uses the full 34,493-person released denominator while conviction and arrest rates use the 34,215-person tracked denominator — a 278-person difference that slightly inflates the return-to-prison denominator relative to the other two. Cross-definition comparisons without explicit denominator disclosure are unreliable.

Program Achievement Associations

Mandatory caveat: All achievement-versus-non-achievement rate comparisons below are associations — they reflect differences between individuals who completed programs and those who did not, without adjustment for prior risk level, offense type, time served, or other factors that influence both program participation and recidivism. Neither the department’s reports nor the legislative analyst’s report publish risk-adjusted estimates. These figures must not be interpreted as causal program effects.

Credential-level gradient: The association between credential level and conviction rate is steep and monotonic — associate degree holders (303 individuals, 5.6%%) and bachelor’s/master’s holders (19 individuals, 0%%) show the lowest rates in the cohort. The small denominators for these groups (particularly 19 individuals for the highest credential) preclude stable rate estimation; these figures should be treated as directional, not precise.

Conviction Rate Trend — Prior Cohorts

Note

Key insight: Four consecutive annual declines in the primary conviction rate are confirmed by independent reconstruction. The FY2019-20 rate of 39.1%% is verified as (7,567 + 5,828) ÷ 34,215 = 39.1%%.

Intervention Sensitivity

Modeled scenarios — not observed outcomes. The base case ROI of 2.59× uses the Legislative Analyst’s 2017 cost figure of $10,000/participant-year applied to FY2019-20 outcomes — a period-mixed, modeled estimate. Scenario parameters are assumptions about structural changes; they are not derived from controlled trial data. All figures should be treated as directional planning inputs, not predicted results. The avoided-conviction benefit of $35,000 is an assumption — it is not sourced from either the department or the Legislative Analyst.

The scenario spread is instructive: risk-needs triage (3.1× ROI, 22%% recidivism) dominates the base case (2.59×, 26.0%%) on all three outcome dimensions, at only marginal cost increase ($10,200/participant-year). A funding reduction of 15% produces the worst outcome profile (1.8× ROI, 28.5%% recidivism), and the 13.5%% completion rate implies the total benefit pool shrinks alongside the unit efficiency loss.

Key Recommendations

1. Commission a risk-adjusted program evaluation before the next budget cycle. The department’s research office, in coordination with the Legislative Analyst’s Office, should contract an independent regression analysis comparing conviction outcomes across program types, controlling for static risk factors available in the department’s own records (offense type, prior commitments, time served). This directly addresses the 46.8%-point definitional spread and the unadjusted gap figures that currently prevent the subcommittee from determining which of the $315M in annual program spend produces the greatest recidivism reduction per dollar. Until risk adjustment exists, the association figures cited in this report — including the 14%-point CTE gap — cannot be used to justify differential funding.

2. Implement risk-needs triage screening to prioritize program slot allocation. The department’s program management office should adopt formal risk-needs assessment–based slot prioritization within the next program year. The sensitivity model projects that triage targeting could move the recidivism rate from 26.0%% to 22%% and the ROI from 2.59× to 3.1×, at an estimated cost of $10,200/participant-year — a $10,200 per-participant figure that is a modeled assumption and should be validated against actual operational overhead. The Legislative Analyst’s 2017 finding of unused slots alongside waitlists indicates that slot allocation, not raw capacity, is the binding constraint.

3. Establish a wage-record data linkage with the state employment development department and publish 12-month recidivism breakdowns. The department’s research office should negotiate a quarterly wage-record match for all release cohorts, enabling post-program employment reporting currently disclosed as unavailable. Publishing 12-month conviction breakdowns (in addition to the three-year primary measure) would allow the subcommittee to assess program effects within a time horizon relevant to annual budget decisions, and would resolve the data gap that prevents any meaningful monitoring of the 36%% next-cohort target on an in-year basis.

4. Update the per-participant cost basis from the 2017 Legislative Analyst figure. The department’s fiscal unit should produce a current per-participant cost estimate for each program type and submit it alongside the next recidivism report. The modeled CTE cost-per-successful-outcome of $13,516 and the base ROI of 2.59× are derived from a $10,000/year figure that is seven years old relative to the FY2019-20 cohort; budget decisions based on that cost basis carry material uncertainty. A current cost estimate would also allow the legislature to evaluate whether program expansion, triage, or staffing investment (sensitivity range: $9,500–$11,500/participant-year) produces the best return at current price levels.

Data Coverage and Disclosure

The following metrics were requested or are conventionally reported in recidivism reduction frameworks but are not available from the department’s published reports or the Legislative Analyst’s 2017 review. They are disclosed here rather than omitted silently:

Metric Status Reason
Employment at release / 90-day Unavailable Wage-record linkage absent from departmental reporting
Time to case resolution Unavailable Court-system metric; not captured in DOC reporting
LSI-R / risk score reduction Unavailable Pre/post risk assessment not published in these reports
Incident rate per 1,000 (in-custody) Unavailable Not published in the recidivism or program achievement reports
12-month recidivism breakdown Unavailable Only three-year rate published; 12-month used as proxy in glide path
Per-program cost breakdown Unavailable LA 2017 figure is aggregate; no program-level cost published
Cognitive behavioral program conviction rate Unavailable 6,509 achievers noted; conviction rate not disaggregated in source

The two-denominator structure — 34,215 for conviction/arrest, 34,493 for return-to-prison — is an inherent feature of departmental reporting methodology and must be disclosed in any cross-definition comparison. The 278 individuals excluded from the tracked cohort lack automated RAP sheets; their outcomes are not captured in the conviction or arrest rates.

Appendix — Where every number came from

Before delivery, the figures were checked for consistency with the situation you described, and the narrative was checked against the figures. Anything that couldn’t be verified is labeled as an assumption above.

Value Amount Source
cohort_released 34,493 34,493 individuals released FY2019-20
cohort_tracked 34,215 34,215 comprise the tracked cohort
cohort_excluded 278 278 excluded lacking automated state rap sheet
felony_convictions 7,567 7,567 felony convictions stated
misdemeanor_convictions 5,828 5,828 misdemeanor convictions stated
no_conviction_count 20,820 20,820 completing window without conviction
conviction_rate_pct 39.1% 39.1 percent department primary measure
arrest_rate_pct 64.2% three-year arrest rate 64.2 percent
return_to_prison_pct 17.4% return-to-prison rate 17.4 percent
conv_rate_1617_pct 47.6% FY2016-17 47.6 percent prior cohort
conv_rate_1718_pct 44.6% FY2017-18 44.6 percent prior cohort
conv_rate_1819_pct 41.9% FY2018-19 41.9 percent prior cohort
tracked_1617 31,792 31,792 tracked FY2016-17
tracked_1718 35,447 35,447 tracked FY2017-18
tracked_1819 36,086 36,086 tracked FY2018-19
felony_1617 7,347 7,347 felony FY2016-17
misdem_1617 7,776 7,776 misdemeanor FY2016-17
felony_1718 7,406 7,406 felony FY2017-18
misdem_1718 8,398 8,398 misdemeanor FY2017-18
felony_1819 7,525 7,525 felony FY2018-19
misdem_1819 7,604 7,604 misdemeanor FY2018-19
arrest_rate_1819_pct 66.7% FY2018-19 arrest rate 66.7 percent
return_prison_1819_pct 16.8% FY2018-19 return-to-prison 16.8 percent
achievers_any 12,850 12,850 earned at least one achievement
academic_achievers 2,113 2,113 academic education achievers
academic_conv_pct 27.7% 27.7 percent conviction rate academic achievers
academic_noachieve_pct 39.9% 39.9 percent conviction rate no academic
cte_achievers 2,026 2,026 career technical education achievers
cte_conv_pct 26% 26.0 percent conviction rate CTE achievers
cte_noachieve_pct 40% 40.0 percent conviction rate no CTE
cbi_achievers 6,509 6,509 cognitive behavioral program achievers
transitions_achievers 6,873 6,873 Transitions achievers
transitions_conv_pct 36.9% 36.9 percent conviction rate Transitions
assoc_degree_n 303 303 individuals associate degree
assoc_degree_conv_pct 5.6% 5.6 percent conviction rate associate degree
bach_master_n 19 19 individuals bachelor’s or master’s
bach_master_conv_pct 0% 0.0 percent conviction rate bachelor’s/master’s
ged_conv_pct 29.7% 29.7 percent conviction rate GED
hse_conv_pct 33.3% 33.3 percent high school equivalency
hs_diploma_conv_pct 36.1% 36.1 percent high school diploma
la_program_budget_m $315 roughly $315 million rehabilitative programming
la_cost_per_participant $10,000 ~$10,000 per participant per year, 2017 report
baseline_recidivism_pct 39.1% department three-year conviction rate as baseline
cost_per_participant $10,000 $10,000 per participant-year, 2017 LA report
Value Amount Basis
target_recidivism_pct 36% next-cohort goal consistent with trend trajectory
incident_rate_per_1000 null not reported in departmental sources; disclosed unavailable
avg_risk_score_reduction null LSI-R reduction not published; disclosed unavailable
employment_at_release_pct null wage-record linkage absent; disclosed unavailable
benefit_per_avoided_conviction $35,000 avoided incarceration cost proxy; assumption, not source-stated
target_completion_pct 85% illustrative departmental improvement target
target_enrollment_pct 45% illustrative capacity-expansion target
target_no_reoffense_pct 80% illustrative outcome improvement target
target_risk_reduction null LSI-R data not published; target undisclosed
target_incident_rate null incident data not published; target undisclosed
target_employment_pct null employment data not published; target undisclosed
capacity_25_recid 25% higher volume dilutes intensive support slightly
capacity_25_completion 17% proportional capacity increase assumption
capacity_25_cost $9,500 modest economies of scale at higher volume
capacity_25_roi 2.85 modeled from scaled cost and benefit assumption
staffing_50_recid 23.5% case manager ratio improvement reduces recidivism
staffing_50_completion 18.5% better case management improves completion
staffing_50_cost $11,500 staffing increase raises per-participant cost
staffing_50_roi 2.6 higher cost partially offsets benefit gain
triage_recid 22% risk-needs targeting concentrates gains
triage_completion 19% better-matched participants complete at higher rate
triage_cost $10,200 modest assessment overhead added
triage_roi 3.1 targeted benefit realization improves ROI
reduced_funding_recid 28.5% fewer slots and support increases recidivism
reduced_funding_completion 13.5% reduced resources lower completion rate
reduced_funding_cost $10,000 unit cost unchanged; fewer participants served
reduced_funding_roi 1.8 reduced scale shrinks total benefit realized
Value Amount Grounding
enrollment_rate 0.3755662721 ✓ re-checked from your inputs
completion_rate 0.15766536965 ✓ re-checked from your inputs
no_reoffense_rate 74.0% ✓ re-checked from your inputs
total_convictions_1920 13,395 ✓ re-checked from your inputs
conv_rate_reconstructed_pct 39.1% ✓ re-checked from your inputs
conv_rate_reconstructed_1617_pct 47.6% ✓ re-checked from your inputs
conv_rate_reconstructed_1718_pct 44.6% ✓ re-checked from your inputs
conv_rate_reconstructed_1819_pct 41.9% ✓ re-checked from your inputs
delta_1617_to_1718_pct -3% ✓ re-checked from your inputs
delta_1718_to_1819_pct -2.7% ✓ re-checked from your inputs
delta_1819_to_1920_pct -2.8% ✓ re-checked from your inputs
delta_total_1617_to_1920_pct -8.5% ✓ re-checked from your inputs
definition_spread_pct 46.8% ✓ re-checked from your inputs
academic_assoc_gap_pct 12.2% ✓ re-checked from your inputs
cte_assoc_gap_pct 14% ✓ re-checked from your inputs
enrolled_count 12,954 ✓ re-checked from your inputs
completed_count 2,042 ✓ re-checked from your inputs
no_reoffense_count 1,511 ✓ re-checked from your inputs
current_recidivism_pct 26.0% ✓ re-checked from your inputs
recidivism_delta_pct 13.1% ✓ re-checked from your inputs
completion_pct 15.8% ✓ re-checked from your inputs
enrollment_pct 37.6% ✓ re-checked from your inputs
cte_not_convicted_count 1,499 ✓ re-checked from your inputs
cte_modeled_cost_per_success $13,516 ✓ re-checked from your inputs
roi_multiple 2.59 leans on: benefit_per_avoided_conviction

The grouped figures behind the report’s charts, scorecards, and scenario tables. Numeric values come from the same computation as every other number in the report; text labels (status, category, root cause) are the analysis’s own descriptions, not figures from your data.

funnel_df

Stage Count Conversion_Pct
Released 34493 100
Enrolled in Program 12954 37.6
Program Completed 2042 15.8
No Re-Offense at 12 Months 1511 74

scorecard_df

Metric Baseline Target Status
Recidivism Rate % (3-yr conviction, primary measure) 39.1 36 Not Met
Program Completion Rate % (CTE sub-program) 15.8 85 Not Met
Incident Rate per 1,000 (not published — unavailable) Data Unavailable
Program Enrollment Rate % (achievers / tracked cohort) 37.6 45 Not Met
Avg Risk Score Reduction (LSI-R; not published — unavailable) Data Unavailable
Employment at Release % (wage-record linkage absent — unavailable) Data Unavailable
No Re-Offense at 12 Months % (CTE completers; 3-yr rate used — no 12-mo breakdown published) 74 80 Not Met

sensitivity_df

Scenario Recidivism_Rate_Pct Completion_Rate_Pct Cost_Per_Participant ROI_Multiple
Base Case 26 15.8 10000 2.59
+25% Program Capacity 25 17 9500 2.85
+50% Case Manager Staffing 23.5 18.5 11500 2.6
Risk-Needs Triage 22 19 10200 3.1
Reduced Funding -15% 28.5 13.5 10000 1.8

trend_df

Month Recidivism_Rate_Pct Completion_Rate_Pct
1 26 15.8
2 26.9 15.8
3 27.8 15.8
4 28.7 15.8
5 29.6 15.8
6 30.5 15.8
7 31.5 15.8
8 32.4 15.8
9 33.3 15.8
10 34.2 15.8
11 35.1 15.8
12 36 15.8

How each number was derived

Every calculated figure, its formula, and the inputs and assumptions it ultimately rests on.

Value Amount Formula Traces back to
enrollment_rate 0.3755662721 achievers_any/cohort_tracked achievers_any (input), cohort_tracked (input)
completion_rate 0.15766536965 cte_achievers/achievers_any cte_achievers (input), achievers_any (input)
no_reoffense_rate 74.0% 1 - cte_conv_pct/100 cte_conv_pct (input)
total_convictions_1920 13,395 felony_convictions + misdemeanor_convictions felony_convictions (input), misdemeanor_convictions (input)
conv_rate_reconstructed_pct 39.1% round((total_convictions_1920/cohort_tracked) * 100, 2) felony_convictions (input), misdemeanor_convictions (input), cohort_tracked (input)
conv_rate_reconstructed_1617_pct 47.6% round(((felony_1617 + misdem_1617)/tracked_1617) * 100, 2) felony_1617 (input), misdem_1617 (input), tracked_1617 (input)
conv_rate_reconstructed_1718_pct 44.6% round(((felony_1718 + misdem_1718)/tracked_1718) * 100, 2) felony_1718 (input), misdem_1718 (input), tracked_1718 (input)
conv_rate_reconstructed_1819_pct 41.9% round(((felony_1819 + misdem_1819)/tracked_1819) * 100, 2) felony_1819 (input), misdem_1819 (input), tracked_1819 (input)
delta_1617_to_1718_pct -3% round(conv_rate_1718_pct - conv_rate_1617_pct, 1) conv_rate_1718_pct (input), conv_rate_1617_pct (input)
delta_1718_to_1819_pct -2.7% round(conv_rate_1819_pct - conv_rate_1718_pct, 1) conv_rate_1819_pct (input), conv_rate_1718_pct (input)
delta_1819_to_1920_pct -2.8% round(conviction_rate_pct - conv_rate_1819_pct, 1) conviction_rate_pct (input), conv_rate_1819_pct (input)
delta_total_1617_to_1920_pct -8.5% round(conviction_rate_pct - conv_rate_1617_pct, 1) conviction_rate_pct (input), conv_rate_1617_pct (input)
definition_spread_pct 46.8% round(arrest_rate_pct - return_to_prison_pct, 1) arrest_rate_pct (input), return_to_prison_pct (input)
academic_assoc_gap_pct 12.2% round(academic_noachieve_pct - academic_conv_pct, 1) academic_noachieve_pct (input), academic_conv_pct (input)
cte_assoc_gap_pct 14% round(cte_noachieve_pct - cte_conv_pct, 1) cte_noachieve_pct (input), cte_conv_pct (input)
enrolled_count 12,954 round(cohort_released * enrollment_rate) cohort_released (input), achievers_any (input), cohort_tracked (input)
completed_count 2,042 round(enrolled_count * completion_rate) cohort_released (input), achievers_any (input), cohort_tracked (input), cte_achievers (input)
no_reoffense_count 1,511 round(completed_count * no_reoffense_rate) cohort_released (input), achievers_any (input), cohort_tracked (input), cte_achievers (input), cte_conv_pct (input)
current_recidivism_pct 26.0% round((1 - no_reoffense_rate) * 100, 1) cte_conv_pct (input)
recidivism_delta_pct 13.1% round(baseline_recidivism_pct - current_recidivism_pct, 1) baseline_recidivism_pct (input), cte_conv_pct (input)
completion_pct 15.8% round(completion_rate * 100, 1) cte_achievers (input), achievers_any (input)
enrollment_pct 37.6% round(enrollment_rate * 100, 1) achievers_any (input), cohort_tracked (input)
cte_not_convicted_count 1,499 round(cte_achievers * (1 - cte_conv_pct/100)) cte_achievers (input), cte_conv_pct (input)
cte_modeled_cost_per_success $13,516 round((cte_achievers * cost_per_participant)/cte_not_convicted_count) cte_achievers (input), cost_per_participant (input), cte_conv_pct (input)
roi_multiple 2.59 round((cte_not_convicted_count * benefit_per_avoided_conviction)/(cte_achievers * cost_per_participant), 2) cte_achievers (input), cte_conv_pct (input), cost_per_participant (input), benefit_per_avoided_conviction (assumption)

Prepared with SectorPrompts