Criminal Justice Cost-Benefit Analysis

Transcription

1 Criminal Justice Cost-Benefit Analysis Michael Wilson Economist and Criminal Justice Research Consultant 4/5/17

2 What is cost-benefit analysis? An approach to policymaking A systematic tool for monetizing public policy A method to weigh options A way for finding out what will achieve the greatest results at the lowest cost

3 Why use cost-benefit analysis Strengths Inform policy Efficient use of resources Common measurement Challenges Limited data in some areas Uncertainty in some estimates Need to decide cost/benefits to include

4 CBA use in criminal justice Pew-MacArthur Results First Initiative Working with jurisdictions to implement a cost-benefit approach to invest in policies and programs that are proven to work Vera Institute s Cost-Benefit Knowledge Bank Resources for understanding cost-benefit analysis within the criminal justice system MW Consulting Creation of local cost-benefit models through the Justice Reinvestment Initiative Completed pretrial cost-benefit model

5 Criminal justice cost-benefit analysis methodology

6 Criminal justice CBA: big picture Estimate the cost of crime in your jurisdiction Examine programs/policies that reduce crime Apply the costs of crime to the programs/policies that reduce crime to estimate their monetary benefits

7 CBA framework Use standard economic tools that have been developed to monetize crime Criminal justice system costs Marginal costs compared to average costs Victimization costs Mapping a jurisdiction s resource use for individuals going through the criminal justice system

8 Taxpayer costs Police Courts, District Attorney and Public Defender Probation Jail Prison Post-prison supervision

9 Taxpayer costs Average Cost is the total program/agency cost divided by the number of individuals served Marginal Cost is the change in the total cost from incremental changes in the number of individuals served Example: When a small number of inmates are added to the jail, certain variable food and service costs increase immediately; however, new staff are not typically hired right away

10 Marginal costs: jail When one individual is added to a jail, certain variable costs increase immediately: food, laundry, medical. Officers would not be added right away, but would be if 100 more individuals were admitted Appropriate because administrative, capital investment, and most utilities are fixed costs Estimated by change in total costs divided by change in average daily population

11 Recoverable vs non-recoverable costs Changes in the population of the jail or prison may result in reduced expenses for food, linens and some staffing and can be removed from the budget Workload reductions to the local police and courts may not result in actual budgetary savings but may improve the level of efficiency

12 Marginal costs of a conviction Multiple system resources used Courts, district attorney and public defender What expenses to include? Direct, overhead, capital, other Can we break costs out by crime type or class?

13 Marginal cost estimate methodology Top down approach Isolate costs directly related to providing the service or activity Determine what expenditure categories are fixed and what are variable Divide operating expenditures by the number of units served

14 Conviction cost estimate: top down

15 Marginal cost estimate methodology Bottom up approach Identify the process behind an activity Estimate the duration of the activity Gather baseline expense value for salaries and materials Multiply costs of direct activities by time

16 Conviction cost bottom up example Iowa estimated the judicial time per crime class They then multiplied the time by the average wages and benefits They also estimated the percent of crime categories from the model in each crime class to estimate the overall weights Another study estimated the marginal cost of drug court hearings by the hourly cost of the personnel involved $3.50 per minute multiplied by an extra 9.7 minutes for a drug court hearing

17 Victimization costs How can costs to victims be estimated? Victimization estimates: K. E. McCollister, M. T. French, & H. Fang (2010). The Cost of Crime to Society: New Crime-Specific Estimates for Policy and Program Evaluation. Drug and Alcohol Dependence, 108(1), Cohen, M. A. & Piquero, A. R. (2009). New evidence on the monetary value of saving a high risk youth. Journal of Quantitative Criminology, 25, Tangible and intangible losses Loss estimates

18 Probability Resource use Likelihood of using each resource Based on a decision tree to show which crimes result in which resources Length of Stay Amount of time using each resource Based on a decision tree to show the length of resource use by crime type

19 Resource use: probability Jail: 60% Probation: 70% None: 30% Felony Conviction Prison: 30% Parole: 85% Probation: 10% None: 15%

20 Resource use: length of stay Felony Conviction Jail: 4 months Prison: 2 years Probation: 1 year Parole: 1 year and 9 months

21 Cost of crime estimate Cost of each criminal justice resource Mapping offender s movement through the system Discounting future dollars to present value Adjusting for inflation Sensitivity analysis

22 Application of cost-benefit analysis

23 Cost-benefit analysis of programs Baseline recidivism rates Measurements of program effectiveness Own evaluation Meta-analysis: Washington State Institute for Public Policy Cost of programs

24 Cost-benefit analysis of programs

25 Cost-benefit analysis of programs

26 Cost-benefit analysis of programs Projected benefits based on long-term recidivism rates and meta-analysis Relies on projections from literature and assumptions about how long effects persist Does not require longitudinal data Estimate benefits based on actual program participants compared to a control group Based on local evaluation Typically only measures benefits over a short horizon

27 Model demonstration

28 Cost-benefit and risk assessment Examine property offenders over three year period Examined their risk of being reconvicted of a felony Overlap where some high risk property offenders were sentenced to prison and some low risk property offenders were sentenced to probation By using risk at sentencing it is possible to save prison beds while keeping crime constant or to reduce crime while keeping beds constant

29 Cost-benefit and risk assessment Risk Scores by Sentence Type Property Convictions % 12.0% 10.0% IN PR 8.0% 6.0% 4.0% 2.0% 0.0% 0-5% 5-10% 10-15% 15-20% 20-25% 25-30% 30-35% 35-40% 40-45% 45-50% 50-55% Risk Score 55-60% 60-65% 65-70% 70-75% 75-80% 80-85% 85-90% 90-95% %

30 Cost-benefit and risk assessment Method 1 Prison bed neutral Move 650 high risk probationer to prison Expect to avoid 350 felony arrests Move 650 low risk prisoners to probation Expect to have 240 felony arrests Savings of 110 felony arrests Using the cost-benefit model we avoid an estimated $3.9 million in costs to tax payers and crime victims

31 Cost-benefit and risk assessment Method 2 Crime neutral Move 780 low risk prisoners to probationer Expect to have 900 felony arrests Move 550 high risk probationers to prison Expect to avoid 940 felony arrests Savings of 230 prison beds and avoids 40 felony arrests Using the cost-benefit model based on Oregon s cost of incarceration we save $3.4 million to tax payers

32 Pretrial cost-benefit analysis Validated risk assessment Pretrial outcomes by risk FTA likelihood New crime likelihood Likelihood of pretrial supervision

33 Pretrial cost-benefit analysis Estimate the system costs of detaining individuals Estimate the crime avoidance/increase and failure to appear avoidance/increase of detaining individuals Monetize the additional/reduced crime and failure to appear

34 Likelihood of Crime Pretrial CBA Example: Likelihood of Crime by Risk Score Distribution 35% 30% 25% Risk of Crime per Defendant 20% 15% 10% 5% 0% 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% Risk Distribution

35 Cost Pretrial CBA Example: Crime Costs by Risk Score Distribution $2,500 $2,000 Cost of Crime per Defendant $1,500 $1,000 $500 $0 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% Risk Distribution

36 Cost Pretrial CBA Example: Crime Costs by Risk Score Distribution $2,500 $2,000 Cost of Detention per Defendant Cost of Crime per Defendant $1,500 $1,000 $500 $0 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% Risk Distribution

37 Cost Pretrial CBA Example: Crime Costs by Risk Score Distribution $2,500 $2,000 Cost of Detention per Defendant Cost of Crime per Defendant $1,500 $1,000 $500 $0 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% Risk Distribution

38 Company LOGO Sample Pretrial System Risk Category % of Bookings by Risk % Released One Day or Less % Released More than One Day Less than One Week Of % Released Defendants One Week or Released, % Longer Supervised Cat 1 13% 37% 37% 4% 91% Cat 2 41% 29% 33% 9% 92% Cat 3 28% 4% 21% 12% 93% Cat 4 18% 2% 12% 11% 93%

39 Company LOGO Policy Change Example #1 70% of Category 1 defendants released in one day or less and 25% released in 2 to 7 days 65% of Category 2 defendants released in one day or less and 25% released in 2 to 7 days 25% of Category 3 defendants released in one day or less and 25% released in 2 to 7 days 0% of Category 4 defendants released in one day or less and 20% released in 2 to 7 days

40 Denver Example #1: Cost Avoidance of Quicker Pretrial Release Total Cost Avoidance FTA Cost Avoidance Crime Cost Avoidance System Cost Avoidance -$2,000,000 $0 $2,000,000 $4,000,000 $6,000,000 $8,000,000

41 Company LOGO Policy Change Example #2 Same jail release timing as Example #1 Reduce pretrial supervision for category 1 defendants from 91% to 60% of the time Reduce supervision length for category 1 defendants from 112 to 75 days and category 2 defendants from 121 days to 90 days

42 Company LOGO Example #1 Results for both examples 273 fewer jail beds and $5.2 million in system cost avoidance 39 fewer misdemeanors and 8 fewer felonies for a total cost avoidance of $0.9 million 249 more FTAs for a total cost of $0.25 million Example #2 Same crime, jail and FTA impacts as above System cost avoidance of $5.5 million or an additional $266,00 in reduced supervision costs

43 Contact information Michael Wilson, Economist and Research Consultant