Cash Transfer at Release and Its Impact on Criminal Recidivism: An Evaluation of the PIA Program.

Last registered on December 06, 2023
View Trial History

Pre-Trial

Trial Information

General Information

Title
Cash Transfer at Release and Its Impact on Criminal Recidivism: An Evaluation of the PIA Program.
RCT ID
AEARCTR-0012251
Initial registration date
November 29, 2023

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
December 06, 2023, 8:21 AM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name
Alejandro Cid
Affiliation
Universidad de Montevideo
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Federico Veneri
PI Affiliation
Iowa State University
Contact Investigator
PI Name
José María Cabrera
PI Affiliation
Universidad de Montevideo
Contact Investigator
PI Name
Edgardo García
PI Affiliation
Ministerio del interior
Contact Investigator

Additional Trial Information

Status
In development
Start date
2023-11-15
End date
2025-03-01
Keywords
Behavior, Crime, Violence, & Conflict
Additional Keywords
cash transfer, recidivism, prision release
JEL code(s)
D04, I38
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
Incarceration results in the disruption of relationships for those in prison, affecting family, work, and community ties. Reestablishing those connections after being released is a complex process. In many cases, recently liberated individuals won't have access to the support system required to reinsert themselves in society after their liberation, resulting in prison recidivism in the short term.
The newly published recidivism statistics for the 2019 Uruguayan liberated cohort suggest that the problem might be more dire. Three of every ten liberated individuals will be re-incarcerated within the first six months after their liberation, 44.1% within the first year, and 65.6% will be reincarcerated within three years.

This project will assess the effectiveness of "Proyecto de Inclusión Asistida" (PIA, assisted inclusion program), a program jointly run by the Ministry of Social Development (MIDES) and the Ministry of Interior (MI) from Uruguay.

The program aims to reduce the recidivism rate among male released inmates who serve a sentence of least two months by providing monthly conditional cash transfers for six months, access to transportation passes (local or inter-state), information about other government programs they may be eligible for, and periodic personalized mentoring from social workers after their liberation.
After an initial evaluation before being released, a specific social worker is assigned to each case, and follow-ups are scheduled outside the prison within the first month, second, and fourth month after being liberated. These visits aim to allow individuals to continue developing their life plans after prison and are a requirement for continuing to receive the program's benefits.

The primary goal of the program is to reduce the rates of recidivism and involvement in criminal activities, as well as to increase employment rates and the take up of social services after liberation for a more at-risk population. The results of this project will contribute to a better understanding of the short and medium-term impact of conditional cash transfer programs and mentoring for released individuals.
External Link(s)

Registration Citation

Citation
Cabrera, José María et al. 2023. "Cash Transfer at Release and Its Impact on Criminal Recidivism: An Evaluation of the PIA Program.." AEA RCT Registry. December 06. https://doi.org/10.1257/rct.12251-1.0
Sponsors & Partners

Partner

Name
Ministerio del Interior
Type
government
URL
https://www.gub.uy/ministerio-interior/
Name
Ministerio de Desarrollo Social
Type
government
URL
https://www.gub.uy/ministerio-desarrollo-social/
Experimental Details

Interventions

Intervention(s)
The PIA is a collaborative effort by the Ministry of Interior and the Ministry of Social Development in Uruguay. This public intervention targets male offenders upon their release from prison. The main intervention is a conditional cash transfer via a debit card, a transportation pass, and a booklet with information about other government programs they may be eligible for (shelters, work programs, etc). The transfers are conditional on follow-up meetings with assigned social workers who will guide them in developing a life plan after prison.
These cash transfers are distributed monthly for six months, contingent on specific requirements. The amount of the monthly cash transfer equals the Tarjeta Uruguay Social (TUS) program, currently about USD 40 (1600 Uruguayan pesos).

Sample Criteria. To participate in the program, individuals must meet the following conditions:
a) They have been incarcerated for at least two months.
b) The crimes they committed are not related to sexual abuse or money laundering.
c) they have a valid national ID number

Intervention Goals. The primary objectives of this intervention are to reduce the rate of recidivism and involvement in any crime while simultaneously increasing employment opportunities and the utilization of social services for disadvantaged individuals.

Control Group. Individuals assigned to the control group will experience the "business as usual" scenario, devoid of cash transfers, information booklets, or personal mentoring. Due to budget constraints, the government cannot allocate monetary resources to provide these components to the control group. Some individuals in the control group may receive an urban bus ticket under the "business as usual" scenario. Moreover, as other citizens, they can go to the MIDES offices and seek information about training and social programs.
Intervention Start Date
2023-12-01
Intervention End Date
2024-08-31

Primary Outcomes

Primary Outcomes (end points)
Prison recidivism.
New arrest (without imprisonment)
Primary Outcomes (explanation)
The sources of both variables are administrative data measured at different time frames since the release date.

Secondary Outcomes

Secondary Outcomes (end points)
Take up of social services for disadvantaged people
Employment

Consumption patterns of the treated group.
Secondary Outcomes (explanation)
The sources of both variables are administrative data measured at different time frames since the release date.

Experimental Design

Experimental Design
We are conducting a randomized control trial (implemented by MIDES and MI) with two arms: a treatment group comprising 500 individuals and a control group with at least 1000 individuals.

Each month, approximately 200 individuals who meet specific criteria are released from the participating prison. Through randomization, 60 are assigned to the treatment group and the rest to the control group. The target sample (500) will be achieved over eight to nine months. While the exact number of prisoners released during the study period is unknown at this point, pre-treatment information from the previous year ensures a control group size of at least 1000 individuals. The size of the treated group is fixed due to budget constraints and will consist of 500 individuals.

We will estimate Intent to Treat (ITT) effects for each outcome of interest.

Heterogeneity Analysis:
The heterogeneity of the ITT effect will be examined across four dimensions, defined by the clusters from the block randomization:
- Number of previous offenses
- Age
- Length of the sentence
- Whether the family visited the inmate during incarceration.
Experimental Design Details
Not available
Randomization Method
Randomization Process: The randomization will be conducted in-office using a computer, the code becoming publicly available upon the final paper's publication.

Stratification. The sample will be stratified into 16 groups based on the following pre-treatment characteristics:
- Number of previous offenses.
- Age.
- Length of the sentence.
- Family visits during incarceration.

Sampling Procedure. For each month, we plan to perform 100 treated and control samples generated by block-randomization. The final selected sample among the candidates will result from a random choice, ensuring balance across additional pre-treatment covariates, such as the type of crime.

Seed for Random Number Generation. To ensure transparency and reproducibility, the seed for the Stata random number generator will be derived from the last results of the national lottery.
Randomization Unit
Individual. No clusters (all the sample comes from one prison).
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
0
Sample size: planned number of observations
1,500+ individuals. As previously mentioned, the number of released prisoners for the next six months remains unknown. Therefore, the final sample size can only be estimated using historical data. The exact number of participants will be determined monthly during the randomization process for each of the 8 to 9 waves of the experiment.
Sample size (or number of clusters) by treatment arms
500 individuals treatment, 1000+ individuals control.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Using historical data for the same context, we can calculate the power for different horizons of recidivism (the main outcome). 3 months recidivism: Mean 0.18; Std. Dev 0.382 Alpha 0.05, Power 0.80, N 1500: we would be able to detect effects of 5.87 pp. If alpha was set at 0.10, the MDE is 5.2 pp, and if the power is 0.90 (and alpha 0.05) the MDE is 6.79 6 months recidivism. Mean 0.31; Std. Dev 0.460 Alpha 0.05, Power 0.80, N 1500 --> MDE = of 7.06pp. Alpha 0.05, Power 0.90, N 1500 --> MDE = of 8.17pp. Alpha 0.10, Power 0.80, N 1500 --> MDE = of 6.27pp. The minimum detectable effects of the intervention will be enhanced by utilizing block randomization, considering that the four strata-forming variables are predictors of the outcome (data from a pre-intervention year). Notably, the number of previous offenses, acting as a proxy for the lagged dependent variable, exhibits the strongest correlation. Moreover, increasing the size of the control group can enhance statistical power (the treatment group size is fixed due to budget constraints). The control group size is estimated as a lower bound, contingent on the unknown flow of released inmates, as previously discussed. A percentage of inmates who will be released in the course of the intervention has not yet committed crimes leading to imprisonment. A measure of MDE for very short-term recidivism (e.g., 1 week) seems implausible at this point, given a mean of 0.01 (1%) and a standard deviation of 0.108.
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Comité de Ética de la Universidad de Montevideo
IRB Approval Date
2023-11-13
IRB Approval Number
CE 2023/9
Analysis Plan