Evidence from a Harsher Law. Coautores: Néstor Gandelman (Universidad ORT Uruguay) e Ignacio Munyo (Universidad de Montevideo) La presentación del Dr. Néstor Gandelman se realizó en la XXI Jornadas de Economía del Banco Central del Uruguay el 19 de agosto del 2016. En el estudio se propone una estimación basada en un quasi-experimento natural producido por la ley 19.055. Se encuentra que el incremento en el tiempo de reclusión de menores infractores está asociado a una menor reincidencia criminal post liberalización.

1. Juvenile Incarceration and Crime after Release
Evidence from a Harsher Law
Néstor Gandelman
Universidad ORT Uruguay
Ignacio Munyo
Universidad de Montevideo
Jornadas de Economía del BCU, Montevideo, 19 de agosto 2016
&

2. Juvenile crime
• In Uruguay, youth crime tripled in the last 15
years
– Minors comprised 8% of total population and accounted for
15% of total offences, 26% of the homicides and more than
40% of the robberies (Munyo 2013)
• …also a problem in almost every country in
Latin America (Frühling and Martínez 2011)
• …and the rest of the world (United Nations 2004, 2007)

3. Objective
• To evaluate the impact of longer time in jail
on juvenile recidivism
How do we do that?
• We evaluate the impact of law 19,055 that
increased time in jail for some juvenile
offenders

4. Identification strategy in brief
• Law 19,055 (Jan-13) increased the minimum
incarceration period for some crimes (robberies)
but not others (theft)
• Theft is defined as depriving a person of property
without the use of violence, whereas robbery is
defined as depriving a person of property with the
use or threat of violence
• This provides a quasi experimental framework
ideal for a difference in difference strategy

5. Literature
– Harsher punishment increases recidivism
• Criminal learning environment (Chen and Shapiro 2007;
Bayer at al 2009; Camp and Gaes 2009; DeLisi et al 2011)
• Interruption in the accumulation of work-related skills
& labor market stigmatization (Western 2002; Holzer 2007)
– Harsher punishment reduces recidivism
• Deterrence given by unpleasantness life in custody
(Smith and Gartin 1989)
• Theory: Increase in the sentence length has
ambiguous effects on recidivism

6. Literature
• Empirical evidence: Mixed
General deterrence
– Criminological studies in the U.S. find no evidence of harsher punishments
(Singer 1988; Jensen and Metsger 1994; Steiner et al. 2006)
– Economic literature all over the world finds that harsher punishments deter
potential juvenile offenders (Levitt 1998; Imai and Krishna 2004; Mocan and
Rees 2005; Oka 2009; Entoff 2011; Ibáñez et al. 2013).
Specific deterrence
– On the one hand, Hjalmarsson (2009) finds that incarceration reduces
recidivism by 13 percentage points after taking advantage on discontinuities in
punishment that arise in Washington State’s juvenile sentencing guidelines
– On the other hand, Aizer and Doyle (2015) examine the outcomes of
comparable youths in the U.S. who are randomly assigned to judges who differ in
their sentencing severity and find that confined juvenile have 25-year
reconviction rates 22 percentage points higher

7. Literature
• Nagin et al. (2009)
– The existing research is limited in size, in quality,
and in its insights into why a prison term might be
preventative or criminogenic.
• Moestue et al. (2013)
– Review the empirical evidence on youth violence
prevention programs in Latin America and conclude
that given the limited number of high-quality impact
evaluations, there is still weak empirical evidence for
determining what works and what does not.

8. Data
• The universe of juvenile offenders (2011- Feb 2016)
• Variables included
– Type of crime
– Detailed information on time spent in each facility and escapes
– Some control variables (age, gender, city location, education, id code for the
mother)
• Anonymity: due to legislative protection we are able to
individualize inmate only by a code-number
– (Limitation): we can estimate recidivism only before adulthood since we
cannot cross this data with the information of the adult penal system
– In order no to bias our estimates we work with individuals up to 17.5 years
old when released to have a 6 months window to measure recidivism

9. • We consider 2 Datasets:
– Only first incarceration and release
– Every incarceration and release
Data

10. Data

11. Data
Mean Difference Between Juveniles Who Committed Thefts and Robberies

12. Empirical strategy
• First step: to establish a time framework to
measure recidivism: 6 months
• Second step: classify offenders in 4 groups
– (i) committed a robbery before the legal modification
– (ii) committed a robbery after the legal modification
– (iii) committed a theft before the legal modification
– (iv) committed a theft after the legal modification
Effect of increase in sentence length in recidivism:
[(ii)-(i)]-[(iv)-(iii)]

13. Empirical strategy
Sentence Length in days by Type of Crime Before and After Law 19,055

14. Empirical strategy
Database Considering only First Release
Sentence Length

15. Empirical strategy
Database Considering Every Release
Sentence Length

16. Empirical strategy
Average Recidivism by Type of Crime Before and After Law 19,055

17. Empirical strategy
Database Considering only First Release
180-Day Recidivism

18. Empirical strategy
Database Considering Every Release
180-Day Recidivism

19. Empirical strategy
• Formally:
–Y is the effective sentence length or a dummy indicating those that
recidivate within a certain period.
–Robbery=1 for those individuals that committed robberies (treated)
–After=1 for all crimes committed after the new law (treatment period)
–X are controls
itjijt
ittiijt
X
RobberyAfterAfterRobberyY



 *3210

20. Impact of the Law 19,055
on Sentence Length
Results

21. Impact of Longer
Sentences on Recidivism
Results

22. Impact of Longer
Sentences on 60-Day
and 30-Day Recidivism
Results

23. Placebo tests
(180-Day Recidivism)
Results

24. Conclusion
• There was a significant increase in sentence for those juveniles that
committed robberies (the new law was effectively applied)
• This longer time in juvenile correctional facilities did not increased
recidivism
• Moreover, in most specifications we find a negative and statistically
significant effect of more time in custody on crime after release
• Limitation: we do not consider long run effects on future adult crime

25. Discussion
• Our result apparently contradicts recent empirical findings in the
literature suggesting that harsh prison conditions increase post-release
criminal activity
• However, our results could favor the literature that suggests a U-shaped
relationship between severity of punishment and future criminal
behavior, with an optimal level of punishment minimizing the likelihood
of recidivism (Pinchler and Romer 2011)
• In this line, the legal modification in Uruguay suggests the increase in the
level of punishment within the juvenile system seems to be on the
downward side of this theoretical “U”
• At the end, this study questions the general perception that the increase in
incarceration rates is not an effective public policy and suggests digging
further on this issue

26. Juvenile Incarceration and Crime after Release
Evidence from a Harsher Law
Néstor Gandelman
Universidad ORT Uruguay
Ignacio Munyo
Universidad de Montevideo
Universidad de Chile, Santiago, 6 de Julio de 2016
&