The Impact of Proposition 47
on Crime and Recidivism

JUNE 2018

Mia Bird, Magnus Lofstrom, Brandon Martin, Steven Raphael, and Viet Nguyen
with research support from Justin Goss

PPIC.ORG

© 2018 Public Policy Institute of California.
PPIC is a public charity. It does not take or support
positions on any ballot measures or on any local, state,
or federal legislation, nor does it endorse, support, or
oppose any political parties or candidates for public office.
Short sections of text, not to exceed three paragraphs,
may be quoted without written permission provided
that full attribution is given to the source.
Research publications reflect the views of the authors
and do not necessarily reflect the views of our funders
or of the staff, officers, advisory councils, or board of
directors of the Public Policy Institute of California.
Cover photo courtesy of the California Department
of Corrections and Rehabilitation.

SUMMARY
CONTENTS
Introduction	4
Incarceration 	

5

Crime Trends	

7

Jail Bookings	

14

Recidivism 	

16

Prop 47 Grant Programs 	

19

Conclusion 	

20

Notes	23
References	26
About the Authors	

27

Acknowledgments	29

Technical appendices to
this report are available
on the PPIC website.

Passed by voters in November 2014, Proposition 47 brought broad and significant changes to California’s criminal justice system. Undertaken in the wake
of public safety realignment in 2011, Proposition 47 reduced the penalties for
certain lower-level drug and property offenses and represented a further step
in prioritizing prison and jail space for higher-level offenders. The policy has
sparked continued debate around two key questions:
Did Proposition 47 increase crime?
ƒƒ We find no evidence that violent crime increased as a result of Proposition 47.
While California saw an uptick in the violent crime rate from 2014 to 2016,
this trend appears to have preceded the reform and is due in large part to
unrelated changes in crime reporting after 2014.
ƒƒ We find some evidence that Proposition 47 affected property crime. Statewide, property crime increased after 2014. While the reform had no apparent impact on burglaries or auto thefts, it may have contributed to a rise in
larceny thefts, which increased by roughly 9 percent (about 135 more thefts
per 100,000 residents) compared to other states. Crime data show that thefts
from motor vehicles account for about three-quarters of this increase.
ƒƒ Despite recent upticks, California’s crime rates remain comparable to the
low rates observed in the 1960s—even with the dramatic reductions in incarceration ushered in by recent criminal justice reforms.
Did Proposition 47 reduce recidivism?
ƒƒ Recidivism rates decreased due to Proposition 47. Using data from 12
California counties, we find that among individuals released after serving
sentences for Proposition 47 offenses, the two-year rearrest rate was
70.8 percent, 1.8 percentage points lower than for similar individuals released
before the reform. The two-year reconviction rate for individuals released
under Proposition 47 was 46.0 percent, 3.1 percentage points lower than their
pre-reform counterparts.
ƒƒ These overall declines were driven by substantial reductions in recidivism
rates for Proposition 47 offenses. Rearrest and reconviction rates for these
offenses were 10.3 and 11.3 percentage points lower, respectively, than for
similar individuals before the reform. Our findings suggest that the measure
reduced both arrests by law enforcement and convictions resulting from
prosecutions by district attorneys. However, we are not able to separate the
reform’s effects on reoffending from its effects on the practices of criminal
justice agencies.
Proposition 47 redirected the savings from reduced incarceration to treatment
interventions, with the goal of reducing recidivism. While it is too early to know
if this shift in funding has affected recidivism rates, in the coming years the state
and counties will be better able to assess the impact of increased interventions
and to identify promising strategies. As California continues to pursue criminal
justice reforms, understanding the effects of Proposition 47 and local treatment
programs will be essential to achieving further reductions in recidivism and
maintaining public safety.

Introduction
Over the past 40 years, California’s corrections system has undergone remarkable changes.
“Tough on crime” policies led to a more-than-sevenfold increase in the prison population
between 1980 and 2006. During the same period, expenditures rose dramatically and the corrections system’s share of the state budget tripled. The growth in the prison population led to
overcrowding and poor conditions, prompting lawsuits alleging inadequate medical and mental
health care. In 2009, a federal court ordered the California Department of Corrections and
Rehabilitation (CDCR) to cut its institutional population by tens of thousands of inmates, to
137.5 percent of design capacity. The state appealed the mandate, but the US Supreme Court
upheld it in May 2011. At that time, the state needed to reduce the prison population by about
34,000 inmates by June 2013.
The court-mandated target prompted significant legislative reforms. First, in October 2011,
California enacted public safety realignment (AB 109), which shifted responsibility for many
non-serious, non-violent, and non-sexual offenders to county jail and probation systems. The
reform cut the prison population by about 27,400 in
its first year—but not by enough to meet the court
mandate (Lofstrom, Bird, and Martin 2016). The
Proposition 47 Offenses
prison population declined further as a result of
Check forgery. Proposition 47 makes forging a check
Proposition 36 in 2012, which revised California’s
worth $950 or less a misdemeanor unless the offender
three-strikes law, but the mandated target still was
commits identity theft in connection with the check
not met.
forgery. Previously, check forgery was a potential
In November 2014, another major reform, Proposition 47 (Prop 47), passed with the support of almost
60 percent of voters. Prop 47 reduced the penalties
associated with a set of lower-level drug and property offenses (see text box). These offenses were
reclassified from felonies or wobblers—which could
be charged as either felonies or misdemeanors at the
prosecutor’s discretion—to misdemeanors that carry
a maximum sentence of one year in jail. The law
also applied retroactively, meaning that individuals
already serving prison or jail sentences at the time
of passage could petition to be released or to have
their sentences shortened. Individuals previously
convicted of certain violent or sexual offenses are
not eligible under Prop 47.1 Importantly, the law
also aims to reduce California’s stubbornly high
recidivism rates by requiring that 65 percent of the
net state savings from the measure go toward grants
and programs for mental health and substance-use
disorder treatment.

4 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

felony offense.

Drug possession. Possession for personal use of most
illegal drugs is always a misdemeanor under Proposition 47. Previously, offenders could be charged with a
misdemeanor or a felony, depending on the amount
and type of drug.
Receiving stolen property. Under Proposition 47,
receiving stolen property worth $950 or less was
reduced from a potential felony to a misdemeanor.
Shoplifting. Shoplifting property worth $950 or less
is always a misdemeanor under Proposition 47.
Previously, shoplifting could be charged as a burglary,
a potential felony offense.
Theft. Proposition 47 limits the circumstances under
which theft of property worth $950 or less can be
charged as a felony.
Writing bad checks. Under Proposition 47, it is a misdemeanor to write a bad check for $950 or less. If the
check is for more than $950 or the offender has three
previous forgery-related convictions, the crime is still
a potential felony offense.

Within months after the reform passed, and for the first
time since the 2009 federal court order, the prison
population dropped below the court-mandated target.2
But Prop 47 remained controversial. Supporters presented
the policy change as an opportunity to rein in correctional spending, while reducing lower-level offenders’
involvement with the criminal justice system and shifting
resources toward treatment programs. In the information
guide provided to voters, proponents of Prop 47 wrote,
“For too long, California’s overcrowded prisons have
been disproportionately draining taxpayer dollars and
law enforcement resources, and incarcerating too many
people convicted of low-level, nonviolent offenses.”3 The
Legislative Analyst’s Office (LAO) predicted substantial
reductions in incarceration associated with Prop 47
offenses and subsequent savings for state correctional
systems (LAO 2016). However, opponents expressed
concerns that this reduction in penalties would shift too many offenders from the state prison system to jails and
communities. More offenders on the street could potentially increase the burden on local criminal justice systems
and compromise public safety. Opponents told voters in the information guide, “This measure will overcrowd jails
with dangerous felons who should be in state prison and jam California’s courts with hearings to provide ‘Get Out
of Prison Free’ cards.” 4

Supporters presented Prop 47
as an opportunity to rein in
correctional spending. However,
opponents expressed concerns
that the reduction in penalties
would shift too many offenders
from the state prison system to
jails and communities.

This report assesses whether Prop 47 is meeting its policy goals and whether it has had an effect on public safety. We
begin by describing incarceration levels under Prop 47. Next, we determine whether the noticeable criminal justice
changes introduced by the initiative affected crime rates. We study possible impacts on violent and property crimes by
examining trends in monthly crime rates and draw comparisons between crime rates in California and those of other
states. We then take a deeper look at contact with the criminal justice system—as measured by jail bookings—and
the demographic characteristics of the population booked into jail under Prop 47. Next, we analyze whether recidivism rates (two-year rearrest and reconviction outcomes) for lower-level drug and property offenders released under
Prop 47 have changed compared to those of similar individuals released before Prop 47. Finally, we discuss the redirection of state savings—which started in August 2016—under Prop 47 toward treatment programs designed to reduce
recidivism, particularly among individuals with behavioral health challenges.

Incarceration
Supporters of Prop 47 argued that California has historically over-penalized lower-level drug and property offenders,
resulting in high incarceration levels and correctional costs. By reducing penalties associated with Prop 47 offenses,
supporters anticipated reductions in jail and prison populations due to fewer individuals entering custody and shorter
custody sentences.
Though other population-reduction measures also played a role, Prop 47 further reduced California’s reliance on
incarceration. By December 2016, the state’s prison and jail populations had dropped by more than 15,000 inmates.
The state’s incarceration rate is now at levels not seen since the early 1990s.

PPIC.ORG 5

Prison and Jail Populations
Both state prisons and county jails saw reduced populations under Prop 47. One year after the reform’s passage, the prison population declined by a total of 8,100
inmates (or 6.0%), with the biggest drop occurring in
the first six months (Figure 1). However, not all of this
decline can be attributed to Prop 47. Around the same
time, CDCR was implementing a number of courtordered population-reduction measures that helped
reduce the prison population.5 The jail population also
decreased sharply following Prop 47. One year after
passage, jails held about 7,000 (or 8.5%) fewer inmates,
easing population pressure in crowded jail systems.

Figure 1. Prison and jail populations quickly
declined after Prop 47
150,000

Prop 47
State prison population

125,000

100,000

75,000

County jail population

50,000

Ja
n
Ap -13
rJu 13
lOc 13
t-1
Ja 3
n
Ap -14
rJu 14
lOc 14
t-1
Ja 4
n
Ap -15
rJu 15
lOc 15
t-1
Ja 5
n
Ap -16
rJu 16
lOc 16
t-1
6

Lower incarceration levels stem from changes to both
releases and admissions under Prop 47. Early releases
from jails and prisons took place as individuals
SOURCE: Author calculations based on monthly jail data from the BSCC
already serving time for Prop 47 offenses qualified
Jail Profile Survey and prison data from the California Department of
Corrections and Rehabilitation, 2013–2016.
for resentencing from a felony to a misdemeanor. The
NOTE: Monthly county jail population is the average daily population
(ADP), and the monthly total state prison population is as of the last day
Legislative Analyst’s Office estimates that Prop 47
of the month.
reduced the state prison population by 4,700 during
the first year (LAO 2016). In addition, those convicted
of a Prop 47 offense after passage of the reform were
released earlier from jail than they would have been before, since misdemeanor sentences are
shorter and can be no longer than one year (Grattet et al. 2016). However, not all jail inmates
experienced shorter sentences under Prop 47. As jail populations declined, counties facing courtordered population caps reduced the number of inmates that they released early due to capacity
constraints. Statewide, capacity releases for sentenced jail inmates dropped from 6,000 in
October 2014 to 3,400 in October 2015, a decline of about 43 percent under Prop 47.6
Incarceration levels were also affected by changes in arrests. The monthly number of arrests
decreased after Prop 47, while the monthly number of cite and releases in the field increased—
especially when measured as a share of total arrests—in the first year following the reform
(Figure 2). The average monthly number of arrests dropped from 124,400 in the year before
Prop 47 (from November 2013 to October 2014) to 118,500 in the year after (from November
2014 to October 2015), a decline of about 5 percent. Meanwhile, the average monthly number of
cite and releases increased, from 30,300 to 32,800, or roughly 8 percent. However, the increase in
the use of cite and releases becomes more apparent when cite and releases are measured as a share
of total arrests. This share increased from a monthly average of 24.3 percent of all arrests in the
year before Prop 47 to 27.7 percent in the year afterward, an increase of about 14 percent. The
data suggest that law enforcement may have responded to the reform with fewer initial contacts
with suspected offenders—as seen in the drop in the total number of arrests—but also, more
noticeably, by deciding to book fewer suspected offenders into county jail. We discuss in more
detail changes in jail bookings in a later section of this report.

6 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Figure 2. Prop 47 led to notable changes in total arrests and the use of cite and releases
150,000

Prop 47

50,000

All arrests

Monthly arrests

125,000
40,000
100,000
30,000
75,000

Monthly cite and releases

Cite-and-release arrests

20,000

6

6
t-1
Oc

6

l-1
Ju

5

5

16

r-1
Ap

nJa

t-1
Oc

5

l-1
Ju

4

4

15

r-1
Ap

nJa

t-1
Oc

4

l-1
Ju

3

3

14

r-1
Ap

nJa

t-1
Oc

3

l-1
Ju

r-1
Ap

Ja

n-

13

50,000

SOURCE: Author calculations based on California’s Monthly Arrest and Citation Register (MACR) data from the California Department of Justice, 2013–2016.

Crime Trends
Critics of Prop 47 have raised concerns that the reduced reliance on incarceration would lead to
more offenders on the street, with more opportunities to commit crimes and less severe consequences for those crimes. Would crime rates increase as a result? Although California’s crime
rates are still historically low, the state witnessed recent upticks in crime in 2015 and 2016. For
violent crimes and most property crimes, we find that recent increases are not due to Prop 47.
However, we do see some evidence that Prop 47 may have contributed to a rise in larceny thefts,
especially thefts from motor vehicles.
Our analysis builds on a recent study examining crime rates and Prop 47 by Bartos and Kubrin
(forthcoming). While one of our approaches—comparing crime rates in California to those of
a “synthetic control” made up of similar states—coincides with Bartos and Kubrin’s research,
there are also important differences: First, we extend the study period to include 2016.7 Additionally, in choosing comparison states, we use a time period (2000–2010) not marked by major
criminal justice reforms in California, such as determinate sentencing in 1976, the three-strikes
law in 1994, and more recently public safety realignment in 2011. We also take advantage of
monthly and county-level data to corroborate and supplement our findings. For example, the use
of monthly data allows us to assess whether any post-reform trends were extensions of trends
that began prior to Prop 47. Moreover, we conduct a preliminary analysis of whether crime
increased more in counties that experienced larger declines in jail population and arrest activity
as a result of Prop 47. Note that the available data do not include one key category of Prop 47
offenses: drug crimes.

PPIC.ORG 7

Long-Term Crime Trends
While crime rates fluctuate from year to year, both violent and property crime rates in California
are on a long-term decline (Figure 3). In fact, despite some increases in recent years, crime rates
are still at historic lows.8 This long-term downward trend comes after decades of significant
growth in crime rates.
Between 1960 and 1980, the property crime rate in California more than doubled. It then
declined in the 1980s and 1990s, and by 2000 had returned to levels similar to those of the early
1960s. The property crime rate then increased slightly in the early 2000s before returning to its
long-term downward trend in 2004. Since 2009, it has stayed relatively flat. The 2014 property
crime rate was the lowest observed since 1960, with the second-lowest rate (about 3.5% higher)
occurring in 2016, which saw 2,545 property crimes per 100,000 residents.
The violent crime rate more than quadrupled between 1961 and 1992. It then dropped dramatically, declining every year until 2006, when it went up slightly. In fact, since 1992, the state has
seen year-over-year increases only four times: 2005 to 2006, 2011 to 2012, 2014 to 2015, and
2015 to 2016. In 2016, the violent crime rate of 444 per 100,000 residents was similar to rates
observed in the late 1960s.

Factors Affecting Recent Changes in Violent Crime
There are two technical factors related to crime classification and reporting—and unrelated to
Prop 47—that contribute to the recent increase in reported violent crime rates in California. Thus
we should be cautious in attributing the rise in violent crime rates from 2014 to 2016 to Prop 47.

Figure 3. California’s violent and property crime rates are still at historic lows
3,000

Property crime

7,000

Violent crime

2,500

6,000
2,000

5,000
4,000

1,500

3,000

1,000

2,000
500

1,000
0
1960

0
1965

1970

1975

1980

1985

1990

1995

2000

2005

2010

2015

SOURCE: Author calculations based on FBI Uniform Crime Report 1960–2002 and the California Department of Justice’s Criminal Justice
Statistics Center, California Crimes and Clearances Files, 2003–2016.
NOTE: Violent crime includes homicide, rape, robbery, and aggravated assault. Property crime includes burglary, motor vehicle theft,
and larceny theft (including non-felonious larceny theft).

8 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Violent crime rate per 100,000 residents

Property crime rate per 100,000 residents

8,000

First, the FBI implemented a change in 2014 that
expanded the definition of sexual crimes that constitute
rape. According to the FBI’s Uniform Crime Report,
the new definition added about 38 percent to the number of reported rapes in California in 2014, increasing
the violent crime rate by about 8 more violent crimes
per 100,000 residents. Accounting for this definitional
change reduces the 2014–2016 increase in the violent
crime rate from 13 percent to 10.7 percent.
A more significant factor is the role of the Los Angeles
Police Department’s (LAPD) underreporting of violent
crimes, especially aggravated assaults, prior to 2014.9
An analysis by the Los Angeles Times in 2014 found that
“with the incidents counted correctly, violent crime in
the city was 7 percent higher than the LAPD reported
in the period from 2005 to fall 2012, and the number of
GETTY IMAGES
serious assaults was 16 percent higher” (Poston, Rubin,
Prop 47 did not affect violent crime, but it may have contributed
to a rise in some property crimes.
and Pesce 2015). In response, the LAPD announced
reforms to improve the accuracy of crime statistics,
including the creation of a Data Integrity Unit (Rubin and Poston 2014). A 2015 report by the
Los Angeles Office of the Inspector General (OIG) estimated that if reported assaults had been
properly classified by the LAPD, on average, the aggravated assault rates between 2008 and 2014
would have been about 36 percent higher each year (OIG 2015). Given the size of Los Angeles,
this significant underreporting raises concerns about comparing California’s reported violent
crime rates, and especially its aggravated assault rates, for the 2008–2014 period to those of subsequent years, as that comparison overstates increases in violent crime.
While it is unclear the precise extent to which more accurate reporting of violent crimes in 2015
and 2016 in Los Angeles contributes to the statewide increase, we can examine statewide trends
with and without Los Angeles County. The statewide increase in the violent crime rate between
2014 and 2016 was 13 percent. This drops to 6.4 percent when Los Angeles is excluded. When we
account for both the definitional change in rape and exclude Los Angeles, the increase drops to
4.7 percent. With these two adjustments, the violent crime rate in 2016 is lower than it was in all
other years since 2010, with the exception of 2014 (see Figure A3 in the technical appendices).

Recent Crime Trends
Focusing on crime trends since 2010, we see some recent upticks after Prop 47 but also considerable fluctuation. While monthly rates of violent crime did increase after Prop 47, this upward
trend—which includes data from the LAPD—appears to have started in late 2013 or early 2014
(Figure 4). Between 2010 and 2012, the violent crime rate stayed relatively constant—although
it is subject to seasonal variation—and decreased somewhat in 2013 and 2014. Nonetheless, the
monthly violent crime rate reaches a seven-year high in July 2016 before dropping back down to
pre-reform levels, reaching 35 violent crimes per 100,000 residents in December 2016.

PPIC.ORG 9

Monthly property crime rates show more pronounced trends. After a rather steady increase in
2012, the monthly property crime rate reached a seven-year high in January 2013. It then started
to drift down, staying relatively steady between February and November 2014, when it reached a
seven-year low. After Prop 47, the monthly property crime rate jumped up somewhat and ranged
between 220 and 230 property crimes per 100,000 residents from December 2014 to January
2016. This increase represents a statistically significant departure from pre-reform levels (see
Technical Appendix A for more details). It then went down and fluctuated around 210 monthly
property crimes per 100,000 residents in 2016, similar to the monthly rates from 2010 to 2013.

Assessing the Rise in Property Crime
Property crimes include burglary, larceny theft (e.g., pick pocketing, purse snatching, shoplifting,
theft from motor vehicles), and motor vehicle theft. We find the increase in larceny theft is driving
the immediate increase in the number of monthly property crimes after Prop 47 (see Figure A1 in
the technical appendices). Despite fluctuations, since January 2013 the monthly rate of larceny
thefts had been steadily drifting down to lows of about 125 per 100,000 residents. This number
then increases by about 10 percent (to around 140 thefts per 100,000 residents) in late 2014 and
2015, before returning to roughly pre-reform levels in 2016.
Burglaries and auto thefts do not appear to play a significant role in the relatively sharp postreform increase in monthly property crime rates.10 Like larceny thefts, the monthly burglary rate
also seems to deviate from its downward trend—which started in early 2013—and increased right
after Prop 47 passed. However, within a few months it appears to trend downward again, possibly leveling out in 2016. The monthly auto theft rate goes up somewhat in 2015 and 2016, after
experiencing seven-year lows in early 2014.
Recent increases in two types of larceny theft—thefts from motor vehicles and shoplifting—have
been widely reported in the media and are often linked to Prop 47.11 These two crimes typically
account for about half of all larcenies. Legislative proposals, such as AB 1326, aim to address
repeat offenders’ thefts of property valued under $950 per incident, which are currently considered misdemeanors under Prop 47. For example, the League of California Cities (2017) says in
support of AB 1326, “This measure addresses increased incidents of both individual shoplifting
and organized retail theft operations responsible for the loss of millions of dollars in retail merchandise statewide, which entail a corresponding hit to local sales tax revenues.”
Monthly crime numbers bear out the reported increases in thefts from motor vehicles and shoplifting (Figure 5). However, reports of shoplifting decreased in 2016 and are now at roughly the
same levels as rates observed in 2010. After increases in 2012, reported thefts from motor vehicles
decreased in 2013 and most of 2014. But starting in December 2014, these reported thefts jumped
up from 16,000–17,000 to 19,000–20,000. There is no sign yet of thefts from motor vehicles
returning to pre-reform levels. Furthermore, the increase of 35,300 thefts from motor vehicles
between 2014 and 2016 accounts for almost two-thirds of the 54,700 increase in the number of
property crimes in California.
The average monthly number of reported shoplifting incidents stayed relatively flat, between
7,800 and 8,100 per month, in the years before Prop 47 passed. It then jumped up to 9,000 in
November 2014 and reached a seven-year high of 9,900 in January 2015. The post-reform oneyear average of 9,050 is about 12 percent higher than the one-year pre-reform average. However, shoplifting decreased notably in 2016, when the monthly numbers dropped to pre-reform
10 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Prop 47

65

Property crime
250

60

Violent crime

55
50

230

45
210

40
35

190

30
25

170

20
150

15

Monthly violent crime rate per 100,000 residents

270

Ja
n1
Ap 0
r-1
Ju 0
l-1
Oc 0
t-1
Ja 0
n1
Ap 1
r-1
Ju 1
l-1
Oc 1
t-1
Ja 1
n1
Ap 2
r-1
Ju 2
l-1
Oc 2
t-1
Ja 2
n1
Ap 3
r-1
Ju 3
l-1
Oc 3
t-1
Ja 3
n1
Ap 4
r-1
Ju 4
l-1
Oc 4
t-1
Ja 4
n1
Ap 5
r-1
Ju 5
l-1
Oc 5
t-1
Ja 5
n1
Ap 6
r-1
Ju 6
l-1
Oc 6
t-1
6

Monthly property crime rate per 100,000 residents

Figure 4. Monthly rates of violent crime did not change much after Prop 47, but property crime rates
showed an initial uptick

SOURCE: Author calculations based on California Department of Justice’s Criminal Justice Statistics Center, California Crimes and Clearances
Monthly Files, 2010–2016.
NOTES: Violent crime includes homicide, rape, robbery, and aggravated assault. Property crime includes burglary, motor vehicle theft, and larceny
theft (including non-felonious larceny theft). Since the number of days in a month varies, the monthly numbers are adjusted accordingly. For
example, February numbers in non-leap years are multiplied by a factor of 31/28.

Figure 5. Both shoplifting and thefts from motor vehicles rose after Prop 47, but shoplifting declined
to pre-reforms levels in 2016
Prop 47

24,000

15,000

22,000

Shoplifting

14,000
13,000

20,000
18,000

12,000
11,000
10,000

16,000
14,000

9,000
8,000
7,000

12,000

6,000
5,000

Ja

n1
Ap 0
r-1
0
Ju
l-1
Oc 0
t-1
Ja 0
n1
Ap 1
r-1
1
Ju
l-1
1
Oc
tJa 11
n1
Ap 2
r-1
2
Ju
l-1
2
Oc
t-1
Ja 2
n1
Ap 3
r-1
3
Ju
l-1
3
Oc
t-1
Ja 3
n1
Ap 4
r-1
4
Ju
l-1
Oc 4
t-1
Ja 4
n1
Ap 5
r-1
5
Ju
l-1
Oc 5
t-1
Ja 5
n1
Ap 6
r-1
Ju 6
l-1
Oc 6
t-1
6

10,000

Monthly shoplifting incidents

Monthly thefts from motor vehicles

Theft from motor vehicles

SOURCE: Author calculations based on California Department of Justice’s Criminal Justice Statistics Center, California Crimes and Clearances
Monthly Files, 2010–2016.

PPIC.ORG 11

levels. It is possible that the 2016 drop in shoplifting reflects increased law enforcement efforts
responding to retailers’ substantiated concerns, retailers’ own increased theft prevention, or
simply a decline in the likelihood of reporting shoplifting. Shoplifting accounts for only 1,300
(2.4%) of the 2014–2016 increase in property crimes.

The Impact of Proposition 47 on Recent Crime Trends
The descriptive analysis above suggests a possible increase in some property crime as a result of
Prop 47. Although violent crime has increased since 2014, this trend appears to have preceded the
reform and is overstated due to a definitional change in the crime data and the underreporting
of aggravated assaults by the LAPD prior to 2014. Furthermore, these patterns in property and
violent crime rates do not account for broader crime trends that may be unrelated to Prop 47 and
may extend beyond California.
To determine whether crime rates have been affected by Prop 47, we identify a combination of
states that best represent what California’s crime rates would have been had the reform not been
implemented.12 We then measure the effect of Prop 47 by assessing the degree to which the difference in crime rates between California and the comparison states widens in 2015 and 2016, above
and beyond any possible differences that emerge in the immediate period preceding Prop 47 due to
earlier reforms. The methods and more detailed findings are presented in Technical Appendix A.
After declining in 2013 and 2014, California’s violent crime rate increased in both 2015 and 2016.
While violent crime also increased in 2015 in the comparison states, California’s increases are
greater, but not by much (see Figure A4 in the technical appendices). The estimated relative increase
in California of about 37 more violent crimes per 100,000 residents is not statistically significant;
nationwide, six other states experienced larger increases
after 2014, relative to their comparison states.13

12 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Prop 47

550

500

450
California
(without LAPD)

400

Comparison states

6

14

20
1

12

20

20

10
20

08
20

4

06
20

2

20
0

20
0

00

350

20

In light of the LAPD’s underreporting of violent crimes
discussed above, we rerun the analysis excluding violent
crimes reported by the LAPD (roughly excluding the city
of Los Angeles). We then match the rest of California
with states that had similar pre-reform violent crime
rates to California without Los Angeles. The results
show that the rest of California saw a post-reform
increase in violent crime, but this increase was smaller
than that of the comparison states (Figure 6).

Figure 6. Excluding LAPD shows the rest of
California did not see a greater increase in violent
crime than comparison states
Violent crime rate per 100,000 residents

When we examine different kinds of violent crimes, we
find that an increase in aggravated assaults is the main
contributor to the observed increase in California’s
violent crime rate. But once again, this increase is not
statistically significant; nationwide, eight other states
saw larger increases after 2014 than their comparison
states did. Furthermore, none of California’s changes
in the rates of murder, rape, and robbery are statistically significant compared to changes in the comparison
states.

SOURCE: Author estimates based on annual state-level data from the
FBI’s Uniform Crime Reports, 2000–2016.

While property crime has been steadily decreasing in California since 2004, the state saw an
increase in 2012 that was driven by a jump in auto thefts caused by realignment (Lofstrom and
Raphael 2013). The gap of approximately 230 to 240 property crimes per 100,000 residents
between California and the comparison states remained roughly constant in 2013 and 2014.
However, in 2015 the gap widens further to around 450 and then declines somewhat to about 380
in 2016 (see Figure A5 in the technical appendices). We estimate Prop 47 increased the property
crime rate by about 190 property crimes per 100,000 residents. However, the relative increase is
not statistically significant. Nationwide, eight other states saw larger increases after 2014 relative
to their comparison states.
When we examine trends in different kinds of property crimes, California’s notable decline in
burglaries since 2012 is very similar to the sharp decrease observed in the comparison states;
there is no evidence that Prop 47 affected this trend. The auto theft rate increased from 391 per
100,000 residents in 2014 to 450 in 2016. However, comparison states also saw recent increases,
and again we find that the post-reform increase in California is not statistically significant
compared to increases in comparison states. In contrast, the increase in larceny theft is marginally statistically significant.14 We estimate that Prop 47 led to a rise in the larceny theft rate of
about 135 per 100,000 residents, an increase of close to 9 percent compared to the 2014 rate
(Figure 7).
We conducted additional analyses to corroborate these findings on Prop 47’s impact on crime
rates. Using monthly California crime data, we estimated and tested for the effects of Prop 47 on
specific crime rates before and after the reform.15 Consistent with the analysis above, this method
finds no convincing evidence that Prop 47 led to higher violent crime rates in California, but property crime increased, primarily driven by higher larceny theft rates.

Figure 7. After Prop 47, California’s larceny theft
rate deviated notably from that of comparison
states
2,100

Prop 47

2,000
1,900
1,800
1,700
1,600

California

1,500

Comparison states

14

2

16
20

20

20
1

10
20

6

08
20

20
0

04
20

20
02

1,400

20
00

Larceny theft rate per 100,000 residents

We also performed a preliminary county-level analysis
to examine whether there is a relationship between
monthly crime data and jail incarceration, resentencing
and reclassification petitions, and arrest activity under
Prop 47. We consider these results preliminary because
we are not currently able to incorporate data on recent
changes in the prison population, a key group when
examining the possible impact of Prop 47 on crime
rates.16 While the results are broadly consistent with
the statewide analysis above, they also suggest that we
should be cautious not to overstate Prop 47’s impact on
property crime. The preliminary county-level analysis
once again finds no convincing evidence of an increase in
violent crime overall or in specific violent crime rates.17
There is some evidence of an impact on property crime
overall and on larceny and auto thefts, but the estimates
are not always statistically significant and vary across
our different modeling scenarios.18 The methods and
results for both of these supplemental analyses are presented in Technical Appendix A.

SOURCE: Author estimates based on annual state-level data from the
FBI’s Uniform Crime Reports, 2000–2016.

PPIC.ORG 13

Jail Bookings
One goal of Prop 47 was to reduce individuals’ involvement with the criminal justice system. In
addition to the reductions in arrests and overall incarceration levels described above, here we
examine another indicator of criminal justice contact: changes in jail bookings, which reflect the
use of one of the most expensive correctional resources at the county level. Examining bookings
over time and under policy reform is also important because individuals booked into custody,
even for short periods of time, can face challenges maintaining employment or providing care for
children. Bookings into jail occur when individuals are arrested for new offenses and brought into
custody pending charges. Individuals may also be booked into jail for reasons other than arrests
for new offenses, including bookings on bench warrants, supervision violations, commitments to
custody, and holds for other reasons.
This analysis draws on criminal justice data collected through the BSCC–PPIC Multi-County
Study (MCS), a collaborative effort between the California Board of State and Community
Corrections (BSCC), PPIC, and a group of 12 California counties.19 The MCS counties comprise
about 60 percent of the state population and were selected to represent the demographic, economic, and geographic characteristics of the state.20 In addition, the California Department of
Justice (DOJ) and CDCR provided essential data to fill out the state-local picture.

Trends in County Jail Bookings
Overall, jail bookings in MCS counties decreased from almost 60,000 in October 2014 (the
month prior to Prop 47) to 55,400 one year later, or a decline of about 8 percent. We see a reduction in bookings among all racial/ethnic groups (Figure 8). However, declines were relatively
larger for whites and Latinos than for African Americans. In October 2014, African Americans
represented 21.6 percent of bookings; one year after Prop 47, African Americans represented
22.3 percent of all jail bookings into MCS counties. Although this change is small, even a small
increase may be of concern given the highly disproportionate representation of African Americans
in the criminal justice system.
The overall decline in jail bookings is driven by a reduction in bookings for Prop 47 drug and
property offenses. 21 Monthly bookings for Prop 47 offenses in the MCS counties declined by
about 38 percent, from roughly 14,600 in October 2014 to 9,100 in October 2015 (Figure 9).
As expected, bookings for Prop 47 offenses shifted from felonies to misdemeanors. In October
2014, more than two-thirds of bookings for Prop 47 offenses were on felony charges; one year
later, felony charges represented less than one-third of these bookings. 22 Individuals booked for
Prop 47 offenses may still be charged with felonies due to their criminal histories or, in the case
of a property offense, the value of the property.
When we examine the racial/ethnic composition of those booked for Prop 47 offenses, we continue to see that declines in jail bookings after Prop 47 were relatively larger for whites and
Latinos than for African Americans. In fact, the disparity is somewhat greater among those
booked for Prop 47 offenses, with African Americans representing 21.0 percent of Prop 47
bookings in October 2014, compared to 22.6 percent a year later.
Because people can be booked into jail multiple times, we also examine jail bookings at the
individual level over a one-year period. The total number of individuals booked into jail in MCS

14 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Figure 8. Jail bookings declined across all racial/ethnic groups
Prop 47

35,000
30,000

Monthly jail bookings

Latino
25,000

White
African American

20,000

Other
Asian American

15,000
10,000
5,000

5

5
t-1
Oc

5

l-1

r-1

Ju

4

4

15
n-

Ap

Ja

t-1
Oc

4

l-1

3

14

r-1

Ju

Ap

nJa

3
l-1

t-1
Oc

3

13

r-1

Ju

Ap

2
t-1

2

2

l-1

nJa

Oc

Ju

12

r-1
Ap

nJa

Oc

t-1

1

0

SOURCE: Author calculations based on data from the BSCC–PPIC Multi-County Study (MCS).
NOTE: These data represent bookings in the 12 MCS counties.

Figure 9. Jail bookings for Prop 47 offenses declined and shifted from felonies to misdemeanors
18,000

Prop 47

16,000

Monthly jail bookings

14,000

Overall
Felony

12,000

Misdemeanor
10,000
8,000
6,000
4,000
2,000

5

5
l-1

t-1
Oc

5
r-1

Ju

n15

Ap

Ja

4

4
l-1

t-1
Oc

Ju

4

3

-14

r-1
Ap

Ja
n

3
l-1

t-1
Oc

Ju

3

13

r-1
Ap

2
t-1

2
l-1

nJa

Oc

Ju

2

-12

r-1
Ap

Ja
n

Oc

t-1

1

0

SOURCE: Author calculations based on data from the BSCC–PPIC Multi-County Study (MCS).
NOTE: These data represent bookings in the 12 MCS counties.

PPIC.ORG 15

counties declined from 466,000 in the year prior to Prop 47 to 417,600 in the year after, a decline
of 48,400, or about 10 percent. This change was largely driven by the decline in jail contact
among those booked for Prop 47 offenses. In the year before Prop 47, 120,000 individuals were
booked into jail for Prop 47 offenses; in the year after, this number declined to 77,300 (a decline
of 35.6%). Importantly, the mean number of bookings per individual within a one-year period did
not increase, meaning that those who have contact with the jail system after Prop 47, on average,
are not likely to be booked more frequently compared to before Prop 47.23
The decrease in jail bookings and the number of individuals with jail contact under Prop 47
reflect a shift from jail custody to alternative responses for Prop 47 offenses. For jail systems that
suffer from overcrowding, reductions in bookings for Prop 47 offenses have relieved crowding
and allowed for a decrease in capacity-based releases from jail for more serious offenders (Bird et
al. 2016; Grattet et al. 2016).

Recidivism
Those in favor of Prop 47 argued the criminal justice system could achieve better recidivism outcomes for lower-level drug and property offenders by reducing incarceration levels and increasing
the use of treatment interventions, particularly for those with mental health and/or substance-use
challenges. However, those opposed to Prop 47 raised concerns that less severe penalties could
increase recidivism and reduce the ability of courts and
prosecutors to induce individuals into treatment programs as an alternative to custody time.
We find that Prop 47 led to lower rates of rearrest and
reconviction among Prop 47 offenders, compared to
similar individuals prior to the reform. Our analysis
indicates that the measure reduced both arrests by law
enforcement and convictions resulting from prosecutions
by district attorneys.
This analysis draws on data collected through the
BSCC–PPIC Multi-County Study (MCS). 24 While the
MCS data provide rich demographic and criminal history information, we are not able to perfectly identify
the population of Prop 47 offenders due to two data
limitations. First, we are not able to exclude offenses for
which the property value was greater than $950; and
second, we are not able to distinguish shoplifting from
other forms of second-degree burglary before Prop 47.
Therefore, our pre- and post-reform offender groups
include all individuals convicted of Prop 47 property
offenses, regardless of the property value, and all
second-degree burglaries. We provide more information
about the Prop 47 offender groups used in this analysis
in Technical Appendix B.

16 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

GETTY IMAGES

Lower recidivism rates may reflect changes in reoffending or changes
in the practices of criminal justice agencies after Prop 47.

The pre-reform group includes individuals released from jail or prison custody, or sentenced to
probation, between November 5, 2011 and October 31, 2012, allowing for a two-year recidivism window before the passage of Prop 47. 25 The post-reform group includes individuals
released from jail or prison, or sentenced to probation, between November 5, 2014 and October
31, 2015, followed for two years after release. We identified 44,985 individuals in the pre-reform
group. From this group we selected a comparison group of 20,029 individuals who most closely
resembled the 28,484 individuals in the post-reform group in their demographic and criminal
history characteristics. 26 After matching, the post-reform group is still slightly older and more
likely to be male than the pre-reform group. The post-reform group also has slightly higher levels
of past offending, suggesting a higher likelihood of reoffending. In order to draw conclusions
about differences in recidivism rates before and after Prop 47, we use a regression model to adjust
for these remaining differences in demographic and criminal history characteristics.
We examine two measures commonly used in recidivism research—rearrest and reconviction
rates. These measures are often used to capture changes in levels of reoffending in response to
a policy change. However, these measures may also reflect changes in the practices of criminal
justice agencies. For example, if a policy change led to a shift in policing strategies, such as
reduced enforcement for drug possession offenses, we may observe changes in rearrest rates even
if there is no change in the underlying behavior of former offenders. Available data do not allow
for an analysis that separates the effect of Prop 47 on reoffending from the reform’s effect on the
practices of criminal justice agencies.

Figure 10 presents estimates of the effects of Prop 47
on the overall rearrest rate for any offense and the
rearrest rate for what we categorize as Prop 47 offenses.
After adjusting for demographic and criminal history
characteristics, we find the post-reform group has an
overall two-year rearrest rate of 70.8 percent, 1.8
percentage points lower than the rearrest rate of the
pre-reform group.
The overall decline in rearrest rates is driven by a
decline in the Prop 47 rearrest rate, and more specifically a decline in rearrests for Prop 47 drug offenses.27
The post-reform rearrest rate for drug offenses is
11.3 percentage points lower than the comparable
pre-reform rearrest rate. However, there is no significant difference between pre- and post-reform rearrest
rates for property offenses. This finding suggests that
reductions in enforcement for Prop 47 offenses likely
play a role in lower rates of rearrest, as law enforcement
has more flexibility in responding to drug offenses than
property offenses.

Figure 10. Two-year rearrest rates declined after
Prop 47, driven by reductions in lower-level drug
offenses
80

72.6 70.8

Before Prop 47

70

After Prop 47
60
Rearrest rate (%)

Findings on Rearrest and Reconviction
Rates

50

45.3

40

35.0

32.3

30
20

19.5 19.6

21.0

Prop 47
property
offense

Prop 47
drug
offense

10
0
Any
offense

Prop 47
offense

SOURCE: Author calculations based on data from the BSCC–PPIC
Multi-County Study (MCS).
NOTES: Estimated differences between the pre- and post-reform groups
are significant at the p<.01 level, with the exception of the difference in
the Prop 47 property offense rates, which is not significant.

PPIC.ORG 17

Figure 11. Two-year reconviction rates declined
after Prop 47, for both lower-level property and
drug offenses
60
50
Reconviction rate (%)

Figure 11 presents estimates of the effects of Prop 47
on reconviction rates. The overall two-year reconviction rate for individuals released after serving sentences
for Prop 47 offenses is 46.0 percent, 3.1 percentage
points lower than the reconviction rate of their prereform counterparts. Reconviction rates for Prop 47
offenses follow a similar pattern to overall rates, but
show a more dramatic decline. The two-year reconviction rate for all Prop 47 offenses declines by 11.3
percentage points, from 25.4 percent of individuals in
the pre-reform group to 14.1 percent of individuals in
the post-reform group. Unlike rearrest rates, in which
we saw greater declines for drug offenses than for
property offenses, reconvictions decline substantially
for both Prop 47 property and drug offenses. The twoyear reconviction rate for Prop 47 property offenses is
4.2 percentage points lower for the post-reform group,
and the two-year reconviction rate for Prop 47 drug
offenses is 7.9 percentage points lower.

Before Prop 47

49.1
46.0

After Prop 47

40
30

25.4

20
14.1

15.4
11.9
7.7

10

7.5

0
Any
offense

Prop 47
offense

Prop 47
property
offense

Prop 47
drug
offense

SOURCE: Author calculations based on data from the BSCC–PPIC

Multi-County Study (MCS).
As was the case with rearrests, these lower reconvicNOTES: All estimated differences between the pre- and post-reform
tion rates may be associated with changes in underlying
groups are significant at the p<.01 level.
offender behavior, changes in criminal justice practices,
or some combination of the two. Changes in criminal
justice practices may include less street-level enforcement by police and sheriff departments, and may extend to lower charging rates among district
attorneys and/or less willingness to accept plea deals among public defenders. To gain insight into
where changes in criminal justice practices may be occurring, we reexamine reconviction rates
separately for individuals who have been rearrested. This approach allows us to isolate the change
in reconvictions from the influence of changing rearrest rates.

We find that overall two-year reconviction rates among those who were rearrested are 3.1 percentage points lower for the post-reform group. When we narrow our analysis to reconvictions
for Prop 47 offenses, we find a sharp decrease of 15.0 percentage points in the two-year reconviction rate among those rearrested for any type of offense. These findings show that changes in
reconviction rates are not driven solely by changes in rearrest rates, suggesting Prop 47 induced
changes in criminal justice processes at both the law enforcement and prosecutorial levels.
Although we find lower overall rates of rearrest and reconviction, it is not clear to what extent
these outcomes are driven by changes in reoffending during the early years of Prop 47 implementation. Over time, and as state savings are reallocated to support increased access to treatment,
we may observe further reductions in recidivism tied to improvements in the underlying behavior
of former offenders.

18 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

Prop 47 Grant Programs
Proponents of Prop 47 emphasized that funding would be better spent on community-based
treatment programs and interventions to reduce recidivism rather than on incarceration. In
particular, focusing on substance-use disorder and mental health needs could help lower-level
offenders. With this goal in mind, Prop 47 requires that 65 percent of state savings go toward
mental health and substance-use disorder treatment for those involved in the criminal justice
system. The remainder goes to K–12 schools (25%) and victim services (10%). 28
The first transfer of savings took place in 2016. Thus, programs using these funds are still quite
new and unlikely to have already had an impact on recidivism. In addition, the total amount of
Prop 47 savings is small relative to the billions of dollars spent on the criminal justice system.
Nevertheless, the program provides an opportunity for local communities to develop and implement targeted programs, while supporting evaluation efforts to help identify successful strategies.
This first transfer totaled more than $67 million (including a supplemental payment) and went to
the three agencies tasked with distributing the grants (DOF 2016).29 Long-term savings associated
with Prop 47 are estimated to be $69 million annually, according to the governor’s recent budget
proposal (DOF 2018). As a point of reference, the Community Corrections Performance Incentive
Grant program created in SB 678 includes $106.4 million in funding for 2018–19.30 It should be
noted that Prop 47 did not require counties to report or redirect any local savings that may come
from the change in sentencing.31

Overview of Treatment Programs
Below, we describe grant programs funded through the Board of State and Community
Corrections. These projects are required to include some combination of mental health services,
substance-use disorder treatment, and efforts to work with individuals before arrest or booking
into jail.32 These programs are administrated by the BSCC, which gives priority to projects that
also provide housing-related support and/or other community-based supportive services. Public
agencies submit grant applications, and at least half of the funds must go to non-governmental
community-based organizations. Proposals must also describe how the program relies on evidence-based practices and if other jurisdictions are already using the program.33
The BSCC is using the Prop 47 funds to provide three-year grants, with an estimated $104 million
available in funding from June 2017 to August 2020. The BSCC has funded 23 projects across
17 counties (see Table C1 in the technical appendices for a complete list of projects).34 Fifteen
large-scale projects (more than $1 million) received funding; for example, the Los Angeles City
Attorney’s Office received $6 million to implement a health-focused drug intervention and
pre-booking diversion program in south Los Angeles. Eight small-scale projects ($1 million or
less) received funding. For instance, the Merced County Probation Department received about
$960,700 to provide mental health and drug treatment for young men in the Los Banos region.
Beyond the guiding principles emphasized in the grant requirements, the actual projects vary
significantly—they serve different ages, provide different types of services, target different kinds
of crimes, and intervene at different points in the criminal justice system. Most projects involve

PPIC.ORG 19

several organizations working together. For example, post-conviction programs provide coordination to make sure individuals comply with court orders while also helping them access housing,
medical and mental health care, and drug treatment.
In some cases, these programs are new to the public agency, while in other cases the funds allow
for expansion of current programming. Because the funding amounts were capped by the BSCC
and residential mental health and drug treatment can be
expensive, the number of individuals served in a year is
on the low end, with most programs serving fewer than
100 individuals per year. However, the various programs
could be useful test cases, and the number of individuals
served can potentially be scaled up for programs found
to be successful. Funding for expanded programs could
come from yearly public safety realignment (AB 109)
funding or SB 678 funding provided to counties, or from
public agency general funds or savings realized from
Prop 47.

Funding from Prop 47 gives
local agencies the opportunity
to expand or create promising
programs to reduce recidivism.

Finally, funding recipients have to spend at least 5 percent of their funding on data collection and
evaluation of the program. At the end of the second year of the grant program, each public agency
has to submit a report that examines program implementation and early outcomes. A final report
is due upon completion of the three-year grant program. While these final reports will not be
available to the public until late 2020, they could provide a chance to learn which strategies are
most successful, given consistent reporting and analysis across projects.
In light of the billions of dollars spent annually on the criminal justice system at the state and
local levels in California, the total funding of Prop 47 programs is arguably small. However, the
funded efforts give local public agencies the opportunity to expand or create new programming
that they believe will help reduce recidivism in their communities. Local and state criminal justice practitioners and policymakers should pay close attention to the programs, implementation
successes and challenges, and the final evaluation reports. If these programs are successful, other
agencies could adopt these strategies to treat similar populations in their jurisdictions and across
the state.

Conclusion
Proposition 47 continues to be the subject of much debate. The reform—which reduced penalties
for certain lower-level drug and property offenses—has undoubtedly played a significant role in
California’s recent efforts to prioritize the state’s prison and jail space for higher-level offenders.
Prison and jail incarceration levels declined substantially under Prop 47. We also observe sudden
drops in arrests and jail bookings.
The reduced reliance on incarceration raised concerns among some observers about Prop 47’s
impact on public safety. We find no convincing evidence that violent crime increased as a result of
Prop 47. Though there has been a recent uptick in violent crime, this trend appears to have started

20 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

prior to the reform. Additional factors unrelated to criminal activity—a change by the FBI in
2014 that expanded the definition of rape, and significant underreporting of violent crimes from
2008 to 2014 by the LAPD—contributed to the observed increase. Excluding violent crime data
from the LAPD shows that recent increases in violent crime rates in California were comparable
to those of similar states.
Our analysis does find some evidence of Prop 47’s impact on property crime rates, which went up
immediately after the law was implemented. This increase has been primarily driven by larceny
thefts, particularly thefts from motor vehicles and shoplifting. We find the increase in the larceny
theft rate in California to be nearly 9 percent higher than that of similar states. In 2016, reported
shoplifting decreased notably, but we do not see signs of a reduction in thefts from motor vehicles.
Considering the high costs of incarceration in California, this highlights the need for alternative
crime-reducing strategies, consistent with our earlier research assessing the impact of realignment
on crime (see Lofstrom and Raphael 2013).
The policy goals of Prop 47 are to reduce contact with the criminal justice system and to reduce
recidivism for lower-level drug and property offenders. Our analysis, using detailed data from
12 California counties, shows declines in jail bookings as well as rearrest and reconviction rates
under Prop 47. We find the policy change reduced jail bookings for Prop 47 offenses by more than
one-third. Prop 47 also lowered the number of people
booked into jail by nearly 50,000 in these counties during
the year following its passage.

Policymakers and practitioners
must work together to counteract
increases in property crimes and
improve reentry outcomes.

Lower rearrest rates for individuals released after serving
sentences for Prop 47 offenses were driven by a reduction
in rearrests for drug possession, while lower reconviction rates were driven by a drop in reconvictions for both
Prop 47 property and drug offenses. We find evidence
that Prop 47 reduced both arrests by law enforcement
and convictions resulting from prosecutions by district
attorneys. Reduced levels of correctional contact—which may allow for better continuity of
employment and improved family and community stability—could be a factor in these lower
recidivism rates. However, given the sudden and noticeable decline in arrests after the reform, we
are not able to separate the effects of Prop 47 on individual reoffending behavior from its effects
on the practices of criminal justice agencies.

Prop 47 aimed to reduce recidivism rates by shifting resources from incarceration to mental health
and substance-use treatment for lower-level drug and property offenders. This redirection of state
correctional savings to treatment interventions has only recently been allocated, and thus our
recidivism analysis does not capture individuals released after the implementation of these programs. A complete assessment of the impacts of Prop 47 will need to account for how increased
interventions may affect crime, criminal justice contact, and recidivism, as well as responses by
law enforcement to the reform.
Substantial reductions in reoffending as a result of the treatment programs funded by Prop 47
savings seem unlikely as this funding represents a very small share of corrections spending in

PPIC.ORG 21

California. However, the initiative offers opportunities for local agencies to create or expand
promising programs. It also requires that these programs be objectively evaluated, in hopes of
identifying and scaling up successful interventions. As local agencies and the state learn more
about which programs are effective in reducing recidivism, Prop 47 could provide a path toward
the use of more cost-effective, evidence-based strategies within the criminal justice system.
One of several far-reaching corrections reforms, Prop 47 further decreased California’s reliance
on incarceration: the state’s incarceration rate is now at levels not seen since the early 1990s.
Importantly, crime rates remain historically low, comparable to those in the 1960s. While
research so far has not revealed convincing evidence that violent crime has risen as a result of
reforms, some property crimes have increased. And though Prop 47 reduced recidivism rates for
lower-level drug and property offenses, it is not clear to what extent this is driven by reduced
reoffending, as law enforcement and prosecutorial changes likely contributed to the declines. To
counteract the increases in property crimes like shoplifting and thefts from motor vehicles—and
to improve reentry outcomes of released offenders—policymakers and practitioners will need to
work together to identify effective programs and policies that will reduce recidivism and maintain
public safety.

22 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

N OT E S
Prop 47 excludes individuals required to register as sex offenders and individuals who have
been convicted of a specific set of violent offenses that carry a maximum sentence of life in
prison or death.

1

In 2014, the court ordered CDCR to implement a number of population-reduction measures,
which also helped lower incarceration rates.

2

Arguments in favor of Proposition 47 from the California Secretary of State “Voter Information Guide” (2014). These were attributed to the following proponents: George Gascon, San
Francisco district attorney; William Lansdowne, former police chief of San Diego, San Jose,
and Richmond; and Dionne Wilson, victim’s advocate, Crime Survivors for Safety and Justice.
Also see Males and Buchen (2014) and Teji and Graves (2014) for further discussion of the
goals and expected impacts of Prop 47.

3

Arguments against Proposition 47 from the California Secretary of State “Voter Information
Guide” (2014). These were attributed to the following opponents: Christopher W. Boyd,
president, California Police Chiefs Association; Harriet Salarno, president, Crime Victims
United; and Gilbert G. Otero, president, California District Attorneys Association.

4

These measures included increased credit earning and early parole for certain non-violent
inmates.

5

Data collected from the Jail Profile Survey conducted by the Board of State and Community
Corrections.

6

The California Department of Justice typically releases crime data for any given year in
August the following year. Thus, 2016 is the most up-to-date statewide crime data available.

7

Although crime rates are difficult to predict, factors that notably drive trends include
economic conditions, the dynamics of illegal drug markets, law enforcement personnel and
policing strategies, changes in sentencing laws and incarceration, and demographic shifts (in,
for example, age and race/ethnicity). The influence of these factors on crime trends varies with
the type of crime and across different geographies.

8

Aggravated assaults are the largest contributor to the violent crime rate, accounting for about
60 percent of all violent crimes.

9

Even though we find no convincing evidence that Prop 47 affected auto thefts, our prior
research shows that public safety realignment, implemented in 2011, led to a rise in auto thefts
but did not affect burglary or larceny theft rates (Lofstrom and Raphael 2013; Lofstrom and
Martin 2015).

10

11

For two examples of news articles, see Debbaudt (2016) and Moriarty (2018).

We used the same approach, known as a synthetic control method, when assessing whether
realignment had an impact on crime rates (Lofstrom and Raphael 2013; Lofstrom and Martin
2015), and it has also been used by Bartos and Kubrin (forthcoming). Note that given the
previous significant criminal justice reforms implemented in California since 2011 (most
importantly realignment in 2011, but also Proposition 36 in 2012, as well as earlier reforms
such as determinate sentencing in 1976 and the three-strikes law in 1994), we use the period
from 2000 to 2010 to identify suitable comparison states. Bartos and Kubrin (forthcoming)
identify comparison states by using the period from 1970 to 2014.

12

PPIC.ORG 23

To test whether the differences between California and the matched comparison states are
statistically significant, we rerun the matching process for all other states, generating a set
of matched states for each, and then compare the observed post-reform year differences to
the pre-reform year differences. A ranking of the magnitude of the estimated changes tells
us whether California’s changes stand out and provides the basis for statistical significance.
California’s post-reform change would be statistically significant at the commonly used 5 percent significance level if it ranked first or second. At a 10 percent significance level, the change
would need to be ranked fifth or higher.

13

Specifically, the increase in larceny theft is significant at the 10 percent significance level.
See Technical Appendix A for more details on this analysis.

14

To assess whether Prop 47 led to changes in crime rates, we estimate and compare quadratic crime rate trends separately for the 24-month period before the reform passed and the
27-month period after passage.

15

We hope to obtain monthly county-level data that will allow us to examine in future research
whether this population had any impacts on public safety.
16

There is some evidence of an increase in robberies as a result of fewer arrests but the finding
is sensitive to the model specification.
17

Specifically, the estimates are sensitive to whether the models are weighted by population
and by the time periods used to calculate the changes in crime and jail incarceration rates. See
Technical Appendix A for more details.
18

19
As described on its website, the BSCC is the independent state agency that “provides leadership to the adults and juvenile criminal justice systems, expertise on public safety realignment
issues, a data and information clearinghouse, and technical assistance on a wide range of community corrections issues.”
20
The MCS includes the following partner counties: Alameda, Contra Costa, Fresno,
Humboldt, Kern, Los Angeles, Orange, Sacramento, San Bernardino, San Francisco, Shasta,
and Stanislaus. Technical Appendix B provides more information on the MCS data and the
study counties.

The analysis of changes in bookings over time categorizes bookings as “Prop 47 bookings”
when the most serious charge at booking is a potentially Prop 47–eligible offense based on the
penal code. At the time of booking, we do not distinguish between offenders who would be
eligible, should charges be filed, under Prop 47 based on the dollar value of the item (in the case
of property offenses) or criminal history. The focus of the analysis is on the change in bookings
over time under Prop 47. Technical Appendix B provides more details on the construction of
Prop 47 booking offenses. See Hunter et al. (2017) for additional discussion on the challenges
associated with identifying the Prop 47 offender population.

21

22
Prop 47 excluded individuals who had certain prior violent or sexual offenses from
eligibility. It also imposed other eligibility restrictions. In addition, prior to Prop 47, we are
not able to distinguish shoplifting from other second-degree burglaries; we retain all seconddegree burglaries in both periods to ensure we do not bias our estimates of change over time in
key outcomes. These issues are discussed in further detail in Technical Appendix B.
23
Overall, the mean number of bookings per individual in a one-year period was 1.68 before
Prop 47 and 1.69 after Prop 47. Among those booked into jail for Prop 47 offenses at some
point during the period, the mean number of bookings per individual declined from 1.44 to
1.35 in a one-year period.
24

See Technical Appendix B for more information on the MCS.

24 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

It is important to note that this group was released from custody during the first and second
year of the implementation of public safety realignment. Therefore, recidivism rates for this
group may reflect the effects of realignment and may not be reflective of recidivism rates for
similar offenders prior to realignment.

25

We use a propensity score matching strategy to select individuals released pre-reform who
had similar characteristics to the post-reform group. This strategy is explained in more detail
in Technical Appendix B. Table B2 shows demographic and criminal history characteristics for
all groups.

26

Rearrest data include all arrests statewide that are reported to the California Department of
Justice. It is possible that some arrests are not reported and that, more importantly, the number
of unreported arrests may have changed over this time period. To the extent that law enforcement has increased the use of citations in the field and to the extent that those types of arrests
are less likely to be reported, part of the reduction in rearrests under Prop 47 could be
explained by a reduction in the reporting of arrests.

27

The California Department of Finance (DOF) determines the state savings from Prop 47
and deducts state costs related to the measure for each fiscal year in comparison to the fiscal
year 2013–14, before the law went into effect. State savings come from a reduction in inmates
in the state prison system, reductions in patients within the Department of State Hospitals,
and reductions in criminal caseloads within the court system. Near-term costs of the measure
occur from an increase in the state parole population and an increase in the state court workload that came from the resentencing and reclassification that was possible after the measure
went into effect.

28

29
The LAO’s February 2016 report gives a closer look at the initial estimates of state savings
from the measure. After the DOF projected lower savings than was originally estimated, proponents of the measure pushed the legislature and governor to agree to a supplemental payment
in the first year. For fiscal year 2017–18, nearly $46 million in savings were transferred (DOF
2017), with the state estimating that an additional $64 million in savings will be transferred
for fiscal year 2018–19.
30
SB 678 (signed into law in 2009) created incentives for counties to keep felony probationers
from being sent to state prison for failing the terms of their community supervision.

Several counties have started to calculate and report savings (and costs) realized from the
measure, including Los Angeles and Riverside.

31

For more information on the other grant programs, see Technical Appendix C. Proposition
47 and AB 1056 (signed into law in 2015) lay out the statutory requirements and authority of
the BSCC to operate the grant program.

32

33
The BSCC provides more detailed information about the Prop 47 funding program, including the requirements of the request for proposals and grant recipients, on its website (“Proposition 47 and the BSCC”).

Fifty-three public agencies submitted proposals, including superior courts; school districts;
district and city attorneys; health and human services; and police, probation, sheriff, and
county education offices. In addition to the 23 projects that received funding, 10 large-scale
projects met the scoring thresholds but no more funding was available. Twenty-five proposals (13 small-scale and 12 large-scale) did not meet the scoring thresholds established by the
BSCC. If during the three-year period funding increases beyond the BSCC estimate, the BSCC
will use the ranked list of projects to fund additional projects.
34

PPIC.ORG 25

REFERENCES
Bartos, Bradley J. and Charis E. Kubrin. Forthcoming. “Can We Downsize Our Prisons
and Jails without Compromising Public Safety? Findings from California’s Prop 47.”
Criminology & Public Policy.
Bird, Mia, Sonya Tafoya, Ryken Grattet, and Viet Nguyen. 2016. How Has Proposition 47
Affected California’s Jail Population? Public Policy Institute of California.
BSCC (Board of State and Community Corrections). n.d. “Proposition 47 and the BSCC.”
BSCC, accessed April 9, 2018.
California Secretary of State. 2014. “Voter Information Guide, Proposition 47, Arguments
and Rebuttals.” California Secretary of State, accessed April 16, 2018.
Couzens, J. Richard, and Tricia A. Bigelow. 2017. “Proposition 47:
The Safe Neighborhoods and Schools Act.” California Courts.
Debbaudt, Marc. 2016. “An Explosion of California Property Crimes—Due to Prop 47.”
San Francisco Chronicle, March 18.
DOF (California Department of Finance). 2016. “Public Safety.” From California State
Budget—2016–17. California Department of Finance.
DOF (California Department of Finance). 2017. “Public Safety.” From May Revision—
2017–18. California Department of Finance.
DOF (California Department of Finance). 2018. “Public Safety.” From Governor’s
Budget Summary—2018–19. California Department of Finance.
Grattet, Ryken, Sonya Tafoya, Mia Bird, and Viet Nguyen. 2016. California’s County Jails
in the Era of Reform. Public Policy Institute of California.
Hunter, Sarah B., Lois M. Davis, Rosanna Smart, and Susan Turner. 2017. Impact of
Proposition 47 on Los Angeles County Operations and Budget. RAND.
LAO (Legislative Analyst’s Office). 2016. The 2016–17 Budget: Fiscal Impacts of
Proposition 47. LAO.
League of California Cities. 2017. “League Supports AB 1326 to Address Increase
in Theft Crimes in the Wake of Proposition 47.” League of California Cities.
Lofstrom, Magnus, Mia Bird, and Brandon Martin. 2016. California’s Historic
Corrections Reforms. Public Policy Institute of California.
Lofstrom, Magnus, and Brandon Martin. 2015. Public Safety Realignment: Impacts
So Far. Public Policy Institute of California.
Lofstrom, Magnus, and Steven Raphael. 2013. Public Safety Realignment and Crime Rates
in California. Public Policy Institute of California.
Males, Mike, and Lizzie Buchen. 2014. Proposition 47: Estimating Local Savings
and Jail Population Reductions. Center on Juvenile and Criminal Justice.
Moriarty, Tara. 2018. “San Francisco Averages 1 Car Break-in Every 17 Minutes.”
Fox 5 Atlanta, January 8.
OIG (Office of the Inspector General). 2015. Review of Crime Classification Practices.
OIG, Los Angeles Policy Commission.
Poston, Ben, Joel Rubin, and Anthony Pesce. 2015. “LAPD Underreported Serious
Assaults, Skewing Crime Stats for 8 Years.” Los Angeles Times, October 15.
Rubin, Joel, and Ben Poston. 2014. “LAPD Announces Reforms to Boost Accuracy
of City Crime Statistics.” Los Angeles Times, December 16.
Teji, Selena, and Scott Graves. 2014. Proposition 47: Should California Reduce Penalties
for Drug and Property Crimes and Invest in Treatment? California Budget Project.

26 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

A B O U T T H E AU T H O R S
Mia Bird is a research fellow in the areas of corrections and
health and human services at the Public Policy Institute of
California. She also serves on the faculty of the Goldman
School of Public Policy at the University of California,
Berkeley. At PPIC, she co-leads a collaborative project
between the California Board of State and Community
Corrections (BSCC), PPIC, and several partner counties,
known as the BSCC–PPIC Multi-County Study. This data
collection and evaluation effort is designed to estimate the
effects of realignment on recidivism outcomes and identify best practices for
recidivism reduction at the local level. She also leads a project focused on the
impact of the Affordable Care Act on enrollment and recidivism outcomes for
the criminal justice population. Her past work has covered topics such as the
allocation of realignment funding, healthcare for the correctional population,
and the use of data to improve policymaking. She holds a PhD in public policy,
an MA in demography, and an MPP from the University of California, Berkeley.
Magnus Lofstrom is a senior fellow at PPIC. His areas of
expertise include public safety, immi­gration, entrepreneurship, and education. His recent work examines crime trends
in California, public safety realignment and recidivism,
and California’s jail capacity and construction needs. He
also holds appointments as research fellow at the Institute
for the Study of Labor (IZA), Germany, community scholar
at the Julian Samora Research Institute at Michigan State
University, and research associate at the Center for Comparative Immigration Studies at the University of California, San Diego. He
serves on the editorial board of Industrial Relations and was a member of
California State Controller John Chiang’s Council of Economic Advisors.
Prior to joining PPIC, he was an assistant professor of economics at the University
of Texas at Dallas. He received his PhD in economics from the University of
California, San Diego.
Brandon Martin is a research associate at the Public Policy
Institute of California. He studies corrections and public
safety, with recent work examining the need for jail
construction in California and the changes in state and
local correctional populations. He has previous research
experience in the area of legislative committees, including
legislative and oversight responsibilities and the electoral
considerations of committee assignments. He holds an MA
in political science from the University of California, Davis,
and a BA in political science from Michigan State University.

PPIC.ORG 27

Steven Raphael is an adjunct fellow at the Public Policy
Institute of California and professor of public policy at the
University of California, Berkeley. His research focuses on
the economics of low-wage labor markets, housing, and
the economics of crime and corrections. He is the author
of The New Scarlet Letter? Negotiating the U.S. Labor
Market with a Criminal Record and Why Are So Many
Americans in Prison? (with Michael Stoll). He is the editor
in chief of Industrial Relations: A Journal of Economy and
Society and a research fellow at the University of Michigan National Poverty
Center, the University of Chicago Crime Lab, and the Institute for the Study of
Labor (IZA), Germany. He holds a PhD in economics from the University of
California, Berkeley.
Viet Nguyen is a research associate at the Public Policy
Institute of California. He conducts research on corrections
policy, including the effects of realignment. He is also the
data manager for a collaborative project between the
California Board of State and Community Corrections
(BSCC), PPIC, and several partner counties, known as
the BSCC–PPIC Multi-County Study. Before joining
PPIC, he was a survey specialist and operations analyst
at NORC at the University of Chicago. He holds a BA in
political science, with a minor in public policy, from the University of California,
Los Angeles.

28 THE IMPACT OF PROPOSITION 47 ON CRIME AND RECIDIVISM

AC K N OW LE D G M E N T S
This report benefited from valuable feedback from our external reviewers Chief
Terri McDonald and Susan Turner. We also appreciate the very helpful comments from our internal reviewers, Ryken Grattet and Shannon McConville, and
the editorial support from Vicki Hsieh and Lynette Ubois. This work would not
be possible without data provided by the California Department of Corrections
and Rehabilitation, the California Department of Justice, and the participating
counties in the BSCC–PPIC Multi-County Study. We acknowledge and thank
these data providers for their continued support of research to better inform
criminal justice policy and practice.

PPIC.ORG 29

B OA R D O F D I R E C TO R S

M A S M A S U M OTO, C H A I R

Author and Farmer

PHIL ISENBERG

Former Chair
Delta Stewardship Council

M A R K B A L DA S S A R E

President and CEO
Public Policy Institute of California

D O N N A LU C A S

Chief Executive Officer
Lucas Public Affairs

RUBEN BARRALES

President and CEO
GROW Elect
MARÍA BLANCO

Executive Director
University of California
Immigrant Legal Services Center

STEVEN A . MERKSAMER

Senior Partner
Nielsen, Merksamer, Parrinello,
Gross & Leoni, LLP
L E O N E . PA N E T TA

Chairman
The Panetta Institute for Public Policy

LO U I S E H E N RY B RYS O N

Chair Emerita, Board of Trustees
J. Paul Getty Trust

G E R A L D L . PA R S K Y

Chairman
Aurora Capital Group

A . MARISA CHUN

Partner
McDermott Will & Emery LLP

KIM POLESE

Chairman
ClearStreet, Inc.

CHET HEWITT

President and CEO
Sierra Health Foundation

G A D D I H . VA S Q U E Z

Senior Vice President, Government Affairs
Edison International
Southern California Edison

The Public Policy Institute of
California is dedicated to informing
and improving public policy in
California through independent,
objective, nonpartisan research.

Public Policy Institute of California
500 Washington Street, Suite 600
San Francisco, CA 94111
T 415.291.4400 F 415.291.4401
PPIC.ORG

PPIC Sacramento Center
Senator Office Building
1121 L Street, Suite 801
Sacramento, CA 95814
T 916.440.1120 F 916.440.1121