Economics Letters 152 (2017) 4649

Contents lists available at ScienceDirect

Economics Letters
journal homepage: www.elsevier.com/locate/ecolet

Prisons, recidivism and the agecrime profile

Anna Bindler a , Randi Hjalmarsson a,b,
a
University of Gothenburg, Sweden
b
CEPR, United Kingdom

highlights
Historical data set of over 120,000 cases at the Old Bailey in 19th century London.
U-shape trend in convict age: steep decrease until 1820, then steady increase.
Use of prisons (instead of death or transportation) led to concept of recidivism.
Propose that the rise of prisons and recidivism mechanically led to older convicts.
First historical data on recidivism.

article info abstract

Article history: This paper provides a descriptive analysis of the economic and non-economic channels that led to a
Received 13 October 2016 U-shaped trend in the average age of male convicts in 19th century London using detailed data from
Accepted 1 January 2017 the Old Bailey central criminal court. In addition to discussing industrialization and changing attitudes
Available online 3 January 2017
towards juveniles as potential mechanisms underlying the initial decrease and subsequent increase in
criminal age, we put forth a new explanation of the latter. Did the abolition of capital punishment and
JEL classification:
penal transportation, which led to the rise of the modern day prison system and the emergence of
J10
K14
recidivism, lead to a mechanical increase in the average age of criminals?
K40 2017 The Authors. Published by Elsevier B.V.
N33 This is an open access article under the CC BY license
N43 (http://creativecommons.org/licenses/by/4.0/).

Keywords:
Agecrime profile
Industrial revolution
Prison
Punishment
Recidivism

1. Introduction
One of the most commonly cited stylized facts about the de-
mographics of criminals throughout the world and history is the
agecrime profile, which typically increases in the teenage years,
peaks around age 19 or 20, and then gradually decreases. Fig. 1
shows the profile for both arrestees in the United States in 1980
This paper would not have been possible without the tremendous efforts of our
Research Assistant Michael Bekele, the generous help with the data extraction by
Florin Maican, and the financial support of Foundation for Economic Research in
West Sweden (no. 2250-242 334), and Vetenskapsrdet (no. 446-2014-1735), The
Swedish Research Council, Grants for Distinguished Young Researchers.
Correspondence to: Department of Economics, University of Gothenburg,
Vasagatan 1, 41124 Gteborg, Sweden.
E-mail address: randi.hjalmarsson@economics.gu.se (R. Hjalmarsson).
http://dx.doi.org/10.1016/j.econlet.2017.01.002
0165-1765/ 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
and almost 120,000 convicts at the Old Bailey in London between
1800 and 1900: despite the different continents and more than
100 years in between, the agecrime profiles are remarkably sim-
ilar.
Though criminologists have extensively studied the agecrime
relationship, economists have only recently contributed to the
debate. Grogger (1998) asks whether market wages explain the
agecrime profile, a natural question in the context of the Becker
(1968) model of crimethe opportunity cost of crime increases
with labor market experience as individuals age. Lochner (2004)
develops a human capital based model of crime, which predicts an
agecrime profile that peaks at or before the age of labor market
entry. More recently, Landerso et al. (forthcoming) study how
school starting age affects the agecrime profile.
While the general shape of the age distribution is constant, its
moments (mean and dispersion) can shift over time. One such
example is the rightward-shift in the age distribution of convicts
 A. Bindler, R. Hjalmarsson / Economics Letters 152 (2017) 4649
Fig. 1. Offender-age profiles. Note: share of convicts at the Old Bailey (18001900)
and share of arrests in the US (1980) by 5-year age intervals.
Source: Old Bailey Online Proceedings; BJS1 ; own calculations.
Fig. 2. Convict-age profiles. Note: age distributions of male convicts at the Old
Bailey.
Source: Old Bailey Online Proceedings; BJS (see footnote 1); own calculations.
at the Old Bailey in the latter half of the 19th century (see Fig. 2).
Similar shifts have been studied in US prisons since the 1970s
(Porter et al., 2016). This paper focuses on understanding the
potential economic and non-economic channels that underlie the
shift in the agecrime profile in 19th century London, with an
emphasis on the Industrial Revolution, changing attitudes, and the
rise of imprisonment as the primary sanction.
2. The age of convicted offenders at the Old Bailey
As the central criminal court for the City of London and sur-
rounding Middlesex, the Old Bailey tried the most serious crimes
in Victorian times. Detailed information on the defendant, charge,
verdict, and sentence of each case were recorded in The Proceedings
of the Old Bailey. The Old Bailey Proceedings Online has digitized this
historical data source and provided xml files for each of the 2000
court sessions. See Bindler and Hjalmarsson (2016) for more details
on the data and context.
The primary variable of interest is defendant age. One data
limitation is that the Proceedings only consistently reported age
1 Snyder, H. and Mulako-Wangota, J., Arrest Data Analysis Tool (29-Sep-16) at
www.bjs.gov. Bureau of Justice Statistics, Washington, D.C.
47
Fig. 3. Average age and share of juvenile convicts. Note: average age and share of
juveniles of male convicts at the Old Bailey (property crime).
Source: Old Bailey Proceedings Online; own calculations.
Fig. 4. Criminal history and prison sentences. Note: share of male convicts for
property crime at the Old Bailey with any criminal history or prison sentences.
Source: See Fig. 3.
for those found guilty; we cannot say anything about the age of
those acquitted (25% of the sample from 1800 to 1900). Prior to
1789, age is missing for 99% of those convicted (in addition to those
acquitted).2
Fig. 3 shows the average age of the 69,537 male offenders con-
victed of property crime, as well as the share of juvenile (below age
18) convicts.3 We focus on male property offenders (i) because at
that time property offenses and male defendants comprise 71% and
83%, respectively, of all convicted cases and (ii) to ease comparison
with Vickers and Ziebarths (2016) analysis of age at the Old Bai-
ley from 1835 to 1913.4 The average age is U-shaped between 1800
and 1900, with a steep (17%) decrease from 29 to 24 years between
1800 and 1820, and a less steep but steady increase thereafter. The
share of juvenile offenders mirrors that pattern with an increase
from 11% to 27%, respectively, and a steady decrease afterwards.
What explains these relatively large and quick changes in convict
age?
2 See: https://www.oldbaileyonline.org/static/Population-history-of-london.
jsp#demography.
3 Property crimes include animal theft, arson, burglary, housebreaking, larceny,
mail, receiving, shoplifting, stealing from master, and theft from place.
4 It is our understanding that Vickers and Ziebarth (2016) only had post-1835
data available for their analysis.
48 A. Bindler, R. Hjalmarsson / Economics Letters 152 (2017) 4649
Fig. 5. Average age by selected offense. Note: average age of male convicts for larceny, pocket picking, housebreaking and burglary at the Old Bailey. (For interpretation of
the references to color in this figure legend, the reader is referred to the web version of this article.)
Source: See Fig. 3.
3. The industrial revolution
Researchers offer conflicting arguments regarding the impact
of the Industrial Revolution on the age composition of criminals in
19th century London. On the one hand, Vickers and Ziebarth (2016)
argue that the Industrial Revolution led to younger generations
gaining the specific skills necessary in the new labor market, and
displaced older workers (especially males) into the criminal labor
market, thereby increasing the average age of criminals. Using the
Old Bailey data from 1835 to 1913, the authors demonstrate an
increase in the average age of male property offenders.
However, Fig. 3 casts some doubt on the Vickers and Ziebarth
(2016) theory: the increase in average criminal age is only seen af-
ter 1820, despite the fact that the Industrial Revolution was fully
under way in the early 1800s.5 That is, their graph starts at the bot-
tom of the U-shaped curve in Fig. 3. In addition, if industrialization
displaced older workers into crime, it is hard to reconcile this with
the persistent increase in the average age of criminals over a time
period of 80 years. As Mokyr (1990) points out, many workers dis-
placed by machinery eventually found employment in the facto-
ries.
On the other hand, historians have linked industrialization
with the decreasing average criminal age in the early 1800s. King
(2000) reviews these arguments, which are typically predicated
5 According to Mokyr (1990), the ndustrial Revolution is usually dated between
about 1760 and 1830.
on the assumption of a rapidly growing and predominantly young
population, a flood of young migrants turning to crime instead
of adapting, or the process of industrialization and the growth
of factories itself, leading to a rise in juvenile crime. However,
King (2000) points out that the capital [of London] witnessed
no fundamental changes. Its service sector remained large and
its manufacturing sector was notable for the absence of major
restructuring . . . .
While we cannot rule out that the Industrial Revolution
contributed to either the fall or rise in average criminal age, a
more nuanced understanding is clearly necessary to reconcile the
patterns observed in Fig. 3. This paper emphasizes two alternative
channels through which the age composition of convicts may have
been affected.
4. Changing attitudes towards juveniles
One possibility is that it is not criminal behavior that is chang-
ing, but rather the behavior of the magistrates and courts. First,
King (2000) suggests that the observed increase in juvenile crime
can be attributed to victims and magistrates becoming more will-
ing to hold juveniles responsible, especially once capital punish-
ment was abolished. Second, one could argue that the subsequent
increase in average age is an explicit reaction to the increasing
number of juvenile offenders in the early 1800s. According to
King (2000), there was an increased use of summary judgments
decisions by magistrates rather than juries. However, as demon-
strated by Vickers and Ziebarth (2016), the average offender age
 A. Bindler, R. Hjalmarsson / Economics Letters 152 (2017) 4649
is increasing in the latter 1800s even when excluding defendants
under 18.
5. Changing sentences: the rise of prison
We offer an additional explanation for the increasing age
of convicts in the latter half of the 19th century: the rise of
the modern-day prison and the emergence of recidivism, which
mechanically led to an increase in offender age.
In 1800 England, capital punishment and transportation to Aus-
tralian penal colonies were the primary forms of punishment; in
fact, there were more than 200 capital offenses at this time. Capi-
tal punishment was abolished by a series of offense-specific Acts in
the early and mid-1800s. This led to an increase in transportation
to Australia, characterized by a 46 month voyage when many be-
came ill or died, hard labor upon arrival, and harsh discipline. Even-
tually, the Penal Servitude Acts of 1853 and 1857 (primarily the
former) abolished transportation. Without capital punishment or
transportation, England turned towards imprisonment as the pri-
mary sanction.6
For the sample of males convicted of a property crime, Fig. 4
shows the share with a known criminal history and the share
sentenced to prison. We manually coded criminal history, which
is recorded (but not tagged) in the Proceedings from the 1830s
onwards, and indicates whether the defendant had been in custody
once before (from 1832), more than once (from 1839) or whether
they were known associates of bad character (from 1835). Albeit
not perfect,7 it is perhaps the first measure of recidivism.
The share of convicted males with a known criminal history is
clearly trending up throughout the 19th century. Likewise, there is
a continual increase in the share sentenced to prisonfrom around
30% in 1832 to more than 80% by the end of the century. Note
that the share of prison sentences increases very sharply upon the
abolition of transportation.
Finally, how did these changes in sentence regimes affect the
age of criminals? Fig. 5 shows the average age of male convicts for
larceny, pick-pocketing, housebreaking and burglary.8 Note that
pick-pocketing is included in the larceny measure, but seems to
be especially important for the increase in juvenile crime at the
beginning of the 18th century. For both larceny and pick-pocketing,
6 See Bindler and Hjalmarsson (2016) for more details.
7 For instance, it is hard to know how offenses committed in other jurisdictions
are tracked.
8 Available for other offenses upon request.
49
the figures show the U-shaped pattern: the average convict age is
sharply falling until 1820, remains relatively low until the 1850s,
and then increases to pre-levels by the end of the century. In fact,
the start of that increase appears to coincide with the abolition of
transportation (red vertical line), supporting our hypothesis above.
Housebreaking and burglary capital eligible until 1833 and
1837, respectively, as marked by the first vertical line follow
similar patterns. The increase in average convict age appears to
be driven by the abolition of transportation (second vertical line)
and not capital punishment, which is consistent with the use of
transportation as a substitute sentence for the death penalty upon
its abolition. That is, an aging recidivist population is not observed
until the offenders were kept alive in the United Kingdom.
6. Conclusion
Understanding the sources of shifts in the age-distribution of
crime is essential from a policy perspective. Which subset of the
population should criminal justice policies target? Though we
discuss the role that penal institutions may play in shaping the
agecrime profile in a historical context, this is clearly a question
of modern day relevance.
References
Becker, Gary, 1968. Crime and punishment: An economic approach. J. Polit. Econ.
76 (2), 169217.
Bindler, Anna, Hjalmarsson, Randi, 2016. The fall of capital punishment and the
rise of prisons. How punishment severity affects jury verdicts. University of
Gothenburg Working Papers in Economics 674.
Grogger, Jeff, 1998. Market wages and youth crime. J. Labor Econ. 16 (4), 756791.
King, Peter, 2000. Crime and Law in England, 1750-1840. Cambridge University
Press, Cambridge, UK, p. 348.
Landerso, Rasmus, Nielsen, Helena Skyt, Simonsen, Marianne, School starting age
and the crime-age profile. Econom. J. (forthcoming).
Lochner, Lance, 2004. Education, work and crime: A human capital approach.
Internat. Econom. Rev. 45 (3), 811843.
Mokyr, Joel, 1990. The Lever of Riches: Technological Creativity and Economic
Progress. Oxford University Press, p. 349.
Porter, Lauren, Bushway, Shawn, Tsao, Hui-Shien, Smith, Herbert, 2016. How the
US prison boom has changed the age distribution of the prison population.
Criminology 54 (1), 3055.
Vickers, Chris, Ziebarth, Nicolas L., 2016. Economic development and the
demographics of criminals in Victorian England. J. Law Econ. 59, 191223.