Description
3 articles/chapters means 9 questions, response or comments. (3) for each reading. Make it relevant to each reading.attached 3 articles in one file
Unformatted Attachment Preview
Eur J Crim Policy Res (2015) 21:425–446
DOI 10.1007/s10610-015-9274-x
Is There a Relationship Between Imprisonment and Crime
in Western Europe?
Marcelo F. Aebi 1 & Antonia Linde 1,2 & Natalia Delgrande 1
Published online: 10 May 2015
# Springer Science+Business Media Dordrecht 2015
Abstract This article examines the evolution of prison populations in Western Europe from
1982 to 2011 and its relation with recorded crime trends in the region. Data are taken mainly
from the Council of Europe Annual Penal Statistics in the case of prison statistics and the
European Sourcebook of Crime and Criminal Justice Statistics in the case of police and
conviction statistics, both complemented with the Nordic Criminal Statistics and Eurostat
Crime Statistics. The results show that prison populations rates (stock) rose constantly until
2005 and seem relatively stable since then. On the contrary, the annual flow of entries into
penal institutions has decreased almost continuously since 1987. This apparent paradox is
explained by the fact that the average length of detention has steadily increased during the
whole period under study. In brief, less people are sent to prison each year, but they remain in
prison for longer periods of time. The upward trend in the average length of detention is related
to the development of tough on crime policies across Western Europe and to the increase of
drug offences and non-lethal violent crime until the mid-2000s. In that context, an analysis by
offence shows similar trends in police, conviction, and prison statistics. These results falsify
the hypothesis of total independence between crime trends and imprisonment rates. They also
suggest that the deterrent effect of imprisonment has often been overestimated, and they cast a
shadow on the validity of criminological theories that place property as the main cause of
crime.
Keywords Average length of detention . Crime trends . Imprisonment rates . Flow . Stock .
Violent offences . Western Europe
* Marcelo F. Aebi
[email protected]
1
University of Lausanne, School of Criminal Sciences, ESC – Sorge – BCH, Lausanne CH-1015,
Switzerland
2
Universitat Oberta de Catalunya (Open University of Catalonia), Barcelona, Spain
426
M.F. Aebi et al.
Introduction
From a theoretical point of view, prison population rates are sometimes seen as the consequence of
crime rates, and sometimes as their cause. When prison population rates are seen as the consequence
of crime rates, the logical reasoning is that countries with high crime rates will have higher prison
population rates than countries with low crime rates. When prison population rates are seen as the
cause of crime rates, the typical reasoning is that an increase in prison population rates could reduce
crime rates. The reasons for that decrease would be (a) that the offenders neutralized can no longer
commit offences (incapacitation); (b) that potential offenders can be discouraged from engaging in
crime (general deterrence) but also, in the opposite sense, (c) that the release of an important number
of prisoners —like the one that took place in Italy in 2006, when an amnesty reduced by one-third
the prison population— can lead to an increase of crime. Indeed, a more rigorous methodological
approach suggests that the linear model of cause and effect is not appropriate to describe this
relationship. Circular causality (see, for example, Morin, 1977/1992) offers a better framework as,
from a longitudinal perspective, crime comes first and is followed by punishment, which in turn
could affect future crime, and so on.
From an empirical point of view, the relationship between crime rates and prison population
rates is still a matter of discussion. Modern criminal justice systems are based on the principles
of specific and general deterrence, trusting that the fear of incarceration will discourage
offenders or potential offenders from committing offences. Policy makers often invoke
incarceration as a panacea when hardening criminal laws. Criminologists, on the opposite,
are extremely cautious when discussing the association between imprisonment and crime. The
purpose of this article is to test whether there is a relationship between prison population rates
and crime trends in contemporary Western Europe. Subsidiary, if such relationship exists, we
will briefly discuss whether it has a causal nature and which is the causal order observed.
In the first section of the article we conduct a brief review of the literature on the
relationship between prison population rates and crime rates. Then we include a methodological chapter that presents the data that will be used in the study. After that, we analyze the
evolution of the stock of inmates and the flow of entries into penal institutions from 1982 to
2011 in Western Europe. In order to explain these trends, the following sections take into
consideration the average length of detention during that period, the distribution of the
sentenced prisoners by type of offence, and the evolution of crime according to police and
conviction statistics. Finally, we discuss our main findings and their implications for criminological research and theory.
Prior Research
The relationship between prison population rates and crime rates can be studied both from a crosssectional and from a longitudinal perspective. A cross-sectional approach implies comparisons
between these two rates in different countries, while a longitudinal approach concentrates on
trends in these rates in one or in several countries.1 The following overview of the contemporary
literature starts with studies that apply a longitudinal methodology and continues with those that
use a cross-sectional one, although there are some studies that combine both perspectives.
1
While in this article the units of analysis are countries, it is also possible to conduct similar studies using states
or cities as units.
Is there a Relationship Between Imprisonment and Crime
427
Among the longitudinal studies, those analyzing the case of the United States are the most
quoted in the recent criminological literature. The reason is that the increase in imprisonment
has been used as one of the explanations of the crime drop observed in that country since the
early 1990s. According to the economist Steven Levitt, the increase in incarceration in the
United States during the 1990s accounted for a reduction of crime of approximately 12 % for
homicide and violent crime, Band 8 % of property crime, or about one-third of the observed
decline in crime^ (Levitt 2004: 179). However, incarceration rates started increasing in the
United States since the early 1970s and violent crime followed an upward trend since the early
1960s to the early 1990s (Tonry 1999). This parallel increase of incarceration rates and violent
crime in the 1970s and the 1980s clearly contradicts the conclusions of Levitt (2004). The
latter is aware of that contradiction and, in a footnote, he considers that Bit is perhaps surprising
that the rising prison population of the 1980s did not induce a commensurate decline in crime
in that period^, and that B[a]mong adults, crime rates were in fact steadily falling throughout
the 1980s^, but that these declines Bwere masked by sharply rising youth crime in the 1980s^,
which appear to be due in part to the crack epidemic […], as well as to falling punishments in
the juvenile justice system over this same time period (Levitt 1998)^ (Levitt 2004: 179).
However, in his 1998 article, Levitt had stated that B[t]he rate at which juveniles were arrested
for violent crime rose 79 % between 1978 and 1993, almost three times the increase over that
time period for adults^ (Levitt 1998: 1156), and the only decrease that he mentions refers to the
murder arrest for adults, which had fallen by 7 %. Reviewing the literature on this topic —
which includes mainly the works of Levitt (2004), Spelman (2000), and Donohue and
Siegelman (1998)— Zimring (2007: 52) arrives at the conclusion that Bincarceration played
a rather modest role in the crime decline^. The reason for the confusion about that role is the
use of the term Belasticity^ —taken from the economic literature and related to the relative
effect of price changes and the quality of goods demanded— in a discussion of incapacitation
effects, because Bthere is no mechanism remotely analogous to supply or demand in the
mechanics of incapacitation^ (Zimring 2007: 53).
In Europe, the relationship between prison population rates and crime rates has been studied
through longitudinal analyses conducted mainly in Belgium, Italy, and France. Thus, analyzing
the evolution of prison population rates in Belgium during almost 170 years (1831–1993),
Vanneste (2001: 185) concludes that there is no relationship between the image of crime
provided by police statistics and the one provided by prison statistics; but that, in terms of
secular trends, the size of the prison population is best measured through the Bstrong signal^
provided by the harsher sentences. For example, from 1843 to 1875, for one life sentence
imposed, the size of the prison population grew by 38 inmates, and for one sentence to forced
labor (ranging from 5 to 20 years), it grew by 31 inmates (Vanneste 2001: 87). In the second
half of the twentieth century, Vanneste (2001: 158) also found a strong long-term correlation
(r=.85) between the sentences to forced labor following a crime against persons and the prison
population size. Melossi (2001) has analyzed long-term trends in Italy —from 1863 to the
1990s— and found a correlation between the murder rate and the prison population rate,
especially for the period 1947 to 1994. He suggests as an explanation that the increase of
murders —often related in Italy to those committed by the Mafia— creates a climate of
generalized moral panic, fuelled by the mass-media, which leads the authorities of the criminal
justice system to increase repression and therefore produces a rise of the prison population
(Melossi 2001). Studying the situation in France from 1974 to 2005, Kensey (2007) considers
that the increase in the length of the sentences imposed is responsible for the upward trend in
the French prison population during that period. Her analysis of the distribution of the
428
M.F. Aebi et al.
sentenced prison population is revealing: prisoners sentenced for simple theft represented
48.8 % of the sentenced prison population in 1975 but only 7.5 % in 2005; while, comparing
the same years, those sentenced for assault increased from 6.1 to 17.6 %, and those sentenced
for rape and sexual assault on minors and adults increased from 4.9 to 21.9 %. At the same
time, homicide and qualified theft —which includes robbery— show stable percentages of
roughly 9 % each (Kensey 2007: 94). French police statistics also reflect an increase in rape,
sexual assault, and assault from 1974 to 2004, while homicide increased from the 1970s to the
mid-1980s, reached a peak in the early 1990s, and decreased until the end of the series in 2004
(Kensey 2007: 104–107).
Researchers that apply a cross-sectional methodology usually compare trends in crime rates and
prison population rates in different countries, and they arrive to the conclusion that there is no
relationship between them. For example, Newburn (2007: 15) points out that, with the noteworthy
exception of homicide, Bthe main crime rates in the big cities in the United States and the United
Kingdom are on a similar scale^, but that B[even] though Britain is the highest incarcerator in
Western Europe, its incarceration rate is only one-fifth (or less) of that in America^. As a
consequence, he arrives to the conclusion that Bthere is no direct link between crime rates and types
and levels of punishment^ (Newburn 2007: 15). The incredibly high incarceration rate of the United
States compared to the ones of other Western democracies, led Tonry (1999) to apply, in a somehow
ironic way, the concept of BAmerican exceptionalism^ which can be traced back to Alexis de
Tocqueville. Tonry considers that a crudely empirical explanation, such as BAmerican crime rates
are higher or have increased more than other countries’, and punishment patterns and policies are no
more than a reflection of that reality […] has virtually no validity^ (Tonry 1999: 420–1). The reason
is that the International Crime Victim Survey (ICVS) has shown that B[c]rime rates in the United
States in the 1990s are, for the most part, no higher than in other countries^ (Tonry 1999: 421).
Referring to the work of Zimring and Hawkins (1997), Tonry concludes that B[w]here the United
States stands out is in gun violence; our [U.S.] rates of robberies and assaults involving guns, and of
gun homicides, are substantially higher than elsewhere^ (Tonry 1999: 421).
At the same time, other international cross-sectional analyses have found that the European
prison population rates were positively correlated with the rates of unsuspended custodial
sanctions imposed for homicide and with the rates of homicide according to police statistics
(Aebi and Kuhn 2000), as well as with the rates of homicide according to the World Health
Organization mortality data (Lappi-Seppälä 2011).2 However, even when a correlation between homicide rates and imprisonment rates is found, researchers are extremely cautious as to
the interpretation of their findings, which go against the mainstream position.3
The general impression that the reader gets after going through the scientific literature on
the relationship between imprisonment and crime is that criminologists are extremely sceptical
about it. The reader can also perceive a nuance between the position of criminologists working
with data from the United States —who, overall, seem willing to accept a weak and indirect
relationship— and criminologists working with European data who, in general, tend to deny
2
Ouimet (2012) also found a positive correlation between incarceration rates and homicide rates for the year
2010 and using a sample of 160 countries.
For example, Aebi and Kuhn (2000: 73) Bhave serious doubts about the veracity^ of their findings; while,
according to Lappi-Seppäla (2011: 308) Bwe cannot rule out the possibility that incarceration rates are partly
influenced by differences in crime, especially in the East European and Baltic countries. This applies especially to
homicide. […] However, it is equally possible that high incarceration and homicide rates are both a product of a
third factor. […] This hypothesis deserves further examination in the future^.
3
Is there a Relationship Between Imprisonment and Crime
429
any relationship in theoretical papers or, when confronted to empirical evidence of it, to treat it
as an exception.
In practice, most contemporary researchers prefer relatively sophisticated models to explain
imprisonment rates across nations, such as the ones developed by Cavadino and Dignan (2006a,
b), Solivetti (2010), or Lappi-Seppälä (2011), which include several independent variables in
the equation. The purpose of this article is much more modest as it concentrates only on the role
of crime trends. In that context, the available research suggest that, in Europe, the hypothesis of
a lack of relationship between crime rates and prison population rates seems valid only as far as
cross-national comparisons of the levels of the total prison population and the total volume of
crime are concerned. Indeed, in spite of the reluctance of most researchers, from a crosssectional point of view, European prison population rates seem partially influenced by crossnational differences in homicide (i.e., in general, countries with the highest homicide rates also
present the highest prison population rates) and, from a longitudinal point of view, there is
evidence from a few countries suggesting that trends in homicide and serious offences have
been historically correlated with trends in prison population rates. Hence, the available literature
seems to support the advise given by Garland (2013) in the sense that Bwhen we compare rates
of imprisonment across jurisdictions, or across time, any inferences we draw about repressiveness or punitiveness should be modified by consideration of the patterns, trends, and rates of
crime to which these penal measures are a response^ because Bthere is generally some
relationship, however mediated and indirect^ (Garland 2013: 487, emphasis in the original).4
This study tests whether the latter affirmation is corroborated in a group of Western European
countries, adopting a longitudinal perspective that covers more than two decades and takes into
consideration not only trends in homicide, but also in five other offences.
Data and Methods
The data on prisons and prisoners used in this article are taken mainly from the Council of Europe
Annual Penal Statistics, which were started in 1983 and are better known by their acronym SPACE5
(Aebi and Delgrande 2012).6 The exceptions concern data on prison populations and entries into
penal institutions taken from the Nordic Criminal Statistics 1950–2010 (von Hofer et al. 2012) for
Finland, Norway, and Sweden, and from the Direction of the Prison Services of the Ministry of
Justice for France (Kensey 2007). Similarly, data on the distribution of sentenced prisoners in
England and Wales are taken from the Home Office and Ministry of Justice series on Prisons and
Probation Statistics and Offender Management Caseload Statistics.7
Data on persons convicted for criminal offences come from the five available editions of the
European Sourcebook of Crime and Criminal Justice Statistics (CoE 1999; Killias et al. 2003; Aebi
et al. 2006, 2010, 2014). The same is true for offences recorded by the police, with the exceptions of
Finland, Norway, and Sweden —whose data come from the Nordic Criminal Statistics 1950–2010
4
In a similar perspective, one of the main critics that Nelken (2009, 2010) addresses to the work of Cavadino and
Dignan (2006b) is that they, Blike most of those comparing a large range of incarceration rates, spend little time
on persuading us that crime rates are really the same in all the countries they are comparing^ (Nelken 2010: 61).
5
The acronym SPACE derives from the French title of this series: Statistiques Pénales Annuelles du Conseil de
l’Europe.
6
The annual SPACE surveys since 2000 are available at www.unil.ch/space (last accessed on 14 December
2014). The previous surveys are available only in paper format.
7
Available at www.gov.uk/government/organisations/ministry-of-justice/about/statistics. Last accessed on 14
December 2014.
430
M.F. Aebi et al.
(von Hofer et al. 2012)— as well as Spain, whose data from 1994 to 2011 come from the database of
Eurostat on crime and criminal justice.8 Police and conviction data are used with the adjustments
described in previous articles that have used the same sources (see Aebi and Linde 2010, 2012a, b).
Offences recorded by the police refer to the criminal offences subject to criminal proceedings
registered by the police forces of a country. Correspondingly, persons convicted refer to persons
found guilty, according to the law, of having committed an offence. This includes court convictions
and sanctions imposed by the prosecutor —or by a court that ratifies a decision of the prosecutor
without a formal court hearing— that lead to a formal verdict (for details and exceptions see Aebi
et al. 2014).
The other indicators used in this research are the stock of inmates (available from 1983 to
2011), the flow of entries into penal institutions (available from 1982 to 2010), the average
length of detention, and the distribution by offence of the prisoners serving a final sentence
(available from 1994 to 2011). The stock of inmates (hereafter referred to as the stock) refers to
the total number of prisoners, including pre-trial detainees, held in prison on a given date,
which in the case of SPACE is the 1st September of each year. When calculated on the basis of
100,000 population, this indicator is indistinctly referred to as the imprisonment rate, the
incarceration rate, the detention rate, and the prison population rate (Kuhn et al. 2000: 16). In
this article, all these terms are used as synonyms. The flow of entries into penal institutions
(hereafter referred to as the flow) corresponds to the total number of entries of persons in
detention facilities —including those for pre-trial detainees— during a whole year. As Tournier
(2004: 2) has shown, it is possible in this context to apply to the study of prison populations the
demographic model of the stationary population. In this model, the stock (S) is the product of
the flow (F) multiplied by the length (L), i.e., S = F x L. This formula can then be transposed to
calculate an indicator of the average length of detention (L) expressed in months by computing
the quotient of the average number of prisoners in 1 year (S) by the flow of entries during the
same period (F), and multiplying the result by 12 in order to express it in months.9
L ¼ 12 S F
This formula produces similar results when the stock refers to a given date of a year X and
the flow to the previous year (X-1), and it is used in this article to calculate the average length
of detention in the countries studied. In the case of stock, flow, sentenced prisoners by type of
offence, police recorded offences and persons convicted for criminal offences, the rates per
100,000 population in each country are used to compute geometric means for the group of
countries under study (Figs. 1, 2, 4 and 5).10 The average length of detention is also calculated
8
Available at http://ec.europa.eu/eurostat/data/database. Last accessed on 14 December 2014.
Imagine a country in which ten persons enter into prison every year, of which two persons are sentenced to
5 years, two to 4 years, two to 3 years, two to 2 years, and two to 6 months. If one places this figures on a
spreadsheet, it can be seen that, from the 5th year onward, the prison population of that country would be stable at
29 prisoners yearly. Thus, the stock of that hypothetical country is 29 and the flow 10, while the average length of
the sentences is 2.9 years (5+5+4+4+3+3+2+2+0.5+0.5=29÷10=2.9). This means that knowing two of these
numbers is it always possible to calculate the third one through the above formula. For example, with a stock of
29 and a flow of 10, the average length can be calculated by dividing the stock by the flow (29÷10) as 2.9 years
or 34.8 months (2.9*12).
10
BGeometric means are often more meaningful than arithmetic means, because they are closer to the central
figure (median). […] To calculate the [geometric] mean of n numbers, […] multiply them, then take the nth root^
(Taagepera 2008: 120). According to Dodge (1993: 248–9), the geometric mean is used in particular to calculate
the average of ratios and reduces the influence of extreme values (outliers). Thus, it seems particularly
appropriate for the data analyzed in this article, which include rates per 100,000 population and some outliers.
In the field of crime trends, the geometric mean have been used namely by Eisner (2003) and Pinker (2011: 64).
9
Is there a Relationship Between Imprisonment and Crime
431
Prison populaon rates per 100,000 inhabitants
110
100
90
80
70
60
17 countries
14 countries
2009
2010
2008
2006
2007
2005
2004
2003
2002
2000
2001
1998
1999
1996
1997
1995
1994
1993
1992
1990
1991
1989
1988
1986
1987
1985
1984
1983
50
8 countries
Fig. 1 Trends in the stock of inmates per 100,000 population in Western Europe from 1983 to 2011
as the geometric mean of the average lengths of detention of the countries included in the
analysis (Fig. 3). Finally, the percentage change between the first and the last year of the series
—based on a direct comparison of the rates for both years— and the average annual percent
change, also known as average annual variation, annual growth rate, and average annual
change rate (Harrendorf et al. 2010: 147) are used to measure the evolution of rates during the
time frame of the analyses.
Rates of entries into penal ins tu ons per 100,000 inhabitants
240
220
200
180
160
14 countries
2010
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
1989
1988
1987
1986
1985
1984
1983
1982
140
8 countries
Fig. 2 Trends in the flow of entries into penal institutions per 100,000 population in Western Europe from 1982 to 2010
432
M.F. Aebi et al.
9
Average length of deten on (in months)
8
7
6
5
4
14 countries
2009
2010
2008
2006
2007
2005
2004
2003
2002
2000
2001
1999
1998
1996
1997
1995
1994
1993
1992
1990
1991
1989
1988
1986
1987
1985
1984
1983
3
8 countries
Fig. 3 Trends in the average length of detention in Western Europe from 1983 to 2010
Countries with a population of less than one million inhabitants are not included in our study
because their presence can affect the reliability of the statistical analyses. For example, they may
show substantial changes in their rates from 1 year to another that are only due to the addition or the
subtraction of a few inmates, offences, or persons convicted. Germany cannot be included because,
until the late 1990s, data are available only for the Federal Republic of Germany, usually known as
Western Germany. Finally, as some countries had been unable to provide data for every year covered
by our data sources, linear interpolation was used to calculate the missing values. When the first
years of the series were not available, their values were extrapolated by repeating the value available
for the first available year. When the last year of the series was not available, its value was
extrapolated by repeating the previous available year.11 In spite of these adjustments, some countries
cannot be included in all the analyses because they did not provide data on all the indicators required
(e.g., in the case of Austria, data are available for the stock, but not for the flow). As a consequence,
and in order to maximize the available data, three data sets are analyzed: (a) stock, (b) flow and
average length, and (c) distribution of the sentenced prisoners by type of offence.
Stock data for the period 1983 to 2011 are available for the following 17 countries: Austria,
Belgium, Denmark, England and Wales, Finland, France, Greece, Ireland, Italy, Netherlands,
Northern Ireland, Norway, Portugal, Scotland, Spain, Sweden, and Switzerland. Flow data for
the period 1982–2010 are available for 14 countries, because Austria, Greece, and Switzerland
did not provide such type of data. As both stock and flow data are needed to calculate the
average length of detention, the latter is available for the same 14 countries.12 Data on the
11
Linear interpolation and extrapolation are the standard procedures for the replacement of missing data, which
are used, for example, by the World Health Organization for the calculation of regional averages of homicide
according to health statistics (WHO 2014).
12
Belgium, Denmark, England and Wales, Finland, France, Ireland, Italy, Netherlands, Northern Ireland,
Norway, Portugal, Scotland, Spain, and Sweden.
Is there a Relationship Between Imprisonment and Crime
433
distribution of sentenced prisoners by type of offence from 1994 to 2011 are available for eight
countries because Belgium, Denmark, Ireland, Italy, Netherlands, and Northern Ireland did not
provide reliable data. In order to improve comparability, the same eight countries are included
in the analyses of offences recorded by the police.13 In the case of persons convicted for
criminal offences, it was impossible to obtain reliable series for Norway and Spain, and
therefore only six countries are included. Whenever possible, the three main datasets are
included in the Figures. For example, Fig. 1 shows the trends in stock for the sample of 17
countries and the subsamples of 14 and eight countries. This kind of presentation of the results
allows establishing whether the subsets show convergent or divergent trends.
Findings
This section presents the results of our analyses. It starts with a general overview of the trends in the
stock and the flow of entries into penal institutions, which are then combined to establish the average
length of detention. This is followed by an analysis of the distribution of sentenced prisoners by
offence. Finally, we present crime trends according to police and conviction statistics.
Trends in the Stock of Inmates
Figure 1 shows the evolution of Western European prison population rates (number of inmates per
100,000 population) since 1983. As mentioned before, in order to compare rates according to different
indicators, the figure presents three clusters of countries. These clusters include the 17 countries that
provided stock data since 1983, a subset of 14 of them that will be used for the comparisons with flow
data, and a subset of eight of these countries that will be used for the comparisons with the distribution
of the sentenced prisoners by offence as well as with offences recorded by the police.
Figure 1 shows an almost constant linear increase in prison populations, interrupted by short
periods of relative stability during the second half of the 1980s, the second half of the 1990s,
and, particularly, in the second half of the 2000s, when prison populations started to stabilize
and even decrease in some countries. For the clusters of 17 and 14 countries the annual average
increase is 1.5 % and for the cluster of eight countries it is 1.6 %. In particular, the 2011 rates
were respectively, 51, 54, and 59 % higher than the ones of 1983. The analogy in the evolution
of these rates for the three clusters of countries suggests that the inclusion or exclusion of a few
countries does not affect the overall Western European trend. This similarity in the evolution of
crime measures at the European level has also been observed in the case of police and
conviction statistics (Aebi and Linde 2010, 2012a). Indeed, a country-by-country analysis
(not presented here) shows that only Austria, Finland, and Northern Ireland show lower prison
population rates in 2011 than in 1983. At the same time, the average annual variation of the
stock from 1983 to 2011 is negative only in the cases of Finland and Northern Ireland.
Trends in the Rates of Entries into Penal Institutions
Figure 2 shows the evolution of the rate of entries (flow) into penal institutions per
100,000 population. Once more, in order to allow comparisons, the figure includes the
13
This means that data on the distribution of the sentenced prisoners by offence as well as offences recorded by
the police are available for England and Wales, Finland, France, Norway, Portugal, Scotland, Spain, and Sweden.
434
M.F. Aebi et al.
14 countries that provided flow data since 1982 and the eight countries for which
information on the distribution of the sentenced prisoners by offence and police
recorded offences are available.
The trends shown in Fig. 2 are almost the opposite to the ones of Fig. 1. With the exception
of an increase in the first half of the 1980s and a period of relative stability in the first half of
the 2000s, the flow of entries into penal institutions has been decreasing almost constantly for
more than 20 years. All in all, in the cluster of 14 countries the 2010 rate is 20 % lower than the
one of 1982 and the average annual decrease is 0.9 %; while in the cluster of eight countries
the rate is 29 % lower in 2010 than in 1982 and the annual decrease is 1.6 %. Once more, both
clusters show a similar evolution (Rho=.94; p≤.001); even if the decrease is more pronounced
in the smaller one. A country-by-country analysis shows that only Ireland, the Netherlands,
and Northern Ireland show a higher flow in 2010 than in 1982, and that these countries and
Scotland also show a positive annual average variation of their flow from 1982 to 2010.
A comparison of the trends in stock (Fig. 1) and flow (Fig. 2) for the years 1983 to 2010
shows negative and significant Spearman’s rank correlations for the cluster of 14 countries
(Rho=−.82; p≤.001) as well as for the cluster of eight countries (Rho=−.90; p≤.001).
Trends in the Average Length of Detention
Figure 3 shows the evolution of the average length of detention for the 14 countries that
provided stock and flow data, and for the eight countries that provided data on the distribution
of the sentenced prisoners by offence and police recorded offences.
The trend in Fig. 3 is similar to the one in Fig. 1 and consists in an almost constant linear increase
in the average length of detention of all persons held in custody. The average annual increase is
2.4 % for the cluster of 14 countries and 3.3 % for t
