Since 2006, the FSO has made available reconviction rates by year of conviction, for both adults and juveniles.
For these figures, we take into consideration the number of persons convicted during a given year and establish a recidivism rate by putting this quota in relation to the number of individuals having committed a new offence in the 3 years following their conviction, which leads to their reconviction.
To take the offender's criminal history into account, the three years prior to the reference conviction are analysed in order to determine whether the person has a previous conviction; i.e. whether they have a criminal history.
Yearly recidivism rate by reference event
Recidivism after a conviction (not involving custodial sentence of longer than six months)
Recidivism after release (involving custodial sentence of longer than six months)
The FSO also calculates a reconviction and reimprisonment rate by year of release. The term re-offender applies to Swiss adults who commit a felony or misdemeanour within three years of release from a custodial sentence and are subsequently reconvicted. The term re-offender applies to Swiss adults who commit a felony or misdemeanour within three years of release from a custodial sentence and are subsequently reconvicted and reimprisoned.
Long-term recidivism rate of persons born the same years
Reconvictions of minors in adulthood from 1999 to 2015: risk factors
This third analysis continues and effectively completes the 2017 study:
- by presenting more detailed statistical tests that enable the influence of each of the risk factors to be estimated when jointly considered;
- by including foreign nationals with a C permit in the persons monitored;
- by considering the speed at which a juvenile delinquent is referred to the juvenile court and the influence of this variable on recidivism in adulthood;
- by defining more precisely the nature of offences committed during childhood and adolescence.
Furthermore, it also expands on the initial research question. In the 2017 publication, the research question was as follows: “If a minor is convicted at least once, what is the probability that they will be reconvicted in adulthood and what are the factors that have an impact on the risk of recidivism in adulthood?” This time, it was decided that in addition to considering this issue in more depth, an extra research question would also be analysed: “According to the trajectory during childhood and adolescence (whether or not a juvenile delinquent), what is the probability that a person may be convicted in adulthood and what are the other factors that influence the risk of being convicted in adulthood?”
Two datasets were created to answer these two questions. It should be noted that the new question has been considered first.
The first question the FSO considered was “What are the factors that have an impact on the risk of being convicted in adulthood?” It based its analyses on a dataset consisting of 95 695 people (including 7 428 judged by a juvenile court).
The logistic regression model showed the following results:
Variables in the equation | Odds ratio | Confidence interval at 95% | p-value | |
---|---|---|---|---|
Sex | 5.41 | 5.04 | 5.808 | <0.0001 |
Judgement as a minor | 4.83 | 4.541 | 5.138 | <0.0001 |
Nationality | 1.206 | 1.117 | 1.303 | <0.0001 |
Constant | <0.0001 |
The three variables considered have a significant effect (at a threshold of 0.05) on the likelihood of being convicted in adulthood.
- Gender is the most predictive variable: boys are 5.4 times more likely to be convicted in adulthood than girls.
- A conviction as a minor increases the risk of a conviction in adulthood by 4.8 times.
- Finally, while nationality has a considerable impact it has only a very slight influence on adult convictions (compared with a Swiss national, a person with a C permit is 1.2 times more at risk).
Nagelkerke’s R2 is an indicator – varying between 0 and 1 – that makes it possible to evaluate the overall quality of the logistic regression model. For criminological analyses, if the R2 is greater than 0.20, the model found can be considered satisfactory; it thus explains the 20% variation from the dependent variable. In this case, the model’s R2 indicates that the proportion of variance explained by the three retained independent variables is 15.27%.
The second question the FSO considered was “What are the risk factors influencing recidivism in adulthood?” It focused on a dataset – originating from the first – including 7 428 individuals already judged at least once by a juvenile court.
The logistic regression model shows the following results:
Variables in the equation | Odds ratio | Confidence interval at 95% | p-value | ||
---|---|---|---|---|---|
Sex | 3.8 | 3.186 | 4.534 | <0.0001 | |
Number of judgements | 4 vs 1 | 1.908 | 1.348 | 2.702 | 0.0003 |
3 vs 1 | 1.44 | 1.098 | 1.89 | 0.0085 | |
Offence against public authority (Title 15 SSC) | 1.882 | 1.329 | 2.667 | 0.0004 | |
Number of offences | 4 vs 1 | 1.566 | 1.195 | 2.053 | 0.0011 |
3 vs 1 | 1.528 | 1.218 | 1.917 | 0.0003 | |
Unauthorised taking of a vehicle RTA (Article 94 RTA) | 1.529 | 1.24 | 1.884 | <0001 | |
Offences against personal honour (Title 3 SSC) | 1.454 | 1.085 | 1.947 | 0.0121 | |
Driving without authorisation (Article 96 RTA) | 1.411 | 1.031 | 1.93 | 0.0315 | |
Offences against life and limb (Title 1 CP) | 1.374 | 1.171 | 1.612 | <0001 | |
Age on committing last judgement (<16 years vs >16 years) | 1.347 | 1.187 | 1.529 | <.0001 | |
Drug dealing | 1.285 | 1.01 | 1.634 | 0.0411 | |
Seriousness of offences | Crime vs contravention | 1.271 | 1.073 | 1.506 | 0.0055 |
Offences against property (Title 2 SSC) | 1.197 | 1.043 | 1.373 | 0.0104 | |
Constant | <0001 |
According to the results of the logistic regression model, the gender of the person convicted is the most predictive variable when it comes to evaluating the risk of recidivism in adulthood; however, compared with the results of the bivariate analyses, its influence is minor. Nonetheless, the value of the “odds ratio” shows that boys are almost four times as likely (3.8) to reoffend in adulthood as girls.
The number of judgements passed by a juvenile court figures in second place in the logistic regression model. However, it should be noted that this variable is only significant from the “three against one” difference onwards. This means that a juvenile delinquent who has been convicted twice does not have a significantly higher risk of recidivism in adulthood than a juvenile delinquent judged once. However, children or adolescents who have been convicted at least four times are almost twice as likely to reoffend in adulthood (1.9) than children or adolescents who have only been convicted once.
Committing an offence against a public authority increases the risk of recidivism in adulthood by more than one and a half times (1.88). Among the considered variables, “committing an offence under title 15 of the SSC” thus figures in third place.
The number of offences committed comes in fourth place. Compared with juvenile delinquents who have only committed a single offence, minors who have been convicted for committing at least three different offences are around twice as likely to reoffend on reaching the age of majority.
Among offences against the RTA, only the unauthorised taking of a vehicle and driving without authorisation figure in the refined logistic regression model. These increase the risk of recidivism in adulthood by 1.5 times.
Committing an offence against personal honour, life and limb and property, drug dealing or having been convicted for an offence after the age of 15 makes recidivism in adulthood more likely (the risk increases respectively by 1.45, 1.37, 1.20, 1.29 and 1.35 times).
Depending on the seriousness of the offence committed, the difference is only significant between committing an offence and committing a crime (p-value = 0.0055). This means that committing a misdemeanour rather than a contravention does not significantly change the probability of reoffending in adulthood.
Publication to be downloaded in German or in French
Young offenders and recidivism in adulthood: Statistics on criminal conviction of minors and Criminal conviction statistics 1999 to 2015
The FSO has examined the number of young offenders appearing in the statistics on criminal judgements of minors and who subsequently reappear in the statistics on adult convictions.
To do this, the FSO followed a group of 6 649 Swiss minors born in Switzerland in 1992 who committed an offence under the Swiss Criminal Code (SCC), the Road Traffic Act (RTA) or the Narcotics Act (NarcA). It emerged that between the ages of 18 and 23, 25% of these 6649 young offenders (1 664 persons) had been re-convicted by an adult court.
Males are not only over-represented among convicted minors (73% boys and 27% girls), they are also more likely to continue on a path of delinquency once they become adults. Boys convicted as minors are, in fact, almost four times more likely than girls to be convicted by an adult court (recidivism rate in adulthood of 31% for men and 8% for women).
The greater the number of judgements passed by a juvenile court, the higher the risk of being convicted as an adult. The recidivism rate in adulthood is 20%, 34%, 49% and 64% among young offenders on whom judgement was passed once, twice, three and four times or more respectively, by a juvenile court.
The results indicated that reoffending in adulthood seems to depend partly on the seriousness of the offences perpetrated as a minor. The reconviction rate in adulthood is 17% for minors who committed only fine-carrying offences compared with 29% for those having committed misdemeanours or felonies.
Recidivism in adulthood seems to be linked to the mention of an offence under the NarcA or the RTA in the judgement passed by a juvenile court. In this context, however, the FSO has observed that variety (i.e. the mention of several offences of a different nature) makes recidivism more probable. But variety is more common when offences have been committed under the NarcA or the RTA than under the SCC alone. This could explain why recidivism increases when drug dealing and traffic offences are involved.
The results concerning the influence of age should be treated cautiously due to some inevitable methodological biases.
Generally speaking, the figures show that minors who are convicted in the two years prior to coming of age are more often re-convicted in adulthood. But data regarding age at first conviction also show that minors who are initially convicted at a very young age demonstrate the highest re-conviction rates in adult life.
Publication to be downloaded in German or in French
Reconvictions after a first judgement as adult: Criminal conviction statistics 1984 to 2014
The limit of three years for criminal history analysis means that some convicted persons will be considered as first-time offenders (persons with no criminal history), although they have offended before. For example, someone convicted in 2009 after a first conviction in 2005 is considered a first-time offender in this type of analysis because more than three years stand between the two judgements.
One must bear in mind, however, that the recidivism rate varies greatly depending on whether recidivism is being analysed after a first or second judgement. Although some results are calculated according to the number of past convictions, the fact that the analysis of these convictions is limited to a period of three years before the reference judgement obscures, at least partly, the predominant weight of this criminal history in assessing the risk of recidivism.
Furthermore, in addition to the figures on recidivism published every year, the FSO offers results obtained from another methodological approach: a cohort study.
In other words, we examine the delinquent path of Swiss nationals born in Switzerland in 1966 who have committed a felony or misdemeanour under the SCC, the RTA or the NarcA during the first 10 years after coming of age. Convictions registered in the criminal register between 1984 (date corresponding to their coming of age) and 2014 (date corresponding to latest update of the current database) are taken into account for this examination. It is therefore a long-term study aimed at monitoring a group of persons from the time of their first conviction passed by an adult court.
Of the persons born in 1966, 8690 of them were convicted in the first ten years of adulthood for the first time and registered in the criminal register for committing a felony or misdemeanour.
This group of 8690 first-time offenders was followed over an observation period of 9 years starting from the first judgement. During this period, 38% committed another offence for which a second judgement was passed and for which another registration was made in the criminal register.
This group of 3306 first-time offenders was then followed for a second observation period of 9 years starting from the second judgement. During this second period, 51% of these offenders committed yet another offence which led to a third conviction and another entry in the criminal register.
Recidivism rates increase, therefore, with the number of judgements passed by an adult court (at the end of the observation period, they were 38% with one previous judgement and 51% after two previous judgements).
Having observed the role played by previous judgements, we turned to the influence of the characteristics of the persons convicted, i.e. their age and gender.
When recidivism rates were compared by age, we observed that the younger someone is when they commit their first offence, the more likely they are to receive a second and a third conviction. This observation was made while respecting the same observation period for each individual, i.e. 9 years.
There is a definite, negative correlation between age and recidivism; whereas the rate of first recidivism is 49% for persons having committed their first offence at the age of 18, it is 29% for persons who committed their first offence at age 26.
With regard to the gender of offenders, we observed that men are not only over-represented among first-time offenders, but also among first-time reoffenders (the rate of the first recidivism after 9 years is 40% among men and 26% among women).
However, once they have been convicted a second time, the recidivism rate of women slightly exceeds that of men (men's second recidivism rate is 51%, that of women 53%).
Consequently, one can consider that the small difference between men and women supports the hypothesis that gender does not have a great influence on people with serious criminal involvement, or one can think that the difference is significant and that female reoffenders find it harder to resist crime more than male reoffenders.
A criminal history, the gender and age of the persons convicted play an important role but the rate of recidivism also depends on the type of offences committed which led to the first conviction: when the first judgement includes an offence under the Narcotics Act (NarcA), the rate of the first recidivism at the end of the observation period is 52% whereas for other offences it is 37%.
A similar difference is seen with regard to the second recidivism; the rates are then 61% and 50% respectively. Starting one's criminal career by committing a felony or misdemeanour under the NarcA, therefore, is a risk factor.
However, if the type of offence mentioned during the first judgement influences recidivism rates, it also influences the type of law broken at the second and third judgement (i.e. for the first and second recidivism).
This graph provides new insight into the role played by the type of offences committed and the specific characteristics of the offence.
Criminal behaviour is said to be specific when a convicted person's criminal register shows a tendency to commit - throughout the delinquent career - offences which come under the same law (specific recidivism).
Recidivism was always said to occur as soon as any new offence (under the SCC, the RTA or the NarcA) was committed during the period of observation. Therefore, we only spoke of general recidivism. We had not yet taken into account the difference in behaviour between road traffic offenders and drug dealers (see the distribution of offences in the above graph).
In order to resolve this matter, we compared - for each of the laws mentioned in the first judgements (SCC, RTA or NarcA) - the rates of specific recidivism and general recidivism.
It is important to distinguish between the two types of recidivism because the difference between the rate of general recidivism and the rate of specific recidivism reveals the specific characteristic of the offence: the greater the gap between the rate of specific recidivism and that of general recidivism, the more diverse the criminal behaviour of the persons concerned.
general | specific | difference 1) | ||
SCC | first recidivism | 39% | 20% | 19% |
second recidivism | 54% | 41% | 14% | |
RTA | first recidivism | 38% | 29% |
9% |
second recidivism | 48% | 37% | 11% | |
NarcA | first recidivism | 52% | 19% | 33% |
second recidivism | 61% | 28% | 33% |
Criminal conviction statistics (SUS), state of the Register of Criminal Conduct: 30.04.2016
On this basis, it was observed that when a reference conviction refers to a felony or misdemeanour under the RTA, three out of four subsequent recidivisms come under this law and that adults whose reference conviction mentions a felony or misdemeanour under the NarcA show, conversely, wide diversification in the offences that they commit afterwards (only slightly more than one out of three subsequent convictions are under the NarcA).
In light of these observations, one can ask whether recidivism also depends on the extent to which the offences committed have a specific characteristic, or whether the number of reconvictions increases with the number of laws that have been broken?
To answer this question, recidivism rates after a first judgement for offences under one, two or three law(s) were compared.
We observed that the higher the number of laws broken in the reference conviction, the higher the rates of recidivism. In concrete terms, 9 years from the first judgement, the rate of the first recidivism is twice as high among persons who were judged for breaking three laws (75%) than among persons who were judged for breaking a single law (37%).
It should, however, be noted that there is a link between diversity and the type of offences committed and that there is certainly a link between the length of criminal career and diversity. Consequently, this last graph perhaps only confirms the results described so far; i.e. that the risk of recidivism increases with the number of offences committed and in particular when the person has committed an offence under the NarcA.
Publication to be downloaded in German or in French
Further information
Contact
Federal Statistical Office Section Crime and Criminal JusticeEspace de l'Europe 10
CH-2010 Neuchâtel
Switzerland