Navigation – Plan du site

AccueilTous les volumesVol. ISystems of Conditional Release ( ...

Systems of Conditional Release ( Parole) in the Member States of the Council of Europe

Between the principle of equality and individualization, pragmatism
Pierre V. Tournier
Cet article est une traduction de :
Les systèmes de libération sous condition dans les Etats membres du Conseil de l’Europe  []


This article is based on the preparatory work for the recommendation of the Council of Europe on conditional release (parole), adopted by the Committee of Ministers on 24 September  2003. It  evidences three models of conditional release (parole) in  member states: a “discretionary release system”, - most frequent on the Continent -, a “Mandatory release system” developed in Sweden, and between these two alternatives, the  “mixed release system” found in England and Wales, for example. This third model, founded on a pragmatic approach, combines automatic early release proceedings for prisoners serving a short sentence  and individual  proceedings for long sentences.

Haut de page

Texte intégral

  • 1  The recommendation was written up by the seven members of the Council for Penological Co-operation (...)

11. On September 24, 2003 the Committee of Ministers of the Council of Europe adopted a recommendation on “Conditional Release (Parole)” (Council of Europe, 2003).1 The decision had been made four years earlier by the Council for Penological Co-operation, following the establishment of a previous recommendation on “Prison Overcrowding and Prison Population Inflation” (Council of Europe, 2000). As will be shown, the attention paid by the European institutions to conditional release proceedings was a logical consequence of the earlier diagnosis pertaining to prison demography.

22. For the purposes of this recommendation, “conditional release means the early release of sentenced prisoners under individualized post-release conditions”. The in-depth preparatory work done in this connection showed the existence of extremely different conditional release systems and gave some indication of what changes these measures might undergo within the European Union.

1. In a context of prison population inflation

  • 2  This work was done in collaboration with Marie-Danièle Barré, CNRS-CESDIP.
  • 3  The same number of entries into prison from one year to the next, a same release agenda for all co (...)

33. When studying the prison population inflation experienced by France since the mid-1970s, we were led, in the early 1980s, to suggest a new, simplified demographic approach2 in which the number of inmates at any particular point in time, called P (statistical indicator of “stock”) is analyzed using the flow index for committals to prison for a given year (E) and the average length of imprisonment (d, duration expressed in years). The reference demographic model is the “stationary population” in which P is the product of E multiplied by d (P = E x d).3 This three-dimensional approach, applied to French data, enabled us to show the existence of a shift, starting in 1981, from prison population inflation due to a rising number of entering prisoners for an unchanged length of detention to a model in which the rising number of inmates is accounted for exclusively by longer stays in prison, with stable or even declining rates of prison-entering (Tournier, 1996).

44. The debate on alternatives to imprisonment, which crops up again time after time, had been focused until then on measures and sanctions whose merit was to avoid prison-entering (pretrial surveillance to avoid imprisonment prior to judgment, community service work sentences to avoid short sentences to personal restraint) had therefore to deal also with means of reducing the length of detention. The question could be formulated as follows: how can we reduce the length of detention while acting more efficiently to prevent recidivism? This was summed up by Nicolas Frize, head of the French section of the Human Rights League, as “Less, but better”. That means shorter periods of imprisonment but better attention during detention (better overall conditions, stricter respect for human rights, maintaining of family ties, development of occupational and training activities, cultural activities, medical and social care when necessary and so forth) and also, better attention for people under supervision within the community, with the possibility of accompanying these measures by more effective, and possibly stricter control measures when needed.

  • 4  At the time of our writing, we do not know what will become of the bill, under discussion in the S (...)

55. The question of prison population inflation therefore logically leads up to the issue of mitigation of sentences to personal restraint, and in particular, of its main measure, conditional release. Now the fact is that in France its use has constantly declined over the last 30 years (Tournier, 1997, Commission Farge, 2000). The subject was widely debated here when the left returned to power in June 1997, and the June 15, 2000 Act reinforcing the presumption of innocence and the rights of victims was enriched with a “mitigation of sentences” section that significantly improved the respect given to the sentenced person’s rights, especially with respect to the granting of conditional release, by extending the criteria for granting release, allowing assistance by a lawyer, demanding that justifications be given for decisions, allowing for appeal, etc.4

  • 5  With André Kuhn (Switzerland) and Roy Walmsley (Great Britain) in particular.

66. During the same period we happened to be working on the Council of Europe recommendation on prison overcrowding and prison population inflation.5 Our studies showed that the French situation, briefly described above (new type of inflation tied to the rising length of detention, decline of conditional release, etc.) is also to be found in most European countries.

77. This recommendation, adopted on September 30, 1999, proposed that these phenomena be combated through a multifactorial approach involving the entire penal process: from examination of “the possibility of decriminalizing certain types of offences or reclassifying them so that they do not attract penalties entailing the deprivation of liberty” to the development of measures which reduce the actual length of the sentence served, and of conditional release in particular (Tournier, 2000, 2002).

88. We feel it is important to reproduce, here, the concluding propositions, 22 to 26, of the resolution: Art. 22. In order to make community sanctions and measures credible alternatives to short terms of imprisonment, their effective implementation should be ensured, in particular through the provision of the infrastructure for the execution and monitoring of such community sanctions, not least in order to give judges and prosecutors confidence in their effectiveness; and the development and use of reliable risk-prediction and risk-assessment techniques as well as supervision strategies, with a view to identifying the offender’s risk to relapse and to ensuring public protection and safety.

99. Art. 23. The development of measures should be promoted which reduce the actual length of the sentence served, by giving preference to individualised measures, such as early, conditional release (parole), over collective measures for the management of prison overcrowding (amnesties, collective pardons).

1010. Art. 24. Parole should be regarded as one of the most effective and constructive measures, which not only reduces the length of imprisonment but also contributes substantially to a planned return of the offender to the community.

1111. Art. 25. In order to promote and expand the use of parole, best conditions of offender support, assistance and supervision in the community have to be created, not least with a view to prompting the competent judicial or administrative authorities to consider this measure as a valuable and responsible option.

1212. Art.26. Effective programmes for treatment during detention and for supervision and treatment after release should be devised and implemented so as to facilitate the resettlement of offenders, to reduce recidivism, to provide public safety and protection and to give judges and prosecutors the confidence that measures aimed at reducing the actual length of the sentence to be served and community sanctions and measures are constructive and responsible options.

2. Comparing different systems

1313. In their attempt to analyze the reality of positive law and practices with respect to conditional release in all member States of the Council of Europe, the writers of the new recommendation were led to distinguish between two very different types of CR: the “discretionary release system” and the “mandatory release system”. These form two poles, the interval between which is occupied by other systems, which may be termed “mixed release systems”.

2.1 – The discretionary release system

1414. The discretionary release system is the system in force in most European countries which practice conditional release. This is the case in France, where “first-time” sentenced prisoners may be granted conditional release halfway through their sentence, whereas recidivists must wait to have served at least 2/3 of their prison sentence before being allowed to do the rest in the community under the authority of a sentence-enforcement judge (SEJ) and under the surveillance of an integration and probation adviser (IPA). This is a necessary but not a sufficient prerequisite. Actually only a small minority receives the benefit of this measure. This system is mostly aptly described by the word individualization.

1515. In discretionary systems, individualization—or personalization—of the decision works at no less than three levels. Level 1 – decision to grant conditional release: an offender sentenced to time in prison may very well leave without having been released on parole (leaving at end of sentence). Level 2 – choice of the date of conditional release once the minimum period of detention has been served (defined in absolute terms and/or as a proportion of the sentence). Level 3 – choice of the conditions imposed on the person following release, during the period of monitoring within the community.

1616. Those European countries in which the discretionary system is applied are often faced with a decline in the granting of conditional release in recent years. The reasons for this have been analyzed in the Council of Europe recommendation on prison overcrowding and prison population inflation. There are many reasons:

  1. Public opinion: generally very badly informed about the issues involved in the execution of sentences, it views measures for early release as a sign of "being soft on crime".

  2. The social and economic context: the conditions for being released on parole are often unattainable for population groups that are increasingly marginalised. In addition, there is the difficulty of finding reasonably stable accommodation and especially employment on leaving prison.

  3. Change in the structure of prison populations according to the types of offence for which prosecutions are brought or sentences pronounced: in many countries, the growing numbers sent to prison for sexual violence or drug trafficking do not help to increase the rates of early release, since the decision to be taken may have serious consequences if considered in terms of recidivism.

  4. Competition from non-individualised adjustment measures:  some countries make use of amnesties and/or collective pardons, sentence reductions for which the conditions are exclusively linked to behaviour during imprisonment (positive criterion) or the absence of any serious incident during imprisonment (negative criterion); granting conditional release then becomes almost systematic and the measure therefore loses any individual character.  These procedures, designed purely to deal with the shortage of places and maintain discipline, are far removed from the true spirit of parole.”

1717. The argument given in (a) of the 1999 recommendation should probably be greatly relativized. In the report appended to the 2003 recommendation on CR, the authors point out that only one opinion survey on the subject was brought to their attention by the member States consulted during the preparatory phase (Tubex, Coll. Tournier, 2003). That survey was performed in France by an association of students doing volunteer teaching in prisons (GENEPI, 1998). The lack of opinion surveys in Europe leaves one perplex, given the fact that policy decisions on CR are often claimed to be based on alleged knowledge of what the public thinks, as is clearly shown in the case of Belgium (Snacken, Tubex, 1999).

1818. On the other hand, argument (d) on the competition of non-individualized adjustment measures is particularly apt in the French situation. There is a convergent trend, over a twenty-five year period, toward the replacement of individualizing practices by mass measures, with the dwindling number of conditional releases, the reduction, in 1986, of the legal possibilities for individualizing sentence-shortening, the introduction and extension of safety periods and conversely, the almost systematic granting of sentence cuts for good behavior (this dates back to 1973), the systematic granting of collective pardons by the President of the Republic—be he on the left or right—each year since 1991 for the national holiday.

2.2 The mandatory release system

1919. The mandatory (or at set period) release  system is in existence in Sweden since 1998 (enforced since January 1, 1999). Our colleague, Norman Bishop, describes it as follows:

20inmates who are sentenced to time in prison must be given conditional release once they have served two thirds of their sentence, with a minimum prison stay of one month. Conditional release may be delayed for a set number of days as a disciplinary measure. Conditional release is not applicable in case of a short prison sentence combined with a probationary measure, or in case of life imprisonment. A sentence to life can be converted into a sentence to time by a pardon, in which case the rule of release when two-thirds of the sentence have been served may apply.

2120. In the mandatory release system, individualization only applies at level 3 (choice of post-release conditions). Its advocates emphasize the difficulty in defining scientific criteria for determining when an inmate should be granted conditional release. To avoid arbitrary and highly diversified decisions depending on who makes them, it is preferable to prescribe the same treatment for all. Efforts should then be focused on personalizing supervision (control and care measures), and defining the conditions to be imposed on the offender following prison-leaving. These conditions must be strictly necessary. In some cases they may even be considered perfectly useless. Release is then granted with no supervision whatsoever. Despite the difference in treatment with respect to control conditions within the community, the mandatory release system claims to be essentially egalitarian.

2221. During recent discussions with Norman Bishop, we learned that this system is not unanimously approved in Sweden, by far, and may be repealed by the legislators, under pressure from . . . public opinion, which views it as overly favorable to convicted offenders (see above).

2.3 The mixed release system

2322. Alongside of these two systems, based on radically different options, mixed release systems have developed in recent years. These combine the discretionary system, for long sentences, and a mandatory release system for short sentences. They may be depicted as essentially pragmatic. It is a fact that for short sentences, the existence, or not, of early release hardly has any practical effect, since it only advances release by a few days, at best by several weeks. So why spend precious resources on numerous cumbersome, individual a priori, non-egalitarian proceedings when they could be put to better use? Thus, the selection process is reserved for the fewer cases (long sentences) for which the decision is fraught with consequences for the offender and for society in case of re-offending.

2423. This is the case for the system in use in England and Wales (Tubex, Coll. Tournier, 2003). A general overhauling of the system was introduced in 1991, calling for mandatory release for sentences to less than four years, with the possibility of supervision, whereas the discretionary system was maintained for sentences to four years or more.

2524. More specifically, as of October 1, 1992, prisoners sentenced to a term of less than one year are automatically released at the halfway point. They are not monitored after release, but may be returned to prison, according to article 40 of the 1991 Act, if they are convicted of another offence punishable by personal restraint before the complete expiry of the sentence.

2625. Inmates serving a sentence of one to four years are automatically released at the halfway point and subjected to a regime of follow-up and authorization, with specific conditions except home confinement, with compulsory curfews, until the three-fourths point. Articles 38 and 39 of the 1991 Act call for a possible return to prison in case of non-respect of the follow-up clauses. As in the previous case, sentenced offenders may also be sent back to prison in accordance with article 40 of the 1991 Act.

2726. Last of all, inmates convicted after October 1, 1992 and serving a term of four years or more are subject to the following CR regime: according to article 35 of the 1991 Act, they are eligible for CR halfway through their term. This possibility exists up to two-thirds of the sentence, and is studied in the framework of annual case examination by the Parole Board. Inmates who do not apply for CR or for whom it is refused are automatically released once they have served two-thirds of their sentence (article 33, par. 2 of the 1991 Act.

3. Is there a future for conditional release in Europe?

2827. In its recommendation, the Committee of Ministers, “recognizing that conditional release is one of the most effective and constructive means of preventing re-offending and promoting resettlement, providing the prisoner with planned, assisted and supervised reintegration into the community”, recommends that governments and member states “introduce conditional release in their legislation if it does not already provide for this measure”. But it clearly refrains from asserting that one system is better than another, even if members of the drafting group had their own preferences. The text simply indicates the advantages and disadvantages of each system.

2928. The explanatory memorandum appended to the recommendation identifies the following potential weaknesses for discretionary release systems:

  • An absence of explicit criteria for granting conditional release rendering decision-making erratic;

  • Disparities in decision-making when more than one body is involved in deciding on conditional release;

  • Assessments of the likelihood of relapse into crime, made without the assistance of scientific risk instruments, may prove to be unreliable;

  • Uncertainty about the date of release making it difficult to make practical release arrangements for prisoners;

  • The possibility that the foregoing factors lead to reduced confidence in the system and reduced motivation on the part of prisoners to co-operate in the observance of conditions and the requirements of supervision.

30Mandatory release systems, on the other hand, risk presenting the following weaknesses:

  • Knowing with certainty the date for conditional release reduces motivation on the part of prisoners to take part in programmes and courses designed to enable them to lead crime and drug-free lives after release from prison;

  • Knowing for certain the date of release leads to worsened behaviour by prisoners during their stay in prison;

  • The lack of the possibility to withhold conditional release leads to a marked increase in crime in the community being committed by conditionally released prisoners;

  • The mandatory release will lead judicial authorities to impose longer custodial sentences.

  • 6  In the French case, for example, there are leaves of absence, sentence cuts, individual and collec (...)

3130. These assertions are mostly hypotheses, and some are very difficult to prove empirically. The recommendation is therefore perfectly correct in stressing the need to develop research on the different systems, and above all, the need to communicate their findings to political officials as well as to criminal justice system actors and to all citizens in the countries involved. One of the difficulties involved in conducting such comparative studies should not be underestimated: a study of conditional release, how it is granted and how effective it is cannot be isolated from the rest: that is, it must take into account the entire scene with respect to proceedings for mitigating sentences.6

3231. Perhaps one positive consequence of the September 2003 recommendation will be to further convergence in the evolution of national systems toward a solution incorporating the best of each of them, so that sentence enforcement will be more convincingly committed to respecting human rights.

Translation : Helen Arnold

Haut de page


1  The recommendation was written up by the seven members of the Council for Penological Co-operation with the help of three scientific experts: Norman Bishop (Sweden), Pierre V. Tournier (France) and Hilde Tubex (Belgium).

2  This work was done in collaboration with Marie-Danièle Barré, CNRS-CESDIP.

3  The same number of entries into prison from one year to the next, a same release agenda for all cohorts of prison enterers. The actual evolution of the prison population is obviously very different from this model.

4  At the time of our writing, we do not know what will become of the bill, under discussion in the Senate, calling for the adjustment of the justice system to trends in crime, which might considerably transform law on sentence-enforcement in France.

5  With André Kuhn (Switzerland) and Roy Walmsley (Great Britain) in particular.

6  In the French case, for example, there are leaves of absence, sentence cuts, individual and collective pardons, amnesties, semi liberty schemes, placement in the community and electronic monitoring (EM)

Haut de page

Pour citer cet article

Référence électronique

Pierre V. Tournier, « Systems of Conditional Release ( Parole) in the Member States of the Council of Europe »Champ pénal/Penal field [En ligne], Vol. I | 2004, mis en ligne le 28 janvier 2006, consulté le 07 décembre 2023. URL : ; DOI :

Haut de page


Pierre V. Tournier

Director of Research at the Centre national de la recherche scientifique (CNRS, Paris), Centre d’histoire sociale du XXe siècle, Université Paris I Panthéon Sorbonne (Center for Social History of the Twentieth Century), President of the AFC. 43, rue Guy Môquet 75017 PARIS,

Articles du même auteur

Haut de page

Droits d’auteur


Le texte seul est utilisable sous licence CC BY-NC-ND 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search