Introduction
Sexual and domestic violence account for a substantial share of criminal offences worldwideFootnote 1 and have, in recent years, emerged as a major policy challenge for many governments. In 2022, they represented in France—the case examined in this article—slightly more than half of all bodily injuries recorded by law-enforcement agencies.Footnote 2 The high number of victims and the associated costs led to the designation of these forms of violence as a “national cause” in 2010 and 2017.
In addressing gender‑based violence (GBV), governments rely on a range of instruments designed to punish offenders, protect victims and remedy the consequences of aggression. Penal responses remain widely viewed as inadequate: sanctions are lenient,Footnote 3 attrition rates are highFootnote 4 and protective measures often fail in practice.Footnote 5 Victims still bear much of the burden of violence, while the structural nature of GBV is obscured.Footnote 6
Their failure has generated growing interest in restorative justice, which has expanded widely over the past three decades across English-speaking, multilingual and francophone jurisdictions. In France, since August 2014, all individuals involved in judicial proceedings—whether victims or perpetrators of criminal offences of any kind—have been granted the right to benefit from measures at any stage of the process. Restorative justice has been actively promoted, often enthusiastically, by successive governments. The Ministry of Justice has implemented a coordinated policy across different branches of the judicial administration—a level of coordination that is far from common. This policy has been supported not only by specialized non-governmental organizations, but also by some organizations dedicated to assisting victims of crime. It has gained public visibility, notably following the success of a 2023 film devoted to the subject and its adoption by the Catholic Church as part of its response to sexual violence within its ranks.
All conditions thus appear to be in place for restorative justice to become a public policy in France. Yet, ten years after its introduction, just over 1,000 individuals have participated in restorative measuresFootnote 7 compared with an annual figure of 230,000 convicted persons and 500,000 victims of offences against persons. The allocated budget remains minimal—approximately 1 percent of a victim-assistance fund amounting to €54 million in 2025—and dedicated human resources are scarce, even as overall resources for the judicial system have increased. The systematic dissemination of information about this right, both to judicial personnel and to litigants, has not been achieved; it continues to be a recurrent subject of debate in the National Assembly. Consequently, restorative justice belongs to a category of deliberately ineffective policies—a “symbolic” policy in which “saying is doing.”Footnote 8 Although now enshrined as a right, restorative justice has yet to reach litigants.Footnote 9
How can such a stark divergence between legal recognition, political celebration and the minimal implementation be explained? Scholarly explanations typically emphasize ideological controversies surrounding the relationship between restorative justice and punitive criminal justice,Footnote 10 as well as feminist concerns regarding the risks of revictimization, gendered power relations and implicit pressures placed on victims, particularly in cases of intimate-partner violence. A further major criticism concerns the risk of “second‑class justice,” which may weaken the state’s public accountability.Footnote 11 While these debates illuminate important theoretical tensions, they offer only limited insight into the French trajectory. Our research shows that neither feminist opposition nor public mistrust nor an inherent incompatibility with penal logics constitutes the principal barrier to implementation. Instead, the marginalization of restorative justice results from political and administrative mechanisms: strategies of blame avoidance, institutional ambivalence, downward delegation to under‑resourced local actors and a quasi‑experimental governance model that depends on the discretionary—and often unpaid—labour of frontline practitioners.
Drawing on a decade‑long empirical investigation, this study examines the political uses of restorative justice in France and the processes through which a formally established right becomes inoperative in practice. The first section traces the legislative and political genesis of restorative justice, showing how a seemingly consensual reform has been contained within a low‑visibility and low‑stakes policy space. The second section analyzes its implementation, focusing on the ways in which responsibility is shifted to local actors and how resource scarcity produces structural inaction. The final section explores how this configuration reshapes the place of restorative justice within a penal system still dominated by retributive and carceral logics, ultimately transforming a statutory right into an optional service and obscuring the structural dimensions of gender‑based violence.
State of the art: A policy of blame avoidance?
The scholarly literature on restorative justice has long emphasized the ideological tensions surrounding its relationship to criminal justice. Two broad interpretive frameworks dominate.Footnote 12 A first body of work portrays restorative justice as fundamentally at odds with retributive criminal law, challenging a system centred on the offence and its punishment. In this view, restorative practices represent a paradigmatic alternative grounded in dialogue, accountability and harm repair, and their marginality is understood as a consequence of criminal justice institutions’ commitment to punitive and carceral logics. A second perspective conceptualizes restorative justice less as an alternative than as a complementary mechanism—a pragmatic tool that may coexist with retributive frameworks by offering additional avenues for victim participation and offender responsibility. Yet, even within this more incremental reading, restorative justice tends to be positioned as subordinate to the central aims of the penal system.
These debates become particularly salient in the context of GBV—a domain in which restorative justice has elicited intense scholarly and political concern. Critics, especially within strands of feminist legal scholarship, stress that GBV is embedded in relations of domination and coercive control, rendering face‑to‑face encounters between victims and perpetratorsFootnote 13 potentially harmful. From this standpoint, restorative practices risk reproducing structural inequalities, exposing victims to re‑victimization and undermining the hard‑won recognition of GBV as a matter of public responsibility.Footnote 14 These concerns have translated into legal safeguards at both national and international levels—for instance, in local moratoriaFootnote 15 or in the Istanbul Convention adopted in 2011 by the Council of Europe, which explicitly prohibits mediation in cases of domestic violence and violence against women.
Other feminist scholars, by contrast, have argued for a more nuanced assessment. Rather than rejecting restorative justice outright, they examine the conditions under which dialogical responses might better meet survivors’ needs—through informed consent, specialized training for facilitators, robust risk assessments and procedural safeguards.Footnote 16 They emphasize that, for some victims, particularly those unwilling or unable to initiate criminal proceedings, restorative practices may offer meaningful forms of recognition, validation and agency.Footnote 17 Such arguments have gained traction in jurisdictions in which restorative justice is used as a pragmatic response when prosecution is time‑barred, when cases are dismissed or when survivors explicitly seek an alternative to adversarial procedures against perpetrators who are often also intimate or close relations.
Although these ideological debates are central to the international literature, they insufficiently explain the limited uptake of restorative justice in France. Neither feminist opposition nor punitive public sentiment accounts for the empirical patterns observed. Parliamentary debates rarely express hostility to restorative justice; political actors seldom frame it as a threat to retributive norms; and feminist organizations have remained largely absent from public controversies on the subject. Moreover, empirical studies—including our own—demonstrate that restorative measures can be, and are, implemented in cases involving GBV when institutional conditions permit.
Understanding the French trajectory requires shifting attention from ideological disputes to public policymaking. A substantial literature in the sociology of public action highlights the prevalence of incremental reforms, low‑visibility adjustments and inaction.Footnote 18 Such inaction is especially likely to arise in politically sensitive domains such as GBV, in which apparent consensus coexists with entrenched gender norms and institutional ambivalence.Footnote 19 At the same time, sociolegal scholarship on the “gap” between law and implementation underscores how rights may be formally enacted yet remain practically inaccessible, and how such gaps shape social expectations and state legitimacy. This hypothesis is consistent with feminist critiques of governmental inaction in this area. According to these analyses, institutions seek to bolster their legitimacy by signalling commitment to victims who have long been underserved by the penal system. Yet, such signalling has not produced substantive change. In the wake of the #MeToo movement, no major transformation has occurred; the existing legislative framework has been taken up only partially and in uneven ways.Footnote 20
We draw on insights from both the sociology of public action and the sociology of law. The former shows that political inaction often reflects strategies of blame avoidance, designed to shield elected officials from electoral sanction. The latter emphasizes not the mere existence of implementation “gaps,” but how these gaps are perceived and how they shape social norms and the state’s legitimacy.Footnote 21 Drawing on this theoretical foundation, we approach the limited development of restorative justice not as a product of ideological resistance,Footnote 22 but as an effect of political strategies of blame avoidance,Footnote 23 administrative delegation and governance by inaction. This perspective situates restorative justice within broader patterns of symbolic reform, in which the creation of rights coexists with mechanisms that ensure their restricted or uneven application. It further allows us to examine how institutional choices—rather than doctrinal disputes—structure the practical possibilities for deploying restorative justice, especially in cases of GBV.
Inaction is a recurrent feature of public policy. Yet, how can we explain a government that, despite increasing recognition of its inability to address GBV, refrains from deploying a tool it repeatedly presents as a complementary solution—especially when the criminal justice system fails to process the vast majority of cases? Insights from the sociology of public action, combined with findings from the sociology of law on the routine nature of “gaps” between law and implementation,Footnote 24 offer a compelling answer. Together, these literatures show that inaction is no less consequential than action in the formation and maintenance of political coalitions.
Weaver’s notion of “blame avoidance” captures this dynamic: elected officials may design policies not to claim credit for reforms, but to shield themselves from potential electoral sanction—an electorate being more likely to reject what is done to it than to reward what is done for it.Footnote 25 This hypothesis requires identifying the specific subgroups within a political and institutional context whose reactions policymakers seek to pre‑empt.Footnote 26 A variety of strategies may then be deployed: preventing an issue from entering the political agenda, shifting responsibility to another institution, designating scapegoats, promoting consensus‑based decisions or introducing automatic mechanisms. Such practices generate inertia and forms of “automatic governance,”Footnote 27 even as policies accumulate over time and fill the political space rather than leaving a vacuum.
Building on these insights and examining the entire chain of policy design and implementation, we approach the limited development of restorative justice not as the outcome of ideological resistance, but as the effect of political strategies of blame avoidance, administrative delegation and governance by inaction. This perspective situates restorative justice within broader patterns of symbolic reform, in which the creation of rights coexists with mechanisms that ensure their restricted or uneven application. It further allows us to examine how institutional choices—rather than doctrinal disputes—shape the practical possibilities for deploying restorative justice, in France as elsewhere,Footnote 28 especially in cases of GBV.
Methodology
Our methodological approach combines multilevel and multi‑method research designs aimed at capturing both the political construction of restorative justice and its practical implementation. It is essential to specify not only the motivations of those implementing this policy, but also in relation to whom these motivations are defined—particularly the links between governments and the bureaucracies onto which they shift responsibility. To do so, we adopted an analytical framework operating across three interconnected levels: the national level, at which political and administrative elites shape the contours of the policy; the intermediate level, at which judicial services and civil‑society organizations translate these orientations into operational guidelines; and the local level, at which facilitators’ day‑to‑day practices and participants’ experiences reveal how the policy is concretely enacted.
To document these different scales, the study draws on a three‑year investigation (2021–23) embedded within a broader ten‑year analysis (2014–24) of the introduction, diffusion and implementation of restorative justice in France.Footnote 29 This longitudinal perspective allows us to trace the evolution, institutional interpretation and practical implementation—or stalling—of policy ambitions.
At the national level, we conducted semi‑structured interviews with ten magistrates and senior officials from the Ministry of Justice, including departments responsible for steering access‑to‑justice and victim‑support policies. These interviews, typically lasting between 1.5 and three hours, provided insights into the strategic orientations, constraints and internal debates shaping the governmental approach to restorative justice.
At the intermediate level, we examined the role of judicial services and specialized organizations responsible for delivering restorative measures. Interviews with nineteen facilitators, organizational leaders and professional evaluators allowed us to understand how national directives are interpreted, adapted and sometimes contested within the field. This work was complemented by the systematic analysis of all parliamentary debates referring to restorative justice between 2014 and 2024, as well as ministerial guidelines, inter-ministerial reports and training materials produced during the same period.
At the local level, we conducted extensive ethnographic fieldwork across five socially diverse regions. This included observations of a launch event at a Court of Appeal, a week‑long facilitator training programme, recruitment sessions in which practitioners contacted potential participants by phone and fourteen local working‑group meetings held between 2021 and 2024. We also observed three complete restorative processes, ranging from indirect encounters and preparatory interviews to full victim–offender dialogues and totalling eleven sessions. These cases included intimate-partner violence, youth violence, homicide, traffic violence, violent theft, sexual violence and incest. Through this approach, we gained access to the discretionary practices, emotional labour and institutional constraints that shape the everyday implementation of restorative justice.
Finally, twenty-two interviews were conducted with former participants—victims, offenders and community members—up to five years after their involvement in a measure. These interviews provided a longitudinal perspective on how individuals retrospectively evaluate their experience, the meaning they attribute to it and the effects it may have had on their trajectories. In addition, a focus group was carried out to explore expectations regarding justice among individuals who had or had not participated in measures, either as offenders or victims, in two of these regions (eighteen participants, four groups).
Across all levels, the research adhered to established ethical standards.Footnote 30 Participation was based on informed consent, all identifying information was anonymized and observations of restorative sessions occurred only with the explicit agreement of all parties and partner organizations.
By triangulating policy analysis, elite interviews, organizational ethnography and participant narratives, this methodological design provides a comprehensive understanding of how restorative justice is framed, delegated, constrained and enacted. It allows us to capture not only the intentions and hesitations of central actors, but also the practical frictions, resource limitations and discretionary decisions that shape the lived reality of restorative practices—especially in cases of GBV.
Results: Who is to blame for the disabling of the right to restorative justice?
Is inaction on restorative justice ideologically driven?
Political framing: An uncontroversial debate
Contrary to the hypothesis favoured in the literature, the French Parliament has not been the site of genuine ideological controversy over restorative justice. The literature on public action invites us to examine the reception of restorative justice and the political calculations of those within the legislative and executive branches who have the power to implement it. This approach aligns with the debates found in the specialized literature on restorative justice: if legislators hesitate to operationalize the right to restorative justice, it may be because they fear the ideological controversies that may jeopardize the support of certain segments of the electorate. The political context in France is, at first glance, unfavourable to the widespread use of the tool, particularly in cases involving serious harm to persons. From the perspective of the “blame-avoidance” hypothesis, policymakers anticipate electoral sanction from citizens whom opinion polls consistently depict as strongly supportive of punitive criminal justice. The promotion of policies favouring custodial sentences by both right- and left-wing governments appears to echo a growing societal desire for punishment.Footnote 31 The rise of the far right over the past forty years is another indicator of this shift. Voters, politicians, media and justice professionals would all seem to endorse a form of “penal populism.”Footnote 32 This dynamic has been partly legitimized through the instrumentalization of the figure of the victim. Even feminists—some of whom were initially reluctant to embrace a penal system that they regarded as patriarchalFootnote 33—appear to have yielded to the temptation of offender punishment. This is particularly true for feminists who entered political institutions and gradually transformed the theme of victim protection, so prominent in the 1980s, into an “imperative of repression.”Footnote 34 While the criminalization of offenders and their removal from victims’ social environments only became the norm in 2022, this followed a period of “legislative overinvestment” in sexual violence, which fueled a steady tightening of the law.Footnote 35
The failure to implement the reform could thus be explained by the anticipation of Members of Parliament (MPs) of a majority preference for a punitive approach to GBV, set in opposition to a restorative approach perceived as more lenient. Restorative justice fundamentally breaks with this punitive logic, as the Minister of Justice Christiane Taubira emphasized in 2014. It involves other actors—such as the victim and the community—in the relationship between the state and the offender. Rather than relying solely on fear of punishment, it empowers the victim to foster the offender’s “accountability” and understanding of the consequences of their actions.
However, an analysis of parliamentary debates shows that this ideological controversy is not reflected at the legislative level. No intervention questions the potential of restorative justice to serve as an alternative or to challenge the carceral logic. The debates appear weakly ideologized—mirroring a public opinion marked by nuance, oscillating between punitive impulses and aspirations for moral rehabilitation.Footnote 36 Neither in the government’s 2014 bill nor in the broader discussions on criminal justice policy is the neutralization of offenders set in opposition to the empowerment of victims; indeed, carceral logics can even be reinforced through the very consideration of survivors’ needs. Restorative justice is, in fact, well suited to such complexity. It is ideologically malleable and sufficiently ambivalent to accommodate both humanistic and punitive approaches.Footnote 37 Notably, the legislation that introduced the tool in France framed it not as an alternative to punishment, but as a complement to the penal system, in favour of victims. This emphasis on victim support is not only compatible with a preference for imprisonment; it was originally championed by its strongest advocates. The creation of a victim-support secretariat in 2004 by former right‑wing president Nicolas Sarkozy is a telling example.
The introduction of a right to restorative justice in legislation was likewise not accompanied by any debate on the reservations expressed by those feminists most strongly committed to carceral responses. Yet, in the wake of the Istanbul Convention, the 2012 EU Directive adopts a notably cautious position, emphasizing that such measures “require safeguards to prevent [the victim] from suffering secondary and repeated victimization, intimidation, and retaliation” (para 46).Footnote 38 Feminist actors have never taken a public stance on restorative justice. For instance, they have never followed up on proposals for exchange made by specialized organizations (head of a specialized organization, February 2025). These movements were, in fact, almost entirely absent from public debates on restorative justice, even when French governments began, from 2021 onwards, to encourage its application to time‑barred sexual offences committed by adults against children.
It was instead an institution created by the government that reminded policymakers that the vulnerability of victims requires stringent safeguards. The Independent Commission on Incest and Sexual Violence Against Children (CIIVISE), established by the French government in 2021, expressed in its final report a categorical rejection of what it termed “restorative injustice”: “the same injunction to submit, to force oneself into denial” and to “renounce self‑assertion.”Footnote 39 From this perspective—defending the necessity for criminal prosecution and the separation of the survivor from the aggressor—restorative dialogue is portrayed as endangering victims and as a policy instrument for managing prison overcrowding. This position, however, gained little traction and the commission’s leadership, which faced considerable criticism, was swiftly replaced.
Blame avoidance: Shifting responsibility onto victims who fail to exercise their rights
The dominant mechanism is not ideological resistance, but strategic depoliticization and the shifting of responsibility onto victims. The debate among different strands of feminism over the possibility of a dialogue free from coercion between survivor and perpetrator did not take the form of an ideological confrontation; rather, it shaped the conditions under which restorative justice could be operationalized. Restorative justice is partly designed to meet victims’ needs, following its codification in European Union law. This framing explains why it is primarily supervised by the Service for Access to Justice and Victim Support (SADJAV) within the Ministry of Justice. The reform has taken the form of an “experimental budget for victim assistance” (head of a specialized organization, February 2025). However, this conception as a victim-oriented policy has sometimes had the unintended effect of hindering its implementation.
Within this context, governmental policies have proven ambivalent. On the one hand, face-to-face encounters between victims and perpetrators are encouraged; on the other, the 2019 “Grenelle” agreement on domestic violence recommends their separation. The “Restorative Justice Methodological Guide,” produced in 2020 through collaboration among various judicial services within the Comité National de la Justice Restaurative (CNJR), reflects this oscillation. Domestic violence is characterized therein as requiring “particular vigilance” given the potential for the aggressor’s control over the victim. That same year, the law of 30 July 2020, following the Istanbul Convention, sought to prohibit family mediation in cases of alleged violence or manifest control. Yet, on 23 November 2021, a decree was adopted establishing that restorative measures—applicable when the offender acknowledges the facts, even when prosecution is time-barred—may be implemented in cases of sexual offences committed by an adult against a minor. This approach, and the possibility of extending it to cases dismissed without prosecution, became the subject of an inter-ministerial mission in autumn 2023.
We encountered this ambivalence repeatedly in the field. It is illustrated, for example, by a statement made at a steering committee meeting in 2021 (observed during the course of a previous studyFootnote 40), by a senior official of the prison administration: “No outright prohibition, but precautions taken to the extreme, so as not to endanger the victim or place the mediator in difficulty.”
Restorative justice has thus remained a difficult public policy issue. The ambiguity of its formulation by legislators and Ministers of Justice is revealing. Consider, for example, the debate that took place in the National Assembly during the vote on the government’s bill on justice policy and planning in the summer of 2023.
Following an amendment proposed by the radical left party La France Insoumise (LFI), aimed at systematizing “the provision of information” on mechanisms, all MPs who spoke in favour of them—that is, all parties on the left and right except the far-right Rassemblement National and the Communist Party—voted in favour. The rapporteur of the parliamentary law commission opposed the amendment on the grounds of insufficient resources “to guarantee access to relevant mechanisms for anyone wishing to use them.” Minister of Justice Éric Dupond-Moretti intervened, stating that interaction cannot be imposed on all victims, but that “every time it is implemented, the results are incredible”:
That is why we cannot—excuse me for saying so—use it as a smokescreen. Offering it systematically, when we are not ready, would amount to deceiving people. Do you know, Madame Untermeier, what is missing in many jurisdictions? Organizations, a range of services! (He draws a comparison with Canada.) We must therefore fill this gap in our own way […]. In the meantime, I repeat, we are not ready: there is no point in encouraging people to choose a procedure that is not yet operational or finalized, in making a proposal that will lead nowhere!
C. Untermeier (Socialist Party) responded: “Everyone loves restorative justice, but when it comes to enshrining it in future legislation, no one wants it.”
The minister’s caution reflects a blame-avoidance mechanism: “some people have no desire to be confronted with the author of their misfortune.” This reasoning exemplifies a classic blame‑avoidance strategy by which the presumed absence of demand is mobilized to justify limited information and provision. It is assumed that victims’ “negative representations” prevent them from taking advantage of the law.Footnote 41 According to this logic, if there is a “gap,” then it is because the law has not adapted to social norms.Footnote 42 If there is public inaction, then it is because there can only be minority recourse, always morally questionable given the risk of influence.
This “observation” of insufficient demand justified the launch in 2022 of a national experiment initiated by the coordinating committee established within the Ministry of Justice. The experiment was designed as a truth-seeking operation aimed at “obtaining the figures” on victims’ requests (interview with a committee member, April 2022). In December 2020, this same person told us that restorative justice was not a priority for the ministry. Our interlocutor expressed strong reluctance on the part of magistrates, but considered this a “false problem,” the real issue being that “we cannot find an audience.” He concluded: “[Since they are so low], if we publish the figures, we will bury it!”
The political and moral discomfort surrounding the possibility that restorative justice weakens criminal law has instead given way to a shift in responsibility onto the victim. Is the “non-take-up” of the right the main cause of the ineffectiveness of this policy? This observation is not without foundation. It refers to victims’ reluctance to engage in arduous judicial procedures, sometimes compounded by rejection from family and friends. The scarcity of prosecutions is thus explained not only by dysfunction within the penal system, but also by societal attitudes toward justiceFootnote 43 and the deep entrenchment of sexual and domestic violence in gender norms.Footnote 44 This reluctance also applies to restorative measures. It is not easy to commit to a restorative process, given the fear of reliving trauma—even though victims’ interest often appears independent of this fear: “It reminded me of something I had not forgotten, but the pain had faded […]. It was difficult, but I really, really wanted to participate” (a victim participant, June 2023).
However, tests conducted during the “national experiment” confirm that interest in dialogues exists when information is provided (including in cases of sexual and domestic violence).Footnote 45 The hostility often anticipated does not materialize. At least 10 percent of those informed would immediately exercise their right. This interest among justice users has been further reinforced by the recent visibility gained by restorative justice.
The measures implemented confirm this. GBV is represented at least as much as in judicial proceedings overall. Seventy-three percent of our respondents were concerned by such acts. A specialized organization estimates that they account for more than 80 percent of measures: “In June 2024, it’s 80–85% of our activity […] about 62% sexual violence, including incest—and incest represents a significant share […], there must be 25% domestic violence […], the rest is 5% homicides and 10% other things, crimes against persons” (head of a specialized organization, February 2025). The concerns expressed by some professionals in the system therefore do not prevent the implementation of restorative justice, including for domestic violence: indirect meetings are preferred, but direct mediations are not excluded. The same applies in other countries, where GBV is widely addressed through such mechanisms.Footnote 46
These findings challenge the conviction of central administration officials that victims do not wish to engage in dialogue. Professionals involved at the local level highlight the same ambivalence: “we are told there is not enough volume, except that when victims come forward, we cannot even respond to their requests because ‘the tools are not ready,’ then we are told: ‘You are not generating enough volume.’ So it’s a vicious circle” (a facilitator, youth judicial protection officer, May 2023).
Victims are not responsible for the limited development of restorative justice; rather, their expectations remain poorly understood and institutionally unexamined. To encourage non-take-up, it is advantageous to leave the definition of the “potential public” vague.Footnote 47 But these justifications for inaction present another interesting feature: the discreet reformulation of the problem and the contradictions they introduce. If restorative justice is not effectively implemented, it is because there is no demand, asserts the Ministry of Justice. It considers that victims are sovereign and that they alone must decide how to exercise their rights. They may therefore legitimately refuse to be placed in contact with “the architect of their misfortune” (Minister of Justice Dupond-Moretti, 2023).
Paradoxically, the disabling of the right is thus justified by the government on the grounds that victims do not exercise this right. The responsibility for institutional failure is shifted onto its “user.”Footnote 48 It is therefore left to citizens to determine the effectiveness of the measures implemented. Recent scholarly literature has drawn attention to the emergence of activation and individualization logics that impose upon citizens an “injunction to autonomy.”Footnote 49 This rationale serves to mitigate the perceived acceptability of the non‑take‑up of criminal complaints. A study revealed that 94 percent of victims of sexual violence and 85 percent of victims of domestic violence do not file a complaint.Footnote 50 The prevailing retributive logic is characterized by the following outcome: while offenders prosecuted in court are rare, they remain liable to sanctions. The decision not to initiate judicial proceedings is thus rendered unproblematic.
This individualization of public action also contributes to the “evaporation” of feminism within legislation and its operationalization:Footnote 51 unlike a security‑based approach that assigns responsibility for prosecution to institutions, recourse to restorative dialogues is left to the discretion of the parties. The victim is portrayed as an autonomous individual, rather than as the target of violence within a relationship marked by domination. This obscuring of the problem posed by the existence of a social norm that facilitates the perpetration of sexual violenceFootnote 52 is likewise reflected in the implementation of the measures themselves, which are conceived as interpersonal relations rather than structurally asymmetrical ones.Footnote 53
An apolitical policy?
Far from reflecting mere disinterest, the structured inaction surrounding restorative justice performs an important managerial function: it mobilizes the commitment of local frontline actors without requiring any meaningful public investment. This dynamic helps explain why the government can claim to pursue a policy that produces “incredible results,” in the words of Minister Dupond‑Moretti, while simultaneously insisting that it cannot fully implement this policy—and need not even inform the public of its existence—on the assumption that its intended audience is largely uninterested. The implementation of this policy halts at the threshold of systematic information, which would likely trigger a surge in requests across France. The core issue in the exchanges between the Minister of Justice and MPs is straightforward: must a right be made known to be exercised? The answer provided by the minister and the majority MPs is negative: the institution lacks the capacity to cope with the consequences that such an information campaign might generate. Socialist MP Cécile Untermaier concludes: “A portion of the funds you are mobilizing could be allocated to this objective. Indeed, if restorative justice works so well in Quebec, it is because it is supported by public funding.”
Yet, this policy of non‑recourse reveals a structural impossibility: if only 10 percent of eligible individuals participated, then it would involve 12,000 offenders and 50,000 victims annually. Existing resources would be far from sufficient to meet such demand. It appears that the public system rests on the assumption of a minority use of this right, rather than on its effective universal accessibility.
Should this inefficiency be attributed to the long-standing trend toward budgetary rationalization? Resources have not increased despite rising demand. Due to insufficient dedicated working time and, consequently, a shortage of facilitators, waiting periods are lengthy and many applicants abandon the process. It would be tempting to interpret this decline because of the “great budgetary petrification.”Footnote 54 However, this hypothesis is contradicted by one fact: the law widely adopted by the Assembly in 2023 projected a justice budget of nearly €11 billion by 2027, compared with €9.6 billion in 2023, along with 10,000 new hires over five years to reduce judicial delays. At the same time, the decision was made to create 15,000 prison places by the end of the presidential term, confirming the primacy of a carceral approach. The justice budget has been increasing since the 1980s.Footnote 55
Why, then, deny restorative justice the resources allocated to many other areas of judicial expenditure? The nature of parliamentary debates surrounding it may shed some light on this phenomenon. Legislative debates on justice are dominated by the assertion of the technical apoliticism of legal professionals.Footnote 56 Yet, restorative dialogue is, in principle, consensual: “Everybody loves restorative justice” (Socialist Party representative, 6 July 2023). This “right,” broadly established for all, does not serve as a pretext for partisan confrontation. There are no genuine actors wielding veto power,Footnote 57 except for a few rare dissenters, such as MP Georges Fenech—a former magistrate and head of a far-right magistrates’ union, who ironically referred to it as “hug therapy” (5 June 2014, National Assembly). Likewise, there are no prominent advocates. No political party has made restorative justice a distinctive element of its political platform.
Yet, this apparent apoliticism, which ostensibly favours the government’s endorsement of restorative justice, occasionally reveals a subtle resurgence of politicization. The rationalization of judicial administration never fully eradicates the profoundly political nature of decisions concerning justice, such as the prioritization of public order over civil liberties.Footnote 58 Restorative justice, however, remains entangled with these major cleavages. A somewhat rhetorical question asked by an adviser to the Minister of Justice after the authors of this article were heard in June 2022 illustrates this point well: “But in the end, restorative justice—isn’t it a left-wing thing?” Organizations’ leaders widely identify it as a “left-wing” policy. They further observe that they have never been interviewed by individuals affiliated with right-wing parties (head of a specialized organization, February 2025). It is perceived as “the product of a racialized and left-wing Minister of Justice” (a facilitator, referring to C. Taubira, July 2023).
Disagreements regarding its application to domestic violence and the diversity of conceptions underscore this implicit repoliticization. Parliamentary debates during the summer of 2023 reflect this fragmentation. For an MP from a radical left party (LFI), restorative justice is essential to combat prison overcrowding and to enable victims to engage in interactions with offenders. For an MP from a right-wing party, it serves to promote “the social reintegration of offenders and to reduce distrust toward the judicial system.”Footnote 59 Minister Dupond-Moretti introduced the concept of “victim resilience.” These definitions are, in fact, political viewpoints—the LFI MP emphasized the far-right’s desire for victims to decide “in place of judges”—and represent an extension of the notions of a “good society” advanced by proponents of restorative justice, who have always been highly heterogeneous. From a theoretical standpoint,Footnote 60 as well as in terms of the political orientation of the governments that have promoted it in various countries, restorative justice can be championed by both the political right and the left.
But, if this form of justice is considered as “left-wing” or “right-wing” beneath the thin veneer of consensus, why does it enjoy such broad support? Or, at least, why is it so rarely criticized in the Assembly and within the judicial system? Why does it elicit strong backing from ministers serving governments of different political orientations? To understand why this humanistic and time‑intensive practice—only marginally compatible with the repressive judicial policies dominant over the past forty years—is broadly praised but rarely put into practice, one must consider the state in its role as employer. The transfer of responsibility is not merely a negative policy—a way to offload the issue of GBV. It is also, albeit unintentionally, a positive measure of human resource management.
A managerial policy of the state: Re-motivating judicial personnel
Let local actors take the lead
Restorative-justice measures require the mobilization of numerous actors within the judicial system: probation officers, youth-protection specialists and employees of victim-support organizations. Restorative-justice policy is a “delegated policy” managed by non-governmental organizations: the significant role it assigns to the community converges here with the broader process of “privatization” of judicial systems observed in many countries.Footnote 61 In France, victim-support organizations receive subsidies from the Ministry of Justice to assist victims of criminal offences. These organizations are accredited by the courts of appeal, which also contribute to their funding. In this context, they form part of a discretionary street-level bureaucracy.Footnote 62
This co-production resonates upstream with a principle of action articulated in the ministry’s guide of November 2020: “allow restorative justice policy to unfold across the entire territory” because “it has become essential not to restrict local initiatives […], which may take different forms.” In practice, measures are implemented by services and their partners according to local alliances.
The resulting public action resembles an iterative policy based on continuous experimentation and certification. The introduction of restorative justice into law in 2014 emerged from a pilot project. This experimental spirit has remained attached to it justice ever since: “every time, they keep saying that it is still in its infancy” (head of a specialized organization, February 2025). A comment by an MP during discussions on the government’s 2023 reform project underscores this characteristic: “The amendment […] entails no obligation. I tabled the same amendment five years ago […] and was told that everyone praised restorative justice and that it needed to be developed. Five years later, the deadlock persists due to a lack of resources.”
This ongoing experiment is subject to constant evaluation. The circular of 15 March 2017 already invited services to assess their projects. An evaluation conducted by SADJAV from spring 2022 extended these efforts to gauge the reception of local programmes, suddenly labeling as “experimental” initiatives and practices that had existed for four years. Paradoxically, this marked both the overall support of the institution for implementing the 2014 law and a broad executive delegation of experimentation to a specialized organization, which became the principal intermediary for the policy.
The precariousness inherent in this policy confirms the hypothesis of a responsibility-transfer phenomenon: primary responsibility for policy implementation is assigned to lower levels. This enables institutions to “treat technically” what has not been settled politically.Footnote 63 To understand action in the field of restorative justice, one must analyze its intermediaries, in line with research on chains of intermediation in public action.Footnote 64
The involvement of judicial authorities varies by locality. The recently adopted justice reform law aims to strengthen this involvement: “The Ministry commits to supporting the conclusion of a greater number of local agreements between courts, victim support organizations, prison services, youth protection services, and bar organizations. The objective is that by 2027, all 164 courts will be covered by such an agreement.”Footnote 65
However, local agreements have not proliferated. This weak institutionalization invites reflection on the place of restorative dialogues within criminal procedures, as well as on its position at the margins of the judicial institution. Is the magistrate—the key actor in criminal proceedings—an obstacle?
Judicial control
Magistrates independently define judicial policies in France, although only a politically engaged minority—seconded to the Ministry of Justice—participates in the formulation of public policy.Footnote 66 Restorative justice has thus been placed under the control of judges and prosecutors, unlike in Quebec, where it was kept outside the judicial system.Footnote 67 The ministry funds non-governmental organizations, magistrates oversee the appropriateness of measures, while the prison-administration and youth-protection services employ most facilitators. Consequently, judges have been entrusted with the power to implement—or to block—a policy embedded within the judicial logic. Yet, restorative justice challenges certain foundational principles and hierarchies within the judicial institution. Rather than defining wrongdoing exclusively through legal categories, it prioritizes a redefinition by those directly involved, thereby responding to what Nils Christie famously termed the “theft of conflict” by professionalsFootnote 68 and correspondingly reducing the centrality of magistrates. Should we infer that it fails to develop because magistrates are reluctant to implement it, fearing interference with their prerogatives or acting on principle?Footnote 69
Their resistance may manifest in two ways. The first is through public criticism, which has intensified over the period under study. However, restorative justice is not the target of these statements. The Syndicat de la Magistrature—a professional organization of judges positioned on the political left—has even expressed support for offender–victim interactions in cases of sexual violence; other unions have remained silent.
“Blame,” to borrow Weaver’s term, could take a second form: active obstruction in daily procedures. This has been observed in other mechanisms addressing domestic violence, such as protective orders for victims, which family-court judges have reinterpreted to allow compromises between parties specific to civil law, to the detriment of female victims.Footnote 70 Although magistrates are only required to exercise the legal oversight of restorative measures, their attitudes vary considerably in practice.Footnote 71 Some express concerns about losing judicial authority, even though restorative justice may appear to them as a useful interim measure in cases in which proceedings are dismissed. The judiciary is a heterogeneous and differentiated entity,Footnote 72 in which positions on public policy reforms vary according to institutional roles. Divergences also arise regarding offences deemed unsuitable for restorative measures. The Chancellery has expressed “reluctance” to implement it prior to a judge’s decision (a CNJR member, April 2021) and in certain contexts such as domestic violence and acts of terrorism.
Influenced by the #MeToo movement, legal professionals remain divided between criticizing existing jurisprudence and defending doctrinal inertia.Footnote 73 The latter translates into denunciations of the “victim ideology” that restorative justice may embody.Footnote 74 Despite the leadership of the Ministry of Justice during this pilot phase, prosecutors maintain their opposition to mediation in cases of domestic violence. Judges are not under ministerial authority; magistrates tasked with supervising the policy are sometimes hierarchically subordinate to their local counterparts.
We have observed the tensions that may arise between restorative measures and judicial decisions. Judges can obstruct implementation, for example, when one refuses to lift a restraining order preventing rapprochement—and thus dialogue—between two consenting parties or when another judge takes the initiative to inform a victim of a mediation process initiated by an offender (observations during project group meetings). These contradictions illustrate the broader uncertainty surrounding the relationship between restorative justice and criminal procedure. Such alternative practices could still be integrated into the standard criminal and carceral system. Restorative justice struggles to develop in France due to these minor obstacles and magistrates’ lack of interest. Yet, if inertia prevails, it is also because of the government’s “laissez-faire” attitude.
Re-motivating subordinates
If the policy can be slowed by magistrates, it is above all left to the main actors involved in its deployment: youth-protection and prison-administration officers, as well as employees and volunteers from victim-support organizations. These actors implement it through a multitude of small local projects that remain precarious, as they are always liable to be interrupted by the transfer of a civil servant to another post, as permitted under French public service law.
We must follow the chain of public action to its very end—that is, to the individuals at the bottom of the judicial hierarchy who implement the measures—to reveal a distinctive feature of the responsibility-transfer process: it is not merely a negative avoidance of responsibility, but also a positive form of personnel management. The role of “street-level bureaucrats” in the application of rights has been widely analyzed.Footnote 75 The paradoxical situation of restorative justice in France—actively promoted yet not fully implemented—can be explained by the role played by these civil servants. While it has not succeeded in creating an ad-hoc professional group, it has been appropriated by those involved in the daily construction of public action.
Field observations reveal the strong involvement of those responsible for sentence supervision and victim support. They lack a dedicated budget and the time they can devote to this work is extremely limited. The dialogues are time-intensive, as each measure is preceded by a lengthy process of individual preparation for participants. Implementation often extends into evenings and weekends—effectively “overtime” (a facilitator during a project group meeting). We observed facilitators transporting participants in their own vehicles, purchasing pastries with their own money to share during sessions and similar practices.
In Belgium, restorative justice initially took shape in prisons in the early 2000s as a “partition” left to the free appropriation of consultants whose role was created alongside the principle itself, deployed through a cascading delegation process.Footnote 76 Similarly, the French judicial institution has turned it into an “art of governing as little as possible.” Yet, it has done so without creating a new profession or reorganizing compartmentalized services. Restorative justice thus materializes differently across the country, depending on the way in which various coalitions interpret it. A “do-it-yourself” approach prevails: “They would take a budget from drug treatment, then from something else […]. In fact, each time, it was about cobbling things together […] depending on the convictions of the management” (head of a specialized organization, February 2025).
This political art of “laissez-faire” employed is also an art of allowing criticism to flourish. The “small civil servants” who invest in restorative justice find renewed energy in it, describing it as “food for the soul.”Footnote 77 Through their involvement and among themselves, they express frustration with a judicial system that, reframed through the lens of New Public Management, manages flows and optimizes staff time in an inhumane manner. They are also the least convinced by technological innovations and the most committed to maintaining dialogue with litigants, which they regard as the strength and quality of their profession.Footnote 78 Delegation to these subordinate civil servants and semi-public organizations’ workers helps to overcome resistance in a context of internal tensions within the civil service. These workers are thus led, while continuing to provide the expected service of flow management, to convert this critical frustration into overtime.
Such delegation can be an effective means of steering policy, as Alexis Spire has shown in the case of French immigration policy.Footnote 79 He describes it as a “policy of administrative counters,” in which rights are asserted upstream but rendered inoperative downstream by subordinate administrative agents tasked with neutralizing them. Creating vague frameworks and delegating practical implementation to frontline agents constitute a “ruse of implementation” that shifts responsibility onto street-level actors,Footnote 80 thereby diluting perceptions of accountability and deflecting discontent.
The constant improvisation required of judicial staff has had a comparable effect: it confines critical actors to a marginal, tightly controlled space marked by “stretched” time. In doing so, restorative justice—alongside the exhaustion of those who implement it—contributes to the evaporation of criticism of dysfunction and the punitive nature of criminal policy, as well as to the concealment of the structural dimension of GBV.
Conclusion
Restorative justice in France illustrates how a right can be formally enacted, yet only minimally realized. Despite political endorsement and growing public visibility, its implementation remains marginal. Our analysis shows that this limited diffusion is primarily the result not of feminist resistance or public skepticism, but of institutional caution, blame‑avoidance strategies and the downward delegation of responsibility to under‑resourced local actors.
Kept in a quasi‑experimental mode, restorative justice depends less on sustained public investment than on the discretionary, often overtime labour of frontline practitioners. This governance configuration reframes a statutory right as an optional service, shifting responsibility onto victims while leaving penal orientations largely untouched. Rather than transforming the response to gender‑based violence, restorative justice has been incorporated as a marginal complement that preserves institutional hierarchies and obscures the structural nature of GBV. This analysis highlights how states produce symbolic rights, the implementation of which ultimately depends on subordinate workers.
Open access
