Let’s start at the beginning: what is a research cluster at the Institute for Public Policy (IPP)?
Antoine Bozio: A research cluster at the Institute for Public Policy (IPP) is a small group of researchers working on a specific topic, who can therefore organise themselves to carry out evaluations of certain reforms or public policy measures. The aim is primarily organisational, to enable us to respond promptly to calls for tenders and to share knowledge regarding the data to be used and the scientific literature on the subject. However, this does not mean that the work is compartmentalised, and each cluster is expected to interact with the other clusters. In the case of the police and justice cluster, it is clear that this can be linked to education (young people in contact with the justice system), the labour market (reintegration), socio-fiscal measures (post-release support), and many other areas.
In 2022 and 2023, we formalised the creation of the integration and vocational training and environment clusters. At the start of 2024, a new police and justice cluster is being launched.
Why was this 10th division set up?
Antoine Bozio: The IPP’s aim is to analyse and evaluate all public policies using scientific tools, in particular (though not exclusively) the quantitative methods developed in empirical economics. Establishing a police and justice research centre forms part of this approach, with a view to providing better coverage of these public policies, which are essential to the State’s remit. It became clear to us that three necessary conditions for the creation of a centre had been met: firstly, the existence of a significant research area where research and évaluation work is insufficient; secondly, the underlying data required to carry out such work; and finally, the last essential condition: a group of researchers motivated by this topic. The opening up of administrative data on internal security and justice opens up whole new areas of research to be explored. And with Arnaud, Aurélie and Camille, we had the ideal team to tackle this challenge!
So is the IPP encroaching on the state’s sovereign powers?
Antoine Bozio: Public policies falling within the remit of the state do indeed lag significantly behind in terms of évaluation. Many people believe they do not lend themselves well to this process, as they deal with unique cases. But this overlooks the fact that there are numerous options for security and justice policies, the costs of which to the public purse vary and the outcomes of which may also differ. It is essential that we advance our knowledge in this field so that public debate is informed by scientific and objective research.
Is a public justice policy evaluated in the same way as the effects of a tax increase or tax cut?
Camille Hémet: Yes! The evaluation of policies relating to security or justice generally raises the same methodological questions as those concerning other public policies or fiscal policies: Who is affected? Do the people concerned have specific characteristics that might also influence the outcomes we are interested in? We will therefore need to use évaluation methods similar to those used in other contexts, in order to estimate, as far as possible, causal relationships free from bias.
Aurélie Ouss: Certain aspects, such as empirical approaches, share similarities. The questions asked are often similar: understanding the average impact of judicial decisions, who is affected by these decisions, and under what circumstances. However, there is one important distinction: the justice system deals with individual cases, and at the very heart of the administration of justice lies the aim of delivering justice in relation to specific cases and specific individuals. Furthermore, the objectives of the justice system often go beyond what can be measured in administrative data. It is essential to bear in mind the limitations of what the empirical approaches we are considering can actually reveal, particularly in the field of justice. This requires us to supplement our work with other relevant considerations.
What contribution do the economic sciences make to the justice system and the police?
Camille Hémet: I would say that our main contribution, in relation to issues concerning the Police and justice, amongst others (but not exclusively!), is to provide quantitative analyses, particularly when it comes to the evaluation of public policies. Nevertheless, we are by no means unaware of the vast body of literature from other social sciences (sociology, anthropology, law, criminology, psychology, etc.) which informs our thinking and can help us formulate our hypotheses or consider the variables required in our econometric models. As economists, we tend to seek to quantify a phenomenon (the effect of a policy) across a population that is as representative as possible, which generally requires access to sufficiently detailed individual-level data. However, this is sometimes at the expense of a more nuanced understanding of the underlying mechanisms, which can sometimes be provided by research based on more in-depth qualitative interviews, albeit with small samples. In short, the various social sciences are highly complementary on these topics, as they are on others.
Arnaud Philippe: I tend to distinguish between the social sciences on the basis of the methodologies used rather than the topics addressed. Consequently, it seems to me that the question of the contribution of economics arises in the same way for all subjects. I do not believe that economists are any more or less useful to the study of justice than they are to the study of education or employment; they simply approach these subjects differently from sociologists, anthropologists or historians. Broadly speaking, their approach is more quantitative than qualitative.
Why is there such a reliance on administrative data and its availability?
Aurélie Ouss: To assess the effectiveness of criminal justice policies, it is essential to quantify their impacts. The objectives of these policies are varied and complex, which often makes their evaluation a challenging task. For example, it remains difficult to grasp concepts such as the sense of insecurity or retribution. However, a number of issues, particularly those concerning the impact of policies on judicial costs, the enforcement of sentences and reoffending, can be analysed using administrative data.
The administration of criminal justice generates a vast pool of detailed administrative data. This data is designed to meticulously record judicial decisions at an individual level, such as sentences handed down following a trial or the length of detention. This precision is crucial to ensuring the fair administration of justice. However, this data can also be analysed at an aggregate level to perform an evaluation of the overall impact of these decisions. In this way, researchers can gain a comprehensive overview of judicial pathways, thereby enabling them to accurately evaluate the effectiveness of the policies put in place.
Arnaud Philippe: As in many fields, public authorities collect vast amounts of data that can be used to document certain social phenomena. In the case of the police and justice, they are virtually the only ones to do so, as the state has a monopoly on sovereign functions and some of the behaviours in question are concealed by those who commit them. Whilst certain surveys may document victimisation amongst the population or certain criminal behaviours such as drug use, it is, for example, virtually impossible to measure recidivism without relying on administrative data.
Antoine Bozio: In a country where it is common to criticise the bureaucracy and lack of transparency within the civil service, we should welcome the fundamental shift, which began a few years ago, towards making administrative data available for research and évaluation purposes. It is this movement that is now making it possible to realise the aspiration, expressed by both citizens and policy-makers, to carry out more analysis and evaluation of our public policies.
Aurélie Ouss: In recent years, the judicial system has significantly improved the collection of administrative data and the scope for data matching, thereby paving the way for a better understanding of many aspects of the administration of justice. However, to provide conclusive answers to causal questions regarding the effects of criminal justice policies, it is important to develop empirical strategies aimed at isolating the specific impact of these policies. This requires a dual set of skills: an in-depth understanding of administrative data and measures, as well as a command of applied econometric methodologies, making collaboration between specialist researchers and judicial authorities all the more promising.
Camille Hémet: Finally, I would add that significant efforts are also being made to better document crime and delinquency. This is being done not only through national surveys, but also through the release of some of the administrative data held by the national police and gendarmerie. We very much hope that this momentum will continue, enabling researchers to access more detailed and comprehensive administrative data on internal security issues.
In your view, what are the main challenges today in terms of public policy for the police and justice?
Camille Hémet: I believe that our role as economists is, above all, to answer positive rather than normative questions: what are the effects of existing or proposed policies? What mechanisms are at play? Can we quantify these effects? Can we weigh them against their costs? Answering these questions is a challenge in itself, in terms of methodology and access to relevant data. It is then up to public decision-makers to take our findings on board and consider which policy to implement, in light of their objectives and the estimated effects.
Nevertheless, as researchers, there are certain questions we might naturally turn to. How effective is the support provided by the Police and justice to women who are victims of violence? Does imprisonment in France fulfil its role in facilitating rehabilitation, and how does it compare with other, less commonly used sanctions? Do the resources allocated to the internal security forces and their training enable them to combat crime effectively? Some research has already been carried out on these issues, both in France and abroad, but there is still so much to explore!
Is public debate on these issues less rational and more sensitive and emotional than in other areas?
Camille Hémet: Issues relating to the Police and justice can indeed touch on highly emotionally charged subjects. Nevertheless, it is essential to be able to document the effects of these policies objectively, to assess their successes or failures, and to weigh them against their costs to the public purse. At present, debates on these issues very often lack arguments backed up by quantified evaluations.
Arnaud Philippe: The debate on Police and justice is indeed characterised by a strong emotional element. This stems in part from the high profile of the most serious cases, which feature prominently in the ‘news in brief’ sections of various media outlets, as well as from the high level of politicisation of this issue over the past several years. This visibility of issues relating to crime and delinquency, as well as the public’s perception of them and their incorporation into the discourse of decision-makers, is in itself a subject for study, as we discussed in IPP Policy Brief No. 22 on the effect of media coverage of crimes on judicial decisions.
Aurélie Ouss: Furthermore, the justice system has many objectives, which can often be at odds with one another. For example, the prison service must ensure security by supervising prisoners and help with their reintegration to prevent reoffending. These two objectives can be at odds with one another, and knowing where to draw the line may depend more on one’s personal values than on the effectiveness of the approaches. Such tensions exist in many aspects of judicial decision-making.
Antoine Bozio: There are also very heated debates on economic issues such as social diversity in schools and taxation, to name but two examples. At the IPP, we aim to contribute to public debate, and to this end we prioritise the rigour of our analyses in order to ensure the credibility of their findings. We deliberately refrain from taking the place of political debate, preferring instead to establish facts and causal links. I believe that évaluation can help to calm public debate, particularly when it contributes to the development of a shared assessment that is largely free from dispute.
Police and justice are distinct areas, ministries and sets of data: why bring them together under a single umbrella?
Camille Hémet: If we consider this in terms of public administrations, we can say that these ministries share some common concerns, even if the tools at their disposal differ. For example, protecting the public can be achieved through the actions of law enforcement agencies or through court orders placing particularly vulnerable and at-risk individuals in care or under guardianship. Furthermore, when it comes to combating criminal activity (in the broadest sense), the Police and justice are still two links in the same chain. This is, in fact, reflected in the economic theory of crime developed by the American economist Gary Becker in 1968. According to this theory, the decision to commit a ‘crime’ is the result of a rational ‘cost-benefit’ analysis. And what increases the ‘cost’ of crime? Both the probability of being arrested – which falls within the remit of the police – and the severity of the punishment – which falls within the remit of the judiciary.
Arnaud Philippe: In criminal matters, the judicial process generally follows that of the police. In theory, it is possible to trace individuals through the data, and, increasingly, public authorities are establishing common frameworks to understand – at least in aggregate terms – how different population groups are treated by these two institutions. Furthermore, the debate on security issues is shaped by the relationship between the justice system and the police. The tension between the various stakeholders – whether between ministers or trade unions – is often key to understanding how a particular policy was implemented and how it can be évalué. It therefore seems natural to consider them together, even though these issues obviously also relate to policies on health (particularly mental health), employment, education and so on.
Your work also draws on data, examples and évaluations carried out abroad: in your view, what are the fundamental or historical differences compared with France? What lessons do you draw from this for this emerging centre?
Camille Hémet: If we take the example of the effect of imprisonment on reoffending and rehabilitation, the findings in the quantitative economics literature are largely drawn from experiences abroad, notably in the United States or the Scandinavian countries (not to mention the United Kingdom or Italy), which are very different contexts from one another and can be regarded as extreme cases. For example, the United States has one of the highest incarceration rates in the world (more than 500 people incarcerated per 100,000 inhabitants, compared with 106 per 100,000 in France). Furthermore, Scandinavian prisons are an exception in terms of prison conditions, which are recognised as being among the best in the world, with a very strong emphasis on rehabilitation. The results achieved in these countries are therefore, in principle, difficult to apply to other countries, which reinforces the need to be able to properly evaluate French policies.
Aurélie Ouss: Much of my work focuses on the justice system in the United States. The scale of criminal justice policies is completely different, both in terms of incarceration rates and the lasting impact of a criminal record beyond the criminal justice system, as it is accessible to all employers, landlords, social services and so on… This particular feature limits the applicability of the findings of many studies conducted in the United States to the French context. Another major characteristic lies in the fragmentation of the American judicial system. Indeed, the administration of justice is often decentralised, operating at municipal, departmental or state level. Matching therefore poses significant challenges due to the multitude of administrative authorities involved.
There is therefore a lack of convincing studies on the links between criminal justice policies and areas such as employment, education and health. In France, the structure of administrative data offers opportunities to address, in the long term, these questions concerning the interactions between criminal justice and social policies. This would open up new and essential avenues for gaining a more comprehensive understanding of the effectiveness and equity of criminal justice and social policies.
Speaking of prospects, what are the next steps for this new centre?
Camille Hémet: The next steps are in line with the objectives outlined by Antoine earlier. These include carrying out research, particularly through calls for proposals, and bringing academic knowledge into the public debate.
I have, in fact, just launched an initial project focusing on the educational and career pathways of young people who have been in prison, which I am carrying out in collaboration with several colleagues (Manon Garrouste, Nina Guyon and Laura Khoury) and two PhD students from PSE (Léa Dousset and Maëlle Stricot). This project was set up following a call for proposals issued by the Defender of Rights, the Directorate for Judicial Protection of Young People (DPJJ) at the Ministry of Justice and the Institute for Studies and Research on Law and Justice (IERDJ). It is a project that presents numerous challenges, particularly in terms of access to administrative data, but which can now get underway thanks to the efforts of the relevant government departments to promote transparency.
As regards the second objective of disseminating knowledge, we will begin by publishing, over the coming weeks and months, several IPP Policy Briefs setting out the results of research carried out by Aurélie, Arnaud and myself, amongst others. The first paper will be published in the middle of next week: it explores the effect of criminal justice policies on court decisions.
In the short term, we also plan to bring together researchers and government bodies for a conference in April: we hope to be able to provide further details (and open registration) shortly.
