“Le procès Paradine”, Hitchcock

Review Society

The Vagaries of the Judiciary

About: Arnaud Philippe, La fabrique des jugements. Comment sont déterminées les sanctions pénales, La Découverte

by Rachel Vanneuville , 11 July
translated by Arianne Dorval

How can we explain the variations between judicial decisions in criminal matters? Using the method of natural experiment, economist Arnaud Philippe sets out to identify the factors that influence decisions and determine criminal sanctions. At the risk of forgetting sociology?

Why is simple theft punished by a prison sentence in some cases and by a fine in others? How can we explain what may appear to be unequal treatment or even laxity of justice? In La fabrique des jugements (The making of judicial decisions), Arnaud Philippe proposes to elucidate how decisions are made in criminal matters in France by uncovering the factors involved in their determination.

The originality of the work lies in the analytical approach adopted. The author is a specialist in “empirical economics” (p. 14), a research current that relies on the testing of hypotheses through experimentation. While experiments can be created from scratch, the protocol used here is based instead on “natural experiments” (p. 15). These are defined as “situations that make it possible to compare fairly similar groups, only one of which is subjected to a ‘shock’ that modifies some of the parameters” of interest to researchers (p. 15). The book proposes several natural experiments using different types of shock. The “shock” can consist of a change in the law that sets new penalties for a specific group of offenses: Here, one can compare the sentences given before and after the change to understand how judges have integrated the law. Another type of “shock” is a change in the context in which judges operate (construction of a local prison, media coverage of a crime, etc.). The “shock” can also consist of a change in judgment modalities via the introduction of a new profile of judges (women, citizens, etc.). All of the natural experiments presented in the book aim to quantitatively measure the impact of changes on the severity of judgments. This method meets a number of scientific requirements: the availability of quantitative data, the formulation of hypotheses regarding “causal links” (p. 14), and the testing of these hypotheses through natural experiments. What knowledge about criminal decisions can this method produce?

A Focus on Imprisonment

One of the difficulties faced by those who work on the French judicial system lies in the nature of the available statistical data. These data come from a variety of sources (Ministry of the Interior, Ministry of Justice, etc.), which do not use the same counting units and are therefore difficult to cross-reference. They are also the subject of political battles, particularly when it comes to measuring “delinquency” and the institutional responses to it. It is therefore unfortunate, in a context with which the author is no doubt familiar (as Chapter 3 demonstrates), that the choice to use the national criminal record as the main database is not further explained. This choice “forces” (p. 20) Philippe to focus on prison sentences and to leave aside criminal proceedings that are not documented in the database and primarily concern minor offenses.

It is thus mainly through the lens of imprisonment that the author assesses, in the first part of the book, the impact of “legal and institutional” factors (p. 29) on criminal decisions. The first two chapters are devoted to the proliferation of criminal laws in the 2000s, a process that entailed the creation of new offenses and an increase in the severity of sanctions (in particular with the introduction of minimum sentences against repeat offenders in 2007). [1] These legislative changes have resulted in an overall rise in the number of prison sentences and in heavier penalties for traffic offenses (via the criminalization of offenses for cases that did not go through the courts prior to 2004) and repeat offenses (the increase in sentences that followed the introduction of minimum sentences continued even after the system was repealed in October 2014). However, the author concludes that the changes introduced in the penal sphere and the creation of new offenses have had “virtually no impact” (p. 117), except in the area of political communication.

This conclusion is somewhat surprising, not least because Philippe points out that the laws in question have broadened the scope of action given to law enforcement officers. This has clearly fueled the process described in Chapter 3, which is devoted to the setting of numerical targets for police from the early 2000s onwards. The “policy of targets” has led law enforcement officers to focus on minor offenses and has thereby increased the workload of the judiciary, itself subject to the injunction to prosecute more cases (and consequently to keep more cases in the criminal justice system). In order to manage the influx of cases and improve the penal response rate, the institution has developed alternatives to prosecution before the courts. [2]

The Forgotten Judicial System

These alternative proceedings are not, for the most part, recorded in the national criminal record, and the author does not analyze their general impact on the determination of judicial decisions. Yet, it seems to me that they can help to shed light on an aspect that is absent from the book: the conditions under which decisions are handed down today. Not only are the courts called upon to make more decisions, but with the level of human resources remaining virtually unchanged, they must also do so more rapidly. Hence the importance of avoiding hearings, and, where one exists, of shortening the time taken to process cases. The author rightly notes that the prerogatives of the public prosecutor’s office (which receives the cases and decides on their procedural direction) have increased since the 2000s. It is doubtful, however, that these changes have remained largely “invisible” (p. 161) to sitting judges, who hold hearings and hand down decisions. Their impact is clearly illustrated by the growing recourse to speedy trial proceedings (comparutions immédiates). [3]

Organizational changes in the judicial system should therefore be taken into account when trying to explain how decisions are made: These changes are indeed both a cause and a consequence of the policies pursued. As a number of studies have shown, the transformation of the penal institution has also been driven by members of the public prosecutor’s office. [4] The problem here is less the database used than the approach selected, for the latter favors an interventionist, exogenous causality [5] that makes it impossible to delve into the depths of decision-making processes. Nevertheless, the first part of the book does provide a good overview of how, since the beginning of the 21st century, punitiveness has increased along with social selectivity and the rate of incarceration.

What “Natural Experiments”?

However, some of the gaps in the available data are clearly problematic. On the one hand, the national criminal record does not document the employment, housing and family situation of defendants (p. 11). As Philippe rightly points out, these elements are particularly important in the determination of sentences, which is based on the principle of individualization. On the other hand, the anonymization of court decisions in French statistics makes it impossible to trace the work of judges. Under these conditions, how can one conduct “natural experiments”—that is, compare groups that are comparable only according to certain criteria and for which a set of data is lacking? How, for example, can one measure the effect of a defendant’s birthday on a judicial decision (after having established in Chapter 5 that “gift giving” on birthdays is a social norm in the West) without taking into account gender, social and ethnic origin, the modalities of decision-making, the presence or not of lawyers, etc.? Alternatively, how can one assess the effect of judges’ gender on their decisions in the absence of data on their social origin or professional career? In the area of criminal justice—but also beyond it [6]—similar cases are not necessarily equivalent. At the very least, more detailed information on the construction of comparisons would have been welcome.

The Factors Influencing Judges

However, one should give credit to the author: In several passages, he insists on the weakness of his analyses or even on the impossibility of conducting an analysis altogether. This is particularly the case in the second, shorter part of the book, which is devoted to judges. Here, the author draws on numerous studies of the judicial system in the United States—where individual data on judges are available—and identifies a mosaic of factors likely to determine judicial decisions independently of the configurations in which they are made. According to these studies, judges are influenced by general “social norms” (such as gift giving on birthdays), local norms (judges adapt their decisions to the jurisdictions in which they sit), and even the media (although professional judges seem less sensitive than jurors to media coverage of crimes and criminal trials) (Chapter 5).

Chapter 6 examines the effect of judges’ social characteristics on their decision making. While the influence of social status has not been addressed due to insufficient data and the effect of ethnic origins has been difficult to measure, a gender effect has been identified: Female judges are more severe with female defendants than their male colleagues. Chapter 7 builds on studies in behavioral economics that reveal the existence of “cognitive biases” in judicial decision-making. Philippe mentions in particular the “systematic modes of reasoning,” which leads “to decisions that always lean in the same direction, without the element considered being deemed relevant” (p. 275). He gives as an example the effect of professional experience: When judges move to a new jurisdiction, they initially make decisions based on the standards in force in their previous place of practice. The author could have connected this result to the studies cited in Chapter 5, which describe the gradual compliance of newly-arrived judges with local norms and attribute this behavior to a learning dynamic. How does this dynamic relate to the analysis of “biases”?

More broadly, this observation raises the question of the epistemological approach underlying the hypotheses to be tested and the interpretation of results. Let us take another example, this time from Chapter 8. Here, the author examines a French experiment conducted between 2011 and 2013, in which citizen assessors were introduced in certain magistrates’ court hearings. Initiated by President Nicolas Sarkozy, this reform was based on the conviction that citizens call for more severe punishment. The experiment allows for comparing the severity of judgments between professional and lay judges. The results indicate that the presence of lay judges makes no difference. To account for this finding, the author offers a series of hypotheses: The assessors were outvoted; they deferred to judges; they shared with judges the same vision of sentencing. The problem is that we do not know what these hypotheses are based on, because the author does not explain the mechanisms governing decision-making: Why do judges appear to be in the dominant position? Why do citizens seem reduced to a position of opposition or conformity? Sociological studies have examined the role played by lay judges in various types of courts. They conclude that in order to understand how lay judges participate in decision-making, one must take into account not only their social characteristics (of which nothing is said in the book), but also the operating space of the jurisdictions in which they sit. [7]

While the book clearly sheds light on the complexity of decision-making in criminal matters, the method employed does not allow to really delve into the “production” of judicial decisions. One could further argue that the book suffers from its failure to determine its target audience. If the aim was “to improve the public’s knowledge of criminal justice” (p. 327), then the author should have clarified the workings of said justice (who does what, when, and how) and should have made some of the methodological explanations more accessible to readers who are not specialists in statistical analysis. On the other hand, if the book was intended for an audience of judges or academics familiar with the judicial system, then he should have provided more in-depth explanations of the methodological protocols and the theoretical underpinnings of the formulated hypotheses, and he should have connected the results to the numerous sociological studies on the topic.

Arnaud Philippe, La fabrique des jugements. Comment sont déterminées les sanctions pénales, Paris, La Découverte, 2022, 342 p., 22 €.

Dossier's Articles

by Rachel Vanneuville, 11 July

To quote this article :

Rachel Vanneuville, « The Vagaries of the Judiciary », Books and Ideas , 11 July 2024. ISSN : 2105-3030. URL :

Nota Bene:

If you want to discuss this essay further, you can send a proposal to the editorial team (redaction at We will get back to you as soon as possible.


[1Minimum sentences (peines plancher) are imposed by the court when the defendant has committed a repeat offense. An offence is deemed to be a repeat offence if it occurs after the same or a similar offence (p. 93).

[2These consist of a set of procedures that lead to a penal response to an offence without a hearing before the criminal court: reminders of the law, therapeutic injunctions, penal mediation, etc.

[3Camille Viennot, “Célérité et justice pénale. L’exemple de la comparution immédiate,” Archives de politique criminelle, 29(1), 2007, pp. 117-143.

[4Benoît Bastard and Christian Mouhanna, Une justice dans l’urgence. Le traitement en temps réel des affaires pénales, Paris, Presses Universitaires de France, 2007.

[5See Sacha Bourgeois-Gironde and Éric Monnet, “Expériences naturelles et causalité en histoire économique. Quels rapports à la théorie et à la temporalité?,” Annales. Histoire, Sciences Sociales, 72(4), 2017, p. 1094: “This type of causality is defined in reference to an idealized research protocol in which a voluntary external intervention leads to an effect that is chronologically posterior to the intervention, independent of the movement of other parameters.”

[6See the lengthy methodological development presented by the team working on family affairs to justify the selection of the types of cases to be compared: Émilie Biland and Sybille Gollac (eds) “Justice et inégalités au prisme des sciences sociales,” Rapport, Mission de Recherche Droit et Justice, 2021, 180 p.

[7Pascale Moulevrier, Jean-Noël Retière and Charles Suaud, “La volonté de juger. Les juges non professionnels du tribunal des baux ruraux, du tribunal pour enfants et de la commission d’indemnisation des victimes d’infractions,” Rapport, Mission de recherche Droit et Justice, 2005, 185 p.

Our partners

© - Any replication forbidden without the explicit consent of the editors. - Mentions légales - webdesign : Abel Poucet