1. A Human Rights Crisis

Over the past decade, Mexico has seen a growing number of human rights violations at the hands of state and non-state actors. According to official data reported by Justice in Mexico, the number of intentional homicides has increased steadily since 2015, claiming 34,588 individual victims in 2019 alone.¹ While the government stopped tracking disappearances in 2018,² previous reporting has suggested that the number of disappearances continues to rise each year.³ These findings have been accompanied by further unquantifiable human rights violations, as documented by international organizations and civil society groups. In particular, human rights advocates have noted the sustained prevalence of torture and mistreatment in Mexico.⁴

While substantial reporting by scholars and civil society organizations has underscored the magnitude of the crisis, there is very little publicly available information documenting the prevalence of institutionalized torture as a whole. Mexico’s national human rights ombudsman, the CNDH, registers complaints of torture and cruel, inhuman, and degrading treatment filed against government bodies, but scholars and nongovernmental organizations (NGOs) have noted the contradictory and inconsistent nature of official data on the practice.⁵ This is the case despite the efforts of civil society groups and international organizations documenting the institutionalized use of torture within Mexico.

For example, in 2003, the United Nations Committee Against Torture (UN CAT) released a report illustrating the systematic nature of the practice. The committee examined hundreds of reports of torture in Mexico and found that victims of torture reported eerily similar experiences. Most reported that their torturers forced them to confess to crimes they had not committed, including homicides, kidnappings, robberies, and sexual offenses. Similarly, victims reported nearly identical methods of torture, including electric shocks, asphyxiation, mock executions, and direct threats of harm to family members.⁶

From 2005 to 2007, the CNDH released 4 official recommendations to government organizations based on complaints of torture filed against them. However, from 2008 to 2010, this figure increased to 28 total recommendations. Similarly, the number of complaints of cruel, inhuman, and degrading treatment presented to the CNDH increased during the same time period. At the start of Calderón’s term, in 2006, the commission received 330 total complaints, and by 2010, the figure had increased to 1,161.⁷

In 2014, the UN conducted a second assessment on torture in Mexico, sending Special Rapporteur on Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment Juan E. Méndez to document the practice’s incidence. Méndez reports that torture continues to be “generalized” throughout Mexico, particularly in the context of a growing security crisis. Similar to the 2003 report, Méndez notes that suspects are often detained for alleged links to organized crime and are tortured using common methods. The 2014 UN report also cites Mexico’s continued indifference to the use of forced confessions.⁸

The UN specifically cites the role of Mexico’s public prosecutors in obtaining forced confessions. While judicial police, or other security officials, are typically responsible for carrying out acts of torture, Mexico’s public prosecutors are often complicit in the practice, accepting forced confessions as evidence in their cases. Furthermore, UN reporting found that some public prosecutors were allegedly present while the accused was tortured, and in some cases, the prosecutors sent the accused back to the police to be tortured after they had refused to confess to committing crimes.⁹ Recent scholarly work has also cited widespread failure by judges to identify when acts of torture have taken place in law enforcement custody. In a particular study on criminal cases in “Estado de México” from 2010 to 2014, researchers found that 97% of cases in which injuries consistent with torture were reported by a doctor, judges refused to exclude evidence obtained during detention. In 100% of such cases, judges failed to liberate the accused from detention.¹⁰

Nongovernmental human rights organizations have substantiated these findings for years, documenting the cases of torture and forced confessions in detail, albeit with limited access to official data. The PRODH Center (Centro de Derechos Humanos Miguel Agustín Pro Juárez A.C) first alerted the UN CAT of these abuses in 1998 and has since released dozens of reports documenting the institutionalized practice of torture and mistreatment. The PRODH Center argues that the practice has become a modus operandi within Mexico’s military and security institutions, particularly within the army; the navy; and police forces at all levels of government. Consistent with UN and Human Rights Watch reporting, PRODH documents government officials detaining and torturing suspects for the purpose of extracting coerced confessions.¹¹

2. A Brief History of Torture in Mexico

The use of torture has been a recurring phenomenon throughout Mexico’s history, from the conquest to the present. However, Mexico is not unique in its use of this abusive practice-both historically and in recent years. As Rejali warns, once the use of torture is legitimized by a state, the corrosive practice roots itself in the judicial, intelligence, and military institutions that employ it, lingering for decades.¹² For instance, for all of its opposition to the practice on the world stage, research has found that the United States propagated torture for decades. Since as early as 1950, the U.S. intelligence community began developing a standard of torture (“interrogation”) techniques to fight communism, which it disseminated to law enforcement and military agencies in Asia, Latin America, and Central America during the Cold War.¹³ After the turn of the century, the United States employed torture as a counter-terrorism measure post-9/11.¹⁴ Thus, while the observed phenomenon in Mexico is grave, it is clear these practices are far from unique.

The historic practice of torture in Mexico closely follows the pattern out-lined by Rejali.¹⁵ As early as the 1920s, torture was exercised as an investigative tool as a form of “energetic interrogation”. Police officers typically employed the practice prior to the criminal indictment to produce a confession, often staging mock executions, administering electric shocks, and directly threating harm to the victim’s family members.¹⁶

During single-party rule under the Institutional Revolutionary Party (Partido Revolucionario Institucional, PRI) from 1929 to 2000, torture was practiced as a means of political and social control. In the 1950s and 1960s, protest movements led by farmers, doctors, railroad workers, professors, and students surged, resulting in a brutal crackdown by the PRI-controlled state. During this time, the government illegally detained, forcibly disappeared, and tortured hundreds, if not thousands, of citizens who were thought to threaten the stability enjoyed during the previous decades.¹⁷

In Mexico’s case, current methods of torture employed by law enforcement are strikingly similar to past methods. As Piccato explains, police investigators in the 1920s employed torture for the same reason that law enforcement officers employ the practice today; if police could obtain a confession of guilt, other forms of investigation became unnecessary, and the officers could successfully close the case.¹⁸ In recent years, public prosecutors have seen an increased volume in criminal cases, resulting in fewer than one in five being resolved satisfactorily. This has aggravated the historic pattern observed by Piccatto, increasing the pressures for prosecutors to extract relevant information during the preliminary inquiry stage, often to the detriment of the accused’s human rights.¹⁹

3. Mexico’s “Mixed Inquisitorial” System: The Roots of Abuses by Law Enforcement

To understand how torture became a modus operandi within Mexico’s criminal justice system, it is important to establish the mechanisms that incentivized and sustained the practice. During the post-revolutionary era, Mexico began to depart from more traditional inquisitorial systems of criminal justice, affording new powers to the public prosecutor. These changes were enshrined in the 1908 Organic Law of the Federal Public Prosecutor (Ley Orgánica del Ministerio Público Federal y Reglamentación de sus Funciones), the 1908 and 1917 Organic Law of the Federal Judiciary Branch (Ley Orgánica del Poder Judicial Federal), the 1938 Organic Law of the Federal Attorney General (Ley Orgánica de la Procuraduría General de la República), and numerous subsequent pieces of legislation passed throughout the twentieth century that gradually enhanced the autonomy of the public prosecutor.²⁰

Thus, the practice of torture as a prosecutorial mechanism can be traced to gradual changes within Mexico’s criminal justice system. As such, González-Núñez frames the contemporary practice of torture by Mexican officials in this historic context, reinforced by mechanisms within the country’s previous “mixed inquisitorial” criminal judicial system.²¹ As a result of “procedural immediacy”, or the judicial practice of accepting criminal suspects’ initial statements over subsequent ones, Mexico’s prosecutors and law enforcement bodies were incentivized to use torture as a means to produce confessions. These coerced statements were often accepted as the sole basis for incrimination, reducing the prosecutor’s responsibility to produce objective scientific evidence against the accused.²² Combined with a high degree of autonomy as a result of twentieth century legislation, public prosecutors were able to continue this practice unrestrained.²³

These practices were reinforced by regulations governing criminal detention. Specifically, Article 16 of the 1917 Constitution allows judicial and preventive police to arrest any person caught “in the act” of committing a crime (en flagrante).²⁴ When a suspect is arrested en flagrante, they are handed over to the state or federal public prosecutor. However, the definition of en flagrante was gradually expanded, and in many cases, arrests were made up to seventy-two (72) hours after the crime was allegedly committed.²⁵ This rule allowed police and prosecutors to operate without oversight, increasing the number of criminal suspects in detention. In fact, one study found that arrests en flagrante may have at one time accounted for up to 60% of total arrests in Mexico City.²⁶ Prisoner survey data from 2002 confirms this finding, with 60% of 1,615 randomly sampled prisoners detained in Mexico City, Estado de México, and Morelos reporting having been arrested en flagrante.²⁷

This reliance on detention reinforced police and prosecutorial confessions using torture and mistreatment. In its 2003 report, the UN CAT found that the incidence of torture was highest during the period between detention and committal for trial, when suspects were held at police or public prosecutor’s offices.²⁸ Thus, police and prosecutors possessed not only the incentive to extract criminal confessions, but they were also provided ample opportunity to do so in the context of criminal detention. Indeed, in the same 2002 survey of 1,615 inmates in Mexican prisons, half of the prisoners reported confessing to a crime due to intimidation or torture.²⁹ Thus, on the whole, Mexico’s former “mixed inquisitorial” criminal justice system possessed numerous institutions and procedural elements that reinforced the practice of employing torture and mistreatment to extract confessions.

4. “Revolution in Latin American Criminal Procedure” ³⁰

While the UN and other international organizations began to shed light on these abuses in Mexico, a revolutionary wave of criminal procedure reforms was sweeping Latin America. These reforms included the introduction of oral, public trials and often sought to resolve longstanding issues relating to a lack of transparency and due process.³¹ By the time that the UNCAT released its report on Mexico in 2003, a dozen Latin American countries had already introduced accusatorial criminal codes.³² Langer argues that this series of successive criminal code reforms was largely a result of “peer pressure” on states that had not yet implemented such changes. Specifically, a group of Latin American activists advocated for the adoption of accusatorial systems, along with the help of USAID and other international organizations. This combination of advocacy and pressure by non-state actors played a crucial role in the judicial reform projects of countries like Bolivia, the Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Nicaragua, Paraguay, and Mexico.³³

5. Mexico’s Reform: A Step toward Judicial Accountability

Although Mexico’s national reform project began in 2008, the judicial reform movement began at a subnational level, with some states transitioning to an accusatorial system as early as 2004.³⁴ Following state-level initiatives, the Mexican Congress passed the 2008 constitutional reform that would seek to strengthen transparency, due process, and accountability throughout the criminal process, transforming Mexican criminal procedure from the traditional “mixed inquisitorial” model to an oral adversarial system.³⁵

The previous system was based in civil law traditions descended from Europe rather than the common law systems of the U.S., British, and Australian judiciaries.³⁶

Nonetheless, Mexico’s “mixed inquisitorial” model differed in several key areas from its ancestral European systems. Throughout the twentieth century, Mexico gradually adopted practices that expanded the role of the prosecutor. Consequently, the public prosecutor began overseeing large portions of the criminal process, including police and detective work during the investigation. The prosecutor also maintained a central role during the accusatory phase, particularly as the defense possessed limited ability to challenge prosecutorial evidence or arguments during the trial and sentencing. Furthermore, it was not uncommon for judges to base sentences exclusively on evidence presented by the prosecutor, resulting in more frequent “guilty” verdicts. This practice was compounded by the fact that the sentencing judge was often the same judge that initially found sufficient cause to proceed with a criminal investigation against the accused.³⁷

The 2008 reform sought to realign many of the aforementioned imbalances in favor of a system that allowed the prosecution and defense to engage in oral, adversarial argument. The reform introduced the Accusatorial Criminal Justice System ( Sistema de Justicia Penal Acusatorio, SJPA) that would institute oral, adversarial criminal trials; alternative sentencing; and alternative dispute resolution mechanisms (ADRs). The introduction of ADRs was meant to relieve congestion in Mexico’s penal system, allowing for increased capacity to appropriately follow procedure. The SJPA would also afford stronger rights to those accused of crimes through the presumption of innocence, proper due process, and adequate legal defense. Lastly, the reform would seek to alter the roles of police and prosecutors under the traditional system.³⁸

Specifically, the reform introduced a procedure that would establish probable cause as the basis for criminal indictment. By reducing the threshold of evidence required for a criminal indictment, the reform limited the public prosecutor’s previously dominant role over the preliminary administrative phase of the criminal proceeding, or the averiguación previa. This diminished the public prosecutor’s incentives to produce an immediate criminal confession, as testimonies and declarations to be considered as evidence would have to be presented later in the criminal process before a judge at trial.³⁹

Under the new system, the axis of oversight shifted from the public prosecutor to the judge, who became responsible for monitoring police and prosecutor activities throughout all stages of the criminal proceeding.⁴⁰ This structural shift was accompanied by an explicit prohibition of the use of torture to produce confessions during pre -trial detention, providing judges a mechanism to dismiss cases when torture is suspected.⁴¹

6. Reducing Torture: Institutions, Incentives, and Norms

To understand the widespread use of torture in Mexico, it is necessary to first examine the political environments in which states employ this form of abuse. According to Wantchekon and Healy, illiberal and liberal states practice torture for different reasons. While illiberal states, such as dictatorships, use torture and mistreatment as a means of social control, liberal states only employ torture to extract information.⁴² Luban identifies specific motivations within these broader categories, citing one reason why liberal states employ torture and four reasons why illiberal states do so.⁴³

Specifically, Luban argues that illiberal states may utilize torture in the context of military victory (what Luban deems “victor’s pleasure”), to incite terror, to punish alleged criminals, and finally, to extract confessions. In this last scenario, actors within the criminal justice system employ the practice as a result of institutionalized norms establishing the legitimacy of confessions as culpatory evidence. Meanwhile, liberal states typically torture in a scenario termed “the ticking bomb”.⁴⁴ In this case, the state employs torture to gather intelligence to prevent future evils, such as terrorist attacks.

However, distinctions based on regime type provide a limited explanation of Mexico’s state-sanctioned torture. While this literature requires the characterization of regimes as dichotomous (illiberal versus liberal), most scholarly work acknowledges that states fall on a continuum from fully authoritarian to fully democratic.⁴⁵ In Mexico’s case, most agree that democracy is hardly a finished project.⁴⁶ While the country has managed to adopt promising frameworks in support of democratic reform, the implementation of such mechanisms often lags behind.⁴⁷ In part, this has resulted in growing concerns regarding human rights abuses, impunity rates, and absences in the rule of law. As Levy, Bruhn, and Zebadúa write, “Mexico’s road toward democratization is lined with potholes, red lights, yellow lights, wrong turns, and very disputed speed limits”.⁴⁸ While Mexico has made major strides toward the consolidation of its democracy since 2000, the country still faces obstacles ahead. As a result, it is somewhat fruitless to classify Mexico’s state-sanctioned torture as fully “illiberal” or fully “liberal”, according to Luban’s framework.

Additionally, scholars have found certain exceptions to democratic states’ behavior. Indeed, previous literature has found that the effect of democratic institutions on reducing torture diminishes when the state is faced with “violent dissent”.⁴⁹ As Gambetta writes, “the bigger and nastier the threat is (or is thought to be) the harsher are the infringements on civil liberties that can be justified and accepted by the public”. ⁵⁰ In other words, the political checks and balances that typically prevent the executive from committing or sanctioning acts of torture tend to erode in the face of violent threat. While this work largely examines the role that terrorist groups play in creating this “violent dissent” in democratic polities,⁵¹ Magaloni & Rodriguez apply this line of reasoning to Mexico, arguing that the activities of criminal organizations have resulted in harsh repression by the state.⁵² Since Mexico’s democratic opening, the country has faced growing levels of insecurity as a result of these criminal groups.⁵³ In response, the government increased its militarized counter-drug operations against trafficking organizations under President Felipe Calderon, resulting in increased levels of violence.⁵⁴ Magaloni, Magaloni, & Razu present empirical evidence demonstrating increased levels of torture during this time period, particularly when criminal suspects were detained or accused of drug trafficking.⁵⁵

However, Davenport, Moore, and Armstrong identify a mediating variable that may predict a state’s repressive response to violent threats. They argue that governments that possess “veto”, or constraints on an executive’s authority as a result of the separation of powers, are less likely to employ torture as a repressive response. Polities with high levels of veto necessarily contain incentive structures that push actors to challenge an executive’s use of torture. These states will consist of competitive legislatures and independent judiciaries, including at the subnational level. As the authors demonstrate, the greater the level of separation of powers, the greater the likelihood that any actor will expose the executive’s use of torture. This acts as an implicit check on the executive’s potential responses to violent threats, reducing the likelihood of torture.⁵⁶

This work is in line with existing literature demonstrating how institutions can restrain state behavior.⁵⁷ As Walker argues, democratic institutions provide the structure for autonomous political actors to pursue their individual interests. This structure includes both incentives and restrictions that guide actors’ behavior.⁵⁸ Mexico’s criminal justice reform seeks to provide such a structure to reshape the behavior of government actors, albeit in a challenging security environment. As outlined above, the reform provides stronger counterweights to the role that the prosecutor played in the former system, increasing the veto power of other judicial actors, such as judges, and reshaping the incentive structures that drive the behavior of these actors. As Zepeda Lecuona explains, reductions in torture cannot be explained by mere changes in attitude; rather, reductions in these types of abuses are the result of changes to the incentive structures that influence how judicial actors operate within the system.⁵⁹ Even in the face of Mexico’s mounting security challenges, empirical evidence suggests that reforms introducing such changes may have a significant effect on state actors’ repressive behaviors.

In addition to institutionalized structures that shape actors’ behavior through a system of incentives (or disincentives), Langer points to the relevance of norms in the context of judicial reform. Specifically, the diffusion of Latin American criminal procedure reforms was accompanied by a corresponding diffusion of norms that aimed to increase transparency and accountability of judicial actors-particularly as public security and due process concerns arose throughout the region at the end of the twentieth century.⁶⁰

An ample body of literature has demonstrated the power of these types of norms on the behavior of state actors⁶¹ and their impact on the behavior of judicial actors, such as police. As Worden and McLean point out, policing is a task that involves a high degree of uncertainty, and as a result, officers must frequently make choices with ambiguous implications and consequences.⁶² In other words, there is no procedural manual that explains in full how policing should be conducted. To fill this information void, police often rely on institutional norms in order to make daily decisions⁶³-from the volume of citations issued to the manner in which interrogations are conducted.⁶⁴ As this analysis features police behavior as a primary subject of investigation, it will consider how Mexico’s judicial reform may have reshaped both institutionalized incentive structures and more informal conduct norms, impacting the incidence of torture and mistreatment.

7. Evaluating the Reform: Challenges and Achievements

The success of democratic reforms in reducing the incidence of torture is dependent on the comprehensive implementation of such reforms. As Zepeda Lecuona argues, “80% of the criminal reform’s success lies in its implementation” [own translation].⁶⁵ As such, this section discusses the trajectory of Mexico’s national reform since its 2008 enactment, examining its successes, weaknesses, and existing challenges to full implementation.

While the constitutional reform was passed in 2008, certain states approved and began the use of oral adversarial proceedings as early as 2004 (Nuevo León in 2004, Chihuahua in 2007, and Oaxaca in 2007). These initiatives provided a precedent for other states’ penal reforms and served as precursors to the constitutional reform.⁶⁶ Thus, in 2008, the Mexican Congress provided the country with an eight-year timeframe to fully implement the changes outlined in the reform. While the deadline of June 18, 2016 has long passed, Mexico’s judicial districts are still in the process of implementing and consolidating these sweeping changes. As México Evalúa noted in its most recent performance review of the SJPA, there is still much work to be done. Specifically, judicial training and professionalization efforts have diminished since the implementation phase (2008-2016), and these efforts often lack inter-institutional coordination that could result in more profound improvements to SJPA functioning. México Evalúa also argues that judicial actors lack the resources and training necessary to conduct thorough criminal investigations that would produce legitimate evidence to be presented in criminal trials. Lastly, the report calls attention to the lack of statistical information that is shared across judicial agencies and with the public. As the authors note, this information void has made evaluation of the SJPA’s performance a burdensome task.⁶⁷

Despite challenges in establishing and implementing proper metrics for the reform, there is early evidence suggesting that in certain states, police and prosecutors continue to employ illegal interrogation practices under the reformed system. In a 2016 of Estado de México’s criminal justice system, Fondevila and colleagues found that between 2010 and 2014,⁶⁸ 18% of legal proceedings analyzed contained statements from a medical professional documenting injuries consistent with torture and/or mistreatment.⁶⁹ Furthermore, the authors found that just 10.3% of individuals for whom a legal proceeding was initiated had access to a defense attorney during their time at the public prosecutor’s office.⁷⁰ Combined with the results of this analysis, Fondevila et al.’s findings suggest that SJPA outcomes may vary significantly at the subnational level, depending on a variety of institutional factors out-lined above. As such, this article examines torture and mistreatment at both national and state levels.

Despite methodological challenges and pessimistic findings at the state level, researchers have been able to demonstrate the reform’s overall positive impact on the incidence of human rights abuses in Mexico. For instance, World Justice Project (WJP) presented data demonstrating a marked difference in the incidence of forced confessions between states that implemented the reform between 2007 and 2012 (Baja California, Chihuahua, Estado de México, Morelos, Guanajuato, Oaxaca, Yucatán, and Zacatecas) and states that implemented the reform after 2012. Specifically, WJP reported that from 2005 to 2016, early implementer states observed a 70% decrease in the number of confessions that were the result of pressure or aggression, while all other states cumulatively observed a 34% decrease during the same period.⁷¹

Magaloni and Rodriguez produce similar findings by analyzing data from the National Survey of the Population Deprived of Liberty (Encuesta Nacional de Población Privada de la Libertad, ENPOL), a survey of 58,127 individuals that were imprisoned in Mexico in 2016.⁷² Magaloni and Rodriguez examined prisoners’ reports of torture (e.g., electric shocks, burns, sexual abuse) and compared reports of individuals arrested before and after the implementation of the National Code of Criminal Procedure (Código Nacional de Procedimientos Penales, CNPP). This code was implemented in a staggered fashion across Mexico’s judicial districts from 2014 to 2016 and standardized criminal procedure under the reform at both state and federal levels.⁷³ The researchers employed 65 distinct dates of implementation to capture a more localized effect of the CNPP. The findings demonstrated statistically significant declines in the reported incidence of torture and threats in the period after implementation.⁷⁴

While initial research suggests that the incidence of torture has decreased since the implementation of the reform, further analysis is needed to confirm the reform’s impact on the incidence of forced confessions. WJP has presented preliminary data supporting the connection between SJPA implementation and a reduced incidence of forced confessions. However, a judicial district-level analysis of these figures pre- and post- reform has yet to be conducted. As such, this study seeks to build on both the findings of WJP and Magaloni and Rodriguez regarding torture, while also providing evidence of the reform’s impact on the use of forced confessions as a pros-ecutorial tool at the level of implementation.

8. A “Disturbing Imbalance”: Criminal Detention under the SJPA

Indeed, the SJPA represents a paradigm shift toward a criminal justice system more sensitive to principals of accountability, transparency, and human rights. Still, the reform contains certain measures that have remained controversial among human rights advocates-namely, the continuance of arraigo. Arraigo is a form a preventive detention that does not require criminal charges. As such, the practice defies the principle of presumption of innocence in Mexico’s criminal justice system.⁷⁵ As Zepeda Lecuona argues, the extension of arraigo under the SJPA represents a “disturbing imbalance” in Mexico’s criminal justice system, as it reduces the standard required to subject an individual to the criminal process [own translation].⁷⁶ While public prosecutors are normally required to present evidence before a judge establishing the need for a criminal suspect’s detention, arraigo suppresses this requirement. Instead, the prosecutor need only demonstrate the possibility of the suspect’s involvement in certain criminal activities.

In the SJPA’s current format, arraigo is now restricted to cases involving organized crime; however, detention is allowed for a continuous period of forty (40) days, which can be extended for up to eighty (80) days.⁷⁷ As previously discussed, the reliance on detention in Mexico’s criminal proceedings has served to reinforce the practice of torture and mistreatment by police and public prosecutors by providing ample opportunity for such acts to occur. In fact, evidence suggests that public prosecutors may intentionally classify certain criminal acts under the umbrella of organized crime to allow for a suspect’s detention. Under this procedure, the prosecutor is then permitted to introduce evidence that has not been formally reviewed and sanctioned during a criminal trial-a step required for all other criminal evidence under the SJPA.⁷⁸ This only serves to dismantle the reform’s incentive structures meant to restrict prosecutorial abuses, such as torture and forced confessions, in the context of criminal detention.

As such, arraigo’s presence in the reformed system is the ultimate paradox; its existence sabotages the very reforms meant to curb judicial misconduct and human rights abuses. Indeed, substantial reporting has confirmed the link between arraigo and increases in reports of torture and forced confessions.⁷⁹ In its most recent review of Mexico, the UNCAT urged the country to permanently halt the use of arraigo in order to reduce the incidence of torture and forced confessions during this type of detention.

Thus, despite the introduction of a sweeping criminal justice reform, institutions that reinforce the use of torture and forced confessions continue to exist within Mexico’s judicial system. As such, in its current form, the SJPA is not a silver bullet capable of abolishing the practice of torture and mistreatment. Substantial opportunity for reform still exists, particularly in the realm of criminal detention. Still, the reform represents a significant step toward the consolidation of Mexico’s democratic institutions and toward the implementation of prosecutorial accountability measures. While far from a complete solution, initial research demonstrates the link between the reform and observed reductions in torture and mistreatment. As such, this study seeks to provide further evidence of the reform’s positive impact on Mexico’s human rights paradigm.

1. Defining Torture and Cruel, Inhuman, or Degrading Treatment

To allow for effective comparison with existing literature, this analysis employs the common definition of torture as outlined in the UN Convention Against Torture and Other Cruel, Inhuman, or Degrading Treatment or Punishment, adopted by the UN General Assembly in 1984:

...any act by which severe pain or suffering, whether physical or mental, is intentionally inflicted on a person for such purposes as obtaining from him or a third person information or a confession, punishing him for an act he or a third person has committed or is suspected of having committed, or intimidating or coercing him or a third person, or for any reason based on discrimination of any kind, when such pain or suffering is inflicted by or at the instigation of or with the consent or acquiescence of a public official or other person acting in an official capacity. It does not include pain or suffering arising only from, inherent in or incidental to lawful sanctions⁸⁰ [emphasis added].

According to this definition, torture encompasses harm inflicted for one of the following explicit purposes: (1) extraction of information or confessions, (2) punishment, or (3) intimidation or discrimination. Furthermore, torture is always carried out with the “consent or acquiescence” of state officials or anyone acting in an official state capacity.⁸¹ While torture has historically been used for all three purposes in Mexico, as outlined above, this study will examine the first use of torture defined under the convention: torture as form of extracting information or confessions.⁸²

Although the convention outlines specific criteria for torture, it does not provide a definition for “other cruel, inhuman, or degrading treatment or punishment”. Consequently, scholars have debated the degree to which these two acts differ. Some argue that the severity of suffering is greater for acts of torture, while others maintain that the threshold for severity of suffering is equal, but that the purpose of the acts themselves differ. Nonetheless, sub-stantial research demonstrates that victims of acts typically defined as “other cruel, inhuman, or degrading treatment or punishment,” such as humiliation, fear, and threats of torture, experience similar levels of psychological pain and suffering as victims of torture. Consequently, there is reason to question the separation of these terms in international and domestic law, as the distinction may imply that “other cruel, inhuman, or degrading treatment or punishment” is a less severe form of torture.⁸³

In its official database of human rights complaints, the CNDH considers torture and “other cruel, inhuman, or degrading treatment or punishment” to be two separate violations. According to Mexico’s Office of Domestic Affairs (Secretaría de Gobernación, SEGOB), the difference in classification may lie in the severity of suffering.⁸⁴ Nonetheless, SEGOB also notes that the Inter-American Convention to Prevent and Punish Torture (IACPPT) specifies that acts do not have to cause grave suffering in order to be classified as torture.⁸⁵ SEGOB concludes that each case must be analyzed individually in order to determine its proper classification. As such, it is not fully known how the CNDH distinguishes between these types of human rights violations.

However, there is evidence to suggest that officials intentionally classify cases of torture as “other cruel, inhuman, or degrading treatment or punishment” in order to reduce the perceived severity of certain incidents. In 2003, the UNCAT reported that police often threaten and beat suspects prior to their arrival at the Public Prosecutor’s office. While many of these cases meet the constitutional threshold for torture, they are frequently categorized as cases of “other cruel, inhuman, or degrading treatment or punishment.”⁸⁶ Thus, the distinction between these cases in Mexico likely fails to capture any difference in the severity of abuse. Consequently, this study takes a comprehensive approach, examining both types of abuse in the context of the judicial reform in Mexico.⁸⁷

2. National Commission of Human Rights Alert System

This analysis first examines torture and mistreatment using a hand-compiled database of torture complaints published by the CNDH on its National Human Rights Violation Alert System (Sistema Nacional de Alerta de Violación a los Derechos Humanos).⁸⁸ It includes complaints filed with the CNDH from January 2014 to December 2019 against institutions at all levels of government (municipal, state, and federal).⁸⁹ Each complaint is classified by the state in which the individual was arrested and also by one of six institutional categories classifying the type of government agency implicated in the report. These categories include (1) public security forces (e.g., police), (2) military, (3) public prosecutor’s offices, (4) penitentiaries, (5) municipal agencies, and (6) “other” institutions.⁹⁰ Additionally, a portion of the complaints included in this data set do not identify an institution responsible for the reported violation (listed at “N/D” in Figure 1).

Figure 1 CNDH Torture & Mistreatment Complaints by Institution Type (Jan. 2014-Dec. 2019)⁹¹  

Of the institutional categories outlined in Figure 1, three operate directly within Mexico’s criminal justice system: Public Security Forces, Public Pros-ecutor’s Offices, and Penitentiaries. Together, these government bodies represent a majority of cases of torture reported to CNDH from 2014 to 2019, with 2,275 of 4,072 total complaints (see Figure 2).

Figure 2 CNDH Torture & Mistreatment Complaints Against Judicial Versus Other Institutions (Jan. 2014-Aug. 2019)⁹²  

In order to examine how the judicial reform impacted the incidence of torture and mistreatment complaints made to CNDH, this analysis exclusively examines cases in which judicial institutions were reported to be responsible for alleged abuse(s) (n = 1,669).⁹³ As previously mentioned, these include complaints implicating Public Security Institutions, Public Prosecutor’s Offices, and Penitentiaries. These cases were summed by state (n= 32) for each year that indicators were available (2015-2018). Next, to compare data geographically, CONAPO (Consejo Nacional de Población) state population estimates were retrieved for each year.⁹⁴ These figures were used to calculate complaints of torture per one million inhabitants for each state, controlling for state population.

Data was also collected from annual reports produced by the Center of Investigation for Development A.C. (Centro de Investigación para el Desarrollo A.C., CIDAC) and México Evalúa measuring the comparative level of judicial reform implementation and performance across Mexico’s thirty-two states. These reports evaluate state reform performance on a variety of measures, including the capacities of judicial institutions and the implementation of reform mandates and programs. CIDAC and México Evalúa aggregate these measures into an annual index with a scale of 0 to 1,000, with 1,000 representing the “ideal standard” of judicial reform implementation in a given state.⁹⁵

For each year from 2015 to 2018, separate correlation and regression analyses were conducted to detect any geographic relationship between the criminal justice reform and the number of CNDH torture complaints. Based on the observations of previous research, this study hypothesized that states with higher scores of judicial reform performance would see decreased CNDH reports of torture by judicial operators.

While the results of this analysis are useful in evaluating the initial relationship between the reform and the use of torture by judicial operators, this methodology has its limitations. One challenge of employing CIDAC and México Evalúa index data is that it is an indirect measure of judicial operator behavior and accountability. The index is a broad measure that considers state resources, capacity, and adherence to reform mandates. While it is probable that police and prosecutors operating in states with higher reform implementation scores are held to higher ethical standards, no data exists to draw this conclusion directly. As such, the results of this analysis are meant to serve as a point of departure for further investigation.

3. National Survey of the Population Deprived of Liberty (ENPOL)

This study also examines torture, mistreatment, and forced confessions, as reported by members of Mexico’s detained population. Specifically, the ENPOL survey conducted by INEGI asks 58,127 participants to report their experiences and interactions with the criminal justice process in Mexico.⁹⁶ In order to assess any significant differences in respondent data before versus after the reform, this analysis employs judicial district-level implementation dates. While the 2008 criminal justice reform set an implementation deadline of June 18, 2016, many judicial districts began operation under the new system prior to this date. As such, the implementation date used to compare torture and mistreatment reports before and after the reform varies by judicial district. In total, this analysis includes 55 separate dates of implementation (December 2004- June 2016) compiled by a group of Justice in Mexico researchers, including the author.⁹⁷ The use of implementation dates at the judicial district-level helps to capture the reform’s localized effect on the incidence of torture and forced confessions.⁹⁸

The ENPOL asked respondents if they were subject to specific types of violence both after their arrest and during their pretrial interactions with the public prosecutor’s office. The instrument specifically asked if the detained individual was: (1) punched or kicked, (2) beaten with an object, (3) burned, (4) electrically shocked, (5) injured as a result of any part of their body being flattened with an object, (6) injured by a knife, (7) injured by a firearm, and/ or (8) forced by threat or physical violence to engage in sexual activities.⁹⁹

This analysis examined responses to items two (2) through eight (8) to determine if a respondent was subject to torture or mistreatment. Item one (1), punching or kicking, was excluded to separate incidents of excessive use of force from cases of torture and/or mistreatment. Participants that responded affirmatively to any of the aforementioned items were included in the pool of cases for analysis. To assess the impact of the judicial reform, a chi-square test for independence was employed to test for a significant difference in reported use of torture before and after the reform’s implementation.¹⁰⁰ This study hypothesized that the use of torture by judicial operators would demonstrate a significant decrease following the reform’s implementation.

The survey also asked respondents to report which types of evidence were presented against them at trial. Categories of evidence included a) the accused’s confession; b) statements made by individuals who claimed to have witnessed the crime; c) statements about the accused’s criminal record made by individuals that knew the accused; d) statements made by accomplices to the crime; e) statements made by other detained persons; f) phone records, recordings, photos, or texts; g) fingerprints, blood, hair, or DNA found at the scene of the crime; and/or h) psychological evaluations conducted at the Observation and Classification Center.

To examine the phenomenon of torture as a prosecutorial tool, a second statistical analysis was conducted using ENPOL response data from detained persons that had already been convicted of a crime and received their sentence. Specifically, this study examined the responses of sentenced participants to items a), f), and g), as outlined above. Together, these items determined the extent to which the prosecution’s case rested on the accused’s confession as culpatory evidence. Respondents that reported the use of their confession a) as culpatory evidence, but no documentation or forensic reporting presented to support these statements [f), g)] were included in the analysis. Respondents that met these criteria and reported being the victims of torture were considered to have been subject to a forced confession.

To examine how the criminal justice reform may have influenced the incidence of forced confessions, a second chi-square test was conducted to detect any significant differences in the phenomenon before and after the reform. In line with recent findings demonstrating a significant reduction in certain types of human rights abuses after the implementation of the reform,¹⁰¹ this study hypothesizes that a significant reduction in forced confessions will be observed following the judicial district-level implementation of the criminal justice reform. Following two generic chi-square tests, this analysis also conducted separate chi-square tests for individual states to examine any changes in torture and forced confession at the state level.¹⁰² The author hypothesized that states with higher SJPA performance scores would demonstrate greater reductions in torture and forced confessions following the reform’s judicial district-level implementation.

1. National Commission of Human Rights (CNDH) Alert System

An initial analysis revealed that the incidence of torture and mistreatment complaints against judicial sector operators varied both temporally and geographically. From 2015 to 2018, the states with the lowest average incidence of CNDH complaints per one million inhabitants were Yucatán (0.36), Querétaro (0.65), and Puebla (0.8). Conversely, Tamaulipas (12.10), Nayarit (6.94), and Guerrero (6.70) demonstrated the highest average rate of torture and mistreatment complaints against judicial sector operators during this time period (See Map 1).

Map 1 CNDH Complaints of Torture and Cruel, Inhuman, or Degrading Treatment against Judicial Sector Operators per 1 Million Inhabitants (Average 2015-2018)¹⁰³  

However, state-level data show that the incidence of CNDH complaints has decreased over time. In 2018, the state with the highest complaint rate was Nayarit (5.60), followed by Quintana Roo (4.70) and Veracruz (4.37). Additionally, four states registered zero complaints in 2018 (Campeche, Tlaxcala, Yucatán, and Zacatecas) (See Map 2).

Map 2 CNDH Complaints of Torture and Cruel, Inhuman, or Degrading Treatment against Judicial Sector Operators per 1 Million Inhabitants (2018)¹⁰⁴  

As outlined above, this study hypothesized that states with higher scores on judicial reform performance would see fewer CNDH reports of torture and mistreatment by judicial operators. An initial analysis assessing the geographic relationship between criminal justice reform performance and torture complaints revealed little to no association between the two variables from 2015 to 2017. However, in 2018, the variables demonstrate a significant negative relationship. In other words, states with higher reform performance scores did indeed demonstrate moderately reduced levels of torture by judicial sector operators that year.

Separate tests were conducted for each year that data for both indicators were available (2015-2018), as shown in Table 1. In 2015 and 2017, a mild negative correlation was observed (r = -0.24), while 2018 data produced a moderate negative correlation (r = -0.43). However, only data from 2018 revealed a significant relationship (p = .01), while analyses conducted using 2015, 2016, and 2017 data were not significant (p > .05).

Annual regression analyses revealed a similar pattern to annual correlation tests. While 2015, 2016, and 2017 did not yield significant results, data from 2018 demonstrated a significant R-Squared value (R² = .18, p = .01). In other words, the level of state SJPA performance accounted for 18% of observed variation in the incidence of torture and mistreatment complaints made to CNDH.

Moreover, consistent with the results presented in Table 1, states identified as having the highest incidence of torture complaints in 2018 (see Map 2) also possessed the lowest reform performance scores. Nayarit, Quintana Roo, and Veracruz not only demonstrated the highest rates of torture in 2018, but they were also ranked among the bottom four performers in terms of state SJPA performance scores (31, 30, and 29 of Mexico’s 32 states, respectively). Furthermore, two states with zero registered complaints in 2018, Yucatán and Zacatecas, ranked in the top eight states in terms of reform performance (5 and 8 of Mexico’s 32 states, respectively).

As shown in Table 1, the relationship between judicial reform performance and the incidence of torture complaints was weakest in 2016 (r = -.04, R2 = .00). While the factors influencing this result have yet to be identified, a frequency analysis revealed that in the same year, the number of torture and mistreatment complaints submitted to the CNDH increased significantly (See Figure 3). This suggests that some combination of factors unrelated to judicial SJPA performance may be associated with the increase observed in 2016. This study discusses these potential factors in detail below.

Figure 3 CNDH Torture and Mistreatment Complaints against Judicial Operator Institutions (2014-2019)¹⁰⁶  

As such, this investigation’s geographic hypothesis that states with greater SJPA performance would see reduced levels of torture by judicial sector operators was only partially substantiated. While annual correlation analyses revealed a mild to moderate relationship, only data from 2018 yielded a significant association. Annual regression analyses demonstrated a similar pattern of results, yielding insignificant and negligible associations from 2015 to 2017. However, a moderately strong relationship between judicial reform performance and the incidence of torture complaints was observed in 2018. These findings suggest that factors unrelated to judicial reform influenced rates of torture complaints, particularly from 2015 to 2017. However, they also indicate that the SJPA may be partially responsible for recent reductions in torture complaints. Nonetheless, analyses conducted using detainee survey data provided more reliable evidence of the reform’s effect on torture as a prosecutorial tool in Mexico.

2. National Survey of the Population Deprived of Liberty (ENPOL)

As previously mentioned, it was hypothesized that a significant reduction in reports of torture would be observed after the implementation of the judicial reform. A chi-square test for independence was conducted using ENPOL data to detect any such difference following the localized implementation of the judicial reform. This test sought to build upon the findings of Magaloni and Rodriguez¹⁰⁷ by examining incidents of torture pre- and post- reform, as reported by members of Mexico’s detained population.

Indeed, a chi-square test examining respondent reports of torture revealed a significant difference in the phenomenon following judicial district-level implementation. Specifically, the chi-square test showed an extremely significant difference in the number of detained persons (presentenced and sentenced) subject to torture pre-SJPA implementation and post-SJPA implementation, χ² (1, N = 30,196) = 37.8, p = .000 (See Table 2).

The observed reduction was in line with the findings of ^{Magaloni and Rodriguez (2019)}. Namely, 54.4% of respondents who were arrested prior to the reform’s implementation reported being subject to torture from the time of their arrest to their time in the Public Prosecutor’s office. However, 50.3% of respondents arrested in judicial districts that had already implemented the reform reported having experienced torture, representing an extremely significant 7.4% decrease pre -to post- reform.

This study also hypothesized that a corresponding decrease would be observed in forced confessions after the reform’s judicial district-level implementation. Indeed, a chi-square test for independence revealed a marginally significant reduction in forced confessions reported by sentenced detainees after the implementation of the reform, X 2 (1, N = 16,098) = 3.6, p = .058.

Specifically, 33.3% of respondents arrested pre-reform reported having been subject to a forced confession, while 31.6% of respondents reported the same post-reform. Overall, these data reflect a marginally significant 5% decrease in the reports of forced confessions after judicial district-level SJPA implementation.

The above findings suggest that the reform’s localized implementation had a significant impact on the incidence of torture and forced confessions. These decreases can also be observed in Figures 4 and 5, during the period of reform implementation from 2008 to 2016. While the data above (Tables 2 and 3) examine cases from 1980 to 2016, the graphs below capture a snapshot of the reform period, during which Mexico implemented the reform in a staggered fashion at the judicial district-level. As observed, the percentage of detainees reporting torture and forced confessions decreases during this period of gradual implementation. Consistent with the above statistical findings, these data provide further evidence of the reform’s impact on the incidence of torture and forced confessions.

Figure 4 Percentage of Detainees Reporting Torture and Mistreatment from Time of Arrest through Stay at Public Prosecutor’s Office by Arrest Year (2008-2016)¹¹⁰  

Figure 5 Percentage of Detainees Reporting Forced Confessions by Arrest Year (2008-2016)¹¹¹  

However, chi-square tests for independence conducted at the state level yielded mixed results, albeit in line with the trend observed in Figure 4. In most cases, significant differences in torture between detainees arrested before versus after the reform were not observed simply due to the small sample size at the state level. As such, a reporting threshold was established at 250 cases for each group (arrest pre-reform versus arrest post-reform), or 500 cases per state.¹¹² This allowed the investigation to examine a more robust pool of survey data before versus after SJPA implementation.

As shown in Table 4, each state that met the threshold for 250 cases in each group (500 per state) demonstrated reductions in reported torture pre- to post-reform implementation. However, these reductions were only significant in four out of five states (Baja California, Durango, Mexico City, and Morelos). Nonetheless, these states saw reductions in reported torture well beyond the national average of a 7.4% decrease. Specifically, Baja California and Durango each demonstrated an extremely significant 39% decrease pre- to post-reform. Mexico City also saw an extremely significant decrease of 24% after the reform’s implementation and Morelos showed a significant 14% decrease. Estado de México, on the other hand, demonstrated a smaller 4.2% reduction that was not significant pre- to post- reform (p=.17).

Nonetheless, only five states met the selection criterion of 500 cases per state, and only four yielded significant results in individual chi-square tests. Thus, it was not possible to evaluate the subnational hypothesis that state with higher SPJA performance scores would see greater reductions in torture. Still, it is worth noting that the state with the highest Hallazgos score saw the greatest significant reduction in torture (Baja California), while states with lower SJPA scores saw the smallest significant reductions of the five states (Mexico City, Morelos) (See Table 4). However, due to the limited number of cases, more data is necessary to evaluate the subnational hypothesis regarding torture.

Results of state-level chi-square tests for independence examining reports of forced confessions by sentenced detainees pre- to post- reform yielded more varied findings (See Table 5). Similar to the methodology employed for reports of torture, a threshold of 75 cases per group (150 total per state) was imposed prior to analysis in order to exclude states that lacked sufficient data for analysis.¹¹⁴ It was necessary to employ a significantly smaller threshold, as the overall data pool was smaller in this case (N= 16,098). Of the nine cases that met this criterion, seven demonstrated reductions in forced confession after judicial district-level implementation (Baja California, Chihuahua, Mexico City, Estado de México, Morelos, Querétaro, and Zacatecas). However, these reductions were statistically significant in just three cases (Mexico City, Morelos, and Querétaro). Furthermore, two states showed increases in reports of forced confessions after the reform’s implementation (Chiapas and Durango), although none of these increases were close to reaching statistical significance.

In fact, all statistically significant results for state-level analyses demonstrated decreases in torture and forced confessions consistent with the trends observed in Figures 4 and 5. While the size of these reductions varied by state, each of these findings supported the hypothesis that reports of torture and forced confessions would decrease following judicial district-level SJPA implementation.

However, as only nine cases met the selection criterion of 150 cases per state, it was not possible to evaluate the subnational hypothesis that states with higher Hallazgos scores would demonstrate greater reductions in forced confessions post-implementation. Furthermore, of these nine states, only three demonstrated significant results in a chi- square test comparing reports of torture pre- versus post- reform (Mexico City, Morelos, and Querétaro) . Still, like the results observed in state analyses of torture, the state with the highest SPJA performance score also saw the greatest significant reduction in forced confessions (Querétaro). Meanwhile, states with lower SJPA scores yielded smaller reductions in forced confessions (Mexico City, Morelos) (See Table 5).

Nonetheless, due to the small number of states (and individual respondents) included in this analysis, additional data is needed to evaluate the sub-national hypothesis regarding the effect of SJPA performance on the incidence of reported forced confessions.

1. The Drug War: Data Implications

While this study’s findings point to the significance of the judicial reform in reducing the incidence of torture, these data also suggest that judicial reform is not the only factor influencing the incidence of these abuses. Evidence supporting the geographic hypothesis was found for 2018; however, as illustrated in Table 1, this pattern was not observed from 2015 to 2017. In particular, 2016 figures demonstrated the weakest association between reform implementation and the incidence of torture complaints. In the same year, a marked increase was observed in the total number of complaints of torture and mistreatment against judicial sector operators (see Figure 3).

In general, scholars have documented the Mexican government’s tendency to react to increased organized criminal activity with militarized enforcement, particularly beginning under Felipe Calderón’s sexenio.¹¹⁶ These enforcement operations often involve violent tactics, which have been associated with increased violence by OCGs as a result of group fragmentation.¹¹⁷ In line with these findings, one potential explanation for the observed increase in torture complaints in 2016 is the impact of such enforcement tactics. Specifically, in 2015, the conflict between the government and the Jalisco New Generation Cartel (Cartel de Jalisco Nueva Generación, CJNG) began to escalate. In March of that year, a series of confrontations began between federal and state officials with the CJNG which resulted in the deaths of numerous police officers. In July, Joaquín Guzmán, head of the then-dominant Sinaloa cartel, escaped from prison and was believed to continue running operations for the cartel until his recapture one year later. During this period, public security forces devoted their resources to both containing the threat of the CJNG and to the recapture of Guzmán.¹¹⁸

While it is not known if increased public security operations directly contributed to the rise of complaints of torture and mistreatment from 2015 to 2016, there are numerous victim testimonies suggesting a relationship. Many report having been detained and tortured until they confessed to associations with specific OCGs.¹¹⁹ Furthermore, Human Rights Watch confirmed that it was common practice among Mexico’s military and security forces to torture individuals to coerce confessions of involvement with specific OCGs.¹²⁰ As such, with the enhanced public security operations that accompanied the rise of the CJNG and the fall of Guzmán, it is possible that officials also increas-ingly employed torture as an investigative and prosecutorial tool. This might help to explain the negligible geographic relationship observed between the implementation of the SJPA and the incidence of torture complaints per 1 million inhabitants from 2015 to 2017, and particularly in 2016. In short, the pressures to investigate and prosecute OCGs may have caused Mexico’s public security apparatus to default to more familiar practices-namely, the use of torture to investigate and prosecute criminals.

2. Measuring Reform Implementation across States

Although a geographic relationship was observed between SJPA implementation and a reduced incidence of torture, this finding was only significant in 2018. One potential explanation for weaker associations between these variables is that the México Evalúa SJPA score employed for this analysis measures a wide range of factors. Specifically, the index examines inter-institutional coordination, judicial planning mechanisms, monetary and infrastructural resources, and public policy surrounding reform implementation.¹²¹

While each of these measures is a crucial ingredient to the SJPA’s successful consolidation, a more direct indicator assessing accountability measures established by the reform may have yielded stronger correlations. As this study examines the behavior of prosecutors and other judicial sector operators, a measurement evaluating state performance in this area would help to confirm the geographic hypothesis.

Nonetheless, this investigation’s finding that reform implementation had a significant effect on the incidence of torture and mistreatment complaints in 2018 is strong evidence of the SJPA’s positive impact on human rights in Mexico. As the constitutional reform was officially implemented in 2016, a large portion of judicial districts had not begun operation under the new system until that year. In fact, 27% of the 2,279 municipalities included in this analysis did not begin operation under the SJPA until 2016. Furthermore, a majority (57%) of municipalities did not begin operation until the last quarter of the implementation period, from 2015 to 2016, despite the reform being passed in 2008. As a result, insignificant associations between implementation scores and the incidence of torture complaints from 2015 to 2017 may reflect the SJPA’s lack of consolidation during this period. Simply stated, it may have taken several years for the new system to enter into force, producing an observable effect on human rights violations in 2018.

This study’s 2018 findings support this assumption. As illustrated in Map 2, the states with the highest incidence of torture complaints were Nayarit, Quintana Roo, and Veracruz. Coincidentally or not, these states also ranked in the bottom four of Mexico’s thirty-two (32) states on México Evalúa’s index measuring SJPA performance. As previously stated, two of the four states with zero registered complaints in 2018 (Yucatán and Zacatecas) also ranked in the top eight on the same index. Although this study did not observe a relationship between judicial reform and decreased torture complaints from 2015 to 2017, the significance of the observed relationship in 2018 should not be overlooked. Particularly as the reform is just entering its consolidation phase, early findings demonstrating a link between the criminal justice reform and decreased torture complaints may point to future advances in human rights protections.

Nonetheless, localized analyses examining the effect of SJPA implementation at the judicial district level yielded evidence suggesting that the reform reduced rates of torture and forced confessions among Mexico’s detained population. While state-level performance indicators may have made it difficult to track the reform’s impact on these abuses, judicial district-level data allowed a more localized analysis of the reform’s effects. Thus, findings at the judicial district-level support the trends observed in 2018 with respect to state level SJPA performance and reduced incidents of torture complaints.

3. Constraining Judicial Behavior: Gradual Improvements

The findings outlined above imply that the reform’s initial implementation has had a significant positive impact on the behavior of judicial actors. Before the reform’s proposal in 2008, Mexico failed to implement measures constraining the behavior of criminal justice system operators.¹²² As a result, following Mexico’s 2000 democratic opening, these actors defaulted to familiar practices that had served as the modus operandi of criminal investigation for decades. Torture continued to serve as an investigative and prosecutorial tool in the twenty-first century, suggesting that Mexico needed a comprehensive reform providing structural incentives to re-shape the behavior of judicial actors.

While the use of torture and forced confessions is still prevalent in Mexico, this study’s results suggest that such institutions, even in their early days of implementation, may create new incentive structures and procedural norms that constrain human rights abuses by state officials, despite mounting security challenges. In the case of the SJPA, this study found that in just the initial years of the reform’s implementation, torture and mistreatment by judicial sector operators had already decreased significantly. While a 7% decrease in reports of torture and a 5% decrease in the incidence of forced confessions following the reform may appear to be negligible reductions, the SJPA has only just begun its consolidation in Mexico. Moreover, certain states have already demonstrated larger, significant reductions in detainee reports of torture and forced confessions, suggesting that there may be valuable lessons to learn from the experiences of those states.

Additionally, according to the most recent México Evalúa report published in 2019, no state in Mexico has reached the “ideal standard” for implementation at 1,000 points on the index, and just five of Mexico’s thirty-two (32) states have reached the “halfway point” of 500 points on the index.¹²³ In short, while Mexico’s SJPA was nominally implemented as of June 2016, there is still much work to be done in terms of actual consolidation. Consequently, relatively small improvements in prosecutorial accountability should be viewed as meaningful steps toward a fully consolidated criminal justice system. As this study’s findings suggest, future advances in the reform’s implementation should be accompanied by further decreases in investigative and prosecutorial abuses, such as the use of torture to extract confessions.

4. The Limits of Official Data in Mexico

While this study provides evidence to support the relationship between Mexico’s criminal justice reform implementation and a reduction in the use of torture by justice sector officials, there are several limitations presented by the data employed. Most importantly, scholars, civil society representatives, and international organizations have repeatedly raised concern that official sources of data reporting state human rights abuses in Mexico are opaque, unreliable, and incomplete.¹²⁴ Specifically, this literature cites the lack of a national registry compiling all complaints of torture.

As Anaya Muñoz explains, the absence of a national registry requires that researchers employ proxy variables to approximate the incidence of human rights abuses. In this case, CNDH complaints published on the organization’s National Alert System were employed for the task. However, the use of any type of proxy variable will introduce alternative explanations for observed phenomena. As Anaya Muñoz notes, the number of complaints submitted to the CNDH for a particular type of abuse may coincide with the public’s level of awareness of these crimes or with their access to oversight organizations within civil society. It is also possible that as CNDH resources increase, they will in turn become more productive, increasing their capacity to receive complaints.¹²⁵

Furthermore, while the CNDH’s National Alert System compiles complaints made to the CNDH, the national ombudsman, it does not include complaints made to state institutions. As the PRODH Center notes, there are roughly four times the number of criminal proceedings at the state level compared to the national level. As a result, a large majority of human rights complaints relating to criminal proceedings would likely be registered with state agencies.¹²⁶ To complicate matters, each state maintains its own complaint records, making it methodologically impossible to analyze the phenomenon on the whole. Furthermore, many cases of torture go unreported altogether due to fear of reprisal and official misclassification of torture to lower level crimes, such as abuse of authority.¹²⁷

As a result, data collected from the CNDH’s National Alert System and employed in the aforementioned analyses possess significant methodological limitations. Had this analysis achieved access to state-level data on torture complaints against judicial sector operators, the observed negative correlations with SJPA implementation may have been stronger. As such, one future avenue of research would be to collect state-level complaint data in order to re-test the geographic hypothesis presented in this study.

In the absence of more accurate official statistics on the phenomenon, this study sought to substantiate initial findings using official data by also employing a publicly available survey data. While the ENPOL survey only included members of Mexico’s detained population, which may not be inclusive of all individuals that experienced torture and mistreatment, its exhaustive list of questions helped to capture all forms of torture and mistreatment from the time of arrest to time spent in the Public Prosecutor’s office. While it is still possible that detainees underreported the incidence of torture for fear of reprisal, the data retrieved in connection with this survey are, at the very least, more comprehensive than any existing source of government data on the practice.

In an effort to understand how the reform may have impacted states differently, this analysis disaggregated ENPOL data by state. These subnational analyses revealed large reductions in reports of torture and forced confessions amongst Mexico’s detained population after SJPA implementation. Significant reductions in torture ranged from 14% (Morelos) to 39% (Baja California, Durango), while significant reductions in forced confessions ranged from 36% (Mexico City) to 84% (Querétaro). As such, these findings provide compelling evidence of the positive impact of the reform’s judicial district-level implementation. Nonetheless, the small sample sizes associated with these analyses require that future research employ more robust sources of data to confirm these results and to assess the relationship between SJPA performance and reports of torture and forced confessions. While state-level data on torture and forced confessions has not been publicly available, perhaps a larger sample size of detainees at the state level could help to fill this data void and provide opportunities for further inquiry.¹²⁸

1. Improving Official Data Sources

As discussed, one of the limitations of this study is the lack of reliable official data on the phenomenon of torture in Mexico. Over the years, scholars, NGOs, and international organizations have advocated for the creation of a national registry on torture that would catalog all cases in the same database.¹²⁹ In the face of these pressures, Mexico passed the General Law against Torture (2017) mandating the creation of such a registry. The law required public prosecutors’ and attorney generals’ offices, public human rights organizations, and victims’ commissions to aggregate complaint data to better understand and analyze incidents of torture.¹³⁰ However, three years after the law’s enactment, Mexico has yet to demonstrate any progress toward the creation of the registry.

In fact, in its most recent review of Mexico, the UNCAT set a deadline of May 17, 2020 for Mexico to create such a system with publicly available data.¹³¹ However, there is no information regarding the extent to which the Mexican government has diverted resources toward the implementation of such a registry. In reaction to this lack of transparency, a group of Mexican civil society organizations recently joined to create the Observatory against Torture (Observatorio contra la Tortura), which provides publicly-available data on specific indicators measuring the law’s implementation.¹³² While the observatory provides substantial data on individual indicators, such as the number of investigations of torture and the number of criminal sentences for the crime of torture, it is inherently limited in scope due to a lack of state transparency.

While civil society has been hugely active in monitoring the practices of torture and mistreatment in Mexico, the government has largely failed in providing accurate and reliable data to complement these efforts. Though the enactment of the General Law against Torture was undoubtedly a necessary step toward the eradication of the practice, it has thus far fallen short of its mandates. In the absence of a national registry or similar tracking mechanism, researchers and civil society organizations will continue to encounter obstacles in measuring how recent reforms, such as the SJPA, have affected the incidence of torture in Mexico. This study managed to employ survey data in order to create a proxy variable for the phenomenon, but future research will require data beyond 2016 in order to measure the SJPA’s impact over time. Thus, Mexico must heed its own legal mandates by working to establish a reliable and effective tracking mechanism.

In the absence of such official sources of data, survey instruments such as the ENPOL provide a crucial source of insight into citizens’ experiences in Mexico’s criminal justice system. As such, Mexico must ensure that this study continues in the years following the SJPA’s implementation. Both the UNCAT and a large network of civil society organizations headed by the CMDPDH have urged Mexico’s government to ensure that this survey instrument continues to be implemented in the coming years.¹³³ Without access to these data, researchers and human rights advocates have few reliable sources of information with which to analyze the prevalence of torture in Mexico’s judicial system.

2. Explicitly Condemning Torture

Mexico has already taken certain steps, albeit delayed, to reduce the prevalence of torture as an investigative practice. As mentioned, Mexico passed the General Law against Torture (2017), which established a common definition for the crime of torture, identified specific institutions to investigate and sanction these crimes, designated minimum sentencing requirements for offenders, and established victim support mechanisms.¹³⁴

Importantly, this law mandates that cases of torture be investigated even in the absence of a complaint; any case in which torture may have occurred must be investigated to the full extent of the law.¹³⁵ Moreover, it states that there is no statute of limitation on the crime of torture¹³⁶ and institutes a minimum sentence of ten years for those convicted.¹³⁷ The law also explicitly prohibits evidence obtained using torture,¹³⁸ placing the burden of proof on the prosecutor to demonstrate that evidence was legally obtained.¹³⁹

While these regulations represent a crucial first step toward the prohibition of torture, the state must ensure that its institutions comply with these newly established regulations. In line with UNCAT recommendations, Mexico’s government must explicitly and publicly condemn torture and other forms of mistreatment, sending a strong message that the practice will no longer be tolerated.¹⁴⁰ Moreover, the government must immediately investigate all instances of torture, placing those accused on administrative suspension in order to reduce the likelihood of coordinated reprisals against complainants.

As the General Law against Torture establishes that there is no statute of limitations for the crime of torture, Mexico must eventually investigate and prosecute all previous reports of torture. This is a tremendous task, as torture has been employed for decades both as a prosecutorial tool and for motivations of social control.¹⁴¹ Nonetheless, if Mexico wishes to comply with its own legal mandates and ensure the consolidation of its fledgling judicial system, this will be a crucial step toward institutional legitimacy.

To complicate matters, Mexico already wrestles with staggering impunity rates. According to a recent study conducted by the Universidad de las Américas Puebla (UDLAP), Mexico currently possesses the highest impunity rate in Latin America and the fourth highest on a list of 69 countries, behind the Philippines, India, and Cameroon.¹⁴² The number of criminal convictions for torture and mistreatment in Mexico supports this finding. In 2016, state attorneys general reported 3,214 complaints of torture and mistreatment affecting 3,569 victims. However, just eight of these criminal cases were adjudicated. Similarly, in the period from 2006 to 2015, there were only 15 federal convictions for torture, despite the submission thousands of complaints to the CNDH in the same period.¹⁴³

As UDLAP explains, Mexico’s impunity rate at the state level is highly associated with low levels of capacity; states with fewer judges tend to possess the country’s highest impunity rates (i.e., Aguascalientes, Baja California, Coahuila, Hidalgo, Estado de México). This lack of capacity has further diminished public confidence in the judicial system, resulting in an increase in the percentage of crimes that go unreported, or the cifra negra.¹⁴⁴

As such, the investigation and prosecution of those accused of torture comes with significant capacity and professionalization challenges. Nonetheless, Mexico must begin by publicly and explicitly backing the General Law against Torture. It must remain firm in its condemnation of torture as an investigative tool and begin to establish mechanisms that track the law’s incorporation into the SJPA. At the very least, this will allow researchers to identify areas for improvement and strategies toward full implementation.

3. Strengthening the Rights of the Detained

While crucial steps toward the eradication of torture, reforms such as the SJPA and the General Law against Torture have fallen short. As discussed, one of the SJPA’s primary critiques is that it contradicts itself. It seeks to guarantee the rights of the accused while also permitting practices that reinforce human rights violations-namely, the use of detention without charge, or arraigo. As such, scholars and international organizations have consistently called upon the Mexican government to outlaw the practice.¹⁴⁵

Under the reform, prosecutorial powers to detain organized crime suspects were expanded, allowing detention for an initial period of 40 days, which can be extended to 80 days. The extension can be granted based on the prosecutor’s argument that the suspect represents a flight risk. However, the prosecutor’s office is not often required to substantiate such claims, affording them ample discretion in determining the length of detention.¹⁴⁶ Thus, as long as arraigo continues in Mexico, police and prosecutors will always possess the incentive and opportunity to continue the longstanding practice of torture. Indeed, arraigo represents an “invitation to torture”.¹⁴⁷

In addition to abolishing the practice of arraigo , the state must also strengthen protections for criminal detainees. Upon a suspect’s arrest, police or other officials must immediately bring the individual to the public prosecutor’s office, reducing the possibility of torture during the initial stages of the criminal process. Additionally, all detainees should be immediately informed of the reason(s) for their detention, granted prompt access to an attorney, and given the opportunity to inform a relative or other person of their detention. In cases where these protections are not afforded, a judge must determine that the accused’s due process rights were violated and take appropriate action.

In cases where torture is suspected, suspects must be granted immediate access to medical professionals trained to examine victims of such abuses. These individuals should be thoroughly trained according to the Istanbul Protocol, a set of international standards for investigating and documenting torture and mistreatment.¹⁴⁸ However, the burden of proof should rest on the prosecutor to establish that torture was not employed while the suspect was in the custody of police or the prosecutor’s office.

Lastly, in line with UN and civil society recommendations, Mexico must establish a national registry of detainees that documents the name of each individual in detention.¹⁴⁹ Such a registry should also record the date and time of a suspect’s detention in order to prevent officials from doctoring data to disguise misconduct or abuse.

4. Countering the Counter-Reform

A lack of official data inhibits the efforts of researchers and policymakers to contest claims that the SJPA has exacerbated human rights abuses in Mexico. Indeed, critics have claimed that corruption and impunity are inherent to the SJPA and that the system has contributed to increased levels of insecurity across Mexico.¹⁵⁰ Critics view the SJPA as a “revolving door” that releases criminal actors from detention while failing to protect victims.¹⁵¹ Recently, these critical voices have gained traction, and experts monitoring the SJPA’s performance agree that the threat of counter-reform grows more credible.¹⁵² On January 15, 2020, Mexico’s National Prosecutor Alejandro Gertz Manero and the president’s chief legal advisor Julio Scherer announced a package of nine proposed judicial reforms after a draft was leaked several days prior. According to Gertz Manero and Scherer, the objective of these reforms was to reduce impunity and recidivism. However, as the proposal is currently drafted, these changes would undermine a significant number of hard-won human rights safeguards established under the SJPA.¹⁵³

Most importantly, the proposed counter-reform would allow prosecutors to present evidence obtained by torture before a judge. The SJPA and the subsequent General Law against Torture specifically prohibited the use of evidence obtained by torture, mandating that judges dismiss such evidence when torture is suspected. Nonetheless, the proposed constitutional amendment would afford judges total discretion over evidence admitted at trial-including evidence obtained through human rights violations.¹⁵⁴

In addition, the proposed counter-reforms would reintroduce many of the prosecutorial incentives to practice torture that the SJPA sought to expel. While Mexico’s constitution mandates that no suspect be held for more than 72 hours without judicial review,¹⁵⁵ the proposed changes would remove both judicial oversight and the time limit. Thus, if a suspect were charged for a crime that mandated pretrial detention, they could be detained throughout the duration of the criminal investigation and trial with no judicial review or opportunity to challenge evidence presented against them. Moreover, the package of counter-reforms would expand arraigo beyond organized crime-related cases, allowing prosecutors to seek prolonged detention without criminal charges in any type of case.

This expansion of the prosecutorial right to detain would reintroduce institutional norms and incentives to practice torture and obtain forced confessions.¹⁵⁶ As esteemed political commentator Denise Dresser writes, the counter-reform would introduce “[the] power to detain arbitrarily, spy legally, torture unconstitutionally, and return to the old system that existed prior to the 2008 reform” [author’s translation].¹⁵⁷

While the results of this study suggest that the SJPA is not a silver bullet solution for eradicating human rights abuse, they also reveal that the SJPA can be credited for small but meaningful reductions in torture and mistreatment among judicial sector actors. This evidence runs counter to claims that the reform encourages further abuse by officials.

According to Shirk and Ríos Cázares, achieving the rule of law in new democracies is often an inherently destabilizing process. During the transition phase, state institutions such as the police and the criminal justice system may adjust too slowly to democratic changes to meet the needs of society. As such, citizens may experience reduced access to justice during this period, negatively influencing their perception of democratic reforms and increasing public demands for justice and accountability. Ironically, this period is marked by decreased public confidence in reforms meant to strengthen the rule of law. In SJPA’s case, it may also result in calls to revert back to known, authoritarian models of criminal justice. However, as Shirk and Ríos Cázares warn, these appeals threaten to erode the very institutions that serve as the foundation for the rule of law.¹⁵⁸

As Mexico finds itself in this transition process, threats to fledgling democratic institutions must be taken seriously. Mexico must be diligent in its SJPA consolidation efforts, despite existing counter-reform proposals. As Irene Tello Arista, Executive Director of Impunidad Cero, argues, the counter-reform does not seek to establish better protections for victims, as it claims. Instead, its objective is to reestablish a regulatory backing for abuse and judicial malpractice, attempting to solve problems through legislative action rather than tangible change or follow-through.¹⁵⁹ Given these concerns, Mexico must ensure that its commitment to the system’s full implementation remains steadfast.