When violence surged in early 2015 in Guatemala, then-President Otto Pérez Molina knew how to handle the situation: Blame the street gangs. 

"One group murdered a guy from another group," he told the press. "And that meant that in January more than 40 percent of the homicides were a result of a fight between these two maras."[1]

Just a few years earlier, then-President of Guatemala Álvaro Colom attributed the rise in violence to "drug trafficking organizations" (DTOs), which he said were "invading" his country.[2] He later said the DTOs were responsible for up to 60 percent of the homicides.[3]

These presidents are not alone in their belief that most of the homicides in the region are related to street gangs and drug trafficking organizations. In 2015, Honduras President Juan Orlando Hernández told a gathering at the Woodrow Wilson International Center for Scholars in Washington, DC that as much as 80 percent of the violence in his country could be attributed to organized crime.[4]

Many US officials cling to the same idea. In a 2014 Military Times op-ed, General John Kelly, the current US Secretary of Homeland Security who was then the head of US Southern Command, said that drug cartels and gangs were largely responsible for the rising violence in the region.

"There are some in officialdom who argue that not 100 percent of the violence today is due to the drug flow to the US," he wrote about the Northern Triangle nations of Honduras, Guatemala and El Salvador. "And I agree, but I would say that perhaps 80 percent of it is."[5]

We at InSight Crime have also fallen into this trap. We routinely cite these officials without questioning their sources or basis for understanding the dynamics of violence. Prestigious multilaterals and academics make the same claims. A 2011 World Bank report, for example, declared that, "Drug trafficking is quantitatively far more important than the other risk factors for violence."[6] (A more thorough critique of this statement appears later in this report.)

banner download Homicides-GuatemalaThe emphasis on street gangs and drug trafficking organizations has a palpable impact on how money is allocated. In its analysis of the allocation of funds through the Central America Regional Security Initiative (CARSI), the Congressional Research Service said that 66 percent of the $1.2 billion that CARSI administered in the region between 2008 and 2015, went through the Bureau of International Narcotics Law Enforcement (INL), the branch of the US State Department that manages hard-side assistance like drug interdiction programs, equipment and infrastructure enhancements for police and military personnel, training and use of special police units that are focused on international narcotics trafficking and gangs. Another 31 percent goes through the Economic Support Fund (ESF) account, most of which is channeled through the United States Agency for International Development (USAID).[7] While most of the USAID monies go to soft assistance -- violence prevention, social and economic programs -- some of this money goes to programs that support police in places like Honduras. Meanwhile, other CARSI money has gone towards the formation and maintenance of special anti-gang units.[8]

The emphasis on street gangs and DTOs also impacts domestic allocations of funds. Local allocations of funding are harder to discern, but from the early 2000s, the Northern Triangle countries have adopted a so-called "Mano Dura," or "Iron Fist," approach towards gangs. New laws with broad definitions of "gangs" were combined with police sweeps that incarcerated anyone who fit those broad definitions. The names of these programs have changed, but the core of the strategy remains the same: incarcerate en masse. Prison populations have swelled, but instead of slowing the growth of the gangs, the gang leaders have reorganized behind bars and have expanded their array of criminal activities.

More recently, the Northern Triangle governments have militarized their approach, implementing various states of siege in places where there was drug trafficking activity in Guatemala, enacting "emergency measures" inside jails and in particularly troublesome states in El Salvador, and using the military police in numerous places in Honduras. All three approaches use as their starting point that street gangs and DTOs are at the center of the violence, and illustrate the belief that the only way to slow their activities is to have a large physical presence to deter them from killing one another or innocent civilians. The results of these approaches have been mixed at best, and seem to represent short-term solutions to deep social and economic problems.

The reality is that we do not know how much of the homicides in the Northern Triangle are related to DTOs and gangs. Criminal organizations have undoubtedly proliferated in recent years in Central America. While they vary from street gangs to DTOs to paramilitary mafias, they share some common characteristics: they take advantage of weak government institutions to control physical territory; they seem to thrive where new criminal economies have emerged and where they can diversify their own portfolios; they employ violence and the threat of violence in order to achieve their aims, be they political, criminal or other goals.

The impact of this criminal activity in the region is profound and most certainly includes violence. Latin America and the Caribbean is the most homicidal region in the world today. For a time, the five most homicidal nations on the planet were in the region; three of those are the ones that make up the Northern Triangle.[9] According to a recent United Nations Development Programme (UNDP) report, Latin America and the Caribbean saw homicides increase by 12 percent between 2002 and 2012; it was the only region in the world where homicide rates rose.[10] Eleven countries in the region had homicide rates that would qualify as "epidemic," the UNDP report said, including the aforementioned Northern Triangle nations.[11]

When faced with an incomplete picture as we frequently are when it comes to rising presence of gangs and DTOs, and increasing violence in the region, the easy conclusion is to say that these events are related, but the fact is that there are few analyses of the homicide data to tell us who the victims are, how they have been killed, what the official investigations tell us about these criminal events, and who the presumed murderers are. Conclusive evidence is nearly impossible to obtain given that impunity rates in these countries are so high and data is sparse, uneven and poorly managed.

What's more, the preliminary conclusions issued routinely by politicians, policymakers and experts alike seem to ignore worldwide and regional studies that indicate homicide is the result of a complex set of interrelated factors that can be both personal and communal.[12] The UNDP's recent Human Development Report, for instance, cited four major motors of violence.[13] First, they noted the high incidence of what they call "aspirational crimes," which they connected to questions of inequality -- not poverty -- and social mobility. Second, they offered a "social fabric" theory, which focuses on upheaval in the household; the number of single parent households has doubled in last 30 years, and there have been rapid urbanization and economic shifts that come with trade liberalization and other macroeconomic changes. Third, they cited "crime drivers," including weapons, alcohol, and illegal drugs. Finally, they looked at "institutional weakness" in countries where they found the most violence.

Certainly the Northern Triangle's violence is just as complex. In a report on organized crime in Central America, the UNODC said that the presence of drug trafficking can lead to upticks in violence but can also lead to declines. It is, in other words, part of a series of factors that contribute to homicides, not the motor of them.

"The key driver of violence is not cocaine, but change: change in the negotiated power relations between and within groups, and with the state," they wrote. "For progress to be made, the risk of aggravating violence in the short term must be taken into account."[14]

This question about what are the drivers of homicides in the region comes at a critical moment. In December 2015, the US Congress passed the "Consolidated Appropriations Act, 2016," which allocated $750 million to assist Central America in its fight against crime and violence.[15] What was not surprising was that most of this aid is headed to the Northern Triangle. What is surprising is that the bill tries to shift the balance more in the direction of soft assistance; INL is managing only 30 percent of the aid, and some of this INL funding is going towards "forensic technology" and the International Commission against Impunity in Guatemala (Comisión Internacional contra la Impunidad en Guatemala - CICIG), two initiatives that are designed to fight structural issues as opposed to specific criminal groups involved in the violence. Still, much of the assistance will be aimed at stemming violence, so it is important we understand more about the complex combination of factors that go into rising homicides in the region.

SEE ALSO: Full Report on Homicides in Guatemala

This project was established to begin to fill that knowledge gap. Our intent was to disaggregate motives and actors -- with a special emphasis on gang controlled areas and drug trafficking corridors -- involved in homicides in Guatemala. The end goal was to provide authorities, citizen security specialists and other stakeholders with more nuanced information regarding the dynamics of violence, so that they can better determine intervention strategies in areas with high homicide rates and areas where there the presumed level of gang- and drug trafficking-related violence is high.

Data collection in these countries is spotty at best and, in most cases, inadequate. Governments put little emphasis or resources towards analyzing the information they do collect.

We also studied the way authorities collected and analyzed information from the crime scenes. Data collection in these countries is spotty at best and, in most cases, inadequate. Governments put little emphasis or resources towards analyzing the information they do collect. The project both analyzes the information we did obtain and develops a better understanding of the data collection methods, processes, and other dynamics limiting collection and/or analysis of the information. It is important to note that this is not a study of impunity, and, although we touch on data collection methods, this study does not cover best practices but rather looks for ways in which the Guatemalan system is deficient.

We have therefore organized this study into those two major parts. The first part of the analysis focuses on determining, with as much precision as possible, how many murders we can reasonably attribute to the drug trafficking organizations and the gangs in two specific areas of Guatemala. This section uses as its foundation an analysis of the crime scene reports generated by the police in one area that is a known drug trafficking corridor and another area that is a known area of gang influence. It also uses numerous interviews that we did with police investigators, government prosecutors and Attorney General's Office analysts about who and what are behind the homicides, both on a local and a national level.

The second part is an analysis of the ways in which the police, the Attorney General's Office and other crime-fighting bodies collect and process the information they obtain during these homicide investigations. This part places a particular emphasis on illustrating just where the system can, and often does, fail, thereby contributing to the discussion about how to improve these processes. 


Based on authorities' statements noted earlier, as well as dozens of interviews with police officials, prosecutors and crime analysts, the project began with one basic hypothesis:

• Gang- and drug trafficking-related activity are behind the bulk of homicides in the Northern Triangle. The potential policy implications for this hypothesis are that these countries may require, at least initially, more of a law and order approach, strengthening of police and judicial actors with an eye towards dismantling these criminal organizations.

In order to test this hypothesis, we:

• Identified one (1) drug trafficking corridor,[16]which had a homicide rate above the national average. 

• Identified one (1) gang controlled territory,[17] neighborhood, district or similar and comparable jurisdiction, which had a homicide rate above the national average. 

• Analyzed all the available data on homicides in those areas during a two-year period using information from police, and in some sporadic cases, information from forensic medicine units, the Attorney General's Office, and other government agencies tracking homicides, as well as from non-governmental organizations, universities, think tanks and media.

• Critically assessed preliminary determinations and corroborated, when possible, assertions regarding drug trafficking- and gang-related homicides in those areas.

As per the current standard classification practice of law-enforcement agencies, we distinguished between drug trafficking- and gang-related homicides by looking for specific markers in each event that tend to be closely related to each type of murder and that differ from homicides associated to personal, cultural or socio-economic violence.[18] These markers include the characteristics of the victim, the type of weapon used, the location of the event and, when available, corroborating information from media or official sources. When two or more of these markers are present, we estimate that there is a higher probability that a specific event can be classified as gang- or DTO-related.

The markers for DTO-related murders include:

1. Use of high-caliber or sophisticated weaponry, particularly automatic and semi-automatic assault rifles and 9 mm firearms.

2. Signs of torture.

3. Post-mortem displacement of the body, particularly when it is dumped in a highly visible location, which would imply the desire for publicity.

4. Presence of threatening messages near or on the victim's body, including symbolic mutilations.

5. Direct or circumstantial evidence of the victim's participation in DTO-related activities, including known criminal associations, prior offenses and witnesses' testimony.

The markers for gang-related homicides include:

1. Use of high-caliber pistols and revolvers, but also sharp and blunt objects.

2. Direct or circumstantial evidence of the victim's participation in gang activities, including tattoos, gang paraphernalia, known gang associations, prior offenses and witnesses' testimony.

3. Direct or circumstantial evidence of the victim's connection to gang activities, such as paying extortion or witnessing criminal activities. 

It is possible that both a DTO- and a gang-related homicide share one or more of these markers. Furthermore, it is also possible that the direct or circumstantial evidence of the victim's involvement in criminal activities does not meet the standard that would be necessary to introduce it at trial. We tried to overcome these limitations by analyzing as much raw data as possible for each case, such as witness statements, police reports and corroborating media sources. We also followed-up on some of the cases with the police and the Attorney General's Office investigators when possible.

We did this research in Guatemala where there are four sources of data on homicides: the National Civil Police (Policía Nacional Civil - PNC), the National Institute of Forensic Sciences (Insituto Nacional de Ciencias Forenses - INACIF), the Attorney General's Office (Ministerio Público - MP) and the National Statistics Institute (Instituto Nacional de Estadísticas - INE).

The PNC records data from the crime scene. Police have information on the victims of violence, both dead and wounded (sex, approximate age, and to a lesser degree, the profession), including where the criminal act occurred and the type of weapon used (knife or other sharp object, by hand, or a firearm). Generally, PNC homicide figures are lower than those of INACIF because the latter receives the bodies of people injured as police search the crime scene, some of whom later die in the hospitals.[19]

On the other hand, in some cases the INACIF has overestimated homicidal violence because they often include suicides and accidents in their homicide totals.[20] INACIF also does not report where the crimes have occurred, but rather where the autopsy took place, which is frequently in the central office in Guatemala or in their various regional offices making it very difficult to map the homicides using this data.

In principle, the MP has the highest quality data on homicides because it is the repository for all the information on the case and can assess all of these sources prior to making the determination of whether or not the person was murdered, died due to natural causes or by accident.[21] In addition to information its own investigators have, the PNC and the INACIF also feed the MP with data. In the case of the MP, however, they rely on criminal complaints, often initiated at the time of the crime or shortly thereafter, in order to begin a homicide investigation. For its part, the INACIF will refer to autopsies that are "possibly related to criminal acts under investigation" but stops short of offering possible motives or actors behind the homicides. The PNC data distinguishes between homicides, suicides and accidents. None of these sources, however, has deep analysis or information regarding motives or presumed actors behind the homicides. (More on this process and the results of our investigation later in the report.)

The INE is responsible for collating and organizing this information. However, the INE also generates its own data, called "Vital Statistics," which follows protocol laid out by the 10th edition of the International Classification of Diseases. The INE gets much of its information from the National Registry of Persons (RENAP), which certifies deaths and indicates whether it was a homicide, suicide or an accidental death. However, this source does not include information on the motive or those responsible for the homicides.

The project sought to get around these problems by obtaining information from the crime scenes themselves, from which we hoped to see certain patterns that we could establish as "gang-related" or "drug trafficking-related" homicides, the two ways authorities most often describe homicides. Specifically, we sought information on two areas that government officials from the Interior Ministry (Ministerio de Gobernación) said fit our criteria: Chiquimula therefore represented a "trafficking corridor"; the Zona 18 of Guatemala City represented a "gang area." This was based on intelligence and law enforcement officials' assessments of these areas, collated at the Interior Ministry.

Over a period of several months, we filed freedom of information requests ("derecho de petición") for national data and subnational data from the PNC, INACIF and the MP on homicides, potential motives and other related data that would help us establish motive and actors involved in the homicides for both those areas and the nation as a whole. The data we received, however, only gave us only a surface understanding of these motives and actors.

Given the limitations of these findings, we requested and obtained access to the preliminary police reports in homicide cases in the Zona 18 and Chiquimula. These files are not digitalized, so we had to go to the police stations, search out the physical files, and take photographs of these reports. We also requested access to the Attorney General's Office case files of these homicide cases. This access was granted but only in cases that were cleared. Unfortunately, these represented but a fraction of the total cases that we found in the police files and were not helpful.

In order to help us fill in the knowledge gaps or help us corroborate what we were finding in the police reports as it related to the dynamics of violence in these areas as well as on a national level, we spoke to police investigators and analysts from the Attorney General's Office. Most of these interviews were off the record but represent an important cross-section of understanding as it relates to violence in Guatemala. Nonetheless, it is important to note that they are these individuals' perceptions and are not based on any analytics or a systematic study of the information available.

Throughout, we chronicled the way the different agencies collect and process data. We studied the amount of resources they dedicate to writing this information on paper and entering it into computers, many of which were several years old. We went to their data headquarters to see how they channeled information and organized it. And we interviewed those who are working in these data collection centers or have overseen efforts to improve and consolidate this process on local and national levels. This, as will be made evident later in this report, is not a smooth process and numerous opportunities to use this information are missed along the way.

Major Findings

• We could not reasonably attribute motive nor assign blame to any one criminal group in over 35 percent of the cases. 

• In the trafficking corridor, we could reasonably attribute 28 percent of the homicides to organized crime-related activities; this is less than authorities normally publicly attribute to organize crime.

• In the gang area, we could reasonably attribute 41 percent of the homicides to gang-related activities. This is in line with what authorities say in Guatemala.

• There were some differences in the homicide dynamics of the two areas studied. The two areas differed in the age of the victims. There was a higher number of homicides with a firearm in the gang area than the drug trafficking area.

• In both areas, the information from authorities was fragmented, disorganized and sometimes missing altogether.

• Data collection suffers due to:

• Multiple, clashing bureaucracies that are not speaking the same language, not administering the data on the same platforms or formats;

• Differing criteria as it relates to how to classify the data;

• Poor training and little emphasis placed on importance of collecting and analyzing data;

• Some degree of technological challenges, including old computers and a lack of internet service;

• A culture in which politically-salient statistical information is prioritized over information that will actually resolve cases or give authorities the ability to analyze criminal dynamics. 

*This incredibly labor intensive process was financed by the United States Agency for International Development via Democracy International. It was led by Steven Dudley, co-director of InSight Crime. He initially worked with Carlos Mendoza, formerly of the Central American Business Initiative; Mendoza is now at the Finance Ministry. Julie López worked tirelessly the last quarter of the project compiling data and interviewing analysts and police. López and Victoria Dittmar went through and catalogued the hundreds of preliminary police reports that we gathered from Chiquimula and Zona 18. Steven Dudley wrote the report, with assistance from Mendoza and López. Jaime López-Aranda, a former Mexican government official who specializes in data collection and analysis, reviewed and commented on the methodology and drafts of the report. InSight Crime would like to thank Guatemala's PNC, the MP's office, and the Interior Ministry for their assistance in this project. Read entire report here.

[1] G. Contreras, "Presidente Otto Pérez justifica violencia," Prensa Libre, 3 February 2015. Available at: http://prensalibre.com/noticias/justicia/Presidente-justifica-violencia-rivalidad-pandillas-inseguridad_0_1297070292.html

[2] Geoffrey Ramsey, "Guatemala Warns of Drug Gang Invasion," InSight Crime, 26 May 2011. Available at: http://insightcrime.org/news-analysis/guatemala-warns-of-drug-gang-invasion

[3] Carlos Mendoza, "¿Cuál es la relación entre violencia homicida narcoactividad en Guatemala? Parte I," Plaza Pública, 19 May 2014. Available at: https://plazapublica.com.gt/content/cual-es-la-relacion-entre-violencia-homicida-narcoactividad-en-guatemala-parte-i

[5] Gen. John F. Kelly, "SOUTHCOM chief: Central America drug war a dire threat to U.S. national security," Military Times, 8 July 2014.

[6] Rodrigo Serrano-Berthet and Humberto Lopez, "Crime and Violence in Central America," World Bank, 2011. Available at: https://siteresources.worldbank.org/INTLAC/Resources/FINAL_VOLUME_I_ENGLISH_CrimeAndViolence.pdf

[7] Peter J. Meyer and Clare Seelke, "Central America Regional Security Initiative:

Background and Policy Issues for Congress," Congressional Research Service, 17 December 2016, p. 17. Available at: https://fas.org/sgp/crs/row/R41731.pdf

[8] Ibid.

[9] "Global Study on Homicide 2013: Trends, Context, Data," United Nations Office on Drugs and Crime (UNODC), March 2014.

[10] "Citizen Security with a Human Face," United Nations Development Programme (UNDP), November 2013.

[11] "Epidemic" in this instance means rates that were above 10 homicides per 100,000 residents.

[12] See: Etienne G. Krug, et al., editors, "World report on violence and health," World Health Organization, 2002.

[13] UNDP, op cit.

[14] United Nations Office on Drugs and Crime (UNODC) "Transnational Organized Crime in Central America and the Caribbean: A Threat Assessment," September 2012, p. 13. Available at: https://unodc.org/documents/data-and-analysis/Studies/TOC_Central_America_and_the_Caribbean_english.pdf

[16] We loosely defined this as an area used to move illicit drugs in large quantities. However, there is no standard definition for this type of area. The proxy used to determine it is normally seizures, but in Guatemala, seizures are small and seemingly random, making it nearly impossible to apply. 

[17] We loosely defined this as an area where gangs hold a monopoly or near-monopoly on the use of force. However, there is no standard definition for this type of area. The proxy used to determine a gang-controlled area is normally complaints of extortion from companies with 50 employees or smaller; from bus and taxi cooperatives; and from small shop owners and individuals. But Guatemala has no systematic accounting of these complaints.

[18] The project based this criteria on a similar developed by the Mexican government during the Felipe Calderón administration. Jaime López-Aranda, a former official in the Mexican agency that developed this methodology, was a consultant on this project.

[19] InSight Crime interviews with Interior Ministry (Ministerio de Gobernación) personnel who manage the inter-agency data collection service, Guatemala City, August and September 2015.

[20] Ibid.

[21] Ibid.