Olfato. It is a term used quite often in law enforcement and judicial circles in Central America (and other parts of the world as well). It refers to the sixth sense they have as they see a crime scene, investigate a murder or plow through the paperwork related to one. There is nothing scientific about olfato, yet it seems as if that is the guiding measure as it relates to determining this crucial question: What is behind the steady stream of homicides in Central America, or in this case, Guatemala?

In this study, we have tried to go beyond olfato. However, we too have run into extreme limitations and have had to rely on subjective measures, albeit slightly more consistent ones. And we do believe our work looking through each case gets us closer to understanding the dynamics of violence in Chiquimula and Zona 18, results that are the beginning of what we hope will be a slightly more analytic process in discerning these dynamics.

The importance of this exercise cannot be overstated. Without a clear understanding of the dynamics of homicides, local authorities and their international partners have a hard time developing clear strategies and effectively allocating resources. 

Yet, the answer remains elusive, at least when we consider the limited data that we studied. Indeed, our first and arguably most important conclusion is that, given the high number of homicides that fall into the category of unknown, it is hard to say with any precision that the statements made by authorities about gang-related or drug trafficking-related homicides are incorrect.

banner download Homicides-GuatemalaBut it equally hard to say that they are correct. In fact, from this data and analysis, our preliminary conclusion as it relates to drug trafficking organizations -- and even our slightly modified characterization "organized crime" -- we can say the reality does correspond with the statements. Of course, this point would require more exploration, but what we can say from the data obtained in Chiquimula is that there are numerous homicides that are related to social and economic conflicts that have nothing to do with organized crime or drug trafficking. And only 28 percent of the cases could be reasonably attributed to organized crime.   

The consistency with which homicides occur in places like Chiquimula also lends support to this thesis. Drug trafficking activities -- often measured by using proxies such as seizures -- vary widely from year to year. But homicides in these eastern states do not shift as readily as the patterns of the traffickers.

Meanwhile, gang violence, at least as it is expressed by statements of politicians and their counterparts in the police and the Attorney General's Office, is very close to what we discovered. In our study of Zona 18, we found that our analysis of the homicides coincided almost exactly with that of the MP's analysts (and that of President Otto Pérez Molina at the outset of this report).

Furthermore, our analysis of the motives also overlapped. The upshot is that we agree with the general perception that gang-activity is related to close to 40 percent of the homicides in these urban areas where the gangs are prevalent. Other factors such as age, type of weapon and presumed motives also coincide with the analysts' and police with whom we consulted.

What we do not know as it relates to gang violence is if this is the pattern in all gang-controlled areas. To be sure, the big question is whether we can extrapolate from this information. Unfortunately, we do not think it is. We studied two years. We would need to study more. And we would need to analyze more data from more geographic areas. We would also need to obtain more information from authorities. Preliminary information from the crime scene is only a snapshot. But given the high rates of impunity, and the fact that so few of these cases get any real attention, it is the single best source of raw information that we have.

To date, the emphasis on the data gathering has been to satisfy a political appetite, to show that there is someone paying attention and, in the best case scenario, that these statistics are moving in a positive direction.

What we can say with authority is that there are few homicides that are neatly packaged. The variety of victims, places, types of weapons, times of crimes, and modus operandi that we encountered made our task extremely difficult, which is probably why both the police and the MP mostly avoid this question altogether; in very few cases, can anyone give an answer with 100 percent certainty whether a homicide is drug trafficking-related or gang-related.

Nonetheless, there is a way to get closer to the answer without swerving into the political generalizations that service the presidents, security forces and others interested in painting an oversimplified picture. That is by improving the way that data is collected, organized and analyzed.

As noted in the second section on data collection, there is a mountain of information that is collected throughout the process. This information could be used for many purposes, beginning with the actual criminal investigation. Despite what some prosecutors say, there is a plethora of information provided in the crime scene report by the responding police officers. This information is quite simply underappreciated, and later, it is often lost. Only pieces of it, usually related to overall crime statistics gathering, remain. This is a huge loss.

SEE ALSO: Full Report on Homicides in Guatemala

There are also many ancillary uses for this information. For example, the police do some geospatial mapping of homicide hotspots, but this could be greatly enhanced. The police, for example, seem to have the information they need to do implement a basic form of CompStat.[1] It could also be crossed with data emerging from the homicides themselves to test hypotheses about whether certain types of victims always appear in the same places or certain weapons are consistently used in the same areas.[2]

It could also help to test correlations between rural murders -- by tracking, for example, the use of knives, machetes or blunt objects as the murder weapons -- and the presence of land disputes filed in the court system. A similar set of factors could be tested as they relate to organized crime- or gang-related homicides. The system could also be useful for cross-referencing suspects, victims, witnesses, cars, license plates, etc.

To date, the emphasis on the data gathering has been to satisfy a political appetite, to show that there is someone paying attention and, in the best case scenario, that these statistics are moving in a positive direction. But this is short-sighted and ignores the underlying issues that lead to this violence in the first place. Data is not a bureaucratic burden to be used for career advancement or political benefit. It is the core upon which strategies are made, resources are deployed and lives are saved.


1. Emphasize the need for information and data analysis that focuses on homicide dynamics. This analysis should be a prerequisite in any area where local and international stakeholders are investing heavily to lower homicide rates. The analysis should take into account the raw data from the crimes themselves, as well as the informed opinions and perceptions of security forces, investigators and others monitoring the situation closely.

2. Workshop the use and importance of crime-scene and other investigative information and data for the PNC, INACIF and the MP, so that there is understanding of how it is used to resolve homicides and analyze criminal dynamics. There is little appreciation for a large amount of information gathered at the initial stages of the homicide investigations. In order for anyone to care about it, they need to better understand its utility. 

3. Train and/or hire more data technicians in the police, INACIF and the MP. Focus efforts and money on getting people to better understand the collection, management, and analysis of the data so that not a single bit of it is wasted during any part of the process. There are some infrastructure issues that need to be addressed (see below), but the main issue is that there are not a lot of specialists in this area who are trained to organize. The few specialists there are also appear to be underappreciated and underutilized.

4. Train more analysts. One of the most consistently cited ways to lower homicide rates is to obtain a better understanding of the dynamics of the area in question. The MP has started this process of creating a team of analysts. The police needs to join them.

5. Invest in infrastructure to allow for:

• Safe movement of data and information over the internet to avoid unnecessary use of scarce resources.

• Collecting and collating information into one single digital case file for the murder investigation to give prosecutors and MP analysts the ability to study the case, cross reference it with other cases, etc.

*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] CompStat stands for "Computer Comparison Statistics." It is utilized by police and other security forces to map crime statistics over long periods of time, allowing them to better deploy and utilize their human resources to prevent crime.

[2] Several countries in Latin America are exploring these possibilities, most notably Colombia. There, the police have implemented the so-called Quadrant Plan (Plan Quadrantes) in several major cities, which utilizes geo-referenced crime data to determine priorities and utilization of resources.