Firearms and Injuries during Home Robberies in Mexico, 2010–2017

Abstract

According to data from Mexico’s National Crime Victimization and Public Security Surveys (ENVIPE), from 2010 through 2017, home robberies involving an interaction between a victim and an aggressor occurred every two minutes and resulted in more than 257,000 injuries. The objective of this study is to explore whether victims of home robberies in Mexico are more likely to report an injury if perpetrators are armed with a gun. This analysis compiles data from ENVIPE and runs a binary logistic regression. In contrast to literature on robberies in the United States and studies on street robberies in Mexico, results show that when perpetrators are armed with guns, victims of home robberies in Mexico are more likely to report an injury, even though these injuries are rarely caused by a gunshot. These discrepancies may be explained by different factors. Unlike in the United States, most gun-related crimes in Mexico are perpetrated by organized criminal groups that obtain firearms through illegal channels. Members of these groups are prepared to engage on physical violence against their victims. Additionally, unlike street robberies, the enclosed setting of home robberies might facilitate the use of physical violence as a strategy to coerce victims and successfully obtain profits without generating public attention. Overall, perpetrators use guns as a tool to exert control and avoid resistance. These findings suggest that policies to address organized crime and halt easy access to guns should be prioritized on Mexico’s national security agenda.

Introduction

Victimization surveys indicate that one in every four Mexicans report feeling insecure in their own homes (INEGI 2018). While alarming, this is not surprising. Home robberies involving an interaction between a victim and an aggressor in Mexico have been on the rise and occur with growing frequency. According to data from Mexico’s National Crime Victimization and Public Security Surveys (ENVIPE), from 2010 through 2017, these types of home robberies occurred every 2 min and resulted in more than 257,000 injuries.

Moreover, 21% of these robberies were committed with a firearm.¹ In this regard, studies have established the role of guns on robbery outcomes in the United States, concluding that victims are less likely to suffer injuries if offenders use a gun (Cook 2009; Cook 1983). Similarly, a 2018 analysis found that victims of street robberies in Mexico were less likely to report injuries if perpetrators used firearms (Heredia Gonzalez 2018). The objective of this study is to explore whether victims of home robberies in Mexico are more likely to report physical injuries when perpetrators use guns. Home robberies in Mexico exemplify an interesting comparative study for two reasons. Unlike the United States, most gun-related crimes in Mexico are perpetrated by organized criminal groups that obtain firearms though illegal channels. Members of these groups are better prepared to engage on physical violence against their victims. Additionally, the enclosed setting of home robberies might facilitate the use of physical violence as a strategy to coerce victims and successfully obtain profits without generating public attention.

To conduct the analysis, we compiled data on home robberies from Mexico’s National Crime Victimization and Public Security Surveys (ENVIPE) from 2010 through 2017. ENVIPE defines home robberies as crimes committed by a person or group of persons in which they invade private property to take other people’s belongings, regardless of whether victims are present (INEGI 2018). This category is slightly different from the way robberies are defined in the US as it encompasses cases where victims are not present. For the purpose of this analysis, we are only including those incidents where there was a physical interaction between an aggressor and a victim, in keeping with the US definition. To do so, we filtered for those cases where respondents reported that they were present and had observed the crime. We ran a binary logistic regression and used a dummy variable indicating whether victims were injured as our dependent variable and type of weapons used by perpetrators as our main independent variables.

In contrast to literature addressing the role of firearms on robberies within the US and studies exploring the use of guns during street robberies in Mexico, our results show that when perpetrators are armed with guns, victims of home robberies in Mexico are more likely to report a physical injury. These injuries, however, are rarely due to gunshot wounds and are instead the result of other forms of physical violence. We conducted further analyses that provided potential explanations for these discrepancies. We found that the likelihood of injury increases when home robberies are perpetrated by organized criminal groups defined as those perpetrtated by a group of two or more offenders that use guns to commit the crime. Similarly, we found that, in contrast to street robberies in Mexico, when victims of home robberies are injured, offenders are more likely to succeed in obtaining profits. This suggests that injuries may be a strategy to coerce victims, forcing them to comply during home robberies without generating public attention.

This study is divided into six sections. Section one discusses literature on guns and robberies while section two addresses the problem of organized crime in Mexico and their access to illegal guns. Section three provides an overview of Mexico’s ENVIPE surveys and the construction of key variables while section four presents descriptive data on home robberies. Our econometric analysis and empirical evidence are presented in section five. Finally, section six presents our discussion and conclusion.

Literature on guns and robberies

Within the US, the phenomenon of robberies has been well studied. Evidence shows that most robberies are committed by men and occasionally involve multiple aggressors. Also, it suggests that aggressors and victims are strangers in most robberies, and that in close to 33% of incidents, perpetrators are unsuccessful in obtaining profits (Cook 2009).

According to studies analyzing robberies in the US, guns are rarely used to commit these crimes (Cook 1983; Cook 2009). However, the use of a firearm in the commission of a robbery significantly changes the outcomes. The likelihood of death of the victim when the perpetrator is armed with a gun is three times greater than the likelihood of death of the victim when the perpetrator is armed with a knife, which in turn is three times higher than armed robberies involving other weapons (Cook 1987). This distinction exists because guns are more lethal than other weapons. They are manufactured to be carried with relative ease, to be used from a distance, and to wound and kill swiftly (Cook 1983). The increased levels of lethality due to the presence of a firearm is not unique to robberies. Evidence shows that guns increase the levels of lethality on other forms of violence, including suicide attempts in the US as well as terrorist attacks in developed nations (Spicer and Miller 2000; Tessler et al. 2017). Even more nuanced, studies have concluded that handguns with higher calibers are statistically linked to higher likelihood of deaths during criminal assaults (Braga and Cook 2018).

The sole threat of a gun has important implications. Victims have two choices when confronted with a robber: to comply or to resist, each of them leading to potentially different outcomes (Luckenbill 1982; Cook 2009). When faced with these two choices, victims will most likely select the choice that will help them preserve their well-being (Luckenbill 1982). When perpetrators are armed with lethal weapons, such as firearms, victims are more likely to comply than to resist (Luckenbill 1982). Even if there are multiple victims, if perpetrators display a gun, victims are more likely to comply than to resist. This is because guns allow perpetrators to control large groups at a time (Cook 2009). Studies suggest that only about one in four victims of robberies perpetrated with a firearm in the US attempt any sort of resistance, indicating that the threat experienced by the display of a gun is effective at forcing victims to comply (Cook 1986). Even those gun robbery victims who are physically attacked tend to avoid resisting the attacker if guns are involved, including those who are seriously injured (Cook 1986). Since victims tend to comply more easily, robberies with guns are linked to lower rates of victim injury (Cook 2009; Cook 1983). This is because there is no need for perpetrators to exert additional violence (Cook 1983).

Moreover, evidence indicates that while overall robberies in the US seldom result in high profits for robbers (Cook 1987), the results change significantly when examining cases where firearms are displayed. The likelihood of a successful profitable robbery is greater when perpetrators use a firearm (Cook 2009). This notion is further supported by studies conducted in Australia where robberies involving handguns yielded greater monetary benefits for perpetrators (Bricknell 2008).

Overall, studies have established that robberies in the US are rarely committed with a gun and are usually perpetrated by strangers. The studies also show that guns during home robberies are linked to higher levels of lethality but lower levels of injury. Also, when robberies are perpetrated with a gun, there is less resistance from victims and higher levels of profits for perpetrators.

While these findings have been instrumental in the US, little research has been conducted to analyze the role of guns in robberies perpetrated in Mexico. This country presents an interesting case study because, unlike the US, organized criminal groups in Mexico have become a major security threat in the past years (Calderon et al. 2019). Moreover, even though Mexico has restrictive firearm legislation, these organizations use various illegal channels to gain access to guns (Weigend and Bhatia 2019). As a result, gun violence in Mexico is directly linked to organized criminal organizations.

Organized crime and guns in Mexico

Organized criminal groups have evolved throughout the years adapting to the constant change in technology, communication, and markets. These groups have been able to move quickly, transform and profit illegally. In Mexico, the traditional threats from organized criminal groups involve transnational crime such as the illicit trafficking of drugs. However, in recent years, organized criminal groups have diverted to other crimes such as kidnappings, robberies and extortions. Organized criminal groups have become a major threat to human security in Mexico (Beittel 2018). They systematically engage in the use of violence affecting the wellbeing of the population and their sense of security.

Academics and NGOS have attributed the rise of organized criminal violence in Mexico to many factors (Zepeda 2018). These include high levels of impunity, corruption, the lack of reliable police forces as well as inadequate social policies. Others have indicated that the rise of violence is an unintended consequence of the aggressive security policies launched by former President Felipe Calderón Hinojosa (2006–2012). Studies show that the strategies of targeting drug cartel leaders resulted in the fragmentation of criminal groups which further lead to disputes and violence (Atuesta and Ponce 2017). While these disputes affected drug trafficking related violence, studies show that they also had a spillover effect on the general population (Calderon et al. 2015). Similarly, after the fragmentation of large drug trafficking organizations, weakened groups diversified to other criminal activities (Guerrero 2011). While larger drug trafficking organizations tended to limit the use of violence, these smaller groups resisted such limitations (Beittel 2018). This suggests that, in efforts to maintain profits, these smaller fragmented groups used violence as a key operational strategy on crimes such as extortions, kidnappings and robberies.

In addition to all these factors, recent shifts in U.S. gun production and federal gun laws facilitated access to firearms for Mexican organized criminal groups. Firearms allowed these groups to adopt confrontational strategies to control drug markets while also enabling their diversification to other criminal activities such as kidnappings and robberies (Perez Esparza and Weigend 2013).

When compared to the U.S., Mexico has stricter gun laws. While people in the U.S. can purchase a firearm from the more than 50,000 federally licensed gun dealers across the country (Bureau of Alcohol, Tobacco, Firearms and Explosives 2018), there is only one gun-store in Mexico. Unlike the U.S., which has no federal requirements for firearm registration, Mexican law requires that all firearms be registered in the Registro Federal de Armas (Ley Federal de Armas de Fuego y Explosivos 2019). Additionally, Mexico’s federal law limits the type of weapons that can be purchased and possessed, including high caliber rifles which are readily available on the U.S. civilian market. These restrictions have resulted in relatively low levels of civilian gun ownership, in stark contrast to the levels of gun ownership in the US. A 2018 survey indicates that only 3% of adults from 9 metropolitan cities in Mexico own a firearm (Perez Esparza and Hemenway 2018). In contrast, surveys conducted within the US indicate that around 22% of people own a firearm (Smith and Sona 2015; Azrael et al. 2017).

Despite laws existing to limit access to firearms in Mexico, criminal networks obtain firearms from numerous countries. Nonetheless, analysis of traced crime guns in Mexico indicate that roughly 70% of guns recovered in Mexico originate in the US (U.S. Government Accountability Office 2016), making this country the biggest supplier of illegal guns to Mexico.

There are multiple ways that criminal organizations in Mexico acquire U.S. firearms. One method is by siphoning weapons that were legally exported from the U.S. to Mexican security agencies; many of these firearms are reported as stolen or diverted to criminal groups. (Baez-Zamudio et al. 2018). Corruption and lack of control on international gun sales are cited as important contributors to this problem (Baez-Zamudio et al. 2018).

Gun trafficking across the U.S.-Mexico border is also a major concern. Estimations suggest that, from 2010 through 2012, an annual average of 212,000 firearms were purchased in the U.S. and trafficked to Mexico (McDougal et al. 2014). One of the factors that contributes to gun trafficking on the US-Mexico border is the large number of firearms produced and imported by the U.S. in recent years. While the annual production of guns went from 3.5 million from 1996 to 2005 to 6.7 million from 2006 to 2015, annual imports went from 1.3 million to 3.5 million during the same periods (Parsons and Weigend 2018). Studies indicate that this increase is attributed to the rise in production of rifles and high caliber pistols (Smith et al. 2017), weapons often used by criminal groups in Mexico (Goodman and Marizco 2011).

In combination with this inventory, studies suggest that U.S. gun legislation has a significant impact on gun trafficking to Mexico. Studies have highlighted that the difference in gun legislation between these countries creates optimal conditions for traffickers to take advantage of a gray market (Weigend and Villarreal 2015).² Other studies have analyzed the effects that specific U.S. gun policies have on gun trafficking. As a result of the 2004 removal of the assault weapons ban in the U.S., researchers found higher levels of gun trafficking and gun homicides in Mexican municipalities that border Arizona, New Mexico, and Texas, as opposed to California, a state that maintained its own local ban on assault weapons (Dube et al. 2013). A 2014 study reported that state laws limiting sales of multiple firearms, requiring background checks, banning assault weapons, and prosecuting straw purchasers³ were significantly linked to lower rates of American gun exports to Mexico (Esby 2014).

Overall, access to guns in Mexico is limited by law. However, due to the impact of weak U.S. gun laws on gun trafficking across the U.S.-Mexico border - the existing legal framework on firearms in Mexico is not enough to restrict gun access to organized criminal groups. As such, the federal law is unable to prevent these groups from using firearms to perpetrate acts of violence.

As a result, gun violence has significantly risen in Mexico. Data from the Presidency of Mexico (Presidential Report 2018) indicate that while the Mexican government seized an average of 8500 guns per year from 2001 through 2008, this average increased to 20,000 per year from 2009 through 2017. Guns used to commit homicides have also augmented at alarming rates. While less than 20% of reported homicides were committed with a firearm during 1997, close to 70% of reported homicides were committed with a firearm by 2017 (Parsons and Weigend 2018). The problem goes beyond homicides. Studies have indicated that gun injuries, unintentional shootings as well as street robberies committed with firearms have been on the rise in Mexico in recent years (Weigend and Bhatia 2019).

Despite rising levels of gun-related violence, few studies have analyzed the role of firearms on robberies. To our knowledge, only one study has attempted to fill this gap by exploring street robberies in Mexico and analyzing compliance levels as well as economic, physical, and psychological costs associated with the use of firearms (Heredia Gonzalez 2018). Similar to literature on gun robberies in the U.S., this study found that victims of street robberies in Mexico were less likely to be injured when perpetrators used a gun (Heredia Gonzalez 2018).

To further contribute to the analysis of guns and robberies in Mexico, we explore home robberies. This crime is different from street robberies in many aspects. The enclosed setting of home robberies facilitates the use of physical violence without generating public attention. Additionally, profits resulting from home robberies are significantly higher than those resulting from street robberies, making the former more attractive to organized criminal groups. According to ENVIPE, economic costs reported by victims of home robberies is, on average, 4000 pesos (200 dollars). In contrast, average economic costs reported by victims of street robberies in Mexico is less than 600 pesos (30 dollars).⁴

Key variables from Mexico’s National Crime Victimization and public security surveys

To analyze the role of guns on home robberies in Mexico, we looked at Mexico’s National Crime Victimization and Public Security Surveys (ENVIPE) from 2011 to 2018. These surveys were conducted by Mexico’s National Institute for Statistics and Geography (INEGI) and collected data on criminal incidence, the perception of insecurity as well as trust in law enforcement agencies across the country. The ENVIPE surveys are used to estimate levels of impunity by looking at the percentage of crimes that are not reported to law enforcement authorities (Toledo 2017; Evalúa and Stiftung 2017). These surveys have also been used by researchers to analyze the perception of insecurity in Mexico, its determinants and its link to public policies (Jasso Lopez 2013; Avila et al. 2016; Vilalta Perdomo 2012).

These surveys are conducted on an annual basis. It should be noted that, ENVIPE captures data about criminal incidents that occurred during the previous year. Therefore, the time frame that we are analyzing - ENVIPE 2011 to 2018 - encompasses crimes committed from 2010 to 2017. We selected this period because INEGI started to conduct these surveys after 2011 and there are no previous compatible surveys before that period.

To conduct these surveys, INEGI adopted various recommendations from experts of the United Nations Office of Drug and Crime (2011)⁵ and used a stratified random sampling of Mexican households. As a result, between 78,000 and 106,000 households were randomly selected each year between 2011 and 2018. The surveys are representative at a national and state level as well as for certain metropolitan areas (INEGI 2018).

Within each household, respondents older than 18 years of age were asked about criminal incidents (INEGI 2018). Each year, around 40,000 households had individuals that reported on a crime. This means that the ENVIPE surveys captured data on more than 320,000 crimes perpetrated from 2010 through 2017. ENVIPE surveys, however, do not collect data about homicides or negligent manslaughter. Also, commercial robberies as well as crimes against business are not captured by these surveys. Within ENVIPE, non-lethal crimes are grouped under the following categories: auto theft, auto-part theft, vandalism & graffiti, home robberies, street robberies, other robberies, bank fraud, consumer fraud, extortions, verbal or written threats, abusive behavior resulting in injury, kidnappings, sexual harassment, rape, and other crimes.⁶

It should be noted that there are some issues and limitations with these surveys. In contrast to the ENVIPE surveys in Mexico, the U.S. National Crime Victimization Surveys interview the same household every six months. This is done to avoid telescoping, a term used for the tendency of victims to report crimes experienced outside a certain period. The ENVIPE surveys do not implement this procedure which potentially results in high levels of telescoping. In addition, a general problem with victimization surveys is that respondents may have problems answering each question correctly simply because they cannot remember the details of the event, because they do not understand the question or because they tend to over report certain information such as value of economic losses.

Despite these limitations, we are certain that the ENVIPE surveys are a useful source for analyzing non-fatal crimes in Mexico. The ENVIPE survey captures detailed information about crimes that are not included in law enforcement reports. Furthermore, these surveys capture data on crimes regardless of whether they were reported to authorities. This is not a minor issue. In many states, more than 90% of crimes are not reported to authorities (Evalúa and Stiftung 2017).

To analyze home robberies in Mexico, we merged the annual ENVIPE surveys into a single dataset. We then filtered for those crimes that victims reported as a home robbery under ENVIPE’s predetermined non-lethal crime categories. From 2011 to 2018, a total of 44,563 individuals reported directly to the ENVIPE surveys that they had been victims of a home robbery. However, we focused on those home robberies where there was a clear physical interaction between victims and aggressors. To address this, we considered two specific questions from the ENVIPE surveys. These were whether respondents were present and whether they observed how the crime was perpetrated. We conducted filters based on these questions and as a result had 6150 home robberies involving a physical interaction between a victim and a perpetrator from 2011 to 2018.

The ENVIPE surveys also provide weight factors which are used to make national, state as well as metropolitan level estimations on the number of perpetrated crimes. While INEGI uses these weight factors to report on estimates based on the 44,563 home robberies, we used these weight factors to make our own estimations for numerous subcategories as well as to run our regressions.

Within the ENVIPE surveys, there are two questions that address injuries. A first question asks about victims being injured with a weapon while a second inquiries about other forms of physical violence suffered during the crime.⁷ We used these questions to construct our dependent variable labelled injury. This variable took the value of “1” if victims responded yes to any of these questions and “0” if they did not suffer any form of physical violence. Before 2014, the ENVIPE surveys did not incorporate a question about whether injuries were the result of gunshot wounds. However, using data from the 2014–2018 surveys, we estimated the percentage of injuries that can be attributed to gunshot wounds.

The type of weapon used during a crime is also specified in the ENVIPE questionnaires. This allowed us to create three binary variables. The first variable labelled firearms had values of “1” if aggressors used firearms, regardless of whether they also used other weapons, and “0” if aggressors were unarmed or if other non-firearm weapons such as knives, or blunt objects were used. The second variable we constructed was labelled nonfirearms and had the value of “1” if aggressors were armed with other weapons such as blunt objects or sharp objects and a value of “0” if perpetrators were armed with a gun or unarmed. Finally, we constructed the variable unarmed. This variable had the value of “1” if perpetrators were unarmed and “0” of they were armed with any type of weapon.

ENVIPE also allowed us to analyze other characteristics related to the crime. We created six dichotomous variables that have been used as controlled variables on other studies (see for example Heredia Gonzalez 2018) and that are potentially associated to the outcome of a home robbery. First, we created the variable unknown that adopted the value of “1” if the aggressor or aggressors were strangers to the victim and “0” if victims knew the aggressors.⁸ This variable was selected to discern if it has an impact on the success of robbery.

Similarly, we created the binary variable multipleaggressors. This variable adopted the value of “1” if more than one aggressor participated on the crime and “0” if only one aggressor participated. Studies have shown that as the number of offenders grow, victims of robberies are more likely to report an injury (Wolf Harlow 1987). Furthermore, the binary variable alone was given a “1” if the victim was alone and “0” if accompanied. The number of people present with the victim might affect the success of the robbery considering it might be more difficult for the perpetrator to control the situation and coerce victims if they are outnumbered.

We created the variable urban that had a value of “1” if respondents lived in an urban setting and “0” if that was not the case.⁹ According to crime statistics, community size does make a difference, as crime rates in countries such as the United States are higher in urban areas. This can be attributed to various factors including less probability of recognition and higher number of female-headed households (Glaeser and Sacerdote 1999).

We generated the variable night. This took the value of “1” if the crime occurred between 6 pm and 6 am and “0” if the crime took place between 6 am and 6 pm. Additionally, we generated the variable effects which took the value of “1” if victims reported that aggressors were under the effects of drugs or alcohol and “0” if it was not the case. Studies indicate that drugs and alcohol impact crime indirectly via the effects they have on users’ behavior (see for example U.S. Department of Justice 1994). Observations where victims could not recall any of this information or simply left the questions unanswered were not included in our statistical analysis on section five.

The ENVIPE surveys allowed us to analyze characteristics of the victims that have often been used as control variables when analyzing robberies (see for example Heredia Gonzalez 2018). We created the dummy variable sex that took the value of “1” if the victims reporting the crime were male. We also created three dichotomous variables based on victims age. AgeOne compiled victims ages 18 to 29 while AgeTwo took the value of “1” if victims were 30 to 45 years old. AgeThree adopted the values of “1” if victims were ages 46 or older. Moreover, we generated three dichotomous variables based on the levels of education reported by victims. This served as a proxy variable to income. The variable LowEducation took the value of “1” if victims had no education or only elementary school education. MidEducation compiled victims that had secondary education, including middle and high school. Finally, the variable HighEducation took the value of “1” if victims had obtained a college degree or higher.

Additionally, ENVIPE asks victims about the monetary losses derived from the crime. This question refers to the value of property lost during the incident and does not include health care expenses.¹⁰ This variable allowed us to report on the economic costs of the population as a result of home robberies. We labelled this variable as economiclosses and used it to construct another variable labelled success. This variable took the value of “1” if victims reported economic losses higher than zero, suggesting that offenders were successful, and a “0” if victims reported economic losses equal to zero.

Data on home robberies in Mexico

According to official reports from INEGI and based on weighted data, more than 234 million non-lethal crimes occurred in Mexico from 2010 through 2017 (INEGI 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018).¹¹ Out of these, 7% were reported as home robberies. As mentioned above, this category differs from robberies described in the U.S. which are defined as theft or attempted theft directly from a person (Bureau of Justice Statistics 2019). We further filtered this category to include those incidents where there was a physical interaction between an aggressor and a victim, making this category compatible with what is defined as robbery in the U.S.

According to our estimations using weighted data from the ENVIPE surveys, from 2010 through 2017, there were close to 2.37 million home robberies in Mexico where there was a clear interaction between perpetrators and victims. In other words, this type of robbery occurs every two minutes. Additionally, these types of crimes have been on the rise in recent years. While an estimated 118,000 incidents occurred during 2010, this figure rose to close to 400,000 by 2017. This data could explain why one in every four Mexicans reports feeling insecure in their own homes (INEGI 2018). Figure 1.

Fig. 1

Home robberies where victims were present and observed how the crime was perpetrated, 2010–2017

Aside the rise in frequency of these crimes, the outcomes are particularly alarming. Out of the close to 2.37 million home robberies, more than 257,000 resulted in injury to victims (ENVIPE 2011–2018).¹² This means that, in Mexico, injuries occur in 11% of home robberies that involve an interaction between victims and aggressors. Close to one third of these injuries were perpetrated with a weapon, while two thirds resulted from physical violence that was not related to weapons. Less than 3% of reported injuries involved both weapons and other forms of physical violence.

Victims reported economic losses higher than zero in close to 66% of home robberies.¹³ Putting it differently, perpetrators are unsuccessful in 34% of home robberies in Mexico, a percentage that resembles the rate of unsuccessful robberies in the U.S. In addition, total economic losses derived home robberies in Mexico resulted in close to 20.7 billion pesos (nearly 1 billion dollars) from 2010 through 2017.¹⁴

Descriptors of these crimes are also included within the ENVIPE surveys. Our analysis shows that, in those cases where a victim-perpetrator relation was reported, aggressors were unknown to victims in close to 75% of home robberies (INEGI 2011–2018). Additionally, while multiple aggressors participated in 49% of home robberies,¹⁵ victims reported being alone in 29% of the cases. Furthermore, out of the home robberies analyzed, 65% were reported by individuals living in urban settings with the majority (64%) being perpetrated at night (INEGI 2011–2018).¹⁶ Additionally, in close to 34% of home robberies, victims reported that perpetrators appeared to be under the influence of drugs, alcohol, or both.

We also analyzed characteristics of the victims. According to information from the ENVIPE surveys, 43% of victims of home robberies were males and close to 57 were females. While 31% of victims reporting the crime were ages 18 to 29, 38% were adults ages 30 to 45. Victims ages 46 or older represented close to 31% of victims. Victim’s level of education is also included within the ENVIPE surveys. 53% of victims reported low education levels while 25% reported having a mid-level education. Close to 22% of victims reported having a university degree or higher.

We were able to determine the type of weapon used or whether aggressors were unarmed in 76% of home robberies that involved an interaction between victims and aggressors. Out of the close to 1.8 million cases, 64% were perpetrated by unarmed aggressors while 15% were perpetrated by aggressors with non-firearm weapons. Finally, an estimated 21% of home robberies were perpetrated with a firearm.

Across the categories of home robberies based on weapons, the percentage of incidents resulting in injuries varies significantly. According to data from Table 1, close to 6% of home robberies perpetrated by unarmed aggressors resulted in injury to victims. This percentage rises to 20% when perpetrators were armed with non-firearm weapons. However, when perpetrators used firearms to commit these crimes, victims reported injuries in close to 32% of the cases. After conducting a means test analysis, we found that these differences are significant at a 95% confidence level. In other words, data suggests that victims of home robberies perpetrated with firearms are more likely to report injuries. However, these injuries are seldom gunshot wounds. According to available data from the 2014–2018 ENVIPE surveys, close to 4% of injuries reported by victims of home robberies that were perpetrated with firearms were the result of gunshot injuries. Overall, less than 2 % of home robberies perpetrated with guns resulted in gun-related injuries.

Table 1. Percentage of injuries by types of weapons used to perpetrate home robberies, 2010–2017

In addition to analyzing differences in injury outcomes across the three categories of weapons used, we analyzed differences in these categories across characteristics of the crimes and victims. In most cases, there was little variation across the groups of home robberies based on weapons. However, one variation stands out.

Table 2 divides home robberies by type of weapon and shows the percentage of incidents according to the number of aggressors present during the crime. Data suggests that those home robberies committed with a gun are linked to a higher number of perpetrators. While 12% of home robberies involving guns were perpetrated by a single aggressor, close to 60% were perpetrated by three or more individuals. This is different when analyzing home robberies involving non-firearm weapons. 78% of these incidents were perpetrated by one or two aggressors, with 42% involving a single perpetrator. Only 10% of these types of home robberies involved four or more aggressors. Home robberies perpetrated without weapons are often committed by a single person. 66% of these home robberies involved a single aggressor and only 4% involved four or more perpetrators. This finding supports the notion that gun-related crimes are often perpetrated by organized criminal group members that operate in large numbers.

Table 2. Weapon categories and percentage of home robberies by number of offenders, 2010–2017

Overall, home robberies in Mexico have been on the rise and occur with staggering frequency. These incidents have resulted in high levels of injuries. Our analysis of weighted data from the ENVIPE surveys suggest that firearms play an important role on home robbery outcomes. Firearms are linked to higher levels of injuries, although these are not gunshot wounds. Our analysis also indicates that home robberies involving guns tend to be perpetrated by multiple aggressors. To deepen this analysis, we conducted further statistical analyses. The results are presented on section five.

Empirical evidence

Based on weighted data of 4643 observations,¹⁷ we ran a binary logistic model to address the main question, are victims of home robberies in Mexico more likely to report an injury if perpetrators use a gun? The model included the variable injury as the dependent variable and the types of weapons used by offenders as the independent variables (firearms, nonfirearms and unarmed). The variable unarmed was left as the base group or reference for the category of weapons used. For this model, we also controlled for characteristics of the crime as well as characteristics of the victims discussed in section three. While AgeThree (victims ages 46 or older) will be the reference group for the age categories, HighEducation will be the reference group for the categories of level of education.

When perpetrators use a firearm during a home robbery in Mexico, the odds of victim injuries are 5.8 times higher than if offenders were unarmed. Similarly, the odds of injury to victims when perpetrators use non-firearm weapons are 3.2 times higher than the odds of victim injury with unarmed aggressors. This means that when perpetrators use a gun to commit a home robbery in Mexico, the likelihood of injuries suffered by victims significantly increases.

Likelihood of injury also increases when home robberies are perpetrated by multiple aggressors. The odds of injury when there are multiple perpetrators is close to two times higher than the odds of injury when there is a single aggressor. Our model further indicates that injuries are more likely to occur when victims are males (odds ratio 1.16), when perpetrators are perceived to be under the influence of either drugs, alcohol, or both (odds ratio 1.37) and when victims were alone (odds ratio 1.06). Likelihood of injury is lower when perpetrators and victims are strangers (odds ratio 0.73) as well as when home robberies are perpetrated during the night (odds ratio 0.66). Table 3 indicates that young people ages 18 to 29 and adults ages 30 to 45 are more likely to report injuries than those victims ages 46 or older. Finally, victims reporting low-level or mid-level education are less likely to report injuries that those victims with high-level education.

Table 3. Odds ratio of injuries during home robberies in Mexico, 2010–2017. (Type of weapons as main independent variables)

To deepen the analysis, we ran two additional binary logistic regressions that separated injuries into two groups. The first regression included non-weapon related injuries as the dependent variable while the second regression included injuries perpetrated with a weapon as the dependent variable. As unarmed aggressors are not included in the later model, the variable nonfirearms was left as the base group. On both additional regressions, we included the same control variables as in our primary regression. Results are presented on Tables 7 and 8 on the appendix. Even though the odds ratio is not as high as in Table 3, the use of a gun during a home robbery is linked to higher likelihood of injuries, even when excluding those injuries perpetrated with a weapon. This suggests that aggressors that use guns are more likely to engage in non-weapon-related physical violence. Overall, results of this additional regression resemble those presented on Table 3.

We also ran a regression that analyzed the likelihood of injuries perpetrated with a weapon, excluding those cases with unarmed aggressors. Results of this model are presented on Table 8 on the appendix. When compared with home robberies perpetrated with non-firearms weapons, the likelihood of weapon related injuries increases when perpetrators use a firearm (odds ratio 1.41). While overall results of Table 8 are similar than those presented on Tables 3 and 7, one variation stands out. Tables 3 and 7 show that the variable Effects (perceived to be under the influence of drugs and/or alcohol) is linked to higher likelihood of injury. However, this was not the case on the regression presented on Table 8 (odds ratio 0.88). This suggests that injuries involving weapons are more likely to be perpetrated by aggressors that are not under the influence of drugs or alcohol. Instead, perpetrators under the influence of these substances tend to be linked to other forms of physical injuries.

Moreover, we ran an additional regression that disaggregated the variable nonfirearms into two groups. One group included blunt objects such as baseball bats while the second included sharp objects such as knives. Table 9 on the appendix shows the results of this additional regression. Like Table 3, when perpetrators use a firearm during a home robbery in Mexico, the odds of victim injuries are over five times higher than the odds of victim injury derived from unarmed aggressors. Nonetheless, the odds of injury during a home robbery when perpetrators use blunt objects is significantly higher than when home robberies are perpetrated with sharp objects.

Overall, evidence suggests that victims of home robberies are more likely to report an injury when offenders use firearms. These results contrast findings in the U.S. as well as those analyzing street robberies in Mexico. To provide potential explanations for these discrepancies, we conducted further analyses.

As mentioned in section two, in contrast to the U.S., access to guns in Mexico is limited. Nonetheless, organized criminal networks in Mexico take advantage of international trafficking to easily acquire firearms. In this regard, a potential explanation is that injuries are more likely to occur during gun-related home robberies because most of these crimes are perpetrated by organized criminal groups with easy access to firearms and use violence as a modus operandi.

To further explore this, we conducted another binary regression analysis to determine whether victims of home robberies perpetrated by organized criminal members were more likely to report an injury. We used the variable injury as the dependent variable. While it is challenging to know the exact number of home robberies that were perpetrated by organized criminal groups, we constructed a proxy variable and labelled it “OrganizedCrime.” This variable is constructed on the assumption that home robberies involving firearms and perpetrated by more than one aggressor are linked to organized crime. Therefore, this proxy variable took the value of “1” if home robberies involved firearms while also being perpetrated by more than one offender and “0” if that was not the case. Similarly, we constructed other 3 dichotomous variables based on the number of aggressors and whether offenders were armed with guns. The variable “LoneGunPerpetrator” took the value of “1” if there was a single perpetrator and a gun was involved. The variable “UnarmedGroup” took the value of “1” if there were multiple offenders but the use of guns was not reported by the victims. Finally, the variable “LoneUnarmedPerpetrator” took the value of “1” if there was a single offender and no firearms were reported. For this regression, the variable “LoneUnarmedPerpetrator” was used as the reference group. We used the same control variables as in Tables 3 and 4.

Table 4. Odds ratio of injuries during home robberies in Mexico, 2010–2017. (Type of perpetrators as main independent variables)

When home robberies are perpetrated by armed organized criminal groups, the odds of victim injuries are 9.2 time higher than the odds of injury to victims when home robberies are perpetrated by a single offender not armed with a firearm. The odds of injury to victims when there is a single offender armed with a gun are more than 2 times higher than the odds of injury to victims during home robberies involving a single offender that is not armed with a gun. Similarly, the odds of injury to victims of home robberies when there are multiple offenders not armed with guns is more than 2 times higher than the odds of injury to victims when home robberies are perpetrated by a single offender not using a firearm. This model indicates that injuries during home robberies are significantly more likely to be perpetrated by members of an organized criminal group.

Nonetheless, home robberies perpetrated by lone offenders with a gun are more likely to result in injuries than those home robberies perpetrated without firearms. This suggests that in addition to violence used by criminal organizations, there may be other factors that explain why injuries during home robberies are more likely to occur if perpetrators are armed with guns.

Unlike street robberies, injuries perpetrated during home robberies occur in closed spaces and potentially yield different outcomes. The enclosed setting of home robberies might facilitate the use of physical violence as a strategy to coerce victims and successfully obtain profits. Like evidence from the U.S. suggests, guns serve as a tool that helps perpetrators control the situation and avoid resistance even when victims are seriously injured (Cook 1986). To explore this, we ran two binary logistic regressions to determine if injuries increased the likelihood of success for offenders during both street robberies and home robberies.

We created a binary variable (success) that took the value of “1” if victims reported losses higher than zero and “0” if victims reported losses equal to zero. Those that took the value of “1” indicated that perpetrators were successful. Our regressions included the variable success as the dependent variable and the variable injury as the independent variable. We controlled for the same variables as those presented on Table 3. Based on information from ENVIPE, we analyzed weighted data of 3397 home robberies and 16,469 street robberies that occurred from 2010 to 2017.¹⁸

As presented on Table 5, home robberies involving injuries are more likely to result in economic losses to victims (odds ratio of 1.27). Similarly, the odds of success during a home robbery when perpetrators use a firearm is close to 1.4 times higher than the odds of unarmed perpetrators. Additionally, perpetrators using non-firearm weapons are more likely to succeed in gaining profits than unarmed aggressors (odds ratio 1.07).

Table 5. Odds ratio of success for aggressors during home robberies in Mexico, 2010–2017

Home robberies involving multiple aggressors are more likely to result in some form of profit for the perpetrators (odds ratio of 2.06). Perpetrators are more likely to succeed in obtaining profits when victims are males (odds ratio 1.41), when crimes are reported in urban settings (odds ratio 1.19) and when victims are alone (odds ratio 1.23).

Results from Table 5 also indicate that home robberies perpetrated by aggressors under the influence of drugs and/or alcohol were less likely to result in perpetrators’ success (odds ratio 0.64). When perpetrators are unknown to victims, their likelihood of success is significantly lower than in cases of home robberies where the aggressors are known to victims (odds ratio of 0.64). Perpetrators are less likely to succeed if the home robbery is committed during the night (odds ratio 0.66). Victims older than 46 years of age are more likely to report economic losses higher than zero, signaling more success for perpetrators. Finally, perpetrators are more likely to succeed in obtaining profits when victims report having low or mid-level education.

Evidence presented on Table 6 indicates that when victims of street robberies report injuries, offenders are less likely to succeed in obtaining profits. This stands in contrast to results reported on Table 5 and supports the notion that injuries are more likely to occur during home robberies because they yield different outcomes, which in this case means higher success rates.

Table 6. Odds ratio of success for aggressors during street robberies in Mexico, 2010–2017

In addition, like home robberies, when offenders use firearms to perpetrate street robberies, they are more likely to succeed. Perpetrators of street robberies are also more likely to succeed if they operate in groups of two or more individuals. Similarly, they are more likely to succeed if victims are males and if victims are alone. Nonetheless, in contrast to home robberies, Table 6 shows that perpetrators of street robberies are more likely to succeed when victims are younger and report higher levels of education.

Evidence presented in this section shows that victims of home robberies in Mexico are more likely to suffer injuries when perpetrators are armed with firearms. This result stands in contrast to previous literature on robberies in the U.S. and studies exploring street robberies in Mexico. Data presented in this section suggest that these discrepancies occur because gun-related home robberies are likely perpetrated by organized criminal groups. Similarly, while perpetrators of home robberies are more likely to succeed if they exert physical violence against victims, this is not the case for street robberies. This suggest that injuries are more likely to occur during home robberies because they occur in closed spaces, without public attention.

Discussion and conclusion

Our findings suggest that the likelihood of injury during home robberies in Mexico depends on numerous factors, including the number of perpetrators, whether victims are alone, whether victims know the aggressors, whether a crime takes place at night, whether they occur in urban settings, and whether perpetrators are under the influence of drugs or alcohol. We also found that the likelihood of injury depends on characteristics of the victims. These included victims’ sex, age, and education levels, the latter being a proxy to income.

Overall, evidence suggests that victims of home robberies in Mexico are more likely to report an injury when offenders use firearms. These results contrast findings in the U.S. as well as those analysis of street robberies perpetrated in Mexico. To provide potential explanations for these discrepancies, we conducted further analyses. A potential explanation is that unlike in the U.S., home robberies involving guns in Mexico are perpetrated by members of organized criminal groups. As discussed on section two, these organizations have become more violent. Therefore, individuals working for an organized criminal group would be more likely to injure a victim to obtain profits. Similarly, individuals working for these groups would have access to firearms that are not as readily available to other offenders. Additionally, these individuals would likely operate in high numbers, a notion that is supported by data showing that home robberies committed with a gun are linked to a higher number of perpetrators. In contrast, perpetrators that are not involved with an organized criminal group would have a harder time obtaining a gun and would instead be unarmed or use other weapons to perpetrate a crime. Home robberies perpetrated by offenders that are not members of an organized criminal group would likely operate unaccompanied. This idea is supported by our findings on section four. 66% of home robberies involving unarmed aggressors and 44% of those involving offenders armed with non-firearm weapons were perpetrated by a lone individual.

Another explanation is that, during home robberies, guns are used to control victims and deter any form of resistance. Once in control, offenders may decide to use other forms of physical force as a mechanism to coerce victims and obtain higher profits. Offenders of home robberies can use additional physical violence without garnering any public attention since these crimes take place in closed spaces and involve different dynamics than, for example, street robberies. This relationship between the presence of firearms and the environment of crimes opens an interesting line of research for further studies.

Another interpretation that could help explain why victims are more likely to sustain injuries during home robberies involving firearms is that unlike street robberies, unarmed home robbery offenders or those armed with non-firearm weapons do not anticipate an interaction with victims and are not prepared to engage in physical violence. Instead, when unarmed aggressors or perpetrators armed with non-firearms weapons encounter a victim during a home robbery, they might decide to flee. In contrast, when aggressors anticipate an interaction with victims, they might decide to use a gun and prepare to engage on physical violence.

Furthermore, while this study adds to the literature discussed in section one, it also contributes to the broad conversation on gun trafficking and violence in Mexico presented in section two. Injuries and economic losses derived from home robberies involving firearms highlight the urgency to incorporate gun-related policies within Mexico’s national security agenda to prevent criminal groups from easily acquiring firearms. Examples of these policies have been developed and compiled by different organizations (see Holst 2019; Baez Zamudio et al. 2018).

Due to data limitations, we were not able to analyze the likelihood of injuries and economic losses of victims when protective actions were taken, particularly those involving firearms for self-defense. Therefore, this study does not assess the efficacy of using a firearm in the home for self-defense purposes. However, studies from the U.S. indicate that guns are rarely used for self-dense and are not effective at preventing crime, theft, or injuries (Violence Policy Center 2019; Hemenway and Solnick 2015; Donahue et al. 2019; Cook and Ludwig 2002). Overall, these studies raise important concerns on the efficacy of implementing more permissive gun laws to address the problem of home robberies in Mexico.

Moreover, gun trafficking and access to firearms for criminal organizations is an issue that cannot be solved by Mexican policies alone. Reports indicate that close to 70% of guns recovered in Mexico originate in the U.S. (United States’ Government Accountability Office 2016). Therefore, if the U.S. is committed to reducing organized crime, it must consider the violence perpetuated in Mexico by organized crime groups armed with U.S.-sourced firearms because of weak U.S. laws and regulations around firearms and ammunition. Therefore, a natural approach must include actions within the U.S. to reduce cross-border gun trafficking. These actions include mandating background checks for all gun sales and transfers, banning assault weapons, and more effective prosecution of straw purchasers (Parsons and Weigend 2018).

Appendix

Table 7 Odds ratio of non-weapon related injuries during home robberies in Mexico, 2010–2017

Table 8 Odds ratio of weapon related injuries during home robberies in Mexico, 2010–2017

Table 9 Odds ratio of injuries during home robberies in Mexico, 2010–2017. (Non-firearm weapons broken down to two groups)