by Carolina Torreblanca and José Merino
Between 2004 and 2016, 26,266 women were murdered in Mexico; we estimate that 34% of them, 8,913 women, were victims of femicide.
The systematic differences between murdered men and murdered women: what we can know
The violence men and women experience daily in Mexico is very distinct. While it is true that men have been most of the victims of the huge surge of violence that has flooded the country since 2006, there is another type of violence which precedes and will probably outlive this recent surge: the constant and persistent violence which affects women. To understand it, we must understand the systematic differences between murdered men and murdered women in Mexico.
To start with, a much larger percentage of women than men are killed at home.
While 32.3% of murdered women die at home, only 12% of murdered men die there. That is, while one in three murdered women was killed at home, only one in every ten men died in the same location. It is important to stress that for almost 25% of homicides, we have no information regarding where the killings took place.
The age of the victims is also markedly different: a larger percentage of elderly women and young girls are murdered. While girls from 0 to 11 years make up 4.4% of all female murder victims, boys of that same range represent only .7% of male victims. Elderly women account for 7.5% of all female victims, while their contemporary males represent only 4% of male homicide victims in the same period.
Male and female victims of homicide also differ by their marital status. A larger proportion of female victims are divorced, 5% versus 1.6% of males respectively. On the other hand, more female than male victims were single, 37% compared to 34% for men.
Recently, a new contextual variable was added to the published mortality data, it captures whether domestic violence was related to the homicide or not. However, in 95% of the cases where the victim was male and in 91% of the cases where she was female, this variable is coded as ‘not available’.
This lack of information notwithstanding, a much larger percentage of female homicides than male homicides was connected to domestic violence. While only .8% of male homicides was reportedly connected with domestic violence, the percentage is 5.2% for female murder victims.
Given the systematic differences in the profiles between male and female homicide victims, it is unsurprising that the specific cause of their death, that is, the precise fashion in which the murder was committed, also varies widely.
The most notorious difference is the frequency of homicides by firearm; while on average 64% of male homicides in the last 13 years has been committed with a firearm, it was only used in 41% of female homicides.
Instead of dying by gunshot, women are more often strangled or suffocated. On average, during this period this method has been the cause of death of almost 20% of female homicide victims, while only 6.3% of male victims died because of strangling or suffocation.
What is a femicide? What we can infer given what we do know
Between 2004 and 2016 26,266 women and 211,436 men have been murdered in Mexico. Both the victims and the perpetrators are mainly men, however comparing the absolute magnitude of the violence each sex faces is not only useless but it also helps to obscure the substantial differences that do exist between them.
If we compare the female homicide murder rate to the male homicide murder rate for the last 13 years we can see that, while the former has remained basically constant throughout the whole period, the latter has followed and driven the ups and downs of the national violence epidemic. We are faced, as we had previously stated, with two different phenomena. Although women are also victims of the violence associated with organized crime, their level of victimization is much less volatile than that of men: women die as a result of a constant violence.
It would be a disservice to argue that every homicide where the victim was female qualifies as a femicide. By doing this we would not only be emptying the word of any specific content, making it a synonym with homicide, but we would also be erasing the devastating specificities of the violence women suffer just for being women.
So, how many victims of femicide have there been in the last years in Mexico? First, we must state what does qualify as a femicide.
In the international arena a femicide is defined as a “homicide for gender reasons”. This definition, while clearly asserting the quid of the issue, makes measuring the phenomenon almost impossible. However, the Mexican federal penal code provides a more actionable definition; it considers that a femicide has taken place if:
I.- The victim presents signs of sexual violence of any kind;
II.- The victim was subjected to degrading or offensive injuries, before or after the homicide or the victim was subjected to acts of necrophilia;
III.- There is a known background of violence in the home, school or work environments against the victim;
IV.- The victim and the perpetrator had had a sentimental, caring or trusting relationship;
V.-There is information that points to there being previous threats relating to the crime, such as harassment or other injuries;
VI.- The victim was kept uncommunicated for whatever period before the murder;
VII.- The victim’s body is exposed or exhibited in a public place.
This means that if we wanted to accurately count the total number of femicide victims we would have be able to know if the victim suffered sexual violence, even when this is not the specific cause of death; if the victim was mutilated or injured before or after death; if the perpetrator was an acquaintance or had a relationship with the victim; if the murder was preceded by harassment or other injuries and if the corpse was exposed in a public place, regardless of where the murder was committed. Mortality data published by the National System of Health Information (SINAIS for its name in Spanish) thorough the National Institute of Statistics and Geography (INEGI for its name in Spanish) is insufficient for this purpose.
What can we know? Where the victim died, who she was and how she was murdered; additionally, we have the domestic violence variable, even though (as we have seen) it contains no information for 90% of the cases. With these tools, we construct our own category of “femicide”.
To include those homicides in which the victim and the perpetrator had a previous relationship of any kind, we count as a femicide every instance where the homicide took place at home, regardless of how the homicide was committed. Of course, this way to account for the previous existence of a relationship between the victim and the perpetrator is imprecise, but it is the best proxy considering the information available.
Along with murders committed at home, we count as a femicide all instances where the specific cause of death was listed as “sexual aggression”, regardless of where the murder was committed. Lastly, we counted as a femicide all murders where domestic violence was reported.
Per this new typology, 34% of 26,266 women who were murdered in Mexico between 2004 and 2016 were victims of femicide.
This typology does not mean to imply that women murdered on the street can’t be the victims of femicide as a result of a systemic violence that extends far beyond the confines of familial relationships; what it does do is recognize that it is impossible to know for sure, unless sexual abuse or domestic violence is documented.
If we now compare the female homicide rate, meaning the women murdered each year that are not counted as being victims of femicide per our typology, with the femicide homicide rate we can see that they follow starkly different patterns; while the femicide murder rate per 100 thousand women is very stable during the entire period, with a minimum of .8 and a maximum of 1.4, the female homicide rate follows the national pattern: an increase starting 2007, peaking in 2011 and a new wave of violence from 2015 onward.
These differences allow us to see that, on the one hand, our typology of “femicide” is capturing a particular sort of violence, constant and permanent, that affects women regardless of what is happening on the national level and that, on the other hand, there is a group of women who is in fact affected by the same rise in violence that affects men since 2006.
Femicides in the states
Distinguishing between homicides and femicides allows us to place special attention to those states in the country where the phenomenon is particularly troubling.
In 2016, Tamaulipas was the state with the highest rate of femicides in the country, with 3.5 femicides for every 100 thousand women, next come Guerrero, Colima and Morelos, while Yucatán and Aguascalientes are the states with the lowest rates of femicides.
Precisely because femicidal violence is different from homicidal violence, there are states which have a very low femicide rate but a high homicide rate or vice versa. For example, Michoacán has a rate of 4 homicides for every 100 thousand women, the seventh highest in the country, but a femicide rate of only 1.2.
In general, states tend to have a higher female homicide than femicide rate, however there are six exceptions: for Tamaulipas, Baja California Sur, Campeche, Tlaxcala and Durango the situation is the opposite.
Since 2011 INEGI has published a State Justice Census which allows us to know the exact number of investigations opened for each crime every year in all the states. Since 2014, those crimes include “femicide”.
In 2014, according to our own definition of a “femicide” implemented with the data provided by SINAIS, there were 802 femicides in the country; however, according to the Census there were only 276 investigations opened for the same crime with 322 women killed.
If we subtract the total number of victims of femicide reported by the Census from the total number of femicides per our own estimations, we get that the Estado de México is the state with the larges disparity. While according to our own methodology there were 127 victims of femicide there that year, the Census reports that there were only 48; a difference of 79 women.
The second state with the biggest disparity is Oaxaca, which shows a difference of 58 women. On the other end of the spectrum there are 5 states which reported more victims of femicide than what our own calculations estimate. In Aguascalientes, the prosecutors investigated the murder of 1 woman more as femicide than we estimated had occurred in that state in the same period; while in Morelos the difference was of 10 women.
These discrepancies show very clearly that what makes a homicide qualify as “femicide” in one state is not necessarily sufficient in another and, therefore, to use the Census, that is the information provided by each state’s state attorney, to tally the total number of femicides in the country results in a biased and inexact measure.
Everything we don’t, but should know: The Bogotá Protocol
The Bogotá Protocol on the Quality of Homicide Data for Latin America lists all the minimal requirements to be included when publishing homicide data. These requirements incorporate publishing not only information regarding the victim, but also information surrounding the homicide. Naturally the date and time of the crime, but also of the discovery of the corpse, the motives of the homicide and (something of the highest importance if we want to be able to count femicides) detailed information on the alleged assailant, including his or her relation to the victim.
Doubtlessly our approximation to counting the number of femicides in Mexico is imprecise, this is the inevitable consequence of the fact that both the data on the number of investigations published in the Census and the data based on death certificates, which is gathered by SINAIS and published by INEGI, don’t include all the information the Bogotá Protocol stipulates.
If we want to know the real magnitude of the femicidal violence in the Mexico we need better quality data. However, what our methodology to count femicides with the information we do already have makes clear is the fact that the violence women and men experience is starkly different. Women suffer a constant and permanent violence.
What we want and what is needed is better data to measure this specific violence more precisely.
The views expressed here are solely those of the authors.
 See the Inter-American Court of Human Rights’ ruling “González y otras (Campo Algodonero) vs. México”, 2009
by Luis Rubio
by Miguel Toro
 The head of the FEPADE which is the Special Prosecutor's Office for Electoral Crimes.
 Note: He is an interim in the job because the old Attorney General resigned without a very substantive explanation of the reasons why, a couple of weeks ago.
 Let’s note that Odebrecht has been convicted in several Latin American countries of bribing different politicians to get special treatment and concessions.
 Note: The combined efforts of all opposition parties had the votes to reject the firing from the Attorney General’s Office.
 Note: I don’t presume this list to be exhaustive as it came of the top of my head and not after extensive journalistic reviews. The other parties also have their own corruption scandals, but in this administration, the number of government cases have been vastly superior.
 Note: I refer to him as the third, because he is the son and grandson of two former State of Mexico governors with the same name. He is also the cousin of President Peña Nieto.
 It must be noted that Miranda has always worked for the public sector whose salaries are high but seems improbable that they would be enough for a house of that magnitude.
 This is a yearly survey done to more than 20 thousand respondents in all the countries of Latin America regarding democracy and its values. The report can be consulted here: http://www.latinobarometro.org/lat.jsp
 His November 3rd article can be found here (in Spanish): http://www.elfinanciero.com.mx/opinion/perfil-presidencial.html
by Luis Rubio
 This term was used to describe closed-doors agreements between the Salinas government and the PAN that allowed the latter to obtain several local positions (governors and mayors of certain states) where the local PRI members had clearly committed electoral fraud in those races in exchange for support in Congress.