The Statistical Absence of Femicide (The Structure of Deaths That Cannot Be Counted)

A problem that does not exist in statistics does not exist in policy either.

The category "femicide" does not exist in Japan's official statistics. Police Agency crime statistics record homicide, negligent homicide, robbery resulting in death, and similar offenses, but no systematic aggregation is conducted of "killings of women motivated by gender-based discrimination" — a category combining the victim's sex and the relationship to the perpetrator.

The UNODC (2023) Global Study on Homicide estimates that approximately 89,000 women worldwide are killed annually by a partner or family member. Japan's corresponding data, however, is limited — Japan has been unable to provide data in the form required for international comparison.

The word "femicide" itself has not entered Japan's official vocabulary. This is not the absence of the phenomenon; it is the absence of the concepts and statistics needed to capture it. What cannot be counted becomes effectively "nonexistent" in policy terms.

Fragments Revealed by "Intimate Partner Violence" Statistics

The most systematic statistics on violence against women in Japan are provided by the Gender Equality Bureau, Cabinet Office survey on "Violence between Men and Women." This survey aims to understand the reality of domestic violence (DV — intimate partner violence), but the systematic aggregation of killing counts lies outside its scope.

The National Police Agency's Crime Statistics report includes partial figures on "crimes between relatives" and "homicides between spouses," but categories such as "killings by a former partner" or "killings of women by unknown men motivated by gender-based discrimination" do not exist. The design of the statistics determines the range of problems that can be captured.

The Formation of an International Definition and Japan's Divergence

The concept of "femicide" was formalized by Diana Russell in the 1970s, and since the 2000s it has been adopted by United Nations agencies. UNODC and UN Women have issued recommendations urging member states to collect femicide data.

Japan's response, however, has been limited. Examples of the term "femicide" appearing in government documents are extremely rare. The Ministry of Justice, the National Police Agency, and the Cabinet Office each record violence against women using different classification schemes — none of which are in a form that can be integrated or compared internationally.

Behind this divergence lies resistance to the concept of "femicide" itself. Political objections to "discussing homicide as a matter of gender equality" have delayed the debate over building the relevant statistics. By opposing the concept, it becomes possible to deny the necessity of measurement. This is one manifestation of the mechanism of silence structuring.

The Statistical Rupture between "Stalker Killings" and "DV Killings"

The related statistics that exist in Japan are the National Police Agency's figures on "stalker cases" and the Cabinet Office's counts of DV consultations. These, however, are "consultation counts" and "arrest counts" — they are not directly connected to "the number of women who were killed."

There is a large gap between the number of cases that are reported and the number of incidents of violence that actually occurred. Cabinet Office surveys have repeatedly shown that many women who have experienced DV do not seek consultation. Victimization that goes unreported does not appear in the statistics. Problems that do not appear in the statistics cannot be addressed by policy.

Three Pathways by Which Statistics Become "Invisible"

There are three pathways through which femicide remains statistically uncaptured in Japan.

The first is the absence of classification. No framework exists in homicide statistics for systematically recording "the combination of the victim's sex and the perpetrator's motive." "Killed by a spouse" can be partially captured, but the determination of "killed because of sex" is difficult. Without classification, there is no aggregation.

The second is the non-adoption of the concept. As long as "femicide" is not used as an official term, the statistical design based on its definition cannot begin. Whether to adopt or not adopt a concept is a political choice. By refusing to accept the term, the measurement of the phenomenon can be indefinitely deferred.

The third is definitional contestation. The definition of femicide remains under academic debate, and how to determine "gender-based motivation" in legal and statistical terms is genuinely difficult. Using this difficulty as a reason to defer the development of statistics allows the condition of non-capture to persist. The difficulty is real, but using it as a reason "not to develop" the statistics is a political choice.

The Chain of Ignorance Produced by the Absence of Statistics

What this table shows is that "not creating statistics" is not policy-neutral. Failing to measure has the effect of making a problem appear "smaller" or "nonexistent."

An Agnotological Analysis

The core insight of agnotology as developed by Proctor and Schiebinger is that ignorance is not "a simple absence of knowledge" but "something that is produced." The absence of femicide statistics is a paradigmatic case. Technical difficulties exist, but stopping thought there is equivalent to choosing not to capture the phenomenon.

The same structure discussed in The Obstructive Effect of Strategic Ignorance in EBPM — the "non-adoption of evidence" — operates here. It is not that policy cannot be formulated because data is absent; it is that data is not collected because the priority for formulating policy is absent. The order of the question is reversed.

McGoey (2012) discussed the concept of "convenient ignorance." When not-knowing functions conveniently for particular actors, that ignorance is maintained. The absence of femicide statistics is compatible with the dynamic of those who do not want the severity of the problem to become clear. If the statistics were developed, demands for countermeasures would acquire a foundation.

Conditions for Counteraction

What is required to counteract the statistical void?

The first step is adoption of the concept. By incorporating the term "femicide" into government documents, debate over its definition can begin. Adopting the term is the starting point for discussion.

The second is re-analysis of existing statistics. Research that combines current police statistics and DV statistics to produce an approximation of femicide is feasible. Research conducted by institutions such as the National Women's Education Center partially compensates for the gaps in government statistics.

The third is alignment with international standards. Developing domestic statistical designs that can respond to UNODC collection formats will, gradually, change the condition of invisibility.

To measure is to acknowledge. Resistance to including femicide in statistics is also resistance to acknowledging the problem. Describing that resistance as "the difficulty of building statistics" is the mechanism that keeps the problem in a state of non-capture.

Related: What Is Silence Structuring? | Strategic Ignorance in EBPM | The Falsification of Monthly Labour Statistics and State-Produced Ignorance