Why Disasters Aren’t Gender-Neutral: What the Evidence Does and Does Not Show

1Background: a claim that needs checking

It is often stated that women and girls die in disasters at much higher rates than men. The question matters because it shapes who gets targeted by preparedness programs. This brief reviews what the mortality evidence supports, where gender clearly shapes exposure and access, and what responses have been recommended or tried.

One problem runs through all of it: there are few reliable sex-disaggregated counts. After the 2004 tsunami, Oxfam noted that there was precious little accurate, disaggregated data showing how many of the dead were women.^[3] More than two decades later, a 2026 review reaches a similar conclusion.^[2]

2The cross-country evidence

The most cited quantitative study analyzed up to 141 countries over 1981 to 2002. It found that natural disasters lower the life expectancy of women more than that of men, so that on average disasters kill more women than men or kill women at an earlier age. The effect grew with disaster strength, measured as deaths relative to population, and weakened where women’s socioeconomic status was higher.^[1] The authors argued that biological and physiological differences are unlikely to explain large-scale differences in mortality, and that what matters most is the everyday socioeconomic status of women.^[1]

That is a statement about life expectancy, not a count of deaths by sex in each event, and it describes a statistical average across countries and years.^[1]

3A contested picture

A 2026 study in the journal Disasters tested the claim directly by compiling new datasets on sex-differentiated fatalities in the twenty-first century across geophysical, meteorological, climatological, hydrological and biological hazards. Its findings do not support the widely cited claim of consistently higher female mortality. Instead, sex-disaggregated data remain very limited and the evidence is largely inconclusive, except for biological disasters, where male fatalities are consistently higher. The author recommends that sex-specific patterns be assessed empirically, not assumed, and that countries introduce mandatory, systematic reporting of gender-disaggregated fatalities.^[2]

The two results are not necessarily in conflict, since they use different outcomes, periods and data. Together they say that women’s greater vulnerability is real in some events and settings and not demonstrated as a universal rule, which is a reason to measure it locally and to collect the data that would settle it.^[1,2]

4What individual events show

Some events show large differences. In village surveys after the 2004 Indian Ocean tsunami in Aceh, Oxfam found that 284 of 366 deaths in four North Aceh villages were female (77%), and in the worst-hit village there were four female deaths for each male death. In four Aceh Besar villages only 189 of 676 survivors were female (Figure 1). Oxfam describes the evidence as abundant but partial, and notes that the wave struck on a Sunday morning when women were at home and many men were away on errands.^[3]

In the 1991 Bangladesh cyclone, a rapid survey of 45 housing clusters and 1,123 people found that mortality was greatest among children under 10 (26%) and women older than 40 (31%). The authors recommended that these groups be given special attention by families, neighbors and the authorities in charge of warnings and evacuation.^[4] The survey was nonrandom, so the figures are indicative.^[4]

5Where gender shapes access to protection

Beyond mortality counts, qualitative evidence points to specific barriers. In a consultation with 50 women leaders in coastal Bangladesh, participants said that limited mobility in public spaces and resulting social isolation can mean women do not receive early warnings, or do not receive them in time, and that a lack of decision-making power over what to do with the information adds to their vulnerability.^[5] The same document recommends that emergency shelters be designed with privacy for women, including secure doors and adequate lighting, and with safe cooking, bathing and toilet arrangements.^[5]

A qualitative study of 24 women in two disaster-prone areas of Bangladesh found that shelters lacked privacy, heightened the risk of abuse, and that women faced social stigma if displaced from their homes. Women also reported health risks from inadequate hygiene, toilets and secure gender-segregated living arrangements.^[6]

A systematic review of research on women as active subjects in the emergency phase identified only 25 articles that met its criteria out of 63,852 studies initially found, and concluded that scientific production on this phase remains scant.^[8]

6What responses are documented

Bangladesh’s Cyclone Preparedness Programme has made gender parity a key principle. Women now constitute about 50% of its volunteer corps, up from a ratio of 1 to 6 in 1992, and each village unit has 10 women and 10 men. The programme’s authors argue that greater involvement of women enables it to reach vulnerable women more effectively, while noting that women still need gains in leadership positions and that women’s security in shelters and reproductive-health care need to be part of humanitarian action.^[7] These are program descriptions and arguments by the authors, not evaluations of their effect on deaths.

Consulted women leaders also asked for participation in designing community-based warning systems and for more informative and participatory programming on disaster preparedness.^[5]

7Implications for disaster risk management

Analysis. This section is Disaster Science’s interpretation of the evidence above. It is not drawn from the cited sources.

Count before you assume. Sex-, age- and disability-disaggregated reporting of deaths and displacement is the most basic and the most neglected step. Without it, programs cannot tell whether a gender gap exists in their setting or whether it is closing.

Target the mechanism, not the label. The documented pathways, such as warning access, caregiving during evacuation, swimming and mobility skills, and shelter safety, are specific and fixable. Programs can address them without relying on a universal claim about who dies.

Put women where decisions are made. Volunteer corps, shelter committees and warning design teams that include women are cheap changes with plausible benefits, but they should be evaluated for outcomes, not just participation.

Avoid claims you cannot support. Overstated statistics invite correction and can discredit sound programs. The case for gender-responsive planning rests on documented barriers and data gaps, and it is strong without exaggeration.

8Uncertainty and limitations

Mixed mortality evidence. The cross-country life-expectancy result and the 2026 fatality inquiry reach different conclusions about a general female excess, and neither is a complete count.^[1,2]

Event surveys. The Aceh figures come from eight villages surveyed by Oxfam, which calls the evidence partial, and the 1991 Bangladesh figures come from a nonrandom survey.^[3,4]

Qualitative evidence. The warning and shelter findings come from consultations and a 24-person interview study in Bangladesh and may not apply elsewhere.^[5,6] This brief found no verified evidence for the original claims that recovery aid paid to women is reinvested more broadly or that safer shelters raise uptake, so those statements were removed.

Scope. This brief synthesizes peer-reviewed papers, an NGO briefing and a UN Women consultation report. It presents no original data and is an editorial synthesis that has not undergone peer review.