Media Briefings

Mortality Risks In The 2004 Indian Ocean Tsunami: The Impact Of Household composition

  • Published Date: September 2011

In Aceh, Indonesia, the Indian ocean region that suffered the most deaths in the 2004
tsunami, people in households with more adult males were less likely to die. This was
particularly true for households with prime age males, who presumably were the strongest
and best able to help family members who struggled in the water.
These are among the findings of a new study of the individual, family and environmental
factors that affected mortality risks in Aceh, conducted by a team of Indonesian and US
scientists led by Elizabeth Frankenberg and Duncan Thomas of Duke University. Their
analysis, published in the September 2011 issue of the Economic Journal, also shows that
prime age males were more likely to die if their wives or sons died.
This evidence indicates that even in the face of imminent death, family members reached
out to help each other, as stronger members exploited their comparative advantage and
attempted to help their weaker relatives. In some cases, they were successful; in some
cases, they were not.
This study of tsunami-related mortality analyses a representative survey of 6,000
individuals who, prior to the tsunami, were living in areas of coastal Aceh that were
subsequently heavily damaged by the force of the water.
The research team has tracked down and interviewed the survivors, many of whom were
living in new locations. They have also worked carefully to determine who died in the
tsunami. Over one quarter of the study population perished. In the worst hit communities,
nearly three quarters of the population died.
The study confirms that mortality risks were greater in populated areas that were closer to
the water and in lower-lying areas. But moving beyond these well-known risk factors, the
study provides several additional insights:
 In many natural disasters, the poorest are hit the hardest. This was not the case for
the tsunami in coastal Aceh, which did not discriminate on the basis of standard of
living or socio-economic characteristics. In part, this is because the tsunami was
largely unanticipated and the waves hit the shore with little warning.
 Mortality risks varied across demographic groups. Females were more likely to die
than males; and children and older adults were more likely to die than prime age
adults. Almost 30% of children under 15 and nearly 40% of older women in the study
population died. This is consistent with strength and ability to swim playing an
important protective role.
 But individual characteristics were not the only factors that affected mortality risks:
the demographic composition of households also affected the probability of survival.
Specifically, individuals in households with more adult males were less likely to die.
This was particularly true in households with prime age males.
 On the other hand, if an older woman was in the household, mortality risks were
higher for prime age women and children – perhaps because they died trying to help
the older woman or because she competed for help from stronger members.
 If a prime age male is the strongest household member, his death should not be
related to the death of other members unless he died trying to help those people. It
turns out that prime age males were more likely to die if their wives or sons died.
In recent years, several large-scale natural disasters – earthquakes, tsunamis, floods and
hurricanes – have caused massive damage around the globe. The 2004 Indian ocean
tsunami, a repercussion of one of the most powerful earthquakes ever measured, caused
more deaths than any tsunami in recorded history. Aceh, Indonesia, was the hardest hit
region with over 130,000 people losing their lives.
Notes for editors: ‘Mortality, the Family and the Indian Ocean Tsunami’ by Elizabeth
Frankenberg, Thomas Gillespie, Samuel Preston, Bondan Sikoki and Duncan Thomas is
published in the August 2011 issue of the Economic Journal.
Elizabeth Frankenberg and Duncan Thomas are at Duke University. Thomas Gillespie is at
the University of California, Los Angeles. Samuel Preston is at the University of
Pennsylvania. Bondan Sikoki is at SurveyMETER, Yogyakarta, Indonesia.
For further information: contact Romesh Vaitilingam on +44-7768-661095 (email:; Elizabeth Frankenberg via email:; or
Duncan Thomas via email: