Missing Women: Son Preference and Its Effects
Abstract: This paper reviews the literature on the “missing women” of the world, including
the estimations of the number of missing, the underlying social causes for son preference
across societies, and the mechanisms by which societies skew their sex ratios.
Starting in the late 1980s and early 90s, social scientists began noticing strange anomalies
coming out of the censuses of many countries in South East Asia and Northern Africa, the
most notable among them being Amartya Sen’s publication in the 1990 New York Review
of Books. Son preferences within Chinese, Indian, and South Korean societies had long
been noted, but the total count of men and women in the population revealed massively
skewed numbers in favor of males (Das Gupta et al. 2003, 4). In most western countries,
the sex ratio at birth stands at around 1.05 boys for every girl born (Sen 1990). This initial
surplus of males only lasts for a short time, however, because women tend to have higher
survival rates in all age brackets of a population, especially during times of war or famine
(Sen 1990). Thus, while males start out by outnumbering females, women begin to
outnumber men and continue to widen the gap as the average lifespan rises, to the point
that the population sex ratio of the United States stood at 0.95 men to women when
Amartya Sen first wrote of the discrepancies. There are several biological mechanisms put
forth as explanations for higher male mortality, but the countries with son preference
exhibit exactly the opposite characteristics of mortality. In China, for instance, the sex ratio
at birth has been recorded as being anywhere from 1.07 (Oster 2005) to 1.138 (Zeng et al
1993), and the superabundance of males continues to manifest throughout the population
such that the total population’s sex ratio stood at 1.06 (Sen 1990), an 11 percent difference
with the United States. Thus, fewer females are being born in these countries than would
occur naturally, and they are dying at a much higher rate than their male counterparts
throughout most age brackets, contributing to an intergenerational lifetime disparity that
appears across the entire population. In his original articles, Sen referred to these
populations as having ‘missing’ women, because the number of men so far exceeded the
expected proportion that women somewhere, somehow, had to be dying to achieve the
observed population’s ratio.
Measuring the Missing
As Stephan Klasen and Claudia Wink put it in their 2003 work on missing women, there
are basically two approaches to calculating the number of women missing. Missing women
can be revealed by looking at the gender mortality rates in each age bracket and comparing
these to the expected rates from similarly developing countries, with the differences being
attributable to gender bias. This approach captures which segments of society prove to be
most discriminatory, however it fails to take into account the disparities created by
sex-selective abortion and other pre-natal methods of changing the sex ratio. Accurate data
of gender mortality is also difficult to find in many countries that exhibit population
skewing, and so measurements tend to be correspondingly imprecise. Another approach is
to follow Sen’s lead, which involves looking at the population sex ratio and calculating the
number of women required to put the country on par with an expected, ‘natural’ ratio of
men to women. This method, when combined with close examinations of the birth
sex-ratio, can capture the effects of both post-natal and pre-natal discrimination, and it has
the added benefit of using more accurate measurements, such as censuses and hospital
birth records. Using this second technique, Sen’s original 1990 paper took the population
ratios of developing Sub-Saharan Africa—countries that are economically similar to
China, India, and other areas with missing women, but that have a population sex ratio of
0.9—and arrived at a figure of nearly 100 million missing women worldwide, with 50
million coming from China alone (Sen 1990). Ansley Coale produced a paper in 1991
disputing Sen’s measurements as being overly simplistic, since they failed to incorporate
all of the possible variations that can affect a country’s population sex ratio. To achieve an
accurate measure, according to Coale, a researcher should use controls based on the birth
sex ratio, the sex-specific mortality in various age brackets, and the effects of population
growth on a population’s ratio. The last bears some further explanation: countries
undergoing high population growth tend to have a higher proportion of males, since female
‘hardiness’ only manifests later in life (Coale 1991). When all of these are taken into
account, Coale argues, the actual number of missing women falls to 60 million worldwide.
Still other researchers argued that Coale’s numbers were equally as flawed as Sen’s.
Stephan Klasen, for one, wrote in a 1994 paper about missing women that Coale’s basis for
the sex-specific mortality control were as unrelated to China and India as Sen’s original
comparison of Sub-Saharan Africa was. Coale also used a static sex ratio at birth that
operated independently of country, though it did adjust for the aforementioned controls.
Klasen argued that a country’s development was likely to revise the birth ratio upwards,
since male infant survival rates rise in countries with more developed health sectors. Using
these new controls, Klasen arrived at a figure of 89 million missing women, fully fifty
percent more than Coale and nearly twenty percent less than Sen. When Klasen revisited
the data eight years later with updated census returns, he also included controls for gender
