In response to:

India's Women: The Mixed Truth from the October 10, 2013 issue

To the Editors:

I spotted an uncanny correlation of figures in Amartya Sen’s article on India [“India’s Women: The Mixed Truth,” NYR, October 10].

Rate of recorded      Female births per
rape per 100,000     1,000 male births
people in state         in state

Mumbai     1.2         902   (Maharashtra)
Bangalore  1.1         944   (Karnataka)
Chennai     0.9         946   (Tamil Nadu)
Calcutta     0.3         947   (West Bengal)

This is almost too neat. But it makes sense, at least superficially: the less women—the more rape.

Lothar Luken
County Cork, Ireland

To the Editors:

I was surprised to see in Amartya Sen’s interesting article about violence against women in India no mention of the dowry system. Violence against women in India is not only sexual. Data regarding the maiming, murder, and suicide of brides resulting from in-law dissatisfaction with the money or motorcycles or other goods delivered [as a dowry] following marriage are similarly gruesome as those for rape, and the poor response by police and the courts similarly disturbing.1

Recent studies on the persistence and even escalation over the last decade of dowry-related expenditures and associated crimes2 owe much to Professor Sen’s own call (in the British Medical Journal in 2003) for a better understanding of sex bias in natality—another form of brutality he highlights in his recent NYR article. Surely there are deeply rooted cultural reasons for son preference. However, the prospect of giving birth to a girl (or even worse, a second one) can spell financial ruin to a family that may have to pay out two thirds of its household assets to marry off a daughter.3 It is no wonder, and certainly no “oddity” in light of simple economics, that “daughter aversion” persists. (In the 1990s before sex-selective abortions were illegal, ads on mobile sonogram units read, “Pay 50 Rupees now to save 50,000 Rupees later.”)4

The effects of this aversion are also felt by living girls, especially if they are the second or third in a family. One study showed that at age six months to forty-seven months, the odds ratio of being severely stunted for a girl with two older sisters compared to a boy with two older sisters was 1.61.5

Professor Sen mentions that women’s education “does not seem to be adequately effective in reducing discrimination against giving birth to girls.” Analyses of National Family Health Survey (NFHS) data regarding son preference have shown that women’s education is, at least, positively associated with more egalitarian preferences. In contrast, wealth and economic development are not.6 Daughter aversion is not just a problem of the poor and underprivileged.

Paying dowry has been illegal in India since 1961. Use of ultrasound to detect the sex of a fetus has been illegal since 1994. But the laws have had little effect. It is time that neither class of crime should ever be mentioned without reference to the other. And it is time that the dowry system—which endangers females at multiple times in the life cycle—be more aggressively targeted in India as an insidious social, economic, and public health menace.

Renata Seidel
Evanston, Illinois

Amartya Sen replies:

Lothar Luken makes a very interesting observation about telltale signs of a correlation between (1) the incidence of rapes in the cities in different regions of India, and (2) the lowness of female–male ratios at birth in the regions in which these cities are respectively situated. Even though there is a more complex picture here than can be captured by the simple formula “the less women—the more rape,” the general concern that Luken’s letter brings out is extraordinarily important: How do the different kinds of gender inequality relate to one other?


Amit Bhargava/Corbis

A sign saying that pre-natal sex ­determination is illegal, St. Stephen’s ­Hospital, New Delhi, 2005

Different aspects of gender inequality may tend to relate to one another through their shared link with disrespect for women, the presence and severity of which vary over different parts of India. Gender bias can be expected to influence the prevalence both of rape and of sex-selective abortion (which in turn influences the female–male ratio at birth).

There are, however, other dimensions of gender inequality that do not seem to move together. One such issue of great importance for policy, as I discussed in my essay, concerns the differential reach of the impact of female education. Female schooling is one of the most liberating factors in reducing gender discrimination in general, including the neglect of girls compared with boys, which is sharply less for children of educated mothers. Yet the effectiveness of this liberating factor seems very weak in preventing the abortion of female fetuses. As I explained:

Educated mothers seem clearly less inclined to neglect girls compared with boys once they have been born; but they seem almost as keen on having boys rather than girls as uneducated mothers are.

Evidently there is a need to supplement formal education, important as it is, with active public advocacy of gender equity, which has been effective wherever it has occurred, in India’s political experience.


In her interesting letter, Renata Seidel seems to dispute the absence of a strong empirical connection between women’s formal education and female-fetus abortion, but unfortunately she misses the distinctions involved. She cites a fine paper by Rohin Pande and Anju Malhotra, which does discuss the effectiveness of female education, but omits to note that the subject matter of the paper is the treatment of girls after being born, which is reflected even in the title of the Pande-Malhotra paper: “Son Preference and Daughter Neglect in India: What Happens to Living Girls?”

Renata Seidel’s other point, about the importance of dowry deaths, is entirely persuasive. They are a serious aspect of gender inequality, particularly for more well-to-do families. Dowry deaths have received more attention than perhaps any other feature of gender inequality in India. The middle-class-dominated media in India, including the social media, are quite attentive to this terrible evil, but pay far less attention to other evils from which large numbers of poorer Indians suffer. I have discussed the evil of the dowry issue elsewhere, but in this essay, I was focusing particularly on the plight of poorer Indian women, who cannot expect to provide much by way of dowries (such as “money or motorcycles”) and who get raped without getting the kind of attention that the rape of a medical student in Delhi last December received, and whose children get captured and trafficked without being able to get much attention from the media.

In my discussion of the regional contrast in gender inequality, I put some emphasis on the gigantic dichotomy between states in the north and west of India where the female–male ratio at birth is sharply below the European range, and the states in the south and the east where the ratio is still well within the European range. (Some readers have asked where I got the figures of European female–male ratios at birth that I quoted. They come from the Population Division of the United Nations Department of Economic and Social Affairs, and in particular from the document World Population Prospects: The 2010 Revision (2011).) There are other sources of similar data as well, and there are some minor variations in the numbers reported by different organizations (partly perhaps related to the reference year). But the important point to recognize here is that the regional dichotomy of India is unaffected by these variations.

In fact, we can even vary the particular European countries against which to measure India’s regional contrast in gender inequality. In taking 93.5 as the cutoff low line of female ratio at birth, I referred specifically to Ireland (which had that ratio for 2000–2010 according to UN statistics), but five other European countries, including for example Serbia, have ratios that are at or just below 93.5. Even if we go a little higher within the European range, for example drawing the cutoff line around 94.0, corresponding to Spain (94.0), Italy (94.1), and Greece (93.9), every southern and eastern state in India (with one marginal exception, Odisha) stands above that higher cutoff line, while the northern and western states, all with ratios well below 92.0 (going down all the way to 84.2), fall even further behind. The division within India does seem clear and robust.

I ended my essay with the confession that I did not know of any “convincing clear-cut” explanation of what causes the identified contrast that splits India sharply into two halves. There are, of course, correlations with language groups and cultural tendencies, and also with economic and social variables (on which I am working), but the causal connections that could explain these empirical relations are not very clear—at least not yet. Any help would be welcome.