I have little to add to Becker's convincing discussion. One small point worth noting, however, is a new technology for sex selection, described in an interesting article by Denise Grady in the February 6 New York Times. It is called "sperm sorting" and enables male or female sperm to be concentrated in semen, greatly shifting the odds in favor of producing a child of one sex rather than the other. The cost is only $4,000 to $6,000, which is much less than in vitro fertilization, since the "enriched" sperm can simply be inseminated in the woman rather than requiring in vitro fertilization. Sex selection by sperm sorting may actually be cheaper than ultrasound plus abortion, the conventional method; if so, and it comes to dominate, the ethics of sex selection will be separable from the ethics of abortion motivated by sex selection.
The key points that Becker makes, both of which I agree with, are, first, that sex selection by U.S. couples is unlikely to result in an unbalanced sex ratio; and, second, that in countries such as China and India in which there is a strong preference for male offspring, girls will be treated better if sex selection is permitted, since there will be fewer girls born to couples who did not want them. Of course, as there will fewer girls, period, the net effect on total female utility is unclear: fewer reduces total utility but happier increases it. Since the net effect is uncertain, feminist opponents of sex selection should consider whether, if unwanted girls are born, there are feasible techniques for improving their treatment so that if sex selection is forbidden (assuming that that is feasible--Becker suggests that it is not), there can be reasonable confidence that net female utility will increase rather than decrease.
I also agree with Becker that there is a tendency to self-selection, since as the percentage of girls and women declines, men's demand for them rises, and observing this couples will tend to shift their reproductive selection in favor of girls. Since there is no reason why this tendency must overcome a preference for boys, an unbalanced sex ratio could persist indefinitely. But this is unlikely in rapidly developing countries such as China and India. A strong preference for male children tends to be found in societies in which there is a great deal of subsistence agriculture, a weak social insurance system, and a reliance on private violence (as in a revenge culture) to protect personal and property rights; all these factors increase the demand for male children. As these conditions (the first two of which are important in China and India, and all three of which are important in Iraq, for example) change, the preference diminishes, as we observe in the wealthy societies of Europe and North America, where there is no longer a net preference for having male rather than female children.
Apparently sex selection is actually more common in urban areas than in rural areas of India. But presumably the reason is that access to ultrasound for detecting the sex of a fetus, and to abortion, is greater in cities, and this effect could dominate the greater preference for sex selection in rural areas. Urban Indians might prefer boys because of a lag in the adaptation of traditional values to urban conditions.
The transition to a 50-50 sex ratio, even if inevitable, is likely to take a long time. Suppose at time 1 there is a large excess of male births, followed at time 2 by a dawning recognition that girls are more valuable than had been realized at time 1. Probably time 1 and time 2 will be separated by 20 or 30 years (or more, if there is a "values lag," as I suggested earlier), and so there will be at least one entire adult generation in which the sex ratio is skewed in favor of males. Should countries that face this imbalance worry about it to the extent of taking measures against it? We have a natural experiment, which can help us to answer the question, in societies that permit polygamy. The effect of polygamy (technically polygyny--multiple wives--but polyandry is virtually unknown) is to raise the effective ratio of men to women, since a number of women are removed from the pool available to the nonpolygamous men. In a society in which there are 100 men and 100 women, but 10 of the women are married to one of the men, the male-female sex ratio, so far as the rest of the society is concerned, is 99 to 90. The result is to raise the average age of marriage for men and reduce it for women, reduce the percentage of married men and increase the percentage of married women, reduce promiscuity by increasing women's bargaining power, and possibly increase male emigration and female immigration. None of these effects seem likely to harm society seriously as a whole.
In contrast, research that I discuss in my book Sex and Reason (1992) finds that the low effective male-female sex ratio of the black population in the United States (due largely to abnormally high rates of imprisonment and homicide of young black males) promotes promiscuity because there is more competition among women for men, and reduces the marriage rate and family formation.
In sum, sex selection, at least in favor of males, appears not to have negative external effects. It presumably confers net private benefits (like other preference satisfaction), or otherwise it would not be practiced. (There are no external effects in societies, such as that of the United States, in which sex selection is unbiased.) The case for forbidding it is therefore unconvincing (at least when sex selection is not implemented by abortion, to which there are independent objections) unless it can be shown to create a net decrease in female welfare.