In many cultures, there is a strong preference for male offspring. While the human rights movement and the emancipation of women has in many cases curtailed preferential treatments, at least the most visible ones, new technologies like ultrasound have allowed this practice to flourish in some parts of the world. China and India, in particular, report unusually high male/female ratios. This surplus of males must have consequences on the marriage market, especially for the less desirable ones: the poor.
Lena Edlund and Chulhee Lee study this with the background of Korea. Sons are appreciated because they have sons, thus a married son is preferred to a married daughter, who is preferred to a single son. When the country was poor, however, single sons were still appreciated because they could provide old age support to their poor parents. As the country, and its poor, got richer and with the introduction of an old-age pension system, the value of the single son drops significantly and the sex ratio gets closer to normal.
What puzzles me in all this is why the poor did not favor daughters. They were relatively rare, thus could garner a substantial bride price. Thus there should have been a surplus of male children in rich families and of female children in poor ones. The paper does not provide any evidence of this in the data, but it seems to be a natural consequence of the model to me.