Here is more on gender bias. In many Southern Asian countries, there is evidence that the ratio of boys to girls is too high (see also Chapter 8). A ratio of 110 boys to 100 girls is not uncommon. One obvious hypothesis that springs to mind is that girls are treated badly relative to boys (or are perhaps even selectively aborted or killed), so that their mortality rates are higher. There could be much truth in these assertions. Nonetheless, it is worthwhile to investigate alternative possibilities.

(a) Begin by looking at the target rule discussed in exercise 14. Simplify the target rule using the following assumption: families have children until they have just one boy, and then they stop. Then show that on average, larger families will have more daughters.

(b) Now use the insight of part (a) to argue that in poor societies, girls might have a higher death rate than boys even if there is no discrimination.

Data from exercise 14

You are gathering demographic data in a village. You suspect that families have a gender bias; that is, they have children until a certain target number of sons are born, but you don’t have direct evidence of this. All you have is information on the sex and birth order of each child born to each family in the village. How would you use the data to test your hypothesis that there is gender bias?