Lecture 3- Child Gender and Parental Investments in India (Barcellos, Carvalho, Lleras Muney 2013) Flashcards
Who wrote, “Child Gender and Parental Investment in India” and when?
Barcellos, Carvalho and Lleras-Muney in 2013
What does the previous literature say about the relationship between parental investment and gender?
Previous literature generally does not support the theory that differing outcomes between males and females in salary, labour force parcipitation and fundamental rights (voting) are due to differential treatment by parents.
Deaton (2003)- showed that levels of vaccination between boys and girls are the same in India
Harriss (1995) showed that the level of nutrition given to boys and girls is the same in India.
What was the mistake of previous researchers in this area?
They assumed that families that had girls had the same characteristics as families with boys. Since this is not true, their results are biased and not robust.
Why might families with girls have differing characteristics to a family with boys?
In India, it is seen as preferable to have sons compared to daughters. There are “son preferences”. Because of this, a family where a daughter is the first born compared to a son is more likely to keep having children to gain a son. So on average, girls are in larger families than boys.
How can girls being from larger families on average skew the results of a study?
Girls being from a larer family can cause the differing treatment of boys and girls on average to be underestimated as if there are returns to scale, and the older children are earning an income, these families will have a larger income, increasing the money spent on girls relative to boys. On the other hand, if the children don’t work, the differing treatment of girls could be overestimated as if most of them are from larger families there may be less money to spend on each child.
How did Barcellos reduce the risk of bias in this paper?
They restricted the sample to families who had just had their first baby
Why did Barcellos restrict the sample to families who had just had their first baby?
Because they assumed that the child’s sex is exogenous in these cases as characteristics of a family who just had their first baby should be the same
What was the main goal of the paper?
To find out if parents invest more time and money on sons compared to daughters and, if these factors can be used as an explanatory variable for different wages and labour force participation.
What is the main regression used by Barcellos and Lleras-Muney?
Zih=ao +a1*Bih+aXih+uih
(Z stands for investment in child I in household h)
B is a dummy equaling to 1 if child I in household h is a boy
U is the error term.
What is the empirical strategy used by Barcellos in “Child Gender and Parental Investment in India”
They assume that for parents who have a baby that is too young to have another sibling (0 to 15 months) the sex of the child is exogenous as gender-biased stopping rules haven’t impacted the characteristics of the family yet.
However, this isn’t true if sex-selective abortions are taking place.
How will Barcellos know if Sex-selective abortions were taking place?
If the treatment of boys and girls that are 0-15 months (too young to have a sibling) are different then this due to differing parental inputs not circumstance.
Barcellos carried out a test with the null-hypothesis being that family characteristics don’t predict the gender of the child.
She doesn’t reject this null hypothesis for children that are very young however it is rejected for children that are “old enough” to have a sibling
What does Barcellos do to try to understand the differing treatment of boys and girls.
Tests to see if this is because boys have greater needs on average than girls.
This has been refutes as girls have the higher mortality rate.
Also, vitamin A supplementation is shown to have an equal effect on boys and girls so prioritizing boys doesn’t make sense.
Also, vaccinations have a slightly better effect on girls.
What data-set was used by Barcellos?
A time-survey was used to record the amount of time spent with sons and daughters.
This was in households where they had one child of age 0-15 months.
Barcellos tested and found that for such households, sex of the child is exogenous so results are robust and can be compared for household with a boy vs girl.
