Computer Use and the Demand for Female Workers Assignment

Assignment Task

Computer Exercise

Problem

1. How do female workers choose between kids and jobs? Under the gender equality, this is a tremendously important issue facing all families. To examine the causal relationship between having kids and labor supply of females, open dataset Kid Labor and run the regression using 1980 census of the United States. The sample is composed of married females between age twenty-one to thirty-five in the U.S. who has at least two children.

a. Run a regression with weekm worked as dependent variable and more Child as independent variable. Interpret your result?

b. Using a test we learned to examine whether there exists the problem of heteroskedasticity? Explain your result.

c. It seems that the standard OLS regression cannot examine the direct impact of having additional child on female labor supply because of endogeneity. What can be the source of endogeneity? Why do you think there exists endogeneity?

d. In our dataset, there is a variable called same equal to 1 when the family's first two kids are same sex (either both are girls or both are boys). Using the OLS to examine whether the family with two kids of same sex is more likely to have additional one compared to those with two kids of different sex. Is the result statistically significant? Is it economically significant?

e. Can we use same to solve the problem of endogeneity? Can it be a good instrumental variable for the regression in (a)? Explain why? Does it satisfy the conditions of relevance and exogeneity? Show the process of examining the conditions.

f. Do you think whether same can be a weak instrument?

g. Run the regression in (a), while using the variable same as instrumental variable. Also briefly explain the process how you used the method. In this new regression, what is the impact of having kids on female labor?

h. Run the regression in (g) again after including other variables like hispanic, age, other_race and black. Does the new result significantly different from the previous one? Briefly discuss.

i. Going back to the regression of (a) without using instrumental variables, what else can you think of as another instrumental variable other than same?

j. If we include black, hispanic, other race and constant, is it problematic? Why do you think we need to include these dummy variables in our regression in our regression model? Can you interpret the estimates of them after running regression?

This Computer Science has been solved by our PHD Experts at My Uni Paper.