The data is posted on Blackboard. Download the data lfs4.dta on your personal computer. This data is from the Labour Force Survey 2003. In STATA, add enough memory to open the data set (Set memory 300) and open the data file. Refer to the Record Layout for reference with the variable names. Note that all your results should be weighted to account for the stratified sampling frame of the LFS. Include a copy of the do file with your output.
a. What is the raw average gender log hourly wage gap? How could you get this number in a regression framework? Use descriptive commands such as summarize
b. What was the union density (% workers covered by union)? Is the dispersion of logwages much larger in the nonunion sector than in the union sector? Use the kernel density estimator to plot the two densities.
kdensity lwage [weight=fweight] if unioncv==1, gauss width(0.05)
Is there a min.wage effect in the non-union sector (The min. wage ranges between $7.00 and $8.00?)
c. Is the dispersion of logwages much larger in the public sector than in the private sector? (cowmain = class of worker)
d. Is the dispersion of logwages much larger among females than men?
In the LFS educational attainment is available in 6 categories and age in 12 categories. Generate categorical variable dummies for age and education (for instance, use tab educ_90, gen(edd)). Run a regression of wages on education, age and gender and save the estimated coefficients for education in a single vector (Tip: after reg, gen betaed=0; replace betaed=_b[edd2] id edd2==1; ...replace betaed=_b[edd7] id edd7==1 ) Save your data.
a. Is the male advantage reduced when one controls for these human capital variables? Explain. Are there other controls in your data set that can help reduce the gender differential?
b. Construct a years of schooling variable. Specify how you assign numerical values to each category. Construct an experience variable (gen exp= age-years of education-6) and its square. Substitute the level of schooling dummies with the years of education c. Has the age-earnings profile of high school educated workers the same shape as for university educated workers? Plot these in the same graph, what is the cross over age?