1. Your parents would like to find out more about the effects of education on wage rates (to check you’re not wasting your time at Uni!). The data set “Q1_wages.xlsx” contains information on wages ($/hour), education and years of experience for 1,000 workers. Download this data from Moodle.
(a) i) Your parents ask you to use Excel to estimate the following model:
where Wage= earnings per hour, Educ = years of education and Exper = years of experience and Exper2 is the squared years of experience. Interpret the estimated regression coefficient of Educ. Is the sign of the coefficient what you would expect? Make sure to copy and paste your Excel regression output here.
x
(b) Now consider the estimated coefficients of Exper and Exper2 and explain whether Exper and Exper2 should be included as explanatory variables, justifying your answer using some of the values reported in the regression output.
(c) Can you think of any other variables that your parents have not controlled for (i.e. not included in the regression model) that may be related to both educational attainment and earnings? What does this imply about how they should interpret their estimate of the relation between wages and education.
(d) i) Now, estimate the following model:
where Migrant = 1 for workers born overseas; 0 otherwise and Female= 1 for female workers;
0 for male workers. Copy and paste your Excel output below
ii) Interpret the estimated coefficients on the dummy variables for Female and Migrant, and for each one test the null hypothesis that the coefficient on the dummy variable equals zero
iii) Explain what would be the economic rationale to include the interaction term between Migrant and Educ
iv) Compared with males, by how much are wages higher or lower for females
v) Carry out a joint test that the two dummy variables have coefficients that all equal zero at the a=0.05 level. (Hint: this is a joint test of three coefficients) What do you conclude? Make sure to show working and clearly state your hypotheses
(e) Are the explanatory variables, in the model estimated under Question (d) (i) as a whole statistically significant? (i.e. R2=0)? Explain