Question

1. Specify a regression in which you try to measure the “return” to school in terms of labor income to an additional year of schooling. Label income as y, schooling as s, the return as β and the unobserved component as ε. (2 points)

B. Take as given the estimation part and suppose that you have an estimate of β in hand. Call it βˆ. Without further assumptions, what is the interpretation of βˆ. (2 points)

C. What assumption do you need to interpret β causally? What does the assumption mean in words? (2 points)

D. What experiment would you run if you could control people’s choices to generate a sample in which you are sure that this assumption holds? (2 points)

Answer 1

Question 1  

y = a+ βs + ε.

is the regression model

(B) here interpreation of β is that if a random person will increae its schooling year by 1 year it will increase his average labour income by β.

(C) Here the assumptionsare

(i) There must be no outliers

(ii) all variables must be multivariate normal

(iii) No multicollinearlity

(iv) No autocorrelation

(v) Homoscedescitcity

D.

Here we would take a random sample of yound people aged 20 to 35 and note down their years of schooling and their annual income ad than we just find the regression equation to evaluate the return