Question

Q1) answer the following

a) Omitting an important variable that is uncorrelated with the included regressors will lead to biased coefficients on the included regressors and incorrect coefficient signs.True or False

b)

Experience (EXP) is positively associated with earnings (EARN), and years of schooling (S) is positively associated with earnings. Also, EXP and S are negatively correlated in younger workers. If a researcher estimates the regression :
log(EARNINGS) = β₁ + β₂EXP + u, on younger workers, the estimated coefficient on b₂ will be

A. unbiased.

B. biased downwards and underrepresent the effect of EXP on earnings.

C. overestimate the effect of EXP on earnings.

D. None of these answers are correct.

c).

A researcher wishes to estimate a Cobb-Douglas production function for a microprocessor company of the form
                                                            Y = β₁ X₂^β2 X₃^β3 X₄^β4u
where Y = #microprocessors; X2 = units of capital; X₃ = units of labour; and X₄ = # computer science staff involved in research and development. The researcher has n=30 observations and estimates the following unrestricted regression with OLS:

                              LnY = 1.2 + 0.6ln X₂ + 0.2ln X₃ + 0.3ln X₄                                   RSS=3157.2

In order to test the validity of the SINGLE linear restriction of constant returns to scale,
Ho: β2+β3+β4 = 1, the researcher estimates a restricted regression and find the following results:

               Ln(Y/X₄) = 0.8     + 2.2ln(X₂/X₄)   + 1.3ln(X₃/X₄)    RSS=3683.4

(1.2)       (1.0)                      (0.6)                   

The restriction is valid at a level of significance of 1%..True or False

Answer 1

a) Omitting an important variable that is uncorrelated with the included regressors will lead to biased coefficients on the included regressors and incorrect coefficient signs.

False.

(b) Experience (EXP) is positively associated with earnings (EARN), and years of schooling (S) is positively associated with earnings. Also, EXP and S are negatively correlated in younger workers. If a researcher estimates the regression :
log(EARNINGS) = β1 + β2EXP + u, on younger workers, the estimated coefficient on b2 will be

overestimate the effect of EXP on earnings.

(c) The restriction is valid at a level of significance of 1%

True.