Search for question
Question

4. Feasible Generalized Least Squares. Consider the linear regression model

price = Bo + B₁lotsize + B₂sqrft + 3bdrms + u,

where price is the house price measured in thousands of pounds, lotsize is the size of lot

in square feet, sqrft is size of house in square feet, bdrms is the number of bedrooms, and

E (u | lotsize, sqrft, bdrms) = 0. An econometrician believes that

u² = vexp {0 +1lotsize},

where E (v|lotsize, sqrft, bdrms) = 1. Which assumption(s) of the Gauss-Markov theo-

rem are violated? Describe main steps of an asymptotically efficient procedure for estimating

Bo, B1, B2, B3.

Fig: 1