Calculating model (Analytically)

#1
Hi, can anyone help me out with this question?

Assume that the true data generating process (the true model) is: y = 20 −30(x + r) + 4w + ê. Suppose that we also know that E(ê|x) = E(ê|w) = E(ê|r) = 0, Cov(x, w) = 20, Cov(x, r) = Cov(w, r) = 0, Var(x) = 10, Var(w) = 8 and Var(r) = 6. What would be the estimate of the coefficient for the variable x if we regress y on a constant and the variable x via OLS and we have a very large sample? Carefully motivate your answer.

Thank you very much! :)
 

fed2

Active Member
#3
good answer, is it a problem that
Cov(x, w) = 20 > sqrt( 10 * 8 ) ~= 9
?
though, or am i crazy (-er than usual)