Questions 13

  1. How nonspherical error terms are influencing the properties of OLS?
  2. Show how do we transform, for known variance matrix of error term {$\Omega$}, the model with nonspherical error terms into the model which satisfies the assumptions of CLRM.
  3. (*) Explain how the weight should be constructed in the case of Weighted Least Squares estimator. Justify your answer.
  4. Explan the differnce between Generalized Least Squares estimator and Feasible Least Squares estimator.
  5. (*) Explain the estimation procedure used in the case of Feasible Least Squares estimator using as an example the problem of taking into account the heteroskedasticity present in the model or estimating model with Prais-Winsten estimator.
  6. Explain why it does often make more sense to use OLS estimator with robust variance matrix rather than to use GLS estimator.