26.05.2024 Multiple Linear Regression II#

Consistency#

estimator is consistent when:

  • approaches true population parameter

  • variance \(\downarrow\)

Asymptotic Normality#

What happens, when samples are not normally distributed? (e.g Lotto)

Solution: large samples (they have asymptotic normality)

Testing: stays the same

Asymptotic Efficiency#

Estimator is efficient if

  • consistency

  • asymptotic normally distributed

  • asymptatic vairance <= smaller than other estimators

OLS = asymptotically efficient!