Question

Which of the following statements is correct regarding the implication of the existence of heteroscedasticity?

A) The OLS estimator is biased.

B) The OLS estimator is unbiased and still best. There is no other estimator with smaller variance.

C) The OLS estimator is biased but the standard errors usually computed for the OLS estimator are correct.

D) The standard errors usually computed for the OLS estimator are incorrect.

Answer 1

Option D.

Regarding the problem of heteroscedasticity in the data set, it can be stated that estimated standard error is wrong. Since the value of standard error is wrong, thus, we cannot rely on condidence interval and hypothesis test which shows whether independent variables considered in model specifications are significant or insignificant. Thus, the standard errors usually computed for the OLS estimator are incorrect.

Option A is incorrect because OLS estimators in the presence of heteroscedasticity remains unbiased. Option B is incorrect because OLS eatimators are not the best in this case because they have large variance problem.Option C is incorrect because standard errors computed for OLS estimators is incorrect.

Thus, Option D is correct.