Heteroscedasticity
Summary
Setup:
- The LRM with random sampling
\[y_i=β_1+β_2x_{i,2}+β_3x_{i,3}+…+β_kx_{i,k}+ε_i, i=1,…,n\]
Heteroscedasticity
- If the error terms are heteroscedastic then the conditional variance of the error terms is not constant and the GM assumptions are violated. We write
\[Var\left( ε_i|x_i \right)=σ_i^2\]
Consequences:
- If the explanatory variables are exogenous then the OLS estimator is still unbiased and consistent (under mild conditions).
- The OLS estimator is no longer efficient .
- The OLS GM variance formula is incorrect . Therefore, the OLS standard errors are inconsistent and all inference will be misleading even when \(n\) is large.