heteroskedasticity、serial correlation、multicollinearity都不会影响coefficient estimate的consistency,因为取决于样本数量,但是omitted variable is correlated with variables already included in the model, 为什么会导致coefficient estimate的inconsistent?
另外multicollinearity为什么会导致error term的标准差升高?
谢谢