1. Running OLS on misspecified linear regression model one may obtain inflated coefficients (characterized with overestimated standard errors) that are possibly indibidually iinsignificant despite a possible global significance.
2. In the context of SUR (seeming unrelated regressions) models, OLS are equvalent to GLS or feasible GLS when the regressors are the same across the individual models, whether or not the eooros are contemporaneously covariating. However, this statement is not ture when the errors driving the different regressions are intra-temporally correlated.
两题都是错的。有人可以告诉我理由么?多谢!