A Sufficient and Necessary Condition on Cramér Asymptotic Efficiency in a Semiparametric Regression Model
-
Graphical Abstract
-
Abstract
The paper is concerned with the semiparametric regression model Y = Xrβ+ g(T)+ε, where(Xr,T, Y) are (k + 2)-dimensional random vector, g(·) is an unknown smooth function to be estimated, ε is variable error with zero mean and finite variance It is shown that the LS estimators of β is asymptotically efficiency in sense of Cramér if and only if ε-N(0, σ2).
-
-