Abstract: Information bounds are developed for estimation of regression parameters in the presence of left truncation and right censoring on the observed responses, assuming that the vectors of covariates and censoring/truncation variables are independent (but possibly non-identically distributed). Under certain regularity conditions, asymptotically efficient estimators that attain these information bounds are also given.
Key words and phrases: Censoring and truncation, regression, Fisher information, regular estimators, asymptotic minimax bounds, semiparametric models, adaptive rank estimators, martingales and stochastic integrals, asymptotic normality.