1: \begin{abstract}

2: The problem to establish not only the asymptotic distribution results for statistical estimators 

3: but also the moment convergence of the estimators has been recognized as an important issue 

4: in advanced theories of statistics. 

5: One of the main goals of this paper is to present a metod to derive 

6: the moment convergence of $Z$-estimators as it has been done for $M$-estimators. 

7: Another goal of this paper is to develop a general, unified approach, based on 

8: some partial estimation functions which we call ``$Z$-process'', to the change point problems 

9: for ergodic models as well as some models where the Fisher information matrix is random 

10: and inhomogeneous in time. 

11: Applications to some diffusion process models and Cox's regression model are also discussed. 

12: \end{abstract}