1: \begin{abstract}

2: This paper studies the threshold estimation of a TAR model

3: when the underlying threshold parameter is a random variable. It is

4: shown that the Bayesian estimator is consistent and its limit distribution is

5: expressed in terms of a limit likelihood ratio. Furthermore, convergence of moments of

6: the estimators is also established.   The limit

7: distribution can be computed via explicit simulations from which testing

8: and inference for the threshold parameter can be conducted.  The obtained

9: results are illustrated with numerical simulations. \\

10: 

11: \noindent

12: {\bf Key words and phrases:}  Bayesian estimator, continuous-time diffusion,

13: compound Poisson process, limit distribution, limit likelihood ratio and nonlinear threshold models. \\

14: 

15: \noindent

16: {\it AMS 1991 subject classifications: Primary 62G30; secondary 62M10.}

17: \end{abstract}