97-F-31. Kunitomo, Naoto, "On Estimation of the Simultaneous Switching Autoregressive Models", October 1997.

The simultaneous switching autoregressive (SSAR) model is a non-linear Markovian time series model, which was originally proposed by Kunitomo and Sato (1996) in Structural Change and Economic Dynamics and some of its statistical properties have been investigated by Sato and Kunitomo (1996) in Journal of Time Series Analysis. Since these papers have omitted some derivations of their theoretical results, this note gives more detailed expositions on them which additional technical remarks. We discuss some sufficient conditions for the geometric ergodicity of the SSAR model and the existence of moments. Also we give some sufficient conditions for the consistency and asymptotic normality of the maximum likelihood estimator for the unknown parameters in the SSAR models. Some corrections in the previous papers are given.