The Identification of a Mixture of First‐Order Binary Markov Chains*
证明了面板波数大于等于2倍成分数时，可全局识别所有参数异质的动态二元选择模型，该模型等价于S个一阶二元马尔可夫链的混合。
Abstract Let S be the number of components in a finite discrete mixing distribution. We prove that the number of waves of panel being greater than or equal to 2 S is a sufficient condition for global identification of a dynamic binary choice model in which all the parameters are heterogeneous. This model results in a mixture of S binary first‐order Markov Chains.