A Generalized Model of Misclassification Errors and Labor Force Dynamics

Feng, Shuaizhang, Yingyao Hu, & Jiandong Sun

Abstract：

We study the U.S. labor market transitions using a latent variable approach, explicitly modeling the persistent misclassification process and the non-Markovian nature of the underlying true labor force dynamics. A closed-form global identification for misclassification probabilities and labor transition probabilities is established through an eigenvalue-eigenvector decomposition. Contrary to existing studies, our empirical results suggest that the observed data have understated the true mobility in labor force statuses after we account for persistence in both the misclassification errors and the latent true labor force dynamics.