How "real" is mobility from temporary to permanent employment in Italy? A focus on measurement error

The aim of this paper is to study the effect of measurement error on mobility between different employment types in Italy. For this purpose, we apply a hidden Markov model with two independent indicators for the employment category (permanent contract, temporary contract, self-employed, not employed). The model takes into account that both sources may not be error-free as well as that measurement error may be correlated over time. The two indicators come from ISTAT administrative data and the Labour Force Survey from 2017 to 2021, linked at the individual level. The results show that neither source can be considered error-free and that measurement error severely biases mobility between employment states.