Hidden Markov regression models (HMRMs) serve as a potent tool for analyzing diverse time-dependent data structures within a regression framework. The literature has witnessed significant progress in recent years, particularly in the context of matrix-variate longitudinal data. In this study, we apply a set of parsimonious matrix-variate HMRMs to evaluate the relationship between unemployment and crime rates across Italian provinces. Leveraging the flexibility inherent in HMRMs and the matrix-variate structure, we gain valuable insights. Various states are identified, and the transitions of provinces between these states are examined over time.