This paper revisits the traditional return-based style analysis (RBSA) in presence of time-varying exposures and errors-in-variables (EIV). We first apply a selection algorithm using the Kalman filter to identify the more appropriate benchmarks for the analysed fund return. Then, we compute...