<heading id="h1" level="1" implicit="yes" format="display">Abstract</heading>We revisit the traditional return-based style analysis in the presence of time-varying exposures and errors-in-variables (EIV). We apply a benchmark selection algorithm using the Kalman filter and compute the estimated EIV of the selected benchmarks. We adjust them by subtracting their...