WMAP and Planck temperature data show evidence of several anomalies hinting that the large-scale anisotropies are not fully compatible with the statistical isotropy assumption, one of the fundamentals of the standard model of cosmology. In particular, a “hemispherical power asymmetry” has been found using different estimators with a significance close to 3σ. In this work we perform an analysis of this asymmetry using the local-variance estimator, and the Planck fourth release (PR4) data, which provides an improved data set with reduced systematics and noise, especially in polarization. Additionaly, we have implemented an alternative inpainting approach, which is necessary to reduce the E/B mixing.