Rough volatility via the Lamperti transform

We study the roughness of the log-volatility process by testing the self-similarity of the process obtained by the de-Lampertized realized volatility. The value added of our analysis rests on the application of a distribution-based estimator providing results which are more robust with respect to those deduced by the scaling of the individual moments of the process. Our findings confirm the roughness of the log-volatility process. •The Lamperti transform is used to estimate the regularity of the log-volatility process.•A novel distribution-based method is applied to the transformed process.•The analysis confirm that the volatility process is rough with values H∈[0.06,0.151].