Efficient and robust estimation for financial returns: an approach based on q-entropy

We consider a new robust parametric estimation procedure, which minimizes an empirical version of the Havrda-Charv_at-Tsallis entropy. The resulting estimator adapts according to the discrepancy between the data and the assumed model by tuning a single constant q, which controls the trade-o_ between robustness and e_ciency. The method is applied to expected re- turn and volatility estimation of _nancial asset returns under multivariate normality. Theoretical properties, ease of implementability and empirical re- sults on simulated and _nancial data make it a valid alternative to classic robust estimators and semi-parametric minimum divergence methods based on kernel smoothing