Reactive global minimum variance portfolios with <i>k</i>-BAHC covariance cleaning

We introduce a covariance cleaning method which works well in the very high-dimensional regime, i.e. when there are many more assets than data points per asset. This opens way to unconditional reactive portfolio optimization not enough calibrate dynamical conditional models, happens, for example, new appear market. The is k-fold boosted version of Bootstrapped Average Hierarchical Clustering procedure correlation and matrices. apply this global minimum variance portfolios find that k should increase with calibration window length. compare performance k-BAHC other state-of-the-art methods, including (DCC) non-linear shrinkage. Generally, we our yields better Sharpe ratios after transaction costs competing filtering despite requiring larger turnover. Finally, Global Minimum Variance long–short positions DCC non-stationary investment universe.