Portfolio Allocation with Co-Skewness and Co-Kurtosis

Summary

This research note describes extending mean-variance portfolio allocation with higher-moment information, especially co-skewness and co-kurtosis among assets. Mean and covariance alone can miss differences in joint tail behavior: portfolios with similar average returns and volatility may behave differently when several holdings experience unusually strong gains or losses together. The note associates more negative third moments with greater risk of extreme negative returns, and larger fourth moments with a greater likelihood of extreme events.

The proposed approach adds these higher moments to traditional allocation inputs, estimating parameters on rolling in-sample windows and evaluating the resulting allocations on out-of-sample data. The summary reports that the higher-moment approach performed reasonably in those tests, but it gives no detailed dataset, implementation, benchmark, or numerical results. Its findings are historical and conditional on the model and estimation choices; the source explicitly cautions that market uncertainty, extreme conditions, and model failure can undermine future performance.

Key ideas

• Mean-variance allocation uses expected returns and covariance but does not fully describe joint tail behavior.
• Co-skewness and co-kurtosis can distinguish portfolios with similar means and standard deviations but different extreme outcomes.
• The approach supplements traditional allocation inputs with third- and fourth-moment information.
• The reported evaluation estimates parameters on rolling in-sample windows and tests allocations out of sample.
• The summary offers no detailed performance figures and warns that historical results may not persist.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.