Markov Chain Sector Rotation with Black–Litterman Portfolio Weights

Summary

The strategy estimates sector return states with a Markov chain, which models each state as dependent on the preceding state. It uses the resulting conditional mean and variance estimates as inputs to a Black–Litterman portfolio process, treating the estimated mean returns as subjective views. At each quarterly rebalance, it selects the three sectors with the highest posterior weights. The report describes using the prior three months of sector index returns and a universe of 28 Chinese industry sectors.

The reported historical test spans 2010 through July 2021. It states that the combined method gained 322.69%, compared with 213.75% for an equal-weight portfolio of the three sectors ranked directly by Markov-chain estimates and 47.75% for the broad-market benchmark. It also reports that posterior-weight allocation outperformed equal-weight and market-cap-weight comparisons. The document provides no detailed implementation, cost assumptions, or risk-adjusted statistics, and cautions that historical patterns may change and invalidate the findings.

Key ideas

• A Markov chain estimates sector return states and their conditional means and variances.
• The conditional mean estimates are used as subjective views in a Black–Litterman allocation.
• The strategy selects three sectors with the highest posterior weights and rebalances quarterly.
• The reported historical comparison favors the combined method over direct ranking and benchmark approaches.
• The report warns that changing historical relationships may cause the model’s conclusions to fail.

Tags

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