Correlation-diversified passive portfolio strategy based on asset swapping

Asset Swap Correlation Diversified Passive Portfolio Strategy – Journal of Investment Strategies

  • We are developing a new index investment strategy that can achieve a higher return with lower volatility than the original index, while its behavior is still similar to that of the original index.
  • The idea is to swap assets in the portfolio based on correlation distances between assets, where assets that are strongly correlated to many other assets should be placed in the middle part of the swap.
  • By reducing the weight of assets placed near the central part, which behave like many other assets, we can hope to diversify the portfolio and make it less risky.
  • We solved this permutation problem by taking a quantum-inspired approach.

In this article, we develop a passive strategy for improving index investing, which we call the Correlation Diversified Portfolio Strategy. The proposed method adjusts the weight vector of the original index according to the permutation of the assets belonging to the original index. We seek the permutation of these assets in such a way that assets that have a strong correlation with many other assets are placed at the center of the permutation. By reducing the weighting of these core assets, we can build portfolios that are more diversified and have better risk-return characteristics than the original index. We solve this asset swapping problem by taking a quantum-inspired approach. Concretely, we convert this permutation problem into an unconstrained quadratic binary optimization problem and use simulated annealing on a personal computer or annealing machine to find a near optimal solution in a reasonable time. To examine the utility and computational feasibility of the proposed method, we apply it to three major indices from the United States and Japan, and we provide numerical experiments that show that portfolios constructed by the proposed method can achieve a return. higher with lower volatility compared to the original index, while their behaviors are still similar to those of the original indexes.

Source link

Don F. Davis

Leave a Reply

Your email address will not be published.