Dynamic portfolio selection with sector-specific regularization
(2024) Econometrics and Statistics — Vol. 32, p. 17-33 (2024)
Files
Details
- Authors
- Abstract
- A new algorithm is proposed for dynamic portfolio selection that takes a sector-specific structure into account. Regularizations with respect to within- and between-sector variations of portfolio weights, as well as sparsity and transaction cost controls, are considered. The model includes two special cases as benchmarks: a dynamic conditional correlation model with shrinkage estimation of the unconditional covariance matrix, and the equally weighted portfolio. An algorithm is proposed for the estimation of the model parameters and the calibration of the penalty terms based on cross-validation. In an empirical study, it is shown that the within-sector regularization contributes significantly to the reduction of out-of-sample volatility of portfolio returns. The model improves the out-of-sample performance of both the DCC with nonlinear shrinkage and the equally-weighted portfolio.
- Affiliations
Citations
Hafner, C., & Wang, L. (2024). Dynamic portfolio selection with sector-specific regularization. Econometrics and Statistics, 32, 17-33. https://doi.org/10.1016/j.ecosta.2022.01.001 (Original work published 2024)