MAINTENANCE EN COURS / SITE UNDER MAINTENANCE

Une opération de maintenance est en cours: les résultats de recherches et les exportations peuvent être incohérent.
Site under maintenance: search & exportation results could be inconsistent.
 

Files

LFIN_DP_2024-04.pdf
  • Open Access
  • Adobe PDF
  • 1.15 MB
Download

Details

Authors
Abstract
Estimation risk in portfolio selection can be mitigated with sparse approaches such as lasso that penalizes for the norm of the portfolio weights and excludes assets from the investment universe. The latter are revealed a posteriori, by identifying which assets receive an optimal weight of zero. We show instead that in the presence of parameter uncertainty, it is desirable to remove assets before computing the portfolio weights. In particular, we show that the optimal portfolio size strikes a tradeoff between accessing additional investment opportunities and limiting estimation risk. Our approach disentangles the determination of the optimal portfolio size from the asset selection rule, making it more easily implementable and robust to estimation risk than alternative sparse methods. Empirically, our restricted portfolios substantially outperform their counterparts applied to all available assets. Our methodology renders portfolio theory valuable even when the full dataset dimension is comparable to the sample size.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Vanderveken, R., Lassance, N., & Vrins, F. (2024). Optimal Portfolio Size under Parameter Uncertainty (LIDAM Discussion Paper LFIN 2024/04). https://hdl.handle.net/2078.5/217271