TY - JOUR
T1 - Portfolio models for optimizing drawdown duration
AU - Vedernikov, Andrei
AU - Liesiö, Juuso
AU - Seppälä, Tomi
N1 - Publisher Copyright:
© 2024 World Scientific. All rights reserved.
PY - 2024/3/1
Y1 - 2024/3/1
N2 - The drawdown duration, which measures the time elapsed since the portfolio obtained its maximum value, is an important criterion in active portfolio management for institutional investors. Although several optimization models exist for controlling portfolio drawdown magnitude (i.e. the percentage drop in portfolio value from its latest peak value), developing similar models for the drawdown duration has received minimal attention in the literature. Therefore, this paper develops a family of models for optimizing average, maximum and tail drawdown duration formulated as mixed-integer linear programming (MILP) problems, allowing the utilization of powerful solvers to identify optimal asset portfolios. We apply the developed models to real data on historical returns to compare their performance against traditional and drawdown-based portfolio selection models. The results indicate that the developed models lead to decrease in drawdown duration levels both in in-sample and out-of-sample tests. The constructed efficient frontiers also show a clear trade-off between minimizing drawdown duration and maximizing expected returns.
AB - The drawdown duration, which measures the time elapsed since the portfolio obtained its maximum value, is an important criterion in active portfolio management for institutional investors. Although several optimization models exist for controlling portfolio drawdown magnitude (i.e. the percentage drop in portfolio value from its latest peak value), developing similar models for the drawdown duration has received minimal attention in the literature. Therefore, this paper develops a family of models for optimizing average, maximum and tail drawdown duration formulated as mixed-integer linear programming (MILP) problems, allowing the utilization of powerful solvers to identify optimal asset portfolios. We apply the developed models to real data on historical returns to compare their performance against traditional and drawdown-based portfolio selection models. The results indicate that the developed models lead to decrease in drawdown duration levels both in in-sample and out-of-sample tests. The constructed efficient frontiers also show a clear trade-off between minimizing drawdown duration and maximizing expected returns.
KW - drawdown
KW - Finance
KW - mixed-integer linear programming
KW - portfolio optimization
UR - http://www.scopus.com/inward/record.url?scp=85202801683&partnerID=8YFLogxK
U2 - 10.1142/S0219024924500146
DO - 10.1142/S0219024924500146
M3 - Article
AN - SCOPUS:85202801683
SN - 0219-0249
VL - 27
JO - International Journal of Theoretical and Applied Finance
JF - International Journal of Theoretical and Applied Finance
IS - 2
M1 - 2450014
ER -