First passage times in portfolio optimization: A novel nonparametric approach

• Published in: European journal of operational research Vol. 312; no. 3; pp. 1074 - 1085
• Main Authors: Zsurkis, Gabriel, Nicolau, João, Rodrigues, Paulo M.M.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.02.2024
• Subjects: First-passage probability; Intra-horizon risk; Markov chains; Portfolio optimization; Portfolio optimization; Intra-horizon risk; G11 (Portfolio choice; C41 (Duration analysis); First-passage probability; Markov chains; G17 (Financial forecasting and simulation); C22 (Time-Series models); Investment decisions; C14 (Semiparametric and nonparametric methods: General)
• ISSN: 0377-2217; 1872-6860; 1872-6860
• DOI: 10.1016/j.ejor.2023.07.044

•New framework for portfolio selection, addressing intra-horizon risk.•Nonparametric method for estimating first hitting time probabilities.•Novel approach to Markov chain order selection.•Optimization method minimizes intra-horizon risk, targets return. This paper introduces a portfolio optimization procedure that aims to minimize the intra-horizon (IH) risk subject to a minimum expected time to achieve a target cumulative return. To estimate the first passage probabilities and the expected time a novel nonparametric method and a new Markov chain order determination approach are developed. The optimization framework proposed allows us to include novel path-dependent measures of risk and return in the asset allocation problem. An empirical application to S&P 100 companies, a risk-free asset and stock indices is provided. Our empirical results suggest that the proposed framework exhibits more consistency between in-sample and out-of-sample performance than the mean-variance model and an alternative optimization problem that minimizes the MaxVaR measure of Boudoukh et al. (2004). Overall, the portfolio optimization approach we introduce results in higher out-of-sample annualized returns for relatively low levels of IH risk.