Robust Value-at-Risk Currency Portfolio Optimization for Net Open Position (NOP) Hedging

Subject Areas : Data Envelopment Analysis

Mirfeiz Fallah ^{1 *} , Yousef Toomari ² , Mozhgan Safa ³ , Hossein Moghadam ⁴

1 - دانشیار، گروه مدیریت بازرگانی، دانشکده مدیریت، دانشگاه آزاد اسلامی واحد تهران مرکز ، تهران، ایران.
2 - Unit 1. No21. Khayam AlleySouth Rudaki St. Tehran
3 - استادیار،گروه حسابداری، واحد قم، دانشگاه آزاد اسلامی، قم، ایران
4 - استادیار، گروه حسابداری، واحد قم، دانشگاه آزاد اسلامی، قم، ایران

Received: 2023-07-22 Accepted : 2025-06-22 Published : 2025-01-13

Keywords: Net Open Position, Currency Portfolio Optimization, Bank,

Abstract :

This article presents a comprehensive study on "Robust Value-at-Risk Currency Portfolio Optimization for Net Open Position (NOP) Hedging," specifically focusing on the effectiveness of value-at-risk models, including VaR, Copula VaR, and Copula Conditional Value-at-Risk (CVaR), in optimizing currency portfolios. The research utilizes quantitative analysis and time-series observations of daily logarithmic returns of four major and commonly traded currencies, namely USD, EUR, AED, and CNY, from April 6, 2013, to September 21, 2021. For the out-of-sample evaluation, the dataset is limited to a one-month period from May 21, 2021, to June 21, 2021, using Sepah Bank's historical data. All calculations in this study are performed using the open-source software R 4.2.1. The results indicate that the Copula GARCH VaR model outperforms both Copula GARCH - CVaR and GARCH VaR methods in terms of Sharpe ratio. Additionally, the research highlights the significant role of the USD as an independent currency compared to others, making it an essential component in all three optimization methods used in the study. On the other hand, due to relatively strong and positive conditional correlations among the EUR, AED, and CNY currencies, minimizing the allocation of any one of these currencies is favored to reduce risk as much as possible. Consequently, in the GARCH VaR and Copula-GARCH VaR methods, higher realized returns are achieved compared to the Copula GARCH - CVaR model.