Title: Dynamic futures margin setting method under state dependence

Authors: Wang Hong; Wen Kun; Shouqian Kang

Addresses: Hangzhou Vocational and Technical College, Hangzhou 310000, China ' Hainan Vocational University of Science and Technology, Haikou 571126, China; Jiangxi Administrative Institute, Nanchang 330018, China ' Central China Normal University, Wuhan, 430079, China

Abstract: Margin is not only a basic risk control system for futures trading, but also an important part of the cost of futures trading, and its fundamental position is very important. This paper presents a dynamic margin setting method for futures based on market state, which considers extreme risk control and opportunity cost. In different market conditions, we choose different margin levels to better balance spillover probability and opportunity cost. Using machine learning, we sample the sugar futures traded on the Dalian Commodity Exchange between January 6, 2006, and May 29, 2020. The market is divided into three categories by the hidden Markov model: highly volatile, volatile, and stable. We compare margin level under VaR, CVaR, MMVaR, EWMA and improved EWMA risk standards. Comparative analysis and retrospective test show that the current fixed margin ratio is unreasonable, and the margin level under the single risk criterion cannot balance risk control and opportunity cost well. We recommend that market regulators dynamically adjust margin setting levels according to different market states, thereby luring more investors to invest and boosting the liquidity of the futures market.

Keywords: dynamic margin level; risk criteria; machine learning; market status.

DOI: 10.1504/IJISE.2024.138892

International Journal of Industrial and Systems Engineering, 2024 Vol.47 No.2, pp.176 - 193

Received: 02 May 2022
Accepted: 20 Aug 2022
Published online: 03 Jun 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article