Abstract
Gold-silver ratio (GSR) is a commonly followed ratio among investors. Investors use the ratio as an indicator to determine the right time to buy or sell silver or gold. This study explores whether the investors could benefit from the forecast capabilities of machine learning (ML) algorithms in trading the gold-silver ratio at multiple frequencies. The study proposes the use of a range of, from simple to highly complex, ML algorithms. The study shows that investors can significantly benefit from the use of ML algorithms in trading gold and silver. The ML strategies display a better return, risk and risk-adjusted return performance than the benchmark strategies at lower rebalancing frequencies (i.e., bi-weekly, three-weekly and monthly) than the benchmark buy-and-hold and the long-only and long-short pair trading strategies based on technical analysis. The findings are robust to estimation sample size, rebalancing frequency and leverage level.
Original language | English |
---|---|
Title of host publication | World Finance Conference 2022 |
Publication status | Published - 2 Aug 2022 |
Event | World Finance Conference 2022 - University of Turin - School Of Management and Economics, Turin, Italy Duration: 1 Aug 2022 → 3 Aug 2022 https://www.world-finance-conference.com/conference.php?id=23 |
Conference
Conference | World Finance Conference 2022 |
---|---|
Country/Territory | Italy |
City | Turin |
Period | 1/08/22 → 3/08/22 |
Internet address |
Keywords
- Machine Learning
- Commodity trade
- Artificial intelligence
- technical analysis
- gold silver