Downloads: 13 | Views: 192 | Weekly Hits: ⮙6 | Monthly Hits: ⮙6
Study Papers | Finance | United States of America | Volume 12 Issue 11, November 2023 | Popularity: 7.1 / 10
Cointegration-Based Pairs Trading Strategy in Commodity Markets: A Novel Approach to Statistical Arbitrage
Gautham Mohandas
Abstract: Purpose: Pairs trading has been a popular strategy used by most hedge funds in exploiting the anomalies in the market. It is often used for trading overvalued and undervalued stocks, but can this method be extended to identify cointegrated pairs in commodities If so, how can we identify these pairs? This paper predominantly focuses on ways to find the cointegrated pairs in commodities and how to make an effective trade incorporating this strategy. Design/methodology/approach: The paper uses the commodities from the Goldman Sachs Commodity Index from the period 2007 to 2018 as a training dataset and from 2019 to 2022 as a test dataset. A rolling regression window of 1 day and 30 days has been implemented to overcome the issue of look-ahead bias for the trading pairs. Following that, a trading strategy has been created, longing the spread when the z-score is below -2, shorting the spread when the z-score is above 2, exiting the position when the z-score nears zero, and a stop-loss is created when the spread crosses positive or negative 4. Findings: 3 cointegrated pairs have been found from the commodity index which has a cointegrating vector and is statistically significant. A successful trading strategy has been established for those pairs. Originality/value: The paper primarily focuses on the commodities index. First, it is a unique statistical approach to finding the cointegrated pairs within the index. Second, a new trading approach has been proposed based on the z-score as when to long, short, and fix stop-loss of the pairs. Third, the paper also checks the cointegrated pairs by having a rolling regression window which helps differentiate whether the pairs have lost cointegration and eliminates look-ahead bias.
Keywords: Hedge Funds, Spreads, Regression, Econometric approach
Edition: Volume 12 Issue 11, November 2023
Pages: 1731 - 1735
DOI: https://www.doi.org/10.21275/SR231123060916
Please Disable the Pop-Up Blocker of Web Browser
Verification Code will appear in 2 Seconds ... Wait