Co-Integration Pair Trading on Swedish Equities: Are returns robust to transaction costs?
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesis
Abstract [en]
Statistical arbitrage pair trading, despite its title, carries inherent risks. This strategy hinges on identifying pairs of securities with a historically established long-term equi- librium relationship, so-called co-integrated. When pairs are co-integrated, deviations from their equilibrium are anticipated to be temporary and mean-reverting, thereby of- fering profit opportunities. Larger deviations imply greater profit potential. By taking a pair trading position, investors can adopt a market-neutral stance, executing the strategy without the need for a directional market outlook.
Since its inception in the 1980s, pair trading initially yielded substantial returns, leading to heightened competition and a reduction in both the number and size of arbitrage oppor- tunities. Consequently, the strategy has witnessed diminishing returns as it has become more widely known and practiced. To address this trend, innovative approaches have emerged, such as integrating the Kalman filter to enhance the traditional co-integration method.
This thesis conducts a replication study to assess the effectiveness of a statistical pair trading strategy. The approach employs the Johansen test to identify co-integrated pairs and integrates a Kalman filter to adapt the co-integration coefficient over time. The cen- tral objective is to ascertain whether the strategy yields returns surpassing transaction costs.
Important insights from this study reveal that cumulative returns do not correlate with the quantity of co-integrated pairs, underscoring the significance of meticulous equity se- lection for the strategy’s success. Moreover, the findings advocate for employing a higher threshold for entering spread trades, diverging from previous research. Lastly, the study validates that shorter-duration trades are prone to deliver more substantial and positive returns.
Future research could investigate novel approaches for in-sample ranking of co-integrated equity pairs, leveraging statistical properties to develop more sophisticated portfolios and improve profitability.
Place, publisher, year, edition, pages
2024. , p. 48
Keywords [en]
Co-Integration, Pair trading, Kalman filter, Statistical arbitrage
National Category
Economics
Identifiers
URN: urn:nbn:se:liu:diva-204898ISRN: LIU-IEI-TEK-A--24/04947--SEOAI: oai:DiVA.org:liu-204898DiVA, id: diva2:1871357
Subject / course
Production Economics
Presentation
2024-06-14, R43, Hans Meijers väg 12, Linköping, 10:15 (Swedish)
Supervisors
Examiners
2024-08-082024-06-172024-08-08Bibliographically approved