Institutional Repository of Vilnius University
Kopulomis paremta kointegruotų kriptovaliutų porų statistinio arbitražo prekybos strategija /
Abstract
dc:descriptionThis thesis investigates a copula-based statistical arbitrage trading strategy for cointegrated cryptocurrency pairs. The main objective of the study is to assess whether an additional pair-selection filter based on the Ornstein–Uhlenbeck half-life criterion improves the performance of a copula-based pairs trading strategy in the cryptocurrency perpetual futures market. The empirical analysis uses hourly OHLCV data from Binance USD-M perpetual futures for the period from 2024-01-01 to 2025-12-31. In each rolling cycle, the most liquid USDT-quoted instruments are selected, spreads against BTCUSDT are constructed, cointegration is tested using the Engle–Granger procedure, and pairs of spreads are further filtered using Kendall’s tau and, in the modified strategy, the half-life of the spread-of-spreads process. Trading signals are generated from conditional copula distribution functions, while copula families are selected using the Akaike information criterion. The results show that the half-life filter can improve the selected strategy configurations. The best-performing specification uses the position-closing threshold α2 = 0.35 and a 56-hour half-life threshold. This strategy achieved a total return after transaction costs of 189.29%, an annual return after transaction costs of 73.05%, a Sharpe ratio of 1.97, and a Sortino ratio of 1.27. Compared with the corresponding baseline strategy, the half-life filter increased the return after costs and improved risk-adjusted performance. The strategy also had a low negative correlation with BTC returns, indicating limited linear dependence on the main cryptocurrency market factor. However, the results should be interpreted cautiously, as the study does not include slippage or funding rates, and some performance metrics are affected by forced position closures at the end of trading periods. Therefore, the half-life criterion is best interpreted as a potentially useful optimization parameter rather than a universally valid pair-selection rule.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Simanavičius, Adomas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308327772&prefLang=en_US