Pairs Trading in Cryptocurrency Markets
|Author:||Bc. Miroslav Fil|
|Year:||2019 - summer|
|Leaders:|| prof. PhDr. Ladislav Krištoufek Ph.D.
|Work type:|| Bachelors
|Awards and prizes:|
|Abstract:||Pairs trading is a trading strategy which tries to exploit mean-reversion among
prices of certain securities. It is market-neutral and self-financing, and has been
shown to produce high excess returns in historical backtests.
We employ the most common distance and cointegration approaches on cryptocurrency data from an exchange called Binance spanning the year 2018. The
strategy is mostly unprofitable under transaction costs, but certain combinations
of hyperparameters can perform well. Overall, the distance method performs far
better, being able to achieve 3% monthly profit even in our baseline real-life conditions while the cointegration method always achieves only a slight loss. We also
found that increasing the sampling frequency of the data from daily to hourly
brings mixed results.
Moreover, since we have to reuse estimates of real-life considerations from
equity markets, it is unclear if our results are truly representative of the cryptocurrency market. The strategy is found to be very sensitive to execution difficulties and transaction costs, making their determination crucially important. It
is somewhat easy to get returns in excess of 5% monthly under ideal conditions,
but whether this could be achieved in real trading conditions is still unclear.