Limits of Informational Efficiency in Cryptocurrency Markets

This work analyzes the efficiency of cryptocurrency markets. The primary goal of the work was to reveal whether substantial profits can be made with the use of triangular arbitrage strategy on different cryptocurrency exchanges. This question is relatively new in the context of examining cryptocurrency markets, and additional research is still needed in this field, as the theme of digital currencies is not yet well-established. The primary hypothesis stated that there were significant violations of no-arbitrage conditions on cryptocurrency markets. For the sake of finding arbitrage opportunities, the 3 top liquid exchanges were chosen: Binance, HitBTC, Bitfinex, as well as 3 most liquid currency pairs: BTC/USD, BTC/ETH, ETH/USD. The results show that after accounting for transaction fees, the majority of arbitrage opportunities, seen on the three markets, vanish, which means that efficiency of the markets can not be rejected. Additionally, the research on predictability of triangular arbitrage returns has been made. Contrary to the first result, there was found an evidence of calendar effects and random walk test main hypothesis was rejected. Therefore, we state that even though substantial profits can not be made with the use of triangular arbitrage strategy, there exist patterns in triangular arbitrage returns that can be predicted and need additional research.