Co-movements in market quality of cryptocurrencies

Using intraday data of the transactions and order book of the five cryptocurrencies with the highest market capitalizations in January 2018 from the Bitfinex exchange between August 2017 and July 2018, we extend the market quality and cryptocurrency literatures by investigating the market quality of cryptocurrencies. We replicate the approaches of Marsh and Mazza (2017) and Beaupain, Giot, and Petitjean (2010) in order to assess market quality commonality, liquidity commonality and the impact of volatility regimes on liquidity commonality. We found significant evidence of liquidity commonality (varying in intensity depending on the types of liquidity proxies chosen) in both quiet and volatile regimes for our set of cryptocurrencies with a positive relationship between commonality and volatility. Our results also suggest the presence of commonality in market quality movements of our set of cryptocurrencies. Counter-intuitively, the intensity of the common movements seems to be more important at the end of our period compared to the first three months of the period. Comparisons of the average explanatory power of Chordia, Roll, and Subrahmanyam (2000)’s market model at given periods leads to mixed results. Most conclusive is that for 16 out of 30 of our liquidity proxies across our set of cryptocurrencies, the market model regression assessed on a weekly basis had a significantly higher average explanatory power during the month preceding the end of the Bitcoin bubble as dated by Liu et al. (2019) compared to the month following it. This partially corroborates Liu et al. (2019), who also had stronger explanatory power in their model during the bubble.