Structural breaks and trend awareness-based interaction in crypto markets

This study aims to test multiple structural breaks using Bai-Perron methodology for the crypto markets. We analyze return and volatility (proxied by absolute and squared returns) series of Bitcoin and the following crypto assets: Bitcoin Cash, DASH, Ethereum, IOTA, Litecoin, NEO, XRP. In the analysis, we consider the BTC markets of those 7 altcoins as well as the USD markets. Empirical findings indicate existence of statistically significant structural changes in terms of returns and volatility. Return and volatility series share different dynamics. On the other hand, time series quoted in BTC perform different structural change behavior than the ones quoted in USD. We also observed a clustering of breakpoints in the periods of February-March 2017 and December 2017-March 2018. For the logarithmic returns, in the mean of BTCUSD and series quoted in BTC no structural breaks are available. The mean of logarithmic returns of DASH, ETH and LTC quoted in USD have structural changes corresponding to the beginning of the downtrend in USD price series of the relevant assets. Absolute returns revealed more of structural breaks in terms of volatility than squared returns. Except for IOTA, volatility last longer in the series quoted in USD than the ones quoted in BTC. For the volatility series of ETH, LTC and NEO, regimes in the time series quoted in BTC is affected by the trend change in the USD quoted price series of the asset. We have defined this regime change and trend change relationship as "trend awareness-based interaction". We also found a specific trend change pattern in the time series. In comparison to the time series of altcoins, BTCUSD is the first one which has the trend change almost in all cases. Besides, the time series of BCH, ETH, LTC and NEO (and DASH partly) follows a specific trend change order as the following: first in USD quoted time series of the altcoin and then in BTC quoted ones, while the time series of IOTA and XRP indicate the opposite. (C) 2020 Elsevier B.V. All rights reserved.