Return-volatility relationships in cryptocurrency markets: Evidence from asymmetric quantiles and non-linear ARDL approach
利用比特币和以太坊的模型无关隐含波动率指数，通过非对称分位数回归和非线性ARDL方法，发现加密货币市场中正负收益冲击均导致波动率上升，且在高波动率下存在非对称效应，对交易和风险管理有启示。
Implied volatility has consistently demonstrated its reliability as a superior estimator of the expected short-term volatility of underlying assets. In this study, we employ the newly constructed robust model-free implied volatility (MFIV) indices for Bitcoin and Ethereum (BitVol and EthVol) to explore the asymmetric return-volatility relationship of these cryptocurrencies through the lens of behavioral finance theories. Utilizing the asymmetric quantile regression model (QRM) and the Non-linear ARDL (NARDL) approach, our results reveal a notable difference from equities. Both positive and negative return shocks in the cryptocurrency market lead to an increase in volatility. However, during high volatility regimes, positive (negative) return shocks exert a more substantial impact on positive innovations of volatility for Bitcoin (Ethereum) compared to negative (positive) return shocks. The degree of asymmetry steadily intensifies as we progress from medium to uppermost quantiles of the volatility distribution. These observed phenomena can be attributed to behavioral aspects among market participants, including noise trading, behavioral biases, and fear of missing out (FOMO). Our findings hold significant implications for various aspects of cryptocurrency trading, portfolio hedging strategies, volatility derivatives pricing, and risk management.