Bubble detection in Bitcoin and Ethereum and its relationship with volatility regimes

Purpose - The aim of this paper is to test the existence of bubbles for the daily prices of cryptocurrencies Bitcoin and Ethereum and verify if there is a relationship between bubbles and volatility regimes. Design/methodology/approach - The authors test the presence of bubbles with the generalized supremum augmented Dickey–Fuller (GSADF) test using critical values simulated by the bootstrap procedures of Gutierrez (2011), Harveyet al.(2016) and Pedersen and Schütte (2020). Also, the authors estimate Markov regime switching generalized autoregressive conditional heteroskedasticity model for these cryptocurrencies. Findings - The GSADF test result indicates the presence of bubbles for both cryptocurrencies. Simulating critical values by wild-bootstrap, which is robust to non-stationary volatility, leads to the highest number of bubbles in both cryptocurrencies. In addition, based on the estimates of conditional variance models with regime changes, the authors find that the bubbles identified are associated with a regime of low returns volatility, indicating a change in the trade-off between risk and return when the prices of cryptocurrencies differ from their fundamental values. Originality/value - To the best of the authors knowledge, there are no studies that test the explosive behavior for cryptocurrencies by the GSADF test using the bootstrap method to simulate critical values from the procedures of Harveyet al.(2016) or Pedersen and Schütte (2020). These bootstrapping procedures are robust to heteroscedasticity and avoid the detection of false bubbles. Further, the advantage of Harveyet al.(2016) procedure is the robustness to non-stationary volatility.