A1116
Title: Exploring noncausal and noninvertible ARMA-GARCH dynamics in the cryptocurrency market
Authors: Daniel Velasquez-Gaviria - Maastricht University (Netherlands) [presenting]
Alain Hecq - Maastricht University (Netherlands)
Abstract: The prices of Bitcoin and Ethereum exhibit non-stationarity, marked by frequent episodes of local explosions that abruptly collapse. This behavior resembles the well-known formation of financial bubbles, motivating the use of the mixed causal-noncausal and invertible-non-invertible autoregressive moving average (MARMA) model, which can replicate these characteristics. Additionally, conditional volatility is evident, where episodes of large variations are followed by further large variations and episodes of small variations are followed by further small variations. The estimation and identification of the new MARMA-GARCH model is proposed. An estimation in both the frequency domain and the time domain is suggested and promising techniques are introduced to remove the price trend and estimate the model in the cyclical component. The results indicate that Bitcoin and Ethereum prices exhibit noncausal behaviors, along with significant conditional volatility.
