Crypto-Coins and Credit Risk: Modelling and Forecasting Their Probability of Death

This paper examined a set of over two thousand crypto-coins observed between 2015 and 2020 to estimate their credit risk by computing probability death. We employed different definitions dead coins, ranging from academic literature professional practice; alternative forecasting models, scoring models machine learning time-series-based models; horizons. found that the choice coin-death definition affected best compute However, this was not critical, turned out be same in most cases. In general, we cauchit zero-price-probability (ZPP) based on random walk or Markov Switching-GARCH(1,1) were for newly established whereas credit-scoring machine-learning methods using lagged trading volumes online searches better choices older coins. These results also held after robustness checks considered time samples coins’ market capitalization.