Log-periodic view on critical dates of the Chinese stock market bubbles

We present an analysis of critical dates of three historical Chinese stock market bubbles (July 2006-Oct. 2007, Dec. 2007-Oct. 2008, Oct. 2014-June 2015) based on the Shanghai Shenzhen CSI 300 index (CSI300). This supports that the log-periodic power law singularity (LPPLS) model can describe well the behavior of super-exponential (power law with finite-time singularity) increase or decrease of the CSI300 index, suggesting that the LPPLS is available to predict the critical date. We also attempt to analyze the fitting parameter α of the LPPLS and the forecast gap which is between the last observed date and the expected critical date, proposing that the forecast gap is an alternative way for advanced warning of the market conversion.