PD-EGARCH model based experimental research of calendar effect of securities market of China

This paper discusses the problem of calendar effect in the securities market of China, which is an abnormality test and data mining paradigm of financial time series from different dimensionlity. On the basis of retrospect of research on calendar effects from domestic and overseas scholars, panel data-exponential generalized autoregressive conditional heteroskedastic (PD-EGARCH) model is put out; then analysis of elementary statistic results of two typical indices (shanghai stock index and shenzhen component index) are offered. The primary focus of this paper will be selective descriptions of PD-EGARCH model based test of the primary statistic result of various performances of calendar effects. Finally, the testing result shows, behaviors of calendar effects in Chinese stock market is obvious but different from former research results, which shows new characteristics with arbitrage opportunity and leverage, and which is difficult to make clear by traditional explanation.