Double Threshold GARCHモデルとその株価変化率への応用 : ベイズ統計学を用いたパラメータ推定とモデル選択 [in Japanese] Double Threshold GARCH model and its Application to Individual Stock Price : Bayesian Estimation and Model Selection [in Japanese]
Access this Article
Search this Article
We apply Double Threshold GARCH (DT-GARCH) models to weekly change rate of Japanese individual stock price(Mitsui Fudosan). For comparison, we apply threshold autoregressive (TAR) model and autoregressive (AR) model, too. After Bayesian estimation, we select DT-GARCH model as most appropriate model by DIC. The estimate of its threshold is of －4.17(%). Then the auto regressive model with GARCH process will change the parameters, if the change rate exceeds the threshold.
- Bulletin of the Yamagata University. Social science
Bulletin of the Yamagata University. Social science 42(2), 17-30, 2012-02