论文部分内容阅读
传统的GARCH模型依然使用一个稳定的GARCH过程来刻画和预测金融资产收益率的波动特性,这种假设显然与现实不符。本文通过剔除异常值对波动建模、结构突变检验的影响,使用修正的ICSS算法检测我国股市收益波动的结构突变性,考查结构突变对股市收益波动性的影响,并运用SPA检验评估几种GARCH模型的预测精度。结果表明同其他新兴市场国家一样,我国股市确实存在收益率波动的结构突变现象,且政策效应较为明显;而基于结构突变的GARCH模型,在不同的预测步长下均具有很高的预测精度,验证了结构突变对波动预测的重要性,而包含非对称性质的GJR-GARCH模型则在长期具有良好的预测精度。
The traditional GARCH model still uses a stable GARCH process to characterize and forecast the volatility of the return on financial assets, which is obviously not in conformity with the reality. In this paper, by eliminating the influence of outliers on the volatility modeling and structural mutation testing, we use the modified ICSS algorithm to detect the structural mutation of the volatility of the stock market returns in China, and examine the effect of the structural mutation on the volatility of the stock market returns. Prediction accuracy of the model. The result shows that, like other emerging market countries, the stock market in our country does exist the structural mutation phenomenon of the return rate volatility, and the policy effect is obvious; while the GARCH model based on the structural mutation has a very high prediction accuracy under different forecasting steps, It is verified that the structural mutation is of importance to the volatility prediction. However, the GJR-GARCH model with asymmetric properties has a good prediction accuracy in the long run.