CAJ | 학술논문

本文以8种农产品期货的高频数据为样本，实证考察了我国农产品期货市场已实现波动率的动态特征，发现农产品期货已实现波动率同时具有长记忆性和区制转换性。在此基础上构建了长记忆马尔科夫区制转换模型来预测农产品期货的已实现波动率，并比较和评价了该模型与其他嵌套模型的预测性能。结果发现，我国农产品期货的已实现波动率具有高波动和低波动两种不同的状态，状态之间的转换概率较小，低波动状态的稳定性比高波动状态强；同时引入长记忆性和区制转换能进一步提高模型的预测性能，长记忆马尔科夫区制转换模型是预测性能最好的模型。
본문이8충농산품기화적고빈수거위양본，실증고찰료아국농산품기화시장이실현파동솔적동태특정，발현농산품기화이실현파동솔동시구유장기억성화구제전환성。재차기출상구건료장기억마이과부구제전환모형래예측농산품기화적이실현파동솔，병비교화평개료해모형여기타감투모형적예측성능。결과발현，아국농산품기화적이실현파동솔구유고파동화저파동량충불동적상태，상태지간적전환개솔교소，저파동상태적은정성비고파동상태강；동시인입장기억성화구제전환능진일보제고모형적예측성능，장기억마이과부구제전환모형시예측성능최호적모형。
This paper empirically analyzes the dynamic properties of the realized volatility in China＇ s agricultural commodity future markets by employing the high - frequency data of eight agricultural commodity futures, and finds that the realized volatility exhibits both long memory and regime switching features. To capture these properties simultaneously, we modeled the realized volatility of agricultural commodity futures in China＇ s future markets by a MS -ARFIMA model, and compared its forecasting performance with that of a variety of competing models which have been encompassed by the proposed model. The estimated results show that the dynamics of realized volatility for agricultural commodity futures are characterized by two levels of long memory： one associated with the low - volatility regime and the other one associated with the high - volatility regime, and the probability to stay in the low - volatility regime is higher than that of the high - volatility regime. The outsample volatility forecast results show that the combination of long memory with switching regimes can improve the out - of - sample realized volatility forecast performance, and the proposed MS - ARFIMA model represents the superior out - of - sample realized volatility forecasts over those obtained using nested models in China＇ s agricultural commodity future markets.