摘要
文章引入一种新的权重函数,并构建新的波动率模型——MIX-GARCH-L模型,新模型能够充分利用高低频数据提炼出更有价值的信息。针对新模型参数估计问题,提出MIX-GARCH-L模型的参数估计方法来分析估计量的理论性质,证明了对应的中心极限定理以及用Service-Boostrap方法模拟检验估计量的数据表现。所提模型具有以下优势:新权重函数能够更好地根据交易特征的变动来自动调整不同交易日的权重,从而使每个高频交易日所分配到的权重与未来波动率产生的冲击效果一致;能够利用同一交易过程中多种高频交易数据,信息利用更加充分,使得MIX-GARCH-L模型具有更好的预测精度和预测优势。实证结果显示:MIX-GARCH-L模型的MSPE值明显小于GARCH-RV模型和GARCH-M模型的MSPE值,说明MIX-GARCH-L模型不仅在模型预测上有更高的预测精度,而且在稳健性上的表现也更好。
This paper introduces a new weight function and constructs a new volatility model—MIX-GARCH-L model.The new model can fully utilize high and low frequency data to extract more valuable information.In response to the problem of param-eter estimation in the new model,this paper proposes a parameter estimation method for the MIX-GARCH-L model to analyze the theoretical properties of the estimator,proves the corresponding central limit theorem,and uses the Service-Bootstrap method to simulate and test the data performance of the estimator.The proposed model has the following advantages:The new weight func-tion can better automatically adjust the weight allocation for different trading days based on changes in trading characteristics,so that the weight assigned to each high-frequency trading day is consistent with the impact effect of future volatility.The new volatil-ity model can utilize a variety of high-frequency trading data within the same trading process,and make better use of information,so that MIX-GARCH-L model has better prediction accuracy and prediction advantages.The empirical results show that the MSPE value of the MIX-GARCH-L model is significantly smaller than that of the GARCH-RV model and GARCH-M model,in-dicating that the MIX-GARCH-L model not only has better prediction accuracy in model prediction,but also has better perfor-mance in robustness.
作者
杨炜明
刘涛
王琴
Yang Weiming;Liu Tao;Wang Qin(School of Mathematics and Statistics,Chongqing Technology and Business University,Chongqing 400067,China;Yangtze River Upstream Economic Research Center,Chongqing Technology and Business University,Chongqing 400067,China;School of Big Data and Intelligent Engineering,Chongqing College of International Business and Economics,Chongqing 401520,China)
出处
《统计与决策》
CSSCI
北大核心
2024年第8期22-27,共6页
Statistics & Decision
基金
国家统计局重大统计专项(2023ZX08)
重庆市自然科学基金资助项目(CSTC2020JCYJ-MSXMX0394)
重庆市社会科学规划项目(2022NDQN23)
重庆对外经贸学院校级重点项目(KYSK202203)。
关键词
混频数据
新权重
波动率
参数估计
数值模拟
mixed frequency data
new weight
volatility
parameter estimation
numerical simulation
