摘要
本文以混频数据模型(MIDAS)的建模理论和分析技术为视角,在结构分析框架下构建多种M-MID AS-AR模型,研究了我国高频资产价格对低频CPI影响效果及样本内预测精度。结果表明:资产价格对CPI水平有显著影响;资产价格上涨会通过财富效应抬高CPI;当高频变量股票价格的滞后阶数变动到30阶时,Beta-M-MIDAS-ARDL模型的拟合结果和预测效果优于ARDL模型及其他混频数据模型。
From the perspective of mixed frequency data's econometric models which are the theory and analytical techniques, this paper builds a variety of M-MIDAS-AR model in structural analysis framework, studies the influence,path and prediction precision of high frequency variable asset prices to the price level in China, Our country high frequency effects on low prices to asset prices, The results show that the asset price has a significant effect on price level; Stock prices will raise prices through the wealth effect; When the lag order of decreasing weight function changed to order 30.,Beta-MIDAS-ARDL model fitting results and prediction effect is superior to the ARDL model and other mixed frequency models.
出处
《价格理论与实践》
CSSCI
北大核心
2017年第4期96-99,共4页
Price:Theory & Practice
基金
国家社会科学基金重大项目"新常态下我国宏观经济监测和预测研究(15ZDA001)
内蒙古自治区自然科学基金项目"结构转换GARCH建模及其在金融资产收益波动中的应用"(2016MS0716)