新冠肺炎疫情冲击导致经济出现结构性变化,对通胀预测提出了新的挑战;而大数据时代的到来,则为提高通胀预测的时效性提供了新的机遇.本文据此围绕基于大数据的通胀“现时”预测(nowcasting)进行探索,提出一个基本的现时预测框架,其核心...新冠肺炎疫情冲击导致经济出现结构性变化,对通胀预测提出了新的挑战;而大数据时代的到来,则为提高通胀预测的时效性提供了新的机遇.本文据此围绕基于大数据的通胀“现时”预测(nowcasting)进行探索,提出一个基本的现时预测框架,其核心是引入新的大数据宏观实时变量或大数据预测方法.本文通过引入宏观实时变量--基于互联网在线大数据的居民消费价格指数(internet-based consumer price index,iCPI),包括总类和大类的iCPI日环比指数、周环比指数、旬同比指数和月同比指数,采用LASSO(the least absolute shrinkage and selection operator)降维法和混频数据抽样模型(mixed data sampling,MIDAS),有效地提高了通胀预测的时效性和准确性.研究发现:不同频率的iCPI均有利于提高通胀预测准确性,其表现优于基准模型和大部分的同频传统指标,当其与传统指标相结合时,可进一步降低预测误差,目前尚不能完全舍弃传统变量和方法;在不同频率下(日度除外),iCPI八大类的预测效果优于iCPI总类;不同频率的大数据指标在通胀预测的准确性和时效性上各有优势,这与其背后反映的信息结构有关,其中高频旬同比iCPI表现尤为突出、其能较好地兼顾预测时效性和准确性.本研究为数字经济时代利用大数据提高通胀预测的准确性和时效性、创新宏观经济监测与预测体系提供了有益参考.展开更多
We develop a recession forecasting framework using a less restrictive target variable and more flexible and inclusive specification than those used in the literature.The target variable captures the occurrence of a re...We develop a recession forecasting framework using a less restrictive target variable and more flexible and inclusive specification than those used in the literature.The target variable captures the occurrence of a recession within a given future period rather than at a specific future point in time(widely used in the literature).The modeling specification combines an autoregressive Logit model capturing the autocorrelation of business cycles,a dynamic factor model encompassing many economic and financial variables,and a mixed data sampling regression incorporating common factors with mixed sampling frequencies.The model gene rates significantly more accurate forecasts for U.S.recessions with smaller forecast errors and stronger early signals for the turning points of business cycles than those gene rated by existing models.展开更多
文摘新冠肺炎疫情冲击导致经济出现结构性变化,对通胀预测提出了新的挑战;而大数据时代的到来,则为提高通胀预测的时效性提供了新的机遇.本文据此围绕基于大数据的通胀“现时”预测(nowcasting)进行探索,提出一个基本的现时预测框架,其核心是引入新的大数据宏观实时变量或大数据预测方法.本文通过引入宏观实时变量--基于互联网在线大数据的居民消费价格指数(internet-based consumer price index,iCPI),包括总类和大类的iCPI日环比指数、周环比指数、旬同比指数和月同比指数,采用LASSO(the least absolute shrinkage and selection operator)降维法和混频数据抽样模型(mixed data sampling,MIDAS),有效地提高了通胀预测的时效性和准确性.研究发现:不同频率的iCPI均有利于提高通胀预测准确性,其表现优于基准模型和大部分的同频传统指标,当其与传统指标相结合时,可进一步降低预测误差,目前尚不能完全舍弃传统变量和方法;在不同频率下(日度除外),iCPI八大类的预测效果优于iCPI总类;不同频率的大数据指标在通胀预测的准确性和时效性上各有优势,这与其背后反映的信息结构有关,其中高频旬同比iCPI表现尤为突出、其能较好地兼顾预测时效性和准确性.本研究为数字经济时代利用大数据提高通胀预测的准确性和时效性、创新宏观经济监测与预测体系提供了有益参考.
基金funding from School of Accounting and Finance,Faculty of Business,Hong Kong Polytechnic University.
文摘We develop a recession forecasting framework using a less restrictive target variable and more flexible and inclusive specification than those used in the literature.The target variable captures the occurrence of a recession within a given future period rather than at a specific future point in time(widely used in the literature).The modeling specification combines an autoregressive Logit model capturing the autocorrelation of business cycles,a dynamic factor model encompassing many economic and financial variables,and a mixed data sampling regression incorporating common factors with mixed sampling frequencies.The model gene rates significantly more accurate forecasts for U.S.recessions with smaller forecast errors and stronger early signals for the turning points of business cycles than those gene rated by existing models.
