摘要
近一二十年来,创新和技术水平的突飞猛进加剧了商品市场的竞争格局,促使产品更新换代频率提高、生命周期缩短、新产品功能和质量显著提升。这些新情况是导致产品样本代表性缺失、价格指数结果存在系统性偏差的主要原因,对我国官方价格统计提出了挑战。本文结合兼具扫描数据和网络爬虫数据优点的电商平台大数据,针对新经济时代传统特征价格指数面临的多重问题,提出一种全面、系统的特征价格指数优化方法。以特征价格估计法为基本框架,就模型设定不足、权重缺失和指数的链式漂移问题,提出了充分考虑特征变量规模效应和模型经济意义的加权非线性特征价格模型,并借助RYGEKS指数化解链式漂移。在此基础上,本文利用京东平台大数据进行试算,结果表明:传统价格指数由于样本轮换滞后、质量调整不足存在系统性偏差;改进的特征价格模型纠正了模型设定和权重缺失问题,对产品特征价格的拟合效果更优;完全匹配数据下的RYGEKS指数能够有效克服链式漂移的问题。
In the past one to two decades,the rapid innovation and technological advancement have intensified the competition of the commodity market,increased the frequency of product updating,shortened the product’s life cycle,and improved the functions and quality of new products significantly.These new conditions are the main reasons why there is a lack of representativeness of samples and the systematic deviation of price index results,which pose a challenge to the official price statistics of China.This paper combines the big data of e-commerce platforms,which has the advantages of scanning data and web crawler data,and puts forward a comprehensive and systematic method of optimizing the hedonic price index to solve the multiple problems in the new economic era.Based on the framework of the hedonic price imputation method,this paper proposes a weighted nonlinear hedonic price model for the insufficient model setting,lack of weights,and chain drift of index.This model takes full account of the scale effect of the hedonic variables and the economic significance of the model,and solves the chain drift with the help of RYGEKS index.On this basis,this paper uses the big data of JD platform for trial calculations.The results show that the traditional price index has obvious systematic deviations due to lagging sample rotation and insufficient quality adjustment;the improved hedonic price model corrects the problem of model setting and the lack of weights,and has a better fitting effect;RYGEKS index with fully matching data avoids the chain drift effectively.
作者
雷泽坤
郑正喜
许宪春
Lei Zekun;Zheng Zhengxi;Xu Xianchun
出处
《统计研究》
CSSCI
北大核心
2020年第8期22-34,共13页
Statistical Research
基金
国家社会科学基金重大项目“大数据背景下我国新经济新动能统计监测与评价研究”(18ZDA124)
国家社会科学基金青年项目“国家投入产出表的改进编制与应用问题研究”(18CTJ002)清华大学自主科研计划文科专项课题“数字经济统计核算问题研究”的阶段性成果。
