一名计量经济学家对大数据的展望  

An Econometrician's Perspective on Big Data

作　　者：萧政[1] 周波 Cheng HSIAO(University of Southern California)

机构地区：[1]美国南加州大学经济系 [2]对外经济贸易大学国际经济贸易学院财政税务学系

出　　处：《财经智库》2019年第1期124-137,144,共15页Financial Minds

摘　　要：本文基于大数据的三个特征,分析大数据方法对计量经济学建模、预测以及评估或模拟社会政策效果的影响,并对大数据带来的一些方法论的挑战进行讨论。文章认为,大数据为经济分析和预测开辟了新的可能,但大数据并不是灵丹妙药,只要现实世界仍存在不确定性,基于概率论的统计分析以及对经济学、制度和历史背景的理解依然是必要的。此外,也必须考虑各种假设所带来的后果。In light of the three features of big data, this paper analyzes the impacts of big data on econometric modeling, prediction and measuring or simulating the impacts of social policies, and discusses some specific methodological challenges in light of the availability of big data. Big data sets shed light on economic analysis and prediction. However, big data is not a panacea. No matter how many digital sensors are placed worldwide and how thorough the data are collected, as long as the world is not a deterministic world, statistical analysis based on probability theory, understanding of economics and institutional and historical backgrounds remain essential.

关 键 词：大数据分析 政策评估 计量经济分析 

分 类 号：B41[哲学宗教—外国哲学] C02[社会学]