数据越多越好吗——对数据要素报酬性质的跨学科分析  

Is More Data Always Better:An Interdisciplinary Analysis of the Nature of Returns to Data

作　　者：王超贤 张伟东 颜蒙 Wang Chaoxian;Zhang Weidong;Yan Meng

机构地区：[1]中国信息通信研究院政策与经济研究所

出　　处：《复印报刊资料（统计与精算）》2022年第6期121-137,共17页STATISTICS AND ACTUARIAL SCIENCE

基　　金：国家社会科学基金青年项目“数字经济时代兼顾个人信息保护和数据要素市场培育的治理机制设计”(批准号21CJL010)。

摘　　要：流行的观点认为数据是新的高级生产要素,具有强大的规模经济、范围经济和数据网络效应,是报酬递增的。但这既不完全符合产业事实,也不完全符合经济规律,是一个巫函需辩论也正在辩论的命题。本文以经济学分析为主线,辅以管理学、信息系统和数据工程科学等多学科的理论和证据,对数据要素报酬性质的基本原理、影响机制、情境证据等进行跨学科分析。研究发现,数据要素报酬性质是状态依赖的,在不同领域、不同技术和不同制度条件下,数据的报酬性质是不同的,既可能递增,也可能递减,还可能不变,需具体情境具体分析。为此,本文回到要素报酬性质决定的根源处,归纳了数据价值创造全过程中影响数据要素报酬性质的五类因素:数据质量、数据参与生产的规模报酬、数据的外部经济、数据驱动的学习效应、数据生产力和产品价值的质量阶梯,并在理论上推导了它们影响数据要素报酬性质的机制,得出了五个待检验的实证命题。在此基础上,本文构建了一个分析具体情境中数据要素报酬性质的综合框架,结合实证证据,确定了当前代表性情境中的数据要素报酬性质。本文提出的综合框架和结论对于未来研究数据如何影响企业战略、产业组织、竞争监管与经济均衡等问题具有基础支撑和重要启示作用。The prevailing view is that data is the new advanced factor,with strong economies of scale,scope,data network effects,and increasing returns.But this is a proposition that needs to be debated and is being debated.This paper takes economics as the main line,supplemented by multi-disciplinary theories and evidence of management,information systems,and engineering science,and conducts an interdisciplinary analysis of the basic principles,impact mechanisms,and contextual evidence of the nature of returns to data.The research finds that the nature of returns to data is state-dependent.In different fields,technologies,and institutional conditions,the nature of returns to data may be increasing,decreasing,or constant.Specific analysis is required in particular situations.Therefore,this paper returns to the origin of the determination of the nature of factor returns,and summarizes five types of factors affecting the nature of data returns in the entire process of data value creation,including data quality,returns to scale for data participation in production,external economics of data,data-driven learning effect,data productivity,and quality ladder of the product value.This paper theoretically deduces the mechanism by which they affect the nature of returns to data,and draws five empirical propositions to be tested.On this basis,this paper constructs a comprehensive framework for analyzing the nature of returns to data in specific contexts and combines empirical evidence to determine the nature of data element remuneration in current representative situations.The comprehensive framework and conclusions presented in this paper have fundamental support and an important implication for future research on how data affects corporate strategy,industrial organization,competition regulation,and economic equilibrium.

关 键 词：数据要素 要素报酬 报酬递增 数据学习效应 数据网络效应 

分 类 号：F49[经济管理—产业经济] TP311.13[自动化与计算机技术—计算机软件与理论]