基于分层图限制博弈Shapley值的大湾区跨境数据要素研究

A Study on Cross-Border Data Elements in the GBA Based on the Shapley Value of Hierarchical Graph-Constrained Games

摘要: 数字经济迅速发展使数据成为重要的生产要素。粤港澳大湾区在“一国两制三法域”之下存在合规摩擦，传统的合作博弈难以刻画通道约束。因此本文提出了一种基于图限制博弈的分层蒙特卡洛Shapley值(SGR-MC-Shapley)。在嵌套CES生产函数中引入数据要素后，利用Myerson值将城市间合规通道表示为合作图边并内化其成本权重，构造图限制特征函数，并进一步设计分层抽样和截断策略来降低大规模估值方差。以2024年深圳、香港、澳门等地数据为基础开展情景模拟。研究表明，深圳体量虽仍居核心，但合规壁垒下降时香港作为跨境“联系人”的边际贡献提升最显著。研究为跨境数据流动机制和生态补偿制度提供量化的依据。

Abstract: The rapid growth of the digital economy has caused data to become a factor of production. Within the Guangdong-Hong Kong-Macao Greater Bay Area, due to institutional differences that arise under the framework of “one country, two systems, three legal jurisdictions”, there are compliance frictions that are unable to fully be accounted for by traditional cooperative games models that do not restrict the formation of coalitions. So this study put forth a graph-restricted Monte Carlo Shapley value with stratification (SGR-MC-Shapley). After introducing the data factor into the nested CES production function, Myerson value on inter-city compliance channels is represented with a weighted edge on a cooperation graph with endogenized regulatory cost to obtain a graph restricted characteristic function. In order to solve for the variance to scale from a large scale valuation, in order to further do a stratified sampling and truncation. Scenarios are set up using Shenzhen, Hong Kong, Macao in 2024 as the data source. Result shows although Shenzhen remains the main source because of its economic scale, due to lower compliance barrier, the biggest increase in marginal contribution from Hong Kong takes place, which shows its role in cross-border connecting. This gives a quantitative basis for designing cross-border data flow mechanism and digital ecosystem compensation programmes.