生物技术产业双重网络演化研究——基于时序指数随机图模型  

Research on the dual network evolution of the biotechnology industry——Based on temporal exponential random graph model

作　　者：王慧 高山行[1] 杨张博[2] 李泞芮 WANG Hui;GAO Shan-xing;YANG Zhang-bo;LI Ning-rui(School of Management,Xi'an Jiaotong University,Xi'an 710049,China;School of Humanities and Social Science,Xi'an Jiaotong University,Xi'an 710049,China)

机构地区：[1]西安交通大学管理学院,陕西西安710049 [2]西安交通大学人文社会科学学院,陕西西安710049

出　　处：《科学学研究》2025年第1期178-188,共11页Studies in Science of Science

基　　金：国家社会科学基金重点项目(20AGL004);国家自然科学基金面上项目(72372127);中央高校基本科研业务费(人文社科类)(SK2023039)。

摘　　要：现代高科技企业同时嵌入在不同类型的网络中,但现有研究较少分析这些网络如何共同演化。论文从双重网络嵌入视角出发,利用2009年至2021年中国生物技术产业338家上市公司的29145名股东、35597名董事数据,构建股东关系网络与连锁董事网络,运用社会网络分析方法、时态指数随机图模型研究组织双重网络嵌入的演化路径。研究发现:股东关系网络演化过程中,内部结构越来越紧密,表现出小世界特征,连锁董事网络内部关系数量较少,整体结构稀疏。两个网络的演化均表现出显著的空间同质性特征;股东关系网络中关系的建立还表现出产权异质性,即国企和非国企间有更高概率建立联系;双重网络嵌入关系形成具有跨网络马太效应,企业在一个网络中拥有的关系数量越多,在另一个网络中越容易建立新连接。本研究从单一网络研究视角迈入企业双重网络嵌入视角,反映了高科技企业在实践中嵌入不同网络的情况,检验了股东关系网络与连锁董事网络的内生形成机制,清晰展现了双重网络的演化路径。Modern high-tech companies are simultaneously embedded in various types of networks,yet existing research seldom analyzes how these networks co-evolve.Measured by the strength of relationships,the frequency of interactions within the board of directors is considerably higher compared to the annual general meeting of shareholders.In terms of relationship duration,directors are bound by tenure limits,whereas chain relationships tend to be long-term.Shareholders,on the other hand,may divest their stocks at any time due to market fluctuations,dissolving their shared ownership.In terms of interaction content,major shareholders do not participate in daily corporate governance and have limited understanding of the business,whereas the board directly engages in corporate governance and strategic decision-making.Consequently,the network of chain directors constitutes a stronger relational network compared to the shareholder relationship network,which is relatively weaker.The paper approaches from a dual-network embedding perspective,utilizing data from 2009 to 2021 on 33,845 shareholders and 35,597 directors from 338 biotechnology companies listed in China.It constructs both shareholder relationship networks and chain director networks,employing social network analysis methods and Temporal Exponential Random Graph Models(TERGM)to study the evolutionary paths of organizational dual-network embedding.TERGM considers the interdependence and dependency relationships of networks at different time points,enabling the modeling of various network characteristics like reciprocity,transitivity,and homophily.This is vital for understanding network evolutionary driving mechanisms and allows researchers to explore how different network features influence the formation and dissolution of ties over time.The study finds,firstly,that spatial homophily plays a significant role in the evolution of both networks.Geographically proximate companies are more likely to form chain relationships and shared shareholder relationships.This corroborat

关 键 词：双重网络嵌入 股东关系网络 连锁董事网络 网络演化 TERGM 

分 类 号：F272.3[经济管理—企业管理]