Abstract: There are intricate supply and demand relationships between various market objects in the economic market, which in turn form an industrial organizations network. This paper uses the transaction data from the bank to present the inter-firm relationships in the city of Yiwu, and overcomes the limitation that the input-output method can only reflect the inter-industry connections. In addition, it also introduces the exponential random graph model to investigate the impact of the diverse social processes in the formation of inter-firm relationships and their role in shaping the industrial organizations network from a more microcosmic perspective. The results show that there is a strong tendency of reciprocity in the inter-firm relationships in Yiwu, and there are more two-way interactions than one-way trading. The firms in Yiwu are more likely to choose transaction objects through middleman, which means that transitivity plays an obviously positive role in the network, while popularity and activity mainly play a negative role, that is, firms are not willing to expand commercial channels. Compared with non-profit organizations, firms play more of the role of suppliers, and firms located in relatively affluent and less populated areas have higher export capabilities.

Key words: industrial organization network, inter-firm relationships, exponential random graph model (ERGM), endogenous structural effects, actor-relation effects

摘要： 经济活动中，各类市场主体之间存在错综复杂的供给与需求关系，进而形成产业组织网络。采用银行转账数据还原义乌市企业间的相互关联，克服了投入产出法只能反映产业层面关联的限制，并引入指数随机图模型，从更微观的层面，更充分和全面地考察影响企业间关系形成的多种社会过程及其对产业组织网络的塑造作用。研究发现：义乌市企业关系中存在强烈的互惠倾向，双向互动多于单向买卖；义乌市企业选择交易对象更多地是通过中间商联络，传递性在网络中起到了较大的正向作用，而聚敛性和扩张性主要起到负向作用，即企业拓展商业渠道的意愿不强烈；与非企业机构相比，企业更多地担任供给方的角色，位于相对富裕、人口较少地区的企业具有更高的对外输出能力。

关键词: 产业组织网络, 企业间关系, 指数随机图模型, 内生结构效应, 行动者-关系效应