Visualisation of Syndicate Networks Using Block Modelling

抄録

Intercompany alliances formed through investment syndicates are an effective risk aversion strategy. In Japan, where the direct financing market is still in its infancy, a high percentage of syndicates are formed. However, some aspects of these alliances are unclear, and questions regarding the relationships that are created when syndicates are formed, the commonalities in their bonding patterns, and alliance trends remain. Identifying the essential structural characteristics of Japanese investor networks triggered by syndicated investments would provide important clues for gaining insight into startup ecosystems. Therefore, in this study, we attempt to construct an image matrix and a reduced investor network structure graph using the block model. This study is novel in this respect, as no previous study has examined the visualisation of investor networks in Japan, focusing on social positions. If the essential structure of the chaotic investor network can be succinctly expressed, it will help us understand the formation patterns of the startup ecosystem. The period from 1 April 2014 to 31 March 2022 was selected for analysis, and a dataset of 7,599 cases related to 4,508 investors was prepared from investment statement data by investment round. An adjacency matrix was constructed and divided into nine blocks using CONCOR software. We depict a reduced graph from the image matrix and interpret the divided blocks. The results suggest that as more investors participate in syndicated investments, the structural characteristics of their networks can be seen as center-periphery.