Communities of Investors and Start-up Companies: An Analysis Using Bipartite Stochastic Block Model

         
Author Name KOJAKU Sadamori (Kobe University) / MIYAKAWA Daisuke (Hitotsubashi University)
Creation Date/NO. December 2019 19-E-101
Research Project Creation and Development of High-tech Startups
Download / Links

Abstract

Using unique investment-level data accounting for around 15,000 individual investments done by various types of investors in start-up companies over the periods from 2000 to 2014, we empirically detect community structures consisting of investors and start-up companies through the bipartite stochastic block model and examine their implications on investment performance. The detected community structure represented by multiple groups of investors and start-up companies suggests, first, large heterogeneity of each community in terms of clustered investor types but less so in terms of start-up companies' industry composition. Second, we observe investment performance is higher when the communities are populated by clusters of VCs or non-financial companies. These results jointly imply the systematic concentration of specific types of investors associated with investment performance.