Variance decomposing accelerator and cohort effects among London startups

Dushnitsky, G and Sarkar, S (2018) Variance decomposing accelerator and cohort effects among London startups. [Conference proceeding]

Abstract

A decade after the launch of Y-Combinator, there are dozens of accelerators all around the globe. And while the nurturing effect afforded by accelerator organizations is the subject of scholarly work, we know less about the dynamics that arise organically among the participants of a focal cohort. We synthesize the literature on accelerators to infer that accelerator participation could affect venture outcomes through mechanisms that operate at two distinct levels: i) the accelerator organization (top-down), and ii) the particular cohort of the accelerator that a venture attended (bottom-up). We report a variance decomposition analysis of a proprietary dataset of all ventures graduating from major London-based accelerators since 2008. We find that cohort membership explains about half of the total explained variance in venture outcomes whereas accelerator membership explains a lower fraction of ultimate funding success. For top accelerators, however, industry effect dominates both accelerator and cohort membership effects. Our findings suggest that nascent venture performance is more sensitive to cohort-level dynamics than it is to nurturing by the accelerator organization. For entrepreneurship scholars, the study underscores the opportunity to supplement our understanding of accelerators by focusing on within cohort dynamics, and how accelerator characteristics moderate investor decision-making.