"Whether differences among accelerators explain differences in the performance of member ventures is an important and underexplored question. Conversely, are the effects of accelerators so isomorphic, because they copy each other, that ventures from different accelerators report little performance differences? We use variance decomposition analysis to test whether variations in characteristics of accelerators explain performance differences in the ventures that belong to them. Using a sample of 1,442 ventures from 117 accelerator programs across 22 countries, we find that 11.13–14.18% variance of venture performance can be attributed to accelerator membership. Accelerator membership also accounted for 3.00, 5.15, and 16.65% in the variance for employee growth, employee costs, and revenue change, respectively. Our findings suggest that between accelerator differences can make a significant economic difference to venture performance."

"To understand the intermediary role of accelerators in the developing regional entrepreneurial ecosystem of Bangalore, we analyze data from 54 interviews with accelerator graduates, accelerator managers, and other ecosystem stakeholders, and from 49 websites, 13 online video interviews, 26 online news sources and 301 pages of policy documents. Specifically, we adopt a socially-situated entrepreneurial cognition approach to theorize how accelerator expertise, existing at a meso-level, intermediates between (micro-level) founders and the (macro-level). ecosystem. In our model, four types of accelerator expertise-connection, development, coordination, and selection-together increase stakeholders' commitment to the entrepreneurial ecosystem, leading to venture validation (success or failure) and ecosystem additionality. These findings indicate that accelerators contribute to ecosystems in a way that is distinct from, but supportive of, building individual ventures."