Working Paper: NBER ID: w30682
Authors: Lisa D. Cook; Matt Marx; Emmanuel Yimfor
Abstract: We analyze determinants of access to venture capital for Black founders of high-growth startups. We combine image- and name-processing algorithms with clerical review to identify race for over 100,000 startup founders “at risk” for venture funding. Black founders raise roughly one-third as much venture capital in the five years after founding vs. other startups formed in the same year, industry, and state. What explains the gap? We attribute about a third of the gap to four factors: Black startups have smaller founding teams; Black founders are less likely to have worked at the same companies or attended the same schools as investors who fund startups in the same industry and geography; Black startups are less likely to be located in geographies with plentiful venture capital, and Black startups are less likely to have a patent. We then bound our estimates of the funding gap and show that, to explain the gap, omitted variables would have to be nearly four times as important as the variables we fix. The funding gap is not statistically different from zero in later funding stages (post Series B), suggesting that some investors initially hold—but later amend—incorrect beliefs about Black-founded startups. We also exploit the hiring of Black partners and show that they are more likely to fund successful Black startups, consistent with segmented networks.
Keywords: Venture Capital; Black Founders; High-Growth Startups
JEL Codes: G24; M13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| smaller founding teams in black startups (L26) | funding gap (I22) |
| lack of shared professional or educational backgrounds (I24) | funding gap (I22) |
| geographic disadvantages (R12) | funding gap (I22) |
| lower patent rates among black-founded startups (L26) | funding gap (I22) |
| black partners in venture capital firms (G24) | funding black-founded startups (M13) |
| omitted variables need to be significantly more impactful than controlled factors (C20) | cautious interpretation of omitted variables role (C20) |
| shift in investor perceptions over time (G41) | funding gap not present in later funding rounds (G19) |