Face distance : unpacking the role of ethnic ties in venture capital investment

Abstract

Venture capitalists have been shown to be more likely to invest in entrepreneurs of the same ethnicity. At the same time, this result rests on assumptions about how shared ethnicity is defined both theoretically and empirically. Current measurement of ethnic ties is problematic due to mis-classifications, mixed heritage individuals, and variation in accuracy by ethnicity. This paper overcomes these limitations by taking advantage of a novel source of data -- face photographs -- and by applying advanced machine learning techniques to compute the facial similarity between investors and entrepreneurs in a large scale dataset of realized and potential investments. Results suggest that previous work has vastly underestimated the relationship between ethnic ties and investment. Moreover, this relationship is more nuanced than previously documented, varies with the stage of investment and the type of investors involved, and is associated with a lower likelihood of securing follow-on funding or achieving an exit.