Some things not mentioned above:
Baum, Joel AC, and Brian S. Silverman. “Picking winners or building them? Alliance, intellectual, and human capital as selection criteria in venture financing and performance of biotechnology startups.” Journal of business venturing 19.3 (2004): 411-436.http://www.library.auckland.ac.nz/subject-guides/bus/docs/PickingWinners2004.pdf
Agrawal, A., Kapur, D., McHale, J., 2008. How do spatial and social proximity influence knowledge flows? Evidence from patent data, Journal of Urban Economics, 64.
Zacharakis, Andrew L., and G. Dale Meyer. “The potential of actuarial decision models: can they improve the venture capital investment decision?.” Journal of Business Venturing 15.4 (2000): 323-346.http://www.sciencedirect.com/science/article/pii/S0883902698000160
Keely, Robert H. Determinants of new venture success before 1982 and after a preliminary look at two eras. Instituto de Estudios Superiores de la Empresa, Universidad de Navarra, 1989.http://www.iese.edu/research/pdfs/DI-0173-E.pdf
Chrisman, James J., Alan Bauerschmidt, and Charles W. Hofer. “The determinants of new venture performance: An extended model.”
Entrepreneurship Theory and Practice 23 (1998): 5-30.http://misweb.cbi.msstate.edu/~COBI/faculty/users/jchrisman/files/autoweb/mgt8123/MGT8123(Chrismanetal.,ETP,1998).pdf
Ross Levine, Yona Rubinstein, Smart and Illicit: Who Becomes an Entrepreneur and Do They Earn More?, The Quarterly Journal of Economics, Volume 132, Issue 2, May 2017, Pages 963–1018,https://economics.uchicago.edu/workshops/Rubinstein%20Yona%20Smart%20and%20Illicit.pdf
J. Robert Baum, Edwin A. Locke, and Ken G. Smith, 2001: A Multidimensional Model of Venture Growth. AMJ, 44, 292–303 http://www.taranomco.com/wp-content/uploads/2013/11/247.pdf
Experimentation and the Returns to Entrepreneurship. Gustavo Mansohttps://www.gsb.stanford.edu/sites/default/files/documents/Gustavo.pdf
Note that many of these are trying to test some model of venture success, and only calculate things related to EV as a subcomponent of that project. So it might not always be easy to answer the question you’re actually trying to answer here.
Also, it’s surprisingly hard to define “startup”, and some of the variance in these estimates comes from using different reference classes.
This is fantastic, thanks!