Founders, employees, suppliers, customers, investors and bankers share an interest in the fate of young private companies.
Capitalists and managers care particularly about return on investment. Forms of investment we can think about include founder shares, angel
and seed capital, institutional venture capital, and common shares
relevant for employee stock options.
Unlike public equity and debt instruments, there is no data set today that contains a
representative record of the invested capital and return fate for
private businesses for any of the primary security instruments: common shares, preferred shares, senior debt and mezzanine debt. Ideally, this data would also be arranged by the amount of time
required for the outcome to be realized and some recording of the
company's capitalization at each observation (funding event). Instead, what we have are company-level databases with incomplete records of returns on invested capital, data contributed by individual managers, aggregate fund manager returns and portfolios of managers
(like Calpers).
Survivorship bias is one of the problems with company-level databases. Having been
polled by representatives from Thomson Venture Economics and the like as CFO of a private, venture-backed company, I can completely appreciate
the limitations in the data: if a company has something it is proud of (like
a funding event, or successful acquisition), the company will talk about it (e.g.;
the data will be recorded somewhere). If a company does not obtain follow on
financing or complete a successful exit, it is more likely than not that the message
from the database provider will not be returned and such records will simply dead
end with no account of the economic outcome.
John Cochrane at the University of Chicago has made one of the most serious attempts
to measure the survivorship bias in venture capital manager returns. Here is a link to John Cochrane's 2005 paper, The Risk and Return of Venture Capital published by the Journal of Financial Economics Jan 2005. The findings will be well understood
by practitioners and for those that are more public market oriented, you have seen a
similar story in the measurement of performance of investment partnerships, popularly
if not always properly, referred to as "hedge" funds. Note Malkiel's 2004 paper and understand the critique: WSJ story about the paper. Thanks to George Van and company for discussing their hedge
fund index return calculation methodology. Van Hedge paper.
Abstract excerpted from John Cochrane's site: Estimates the mean return, standard deviation, alpha and beta of venture capital
investments, correcting for selection bias that we only see returns for successful
projects. Even if you don't like venture capital, the selection bias correction
is interesting. Updated again, completely rerun, completely rewritten and much
better. Original December 2000. Appendix containing data and program descriptions plus extra algebra. See above data
and programs link for data and programs.
|
