Recent Working Papers: Risk and Return of Venture Capital

Many tech-based economic development programs recognize the importance of having seed and venture capital accessible to their start up tech firms and entrepreneurs. Some practitioners, though, see a challenge in encouraging equity investment in more conservative, risk adverse regions and localities. The dot-com “correction” of last year probably did not help.

Accurate estimates of the average return on venture capital investment (VC) may help to open the purse strings of hesitant angel and seed funding sources. But what is a valid estimate of the return on individual venture capital investments?

Obtaining an answer to that question has not been easy in the past. In the new working paper, The Risk and Return of Venture Capital, John Cochrane, Professor of Finance with the Graduate School of Business of the University of Chicago, takes on the issue.

Attempts by economists in the past have been plagued by a lack of accurate data on the market value of individual VC deals between initial investment and disposition through an initial public offering (IPO), acquisition/merger, or failure. Also, complete information on individual deals is difficult to obtain because of their private or proprietary nature. The staggered nature of deals going public or failing presents additional challenges for estimating return on investment.

To overcome these problems, most past estimates of return on investment have either made assumption of interim market values to develop annualized snapshots of deals or looked at the return on whole VC portfolios rather than individual deals. Other past papers measuring VC from initial investment to IPO have been limited to working with small samples of completed deals.

Cochrane explains that a selection bias occurs by looking at return to IPO in that most IPOs will be made only when the project has achieved a good return or anticipates a successful offering. IPOs can be delayed almost indefinitely until the value of the project has risen to a satisfactory level, the company is acquired or merges, or the initial investors calls it quits. As a result, only measuring return on VC investment based on IPOs skews the results to the most profitable deals.

Cochrane’s research looks to overcome this selection bias by working with the  VentureOne 1987-2000 database of 16,852 individual financing rounds involving 7,700 companies and $114 billion of investments. The data revealed 21.7 percent of the early 17,000 financing rounds resulted in initial public offerings. Another 3.7 percent had registered their IPOs. Slightly more than 20 percent resulted in mergers or acquisitions while 8.9 percent were no longer in business due to business failure. The balance remained private.

A review of the data set by date of initial investment and date of disposition finds that the whole VC investment cycle is quickening. The entire deal process (entrance to exit) is happening about one year faster now than during the earliest cycle of 1988-1992. Interestingly, this does not necessarily translate to VC’s making more money since those deals going out of business are happening that much earlier as well – and the percentage remains relatively constant.



Without any correction for the selection bias, the net returns for firms that go public or are acquired reveals 15 percent of the deals return less than 0 (lose money) and 35 percent of the returns are less than 100 percent of the initial investment. While the average return is a staggering 698 percent, thanks to a few astounding results, the “most probably” outcome is a more modest 25 percent.

The statistical correction applied for the selection bias moderates the results dramatically. Cochrane found arithmetic average returns of approximately 53 percent compared to the 698 percent. Second, third, and fourth rounds of financing find lower arithmetic average returns as they are traditionally less risky.



Cochrane’s paper, The Risk and Return of Venture Capital, can be purchased from the National Bureau of Economic Research at http://www.nber.org