Opportunity re-evaluation: how risk dimensions influence post-investment venture capital decisions.

Abstract

Why do managers embrace some risks while they reject others that appear to be equally risky? This dissertation examines how risk dimensions influence the decision to hold, reinvest in, or terminate investment in companies within a venture capital portfolio. I draw on prototype theory to argue that reinvestment is more likely in portfolio companies that "look like" successful VC-backed firms, while investment termination is more likely for firms that "look like" losing portfolio companies. Financial risk is an equivocal signal because it can indicate a problem or recur by design in the VC-backed portfolio company. Different levels of market, technology and management risk are all central characteristics of portfolio winners or losers. I predict that increasing financial risk will increase the likelihood of both reinvestment and termination compared to holding the investment. However, market, technology and management risk will negatively influence reinvestment. Regarding investment termination, I argue that less controllable risks will be rejected, while firms with more controllable risk will be retained. Therefore, I predict that market risk and technology risk will positively influence termination, while management risk will have no significant effect. The sample is composed of 542 quarterly observations of 57 companies in an early-stage VC firm's portfolio. The independent variables are the VC's financial, market, technology and management risk assessments. Control variables address fund differences, portfolio company characteristics, cognitive biases, and economic context. The analysis uses multinomial time series logit to compare the likelihood of follow-on investment, investment termination, or profitable sale vs. holding the investment. With the exception of technology risk, the hypotheses are supported. I contribute to the management and entrepreneurship literatures by examining the under-researched area of risk-based decisions subsequent to an initial commitment. I also use a novel combination of prototype theory and behavioral decision theory to show how risk can be decomposed into multiple dimensions with differing effects on decision-making. Furthermore, my analysis goes beyond binary decisions to incorporate several discrete choices. Finally, this study breaks new ground by analyzing nonpublic, contemporaneous records of actual VC post-investment risk assessments and decisions.