Venture capital: risk and return : optimal asset allocation in a venture capital portfolio

Abstract

This thesis examines the impact of including higher moments than the mean and variance when optimizing an investment portfolio. As prior research on venture capital portfolio strategy has focused on diversification across industries and the optimal number of investments, this thesis adds insight to portfolio prioritization by focusing on the effect of portfolio diversification across different risk levels. More specifically, this study uses Monte Carlo to simulate returns from different risk levels and then determines how a “higher moments” optimal allocation change if the returns come from a non-normal as opposed to a normal distribution with everything else being equal. Although this study may provide insight for asset allocation in general, the relevance for the venture capital setting is recognized as high because extreme outcomes are more often observed in these portfolios compared to portfolios of ordinary noted stocks. I review some of the current literature on venture capital returns and find that the individual returns seem very well explained by a lognormal probability distribution. In my analysis, I find that constructing an optimal venture capital portfolio based on the skewness and kurtosis of the distribution, in addition to the mean and variance, should indicate minimal degrees of diversification between different risk profiles. This result does not align with the fact that many venture capital practitioners use stage diversification as a risk reduction strategy. The result can probably also explain some of the differences in performance between U.S. and European venture capital funds.