An analysis of compensation in the U.S. venture capital partnership 1 Helpful comments were provided

In this paper, we analyze the structure of compensation in US venture capital partnerships. We contrast a learning model that extends Gibbons and Murphy (1992) to a situation in which a venture capitalist and an investor split the expected gains from investment with a signalling alternative. Empirical evidence from 419 U.S. venture capital partnerships formed between January 1978 and December 1992 is generally consistent with the four primary predictions of the learning model. Compensation is bunched, with 81 percent of the sample funds sharing between 20 and 21 percent of the profit. The compensation of new and smaller funds shows considerably less variation than older and larger funds. Compensation of older and larger funds is significantly more sensitive to performance than compensation of newer and smaller funds. Finally, funds that focus on early-stage and high-technology investments have higher base compensation, consistent with the greater effort required to monitor these projects.