certain profits, sales, or performance milestones. We now extend our previous
model to explain these facts. In this we bring a new insight to the study of
venture capital finance. Previous models have focussed on the concave-convex
shape of the return stream giving the entrepreneur desirable incentives not to
engage in excessive risk taking (Gompers 1996, Biais and Casamatta 1999) or
signal manipulation (Cornelli and Yosha 1997).25 But in these models, the
conversion of debt into equity is done to improve entrepreneurial incentives. It
is in the venture capitalist’s interest to convert when the good signal arrives
(because he wants to constrain the entrepreneur), so it is not at all obvious why
he should write an agreement constraining himself to do something which he
will anyway find optimal ex post. Automatic conversion does not add anything
to standard voluntary conversion agreements. Our model is the first which
explains automatic conversion. As will become clear in the following analysis,
the time-inconsistency feature of the Coase problem means that conversion
must be agreed ex ante, because it may no longer be optimal ex post. Our
explanation draws upon the feature of venture capital funding outlined above
- that through early stage investments, the venture capitalist learns about the
prospects for profitable entry into the industry.26
6.2.2 A Model of Convertible Debt
We use the same basic model as in section 4 above, but assume in addition
that further information about Firm 2’s profitability is revealed to all parties
after the first stage of Firm 1’s financing.27 We show that even if there is no
new information about firm 1 itself, it is optimal to make Firm 1’s contract
contingent on this information because of the need to solve the Coase problem.
Our assumptions are specified in what follows.
Project
Firm 1’s payoff structure is the same as in the basic model. Firm 2’s project
may be of two types. With probability α the project is ”good”, in which case
it yields the payoffs RH and RL with the probabilities specified in the basic
25We focus here only on those papers which study the cash flow allocation created by
convertible debt contracts, since this is also our approach. A second set of papers stresses
the allocation of control rights induced by convertible securities. An earlier literature looks
at how one might design claims to motivate venture capitalists (in particular, to ensure
adequate monitoring). But this early literature does not explain the use of convertible debt.
26It is arguable that late stage investments do not involve much learning about industry
prospects, because (a) usually by this late stage, the success of the industry is already
assured and (b) late stage investments are usually undertaken by venture capitalists without
specialized knowledge of the industry, with little monitoring. Thus the model of section 4
fits this situation without modification.
27More generally, one might suppose that new information about Firm 1 ’s profitability is
also revealed as Firm 1’s business plan proceeds, but we ignore this additional complication.
23