Different Problem, Same Solution: Contract‐Specialization in Venture Capital

• Author: Dan Bernhardt,Ola Bengtsson
• Published date: 01 June 2014
• DOI: http://doi.org/10.1111/jems.12055
• Date: 01 June 2014

Different Problem, Same Solution:

Contract-Specialization in Venture Capital

OLA BENGTSSON∗

Knut Wicksell Centre for Financial Studies,

Lund University School of Economics and Management

Research Institute of Industrial Economics,

Lund, Sweden

ola.bengtsson@nek.lu.se

DAN BERNHARDT

Department of Economics,

University of Warwick,

Coventry CV4 7AL,

United Kingdom and Departments of Economics and Finance,

University of Illinois,

David Kinley Hall, Urbana, IL 61801

danber@illinois.edu

Real-world financial contracts vary greatly in the combinations of cash flow contingency terms

and control rights used. Extant theoretical work explains such variation by arguing that each

investor finely tailors contracts to mitigate investment-specific incentive problems. We provide

overwhelming evidence from 4,561 venture capital (VC) contracts that this tailoring is overstated:

even though there is broad variation in contracting across VCs, each individual VC tends to

specialize, recycling familiar terms. In fact, a VC typically restricts contracting choices to a small

set of alternatives: 46% of the time, a VC uses the same exact cash flow contingencies as in one of

her previous five contracts. We document specialization in both aggregated downside protection,

and in each individual cash flow contingency term. Such specialization remains economically

and statistically significant even after controlling for VC and company characteristics. We

also find that VCs learn to use new contractual solutions from other VCs in her syndication

network. Our findings challenge the traditional premise that each investor selects from the

universe of combinations of terms to match an investment’s unique contracting problem. Rather,

the cumulative evidence indicates that contract-specialization arises because investors better

understand payoff consequences of familiar terms, and are reluctant to experiment with unknown

combinations.

1. Introduction

Specialization is a wide-spread and economically important phenomenon in capital

markets. Existing research has established that investors specialize investment along

some shared characteristic, for example, industry, development stage or geographical

We are grateful to VCExperts and Joseph Bartlett for help with the contract data. This project was started

when Ola Bengtsson was at Cornell University.We thank Brian Broughman, Charlie Kahn, Larry Ribstein, and

participants at the Midwestern Law and Economics conference for helpful comments. All remaining errors

are our own.

∗After this paper was accepted for publication, Ola Bengtsson passed away on January 5, 2014. Everyone who

knew Ola will miss him very much.

C2014 Wiley Periodicals, Inc.

Journal of Economics & Management Strategy, Volume23, Number 2, Summer 2014, 396–426

Contract-Specialization in VC 397

location, because they want to be familiar with the types of companies in which they

invest. We document a previously-unexplored dimension of investor specialization:

financial contracting. Using a large and detailed sample of contracts between venture

capitalists (VCs) and entrepreneurial companies, we document how investors recycle

cash flow contingency terms from prior investments.

It is well documented that VCs use financial contracts that span a large contractual

space, using several state-contingent and interrelated terms (Kaplan and Stromberg,

2003). In our data, we find that collectively VCs employ hundreds of distinct combina-

tions of terms, each of which implies a unique payoff split with the entrepreneur.At first

impression, this enormous variety seems to suggests that the standard optimal contract-

theoretic approach may well describe real-world practice. That is, entrepreneurial firms

are highly differentiated, each with its own idiosyncratic moral hazard and adverse

selection issues to address; and VCs are sophisticated investors who stand much to

gain—compensation, reputation, and future fund-raising ability—from designing ap-

propriately tailored financial contracts. Accordingly, standard optimal contract theory

would suggest that the enormous variation in economic primitives of different compa-

nies should drive enormous variation in contracting.1

This impression is wrong. In fact, we document that a given VC selects contracts

from a sharply limited universe of combinations. That the entire universe of contract

combinations employed by VCs is so large just reflects that different VCs select from

different limited universes of contract designs. We find that each VC employs neither

boiler-plate contracts nor finely tailored contracts, but rather specializes in contract

designs with which she is familiar.

One might posit that the VC contract-specialization we find is still consistent

with standard optimal contracting approaches: It could be that each VC uses a limited

subset of contracts simply because she specializes in certain investments (e.g., indus-

tries, locations and stages of financing) for which such terms are optimal. We present

strong evidence against this explanation. Contract-specialization remains statistically

significant and economically important even after controlling for company and investor

characteristics, including 340 industry dummies that allow us to control for narrow

industry distinctions such as “Voice Recognition” versus “Voice Synthesis,” or “Legal

Info/Content” versus “Legal Products.” Moreover, we show that VCs recycle contracts

from one industry to other industries. We also show that VCs learn to use new contract

designs from syndication partners, a finding difficult to reconcile with investment-

specialization explanations.

One might alternatively posit that contract-specialization reflects that VCs differ

along some characteristic that explains how contracts are written, for example the ability

to monitor entrepreneurs. But our tests include many controlvariables, including proxies

for a VC’s expertise and experience, that capture such characteristics. Moreover, the

omitted investor characteristic explanation is difficult to reconcile with our result that

there is contract-specialization for separate cash flow contingency terms and our result

that VCs frequently use identical combinations of terms. We also present results on

learning that are hard to reconcile with the omitted investor characteristic explanation.

Collectively, the evidence instead indicates that VCs face costs from experiment-

ing with unfamiliar contract terms, much as VCs incur costs experimenting with new

investment opportunities (Sorenson, 2008), or firms do when experimenting to learn

1. Consistent with this perspective, Kaplan and Stromberg (2003) evaluate VC contracting under the

premise that “VCs are real world entities who closely approximate the investors of theory.” (p. 281)