Is a VC Partnership Greater Than the Sum of Its Partners?

• Date: 01 June 2015
• Author: MICHAEL EWENS,MATTHEW RHODES‐KROPF
• DOI: http://doi.org/10.1111/jofi.12249
• Published date: 01 June 2015

THE JOURNAL OF FINANCE •VOL. LXX, NO. 3 •JUNE 2015

Is a VC Partnership Greater Than the Sum

of Its Partners?

MICHAEL EWENS and MATTHEW RHODES-KROPF∗

ABSTRACT

This paper investigates whether individual venture capitalists have repeatable in-

vestment skill and the extent to which their skill is impacted by the venture capital

(VC) firm where they work. Weexamine a unique data set that tracks the performance

of individual venture capitalists’ investments over time and as they move between

firms. We find evidence of skill and exit styledifferences even among venture partners

investing at the same VC firm at the same time. Furthermore, our estimates suggest

the partners’ human capital is two to five times more important than the VC firm’s

organizational capital in explaining performance.

VENTURE CAPITAL (VC) INVESTMENTS are an important engine of innovation and

economic growth, but extremely risky from an individual investor’s point of

view. For instance, Sahlman (2010) reports that 85% of returns come from just

10% of investments, and over the period 1987 to 2012 only 12.8% of investments

achieved an initial public offering.1Furthermore, there are large differences

in fund performance between top-quartile and bottom-quartile VC funds. In

spite of the rarity of top-performing investments, however, Kaplan and Schoar

(2005) report persistence in fund performance. They show that, in contrast

to other asset classes such as mutual funds, VC firms that have a fund that

outperforms the industry are likely to outperform with their next fund.

The ability to consistently produce top-performing investments implies that

there is something unique and time-invariant about VC firms. For example,

deal flow (Sorensen (2007)) and networks (Hochberg, Ljungqvist, and Lu (2007))

can explain much of the cross section of VC fund performance. Hellmann

and Puri (2002) report that higher quality VCs have industry experience and

∗Ewens is with Carnegie Mellon University, Tepper School of Business, and Rhodes-Kropf is

with Harvard Business School. We thank Viral Acharya, Joshua Coval, Peter DeMarzo, Joan

Farre-Mensa, Thomas Hellmann, Bill Kerr, Josh Lerner,Ramana Nanda, David Robinson, Merih

Sevilir, and Morten Sorensen for fruitful discussion and comments as well as participants at the

American Finance Association meetings, Fed/NYU Conferences on Private Equity, Washington

University 9th Annual Corporate Finance Conference, Stanford/JOIM Private Equity conference,

Jackson Hole Finance Group conference, Harvard University, Stanford University GSB, Case

Western University, as well as Correlation Ventures and VentureSource for access to the data.

Both authors are advisors to and investors in Correlation Ventures. All errors are our own.

1Of the investments included in the VentureSource databases, 12.8% eventually complete an

IPO.

DOI: 10.1111/jofi.12249

1081

1082 The Journal of Finance R

Gompers, Kovner, and Lerner (2009) find that VC partner specialization can

explain cross-sectional differences in performance. There could also be firm

policies or complementarities among partners or other attributes that facili-

tate consistent top performance.

However, the extent to which the attributes important to performance are

part of the VC firm’s organizational capital versus embodied in the human

capital of the people inside the VC firm is an open question. An extreme pos-

sibility is that such attributes are embedded in the firm and the people are

substitutable. At the other extreme, a venture firm is simply a collection of

people. To illustrate, consider another human-capital-intensive environment

we all know well, namely, universities. The question we explore is similar to

asking whether an academic performs better at a top institution or whether

top institutions are just collections of top academics. The greater resources, re-

duced teaching, better students, better colleagues, etc. of top institutions could

make any researcher more productive, which would imply a large effect from

organizational capital. Alternatively, better research could come from human

capital differences, implying that researchers would perform equally well at

any school. In the context of VC firms, features such as brand, resources, rep-

utation, firm deal flow, firm network, investment processes, better colleagues,

etc. could all help a partner perform better. Alternatively, an individual might

have the reputation, network, deal flow, and ability to find, identify, or make

investments. Furthermore, just as university quality may be more important

to researchers who do particular types of research, the firm may be more im-

portant to partners involved in IPOs rather than acquisitions.

Shedding light on the sources of performance in VC firms improves our

understanding of whether a firm is more than the sum of its parts. Williamson

and Winter (1993) credit (Klein, 1988) with distinguishing physical from human

asset specificity. They note that Klein (1988, p. 220), in a response to Coase

(1988), was the first to argue that, while an “organization is embedded in the

human capital of the employees” at the firm, it is “greater than the sum of its

parts. The employees come and go but the organization maintains the memory

of past trials and the knowledge of how to best do something.” Under this

hypothesis, the venture firm holds some of the knowledge of how to make great

investments. Hart (1989, p. 1772) further argues that “the observation that the

whole of organizational capital is typically greater than the sum of its parts

is equivalent to the observation that the total output of a group of workers

typically exceeds the sum of the workers’ individual outputs, to the extent that

there are complementarities.” Complementarities with a VC firm would imply

that partners should match on quality and thus firms should contain partners

of similar ability (see Becker (1981), Kremer (1993), Burdett and Coles (1997),

and Shimer and Smith (2000) for work on complementarities and matching).

VC provides an opportunity to study this question because we can assign

individual investments to partners and follow the latter over time as they

move between firms, which allows us to analyze the relative importance of

partners and firms in performance. On the other hand, focusing on VC limits

the generalizability of our results.

Is a VC Partnership Greater Than the Sum of Its Partners? 1083

We begin by examining persistence at the individual partner-investment

level. To do so, we use the VentureSource database of VC investments in en-

trepreneurial firms founded between 1987 and 2005 (to allow time to observe

outcomes) augmented with hand-collected data. We find evidence of venture

partner skill. For example, controlling for observable firm, partner, and invest-

ment characteristics such as time, industry, dollars invested, VC experience,

investment round number, firm founding date, etc., we find that, among part-

ners who made at least three investments, a one SD increase in past IPO

rate implies an 18% higher probability of an IPO in their third investment.2

Given the rarity of IPOs, the strength of persistence at the partner-investment

level is quite high, particularly in light of recent work by Phalippou (2010),

who argues that ex ante persistence comes only from low performance and

that the Kaplan and Schoar (2005) results are exaggerated. Our evidence of

strong persistence in IPOs, even with numerous deal- and partner-level con-

trols not before possible, provides support for Kaplan and Schoar’s (2005) origi-

nal fund-level persistence results as well as recent work on fund persistence by

Hochberg, Ljungqvist, and Vissing-Jørgensen (2014) and Harris et al. (2014).

We also investigate persistence in other types of exits. On average, the same

partners who have IPOs will continue to IPO, those who achieve top exits

through a merger or acquisition will continue to do so, and those who fail will

continue to fail. Overall, it seems that partners have exit “styles” insofar as

they make investments that tend to exit in the same way.3

Next, we include the past performance of the firm’s other partners to study

the role of the firm in partner performance persistence. We find that a firm’s

past IPO success rate also correlates with a partner’s probability of achieving

an IPO on his next investment. Of course, we cannot tell if this is because

similar-quality partners join together to form a firm (as implied by assortative

matching), in which case past firm performance is just more information about

partner quality, or if better firms make it more likely that a partner will have

an IPO.

When we include a measure of VC fund fixed effects using five-year invest-

ment windows, we continue to find significant persistence. That is, even com-

paring partners in the same firm investing at the same time, we find persistence

in their relative ability to go public or achieve top merger and acquisition exits.

This finding demonstrates the strength of the persistence. Assortative match-

ing should have driven like-quality partners to join together, but on average

VC firms do not seem to be collections of similar-quality partners.

The performance persistence of the average VC partner highlights the po-

tential importance of the partner as well as the firm but does not reveal their

2In a complementary paper,Gompers et al. (2010) examine whether the entrepreneurs receiving

VC demonstrate performance persistence. They find an explanation for the source of persistence,

while we attempt to separate the importance of the firm and the person in outcomes.

3Venture partners determine which deals they want to do and then seek approval from their

other partners. Partners are not assigned deals, nor do they have to give deals they like to other

partners. Deals are typically passed between partners if the partner who finds a particular deal

does not feel he has the time or expertise to investigate or manage the deal. Thus, this style

persistence seems to be a quality of the VC partner.