Content uploaded by Ralph Hybels
Author content
All content in this area was uploaded by Ralph Hybels on May 01, 2023
Content may be subject to copyright.
Interorganizational Endorsements and the Performance of Entrepreneurial Ventures
Author(s): Toby E. Stuart, Ha Hoang, Ralph C. Hybels
Source:
Administrative Science Quarterly,
Vol. 44, No. 2 (Jun., 1999), pp. 315-349
Published by: Johnson Graduate School of Management, Cornell University
Stable URL: http://www.jstor.org/stable/2666998 .
Accessed: 17/05/2011 12:14
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .
http://www.jstor.org/action/showPublisher?publisherCode=cjohn. .
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Johnson Graduate School of Management, Cornell University is collaborating with JSTOR to digitize, preserve
and extend access to Administrative Science Quarterly.
http://www.jstor.org
Interorganizational
Endorsements and the
Performance
of
Entrepreneurial
Ventures
Toby E. Stuart
University of Chicago
Ha Hoang
Case Western Reserve
University
Ralph C. Hybels
Leadership in Medicine, Inc.
? 1999 by Cornell University.
0001-8392/99/4402-031 5/$1 .00.
We thank Ron Burt, Rob Gertner, Steve
Kaplan,
Jesper Sorensen, and Olav So-
rensen for many helpful suggestions. We
also thank Satomi Degami and Mark Ed-
wards of Recombinant Capital for supply-
ing us with information on private bio-
technology firms, and Joshua Lerner for
providing us with the equity index used
in our analysis and for his extensive com-
ments on this paper. Seminar participants
at the Sloan School of Management, Har-
vard Business School, Haas School of
Business, and Columbia University of-
fered many valuable suggestions for im-
proving this paper. This research was
generously supported by a grant (to Stu-
art) from the Kauffman Center for Entre-
preneurial Leadership at the Ewing
Marion Kauffman Foudation, Kansas City,
MO. Please direct all correspondences to
Toby Stuart, University of Chicago,
Graduate School of Business, 11 01 East
58th Street, Chicago, IL
60637.
This paper investigates how the interorganizational net-
works of young companies affect their ability to acquire
the resources necessary for survival and growth. We pro-
pose that, faced with great uncertainty about the quality
of young companies, third parties rely on the promi-
nence of the affiliates of those companies to make judg-
ments about their quality and that young companies "en-
dorsed" by prominent exchange partners will perform
better than otherwise comparable ventures that lack
prominent associates. Results of an empirical examina-
tion of the rate of initial public offering (IPO) and the
market capitalization at lP0 of the members of a large
sample of venture-capital-backed biotechnology firms
show that privately held biotech firms with prominent
strategic alliance partners and organizational equity in-
vestors go to lP0 faster and earn greater valuations at
lP0 than firms that lack such connections. We also em-
pirically demonstrate that much of the benefit of having
prominent affiliates stems from the transfer of status that
is an inherent byproduct of interorganizational associa-
tions.'
Mobilizing resources to build a new organization is an under-
taking laden with uncertainty and unforeseeable hazards
(Stinchcombe, 1965; Aldrich and Auster, 1986; Freeman,
1997). It is also inherently a social process, because entre-
preneurs must access financial and social capital and other
types of resources via business relationships with parties
outside of the boundaries of their organizations. Because the
quality of a new venture is always a matter of some debate,
however, the decision of external resource holders to invest
their time, capital, or other resources in a new organization
is one that must be made under considerable uncertainty
about the embryonic enterprise's life chanrces and its finan-
cial prospects. This paper investigates how interorganiza-
tional relationships, by shaping potential investors' assess-
ments of the quality of young companies, affect those firms'
ability to obtain the resources to survive.
Interorganizational exchange relationships can act as en-
dorsements that influence perceptions of the quality of
young organizations when unambiguous measures of quality
do not exist or cannot be observed. As a result, the valua-
tions of young firms are at times attributions influenced by
the characteristics of the affiliates of the companies under
scrutiny. Because strong relationships with prominent organi-
zations convey the fact that young companies have earned a
positive evaluation from experienced and influential actors,
associations with high-status organizations elevate the repu-
tations of new ventures. This paper documents how the per-
formance of young biotechnology firms is affected by such
an interorganizational certification, or endorsement process,
as it operates in the industry's strategic alliance and equity
ownership networks, as well as through the connections be-
tween new ventures and the investment banks that under-
write their securities offerings. Our empirical analyses focus
on the degree to which the prominence of the business part-
ners of young biotechnology companies affect their ability to
acquire a crucial resource: capital.
315/Administrative Science Quarterly, 44 (1999): 315-349
INTERORGANIZATIONAL ENDORSEMENTS
Many obstacles confront young companies. New organiza-
tions often lack the commitment of their employees, knowl-
edge of their environments, and working relationships with
customers and suppliers (Stinchcombe, 1965). Similarly, un-
seasoned enterprises have little production experience, and
so operate under the guidance of immature and unrefined
routines (Sorensen and Stuart, 1999). Because young com-
panies often are unable to produce outputs of consistent
quality, they face a high probability of dissolution (Hannan
and Freeman, 1984). Moreover, new organizations tend to
be small organizations. In part because they do not have the
financial and other resources to withstand a sustained period
of poor performance, the rate of disbandment among small
organizations is quite high (Aldrich
and Auster, 1986;
Levinthal, 1991). These perils have led organizational sociolo-
gists to argue that young (or small) organizations are highly
vulnerable to environmental selection, a notion succinctly
portrayed as a liability of newness (or smallness).
Because young and small companies encounter so many
potential hazards and because they have short track records
by which outsiders can evaluate their quality, there is consid-
erable uncertainty about the value of new ventures. This un-
certainty is compounded for certain types of organizations,
such as those established to pursue commercial applications
of new technologies (Aldrich and Fiol, 1993). Added to the
usual hazards of inexperience, young technology companies
often require substantial resources to fund early-stage and
speculative development projects, while revenues cannot be
expected until well into the future. Moreover, new technol-
ogy is by its very nature highly uncertain: undeveloped mar-
kets follow unforeseen turns; hyped-up technologies disap-
pear far more often than they engender promised
technological shifts; technologies obsolesce extremely rap-
idly; and unanticipated kinks derail once-promising develop-
ment projects (Tushman and Rosenkopf, 1992). For these
reasons, new technology companies are extremely risky.
The case of Microcide Pharmaceutical illustrates the chal-
lenges in evaluating young companies. Founded in 1992 to
develop novel antibiotics, Microcide is a genomics company.
When the company filed for an initial public offering (IPO) in
1996, it had made progress in the development of a gene
function-based technology platform to identify and commer-
cialize new antibiotics. This technology enables the identifi-
cation of the genes in a microorganism that cause pathoge-
nicity in the body. The company had already filed for 30
patents, although all of them were still being reviewed by
the Patent Office in 1996. On the heels of these potentially
significant achievements, Microcide sought to raise $40 mil-
lion for a 24-percent stake in the company in an IPO (Micro-
cide Pharmaceuticals, SEC form S-1, 1996). At the time of
the offering, however, Microcide had yet to commercialize a
product, and it garnered no revenues other than milestone
payments from its strategic alliance partners. In addition, the
antibiotics market was hotly contested; many of the largest
drug companies and a significant number of upstart biotech
firms were competing in this domain. Because of the in-
tense competition in this area and the scientific uncertainty
31 6/ASQ, June 1999
Endorsements
associated with Microcide's research approach, it was diffi-
cult to predict whether any of Microcide's discoveries would
lead to the development of a clinically viable therapeutic. In
fact, the only certainty at the time of the Microcide IPO was
that the company was many years away from generating a
significant revenue stream from sales of internally developed
products.
In this paper, we examine how perceptions of the value of
early-stage companies such as Microcide are affected by the
relationships that position those firms in interorganizational
exchange networks. Because the quality of young compa-
nies often cannot be observed directly, evaluators must ap-
praise the company based on observable attributes that are
thought to covary with its underlying but unknown quality.
Resource holders therefore assess value by estimating the
conditional probability
that a firm will succeed, given a set of
observable characteristics of the organization.
There are two qualitatively distinct categories of information
that influence perceptions of the probability that a young
company will succeed. First, important constituencies such
as potential investors and customers make quality judg-
ments through careful consideration of the previous accom-
plishments of the organization. In addition to having intrinsic
value, these attainments are often signals (Spence, 1974) of
a young company's abilities. For example, one of the most
important achievements of a new venture in a technology-
intensive industry may be the number of patents that the
organization has accrued. A proven ability to patent new
technologies is important not only because patents are prop-
erty rights to potentially revenue-generating inventions but
also because this track record signals the depth of the firm's
underlying technological capabilities: firms with many pat-
ents are likely to have high-quality scientific and engineering
staffs. In industries in which innovation is one of the pivotal
bases of competition, firms with many patents signal their
ability to create future advances and to capitalize on external
scientific developments that may be relevant to the firm's
commercial interests (Cohen and Levinthal, 1990; Henderson
and Cockburn, 1994; Kogut and Zander, 1996).
The second category of information evaluators can use to
estimate the worth of a young company is the attributes of
the exchange partners of a focal organization. Evaluators rou-
tinely take into account the characteristics of a young com-
pany's exchange partners when they estimate the likelihood
that a new venture will succeed. Therefore, the identity of
exchange partners becomes a primary consideration when
potential investors, customers, employees, suppliers, and
other exchange partners decide whether to commit their re-
sources to a new enterprise. Because the prior
accomplish-
ments of an entrepreneurial venture are rarely sufficient to
resolve the uncertainty about its quality, the social structure
of business relationships can significantly affect perceptions
of the quality (and hence value) of new ventures.
The general argument that the characteristics of affiliates
serve as discernible guides for resolving uncertainty about
the quality of a young or unknown entity follows directly
from the notion that actors' reputations
are constructed in
31 7/ASQ, J une 1 999
part from the identities of their associates (Blau, 1964). Net-
work theorists have long asserted that relationships implicitly
transfer status between the parties in an association (while
also serving as the channels of material and social ex-
changes). For example, a young scholar's professional pros-
pects may be greatly enhanced when he or she has the
backing of a prestigious researcher (Merton, 1973; Goode,
1978). Latour (1987) argued that professional evaluations of
scientific work are influenced by the prestige of the scien-
tist's affiliates, particularly in uncertain research areas where
there is dissension over what constitutes an important con-
tribution. In a study of the diffusion of a new drug through a
community of physicians, Burt (1987) showed that doctors'
perceptions of the therapy improved after prominent physi-
cians chose to adopt it. Generalizing this idea to a corporate
context, Podolny and Stuart (1995) demonstrated that inven-
tions in uncertain technological areas were more likely to
become widely important when they had been previously
adopted by high-status organizations. In a related line of re-
search, Baum and Oliver (1991, 1992) demonstrated that or-
ganization-to-institution ties signal conformance to institu-
tional prescriptions and thereby facilitate young organizations
in their attempts to acquire legitimacy and other resources
(see also Aldrich and Auster, 1986; Singh, House, and
Tucker, 1986; Rao, 1994).
The starting point of much of this work is the observation
that social or industrial structures can be represented as a
set of positions that are arranged hierarchically according to
the prominence of their occupants. In addition, many of
these scholars argue that when there is uncertainty about
the quality of an individual, an organization, or a product, as-
sociations with the occupants of positions of prominence
enhance the regard paid to the "connected" actor and its
endeavors. There are three possible social mechanisms that
may lead would-be investors, customers and other potential
exchange partners to take into account the characteristics of
a focal new venture's affiliates as they strive to assess its
unobserved and uncertain quality: (1) relationships have re-
ciprocal effects on the reputations of those involved, (2) the
evaluative capabilities of well-known organizations are per-
ceived to be strong, and (3) relationships with prominent or-
ganizations signal a new venture's reliability, and, thus, its
high likelihood of survival.
Reciprocal relationships. The first of the three mechanisms
assumes that relationships have reciprocal influences on the
reputations of actors. Thus, even when an association is be-
tween a prominent organization and a young enterprise, the
prominent organization's reputation may be damaged if the
new venture is of very low quality. In general, as long as ex-
change relations create the possibility of a loss of status,
those held in high regard will have a commensurately strong
incentive to avoid low-quality exchange partners. Therefore,
prominent organizations will be exclusive in their selection of
associates: to do otherwise would be to risk dissipating the
economic and social rents generated by a good reputation
(Blau, 1964; Goode, 1978; White, 1985; Podolny, 1994; Pod-
olny and Phillips, 1996). Through this dynamic, relationships
318/ASQ, June 1999
Endorsements
with prominent actors may raise third parties' estimates of
the quality of the affiliated enterprises.
Quality assessments. The second mechanism posits that
resource holders are influenced by the evaluations of promi-
nent actors because they trust the ability of prominent orga-
nizations to discern quality under conditions of uncertainty
(Stuart, 1998). If there is a perceived association between
prominence and evaluative ability, then third parties will in-
terpret a connection to a prominent organization as an en-
dorsement of the initiatives of a young venture. Because
prominent organizations are viewed as experts at the due
diligence process (at least in the domain in which they have
garnered recognition), the fact that one of them has deter-
mined that a new venture is of sufficient quality to merit
transacting with it is, in and of itself, a valuable endorse-
ment. Unlike the first mechanism, however, this process
does not depend on the assumption that the reputational
capital of prominent organizations is at stake in each of their
associations. It is not necessary to assume that an occa-
sional, low-quality exchange partner will meaningfully dam-
age an organization's reputation.
Reliability. Organizational ecologists argue that organizations
that are thought to be reliable, accountable, and trustworthy
have higher chances of survival and better performance
(Stinchcombe, 1965; Hannan and Freeman, 1984). Two as-
sumptions are needed to conclude that new ventures with
prominent partners will be perceived as reliable and account-
able. First, we must assume that gaining a partnership with
a prominent organization draws attention to a new venture.
This premise is likely to hold because the initiates of promi-
nent organizations are focal points that attract the attention
of industry analysts and the business press, as well as po-
tential employees, suppliers, and customers. Second, if we
assume that organizations in general will eschew relations
with firms that may be unreliable, then the mere fact that a
young company was previously selected as a partner in a
consequential exchange relation is, in and of itself, a signal
of reliability. Based on these two assumptions, prominent
associates augment the reputation of young companies
more than do run-of-the-mill partners because the signal of
reliability
and trustworthiness implicit in exchange relations is
most widely disseminated when a new venture's associates
are particularly
well known.
Together, these three social processes suggest that, for a
new venture, gaining a prominent affiliate serves to enhance
perceptions of its quality. When a young company obtains a
prominent associate, it has earned a form of certification by
virtue of the fact that it has withstood the due diligence pro-
cess of a selective and highly capable evaluator. In much the
same way as certain types of organizations can benefit from
the endorsements of licensing agencies (Baum and Oliver,
1991) and from winning certification contests (Rao, 1994),
associations with prominent organizations benefit young
companies in the competition to mobilize resources. Before
formalizing our theory as specific predictions, however, it is
important to note that none of the three mechanisms dis-
cussed above implies that the prominence of a new ven-
ture's affiliates
is perfectly
or even highly correlated with ac-
319/ASQ, June 1999
1
Economists have investigated the effect
of investment bank prestige on underpric-
ing, defined as the spread between the
subscription price of a security (the price
at which it is sold by the underwriter) and
the value of the security after it is traded
on the public markets for one or more
days. Underpricing is costly to issuing
firms because it implies that firms sell
their shares at a low price relative to mar-
ket value and so garner smaller proceeds
from an IPO. Economists argue that low-
risk firms attempt to signal their risk posi-
tion to the market by using high-prestige
underwriters. Since investment banks
must be concerned about their ability to
sponsor new issues in the future, their
prestige is a credible signal of the issuing
firm's quality because they have a strong
incentive to preserve their status levels
by avoiding low-quality securities.
tual differences in new venture quality. Unlike signaling
models of reputation, our argument is only that perceptions
of value are shaped by patterns of affiliations. We have been
agnostic with respect to the tightness of the coupling be-
tween the signal of quality conveyed by the prominence of a
new venture's exchange partners and actual differences in
firm quality, although being perceived as high quality is, in
and of itself, an extremely valuable resource that can be par-
layed into myriad advantages. As a result, evolving differ-
ences in firm quality may closely mirror
differences in the
prominence of exchange partners. Therefore, if it is ob-
served, a high correlation between partner prominence and a
focal firm's quality may reflect the causal influences of the
former on the latter (Podolny, 1993).
The Effects of Endorsements on Performance
We investigate the effect on performance of endorsements
through three types of intercorporate affiliations. A young
technology company's performance can be affected by the
characteristics of its strategic alliance partners, including joint
venture partners and collaborators in research, marketing,
and product development. Our arguments above suggest
that young companies with prominent alliance partners
should garner higher attributions of quality than they would
otherwise because of the characteristics of their alliance
partners, which will be reflected in performance:
Hypothesis 1: The greater
the prominence
of the strategic alliance
partners of a young company,
the better the performance of the
new venture.
The composition of ownership of a young company may also
have a significant effect on its performance. In particular,
the
characteristics of the organizations that acquire equity stakes
in young companies are likely to affect performance through
their impact on the reputations of entrepreneurial ventures.
We separate the affiliative effect of equity investors from
that of nonequity alliance partners because evaluators may
perceive an equity tie as a stronger commitment from the
partner than the commitment implicit in a strategic alliance.
An equity investor signals to a broader community that an-
other organization is impressed enough with a young com-
pany to put up a stake in it. Equity relations are akin to
"strong" ties (Granovetter, 1973) and may impart an addi-
tional level of confidence in the quality of young companies.
Thus:
Hypothesis 2: The greater
the prominence
of the organizations
that
have acquired ownership
stakes in a young company,
the better the
performance
of the new venture.
A final relationship-based determinant of new venture perfor-
mance is the prestige level of one type of service firm that
works with young companies-the investment banks that
manage initial securities offerings for private companies.
Economists have argued that professional service firms can
certify the quality of young companies (e.g., Beatty and Rit-
ter, 1986; Carter and Manaster, 1990; Megginson and
Weiss, 1991). In particular, much of the work in this area has
demonstrated that firms that use prestigious investment
banks are able to garner higher share prices when they first
issue securities.1 According to this work, because presti-
320/ASQ, June 1999
2
We acknowledge an anonymous re-
viewer for encouraging us to include in-
vestment bank prestige in our analyses
of relationship-based determinants of
new venture performance, although two
factors may confound the prediction in
hypothesis 3. First, Bygrave and Tim-
mons (1992: chap. 8) reported that new
companies backed by prominent venture
capitalists were more likely to use lead-
ing investment banks when they con-
ducted an IPO, due to long-standing rela-
tions between leading venture capital
firms and prestigious investment banks.
Unfortunately, we do not know the iden-
tities of the venture capital firms that
backed the companies in our sample and,
so, cannot explicitly control for the pres-
tige of venture capital firms. Second,
leading investment banks may choose
not to handle small public offerings, and
there may be a positive relationship be-
tween venture size and performance (see
below). Both of these considerations
open the main effect on investment bank
prestige to alternative interpretation.
Endorsements
gious investment banks jeopardize their reputation if they
underwrite low-quality securities, high-status banks will avoid
leading syndicates to place the shares of low-quality firms.
Therefore, having a prestigious investment bank is an en-
dorsement of a new venture's quality. In much the same
way that well-regarded equity and alliance partners will instill
confidence in the quality of a new venture, high-status in-
vestment banks reduce investors' uncertainty about the
value of a new venture, leading to our third prediction:
Hypothesis 3: The greater the prominence of the investment bank
of a young company, the better the performance
of the new ven-
ture.2
The Contingency of Endorsements
The arguments we have developed up to this point suggest
that young companies with prominent business partners per-
form better than firms that lack such partners. The mecha-
nism that we have emphasized as creating the hypothesized
performance effects is the implicit transfer of status across
interorganizational exchange relations (such as intercorporate
equity and alliance ties), which builds confidence about the
quality of a new venture among potential customers, suppli-
ers, employees, collaborators, and investors. Through this
process, young companies with prominent exchange part-
ners gain an advantage in the competition for resources. But
there is an alternative process that could create the pre-
dicted performance effects. It is well known that relation-
ships serve as channels of access to resources of various
kinds (Lin, Ensel, and Vaughn, 1981; Marsden, 1983; Burt,
1992). Under two conditions, the fact that exchange relation-
ships create access to resources might, in and of itself, pro-
duce support for the predictions relating exchange partners'
prominence to a focal firm's performance. In particular, if
prominent organizations have greater or higher-quality re-
sources than other firms and if their exchange partners enjoy
access to some of those resources, then our claim that ties
to prominent actors are a competitive advantage for young
companies could be upheld, even if the underlying social
mechanism is not endorsement or certification. Our belief,
however, is that reputation and resource-access effects
work in tandem to create advantages for young companies
with prominent exchange partners. Thus, our objective is
neither to challenge the idea that durable exchange relations
function as channels for resource flows nor to dispute the
presumed correlation between the prominence of an actor
and the amount of resources in its possession. Rather, we
wish to examine whether interorganizational exchange rela-
tions, in addition to conveying corporeal, financial, and
knowledge-based resources, also affect reputations, which
we should be able to demonstrate by testing for contingen-
cies in the endorsement process. In particular,
our empirical
strategy for determining whether a reputation transfer pro-
cess does in fact occur is to investigate whether the magni-
tude of the effect of an exchange partner's prominence is
contingent on the amount of uncertainty about the quality of
a new venture.
The theoretical work from which we derived the first three
hypotheses suggests that the benefits of interorganizational
321/ASQ, June 1999
endorsements vary according to the level of uncertainty
about the quality of a focal firm. The reason for this is that
evaluators will rely on once-removed signals, like the promi-
nence of an organization's exchange partners, when they do
not have sufficient information about the firm to reach an
independent conclusion about its future prospects. As a re-
sult, the prominence of a focal organization's exchange part-
ners should have a particularly strong effect on assessments
of its value when there is considerable uncertainty about the
young company's quality, a point emphasized in both the
signaling literature (Spence, 1974) and in sociological work
on status (Podolny, 1994). When evaluators are confident of
their ability to assess a firm's quality based on its record of
prior achievements, there is little need to infer its quality on
the basis of the identity of its exchange partners and, hence,
a minimal impact of partners' prominence on a focal firm's
performance. In short, endorsements are contingent: the
less that is known about a focal organization, the greater the
influence of the identities of its affiliates on appraisals of its
value. Therefore:
Hypothesis 4: The greater the uncertainty
about the quality
of the
company,
the larger
the impact
of the prominence
of the firm's
ex-
change partners
on its performance.
The Biotechnology Industry
We tested the four hypotheses on a sample of young, ven-
ture-capital-backed biotechnology firms. Two radical innova-
tions-recombinant DNA (rDNA)
and hybridoma (cell fu-
sion)-launched contemporary biotechnology. The
recombination of DNA first occurred in 1973, when Herbert
Boyer and Stanley Cohen successfully introduced genetic
material from one cell into the DNA structure of another.
Two years later, the groundwork for hybridoma technology
was laid by Cesar Milstein and Georges Kohler, who suc-
cessfully employed cell fusion to create monoclonal antibod-
ies. Hybridoma technology involves fusing two different cells
to create a hybrid cell capable of producing highly purified
proteins, called monoclonal antibodies, which serve as the
body's defense against disease-causing bacteria, viruses,
and cancer cells (see Kenney, 1986 for an overview of the
core biotechnologies).
At the scientific advent of contemporary biotechnology, es-
tablished chemical, pharmaceutical, and agriculture firms
were poorly positioned to enter the new field. Biotechnology
required skills in molecular biology and biochemistry, which
were quite distinct from those demanded by the chemistry-
based technologies for which they were expected to substi-
tute. Hence, biotechnology represented a competence-
destroying development (Tushman and Anderson, 1986): it
relied on a novel set of techniques, which existing chemical
and pharmaceutical firms found difficult to acquire. As a re-
sult, the commercialization of biotechnology was shepherded
by start-up, dedicated biotechnology firms.
While 1973 brought the watershed event in the scientific
birth of contemporary biotechnology, 1980 marked the first
true spurt of entrepreneurial activity in the young field. A
considerable increase in the number of foundings of new
biotechnology firms in 1981 can be attributed to a number of
322/ASQ, J une 1 999
Endorsements
pivotal events in 1980. First, there was a clarification of pat-
ent law as it pertained to living organisms: the 1980 U.S.
Supreme Court decision in Diamond v. Chakrabarty estab-
lished the patentability of a new life form, which insured that
biotechnology innovators would be able to appropriate the
returns from patentable discoveries. Also, the passage of the
Patent and Trademark Amendment Act of 1980 enabled uni-
versities to apply for patents on discoveries that grew out of
federally funded research programs. Finally, Genentech's as-
tonishingly successful initial public offering in 1980 set a rec-
ord for the fastest increase in stock price for an IPO, from
$35 at offering to $89 in only 20 minutes.
Given high scientific entry barriers as well as the usual
gamut of inertial forces that constrain the innovative initia-
tives of established firms, existing pharmaceutical and
chemical companies were slow to enter biotechnology (Ca-
hill, Caligaris, and Williams, 1992; Burrill and Lee, 1993). Di-
versified corporations did recognize the commercial potential
of biotechnology, however, and chose to pursue it primarily
through strategic alliances with start-up firms (Barley, Free-
man, and Hybels, 1992; Burrill
and Lee, 1993). In addition,
following the 1980 Amendment to the Patent and Trademark
Act, biotechnology firms entered into a series of alliances
with universities to license and co-develop technology.
Moreover, because innovative capabilities were widely dis-
persed across the players in biotechnology, firms entered
into an extensive set of alliances to gain access to different
technologies and capabilities (Powell, Koput, and Smith-
Doerr, 1996). As a result, there have been many thousands
of vertical and horizontal alliances involving dedicated bio-
technology companies, universities, and pharmaceutical,
chemical, and food and agricultural
companies. The upshot
of this development is that biotechnology firms became im-
mersed in networks of interorganizational equity and alliance
ties; for this reason, the industry offers an ideal setting for
investigating how the structural positions of young compa-
nies affect early life course performance.
METHOD
Sample and data. Our data describe 301 dedicated biotech-
nology firms specializing in the development of human diag-
nostics and therapeutics; excluded, therefore, are firms spe-
cializing in agricultural
biotechnology, the production of
biotechnology-related equipment, and so on. While private,
all 301 companies in our sample were funded by venture
capital firms.
For young venture-backed companies, two crucial perfor-
mance variables are the speed at which the company
reaches an initial sale of securities on the public equity mar-
kets (the time from incorporation to initial public offering)
and the valuation of the firm when it goes public. An initial
public offering (IPO) is the event that transforms a privately
held entrepreneurial venture into a publicly owned company.
As a rule, venture capitalists wish to take public the compa-
nies in which they invest as soon as they anticipate favor-
able valuations (Freeman, 1997). From the new venture's
perspective, selling equity to the public often generates
much-needed capital as well as the opportunity for equity
323/ASQ, J une 1 999
holders to exchange stock for cash, so they too routinely
wish to undertake an IPO quickly and always hope for favor-
able valuations. Therefore, the firm-level performance out-
comes that we examine here are the rate at which new ven-
tures undertake an IPO and the market capitalization of the
firms that experience IPOs.
Assessing the first of our performance measures-the rate
at which new ventures undertake an IPO-required informa-
tion on early life events for the firms in our sample. Securi-
ties and Exchange Commission (SEC) filings contain informa-
tion on the early-round financial history of the firms that
undergo IPOs, necessary for our analyses, but limiting our
analyses to firms that experienced an IPO would introduce
severe survivor bias into the parameter estimates. Therefore,
we obtained data from the biotechnology consulting firm Re-
combinant Capital, which we supplemented with information
from Securities Data Corporation's New Ventures database.
These sources were critical for our project because they
contained information on venture-backed companies even if
they did not have an IPO. We relied on these databases to
acquire the early financial histories for firms that were ac-
quired or that failed prior
to an IPO or that remained private
at the end of the time series.
The time period of our analysis is from January 1, 1978
through December 31, 1991. We concluded the analysis in
1991 because we were unable to gather some of the data
that we required for the study beyond that time. A relatively
small number of biotechnology companies were founded
before 1978, so we begin the analysis very shortly after the
emergence of the industry.
In addition to the sources of information on venture capital
funding, we drew on two other databases. First, we utilized
an extremely comprehensive database on strategic alliances
in biotechnology. The primary sources for the alliance data
was the Bioscan directory (published by Oryx Press), but we
also consulted Genetic Engineering and Biotechnology Firms
Worldwide Directory 1985, Genetic Engineering and Biotech-
nology Yearbook 1985, The Biotechnology Directory: 1986
(Coombs, 1986), World Biotech Company Directory: 1985-
1986 (published by Bioengineering News), Sixth Annual GEN
Guide to Biotechnology Companies (1987), Biotechnology
Guide: U.S.A. (Dibner, 1988, 1991), The Biotechnology Direc-
tory: 1989 (Coombs and Alston, 1989), and 1993 GEN Guide
to Biotechnology Companies. Finally,
we incorporated infor-
mation from an unpublished database provided by the North
Carolina Biotechnology Center. All alliances involving at least
one biotechnology firm (including firms outside of our
sample) were coded from the sources cited above. The ma-
jority of the partnerships in our database are joint ventures,
research and development alliances, and license agree-
ments. One of the authors verified the accuracy of the alli-
ance data through interviews at 10 biotechnology firms in
California and Oregon.
Finally,
we also collected detailed information on more than
30,000 U.S. biotechnology patents with application dates be-
tween 1975 and 1991. The source for these data was the
Micropatent Biotechnology Patent Abstract CD (1994), which
324/ASQ, J une 1999
Endorsements
contains all biotechnology patents granted in the U.S. We
use the application date on each patent, which marks the
time when a patent application first arrived at the U.S. pat-
ent office. For each biotechnology patent, we coded the pat-
ent number, its organizational assignee, the application date,
and the numbers of all patents that were cited by it. We
used the patent data both to control for the intellectual prop-
erty positions of the firms in the sample and to assess the
technological capabilities of their strategic partners and eq-
uity investors.
Using the sources cited above as well as the Directory of
Corporate Affiliations and Lexis/Nexis, we also constructed
detailed, time-varying ownership trees for all of the organiza-
tions that appeared in our data. While most of the dedicated
biotech firms did not have complex corporate structures,
many of their strategic alliance partners did. For the empiri-
cal analysis, we have aggregated all of the alliances and the
patents of subsidiaries or divisions up to the level of the cor-
porate parent, and we have updated all variables as owner-
ship relationships change (e.g., when Smith Kline acquired
Beecham, we assign the alliances and the patents of Smith
Kline and Beecham to the combined entity).
Modeling time-to-IPO. Our two measures of new venture
performance were the rate of going public and the market
capitalization of the firms that undertake IPOs. From the
date of founding onward, the venture-backed biotechnology
firms in our sample are "at risk" of experiencing an IPO. We
have estimated the instantaneous rate of going public (the
hazard), denoted r(t):
r(t)i
= lim [qi(t, t + At)/At],
At->O
where q, is the probability
that a firm goes public between
two discrete time points. To estimate the effects of indepen-
dent variables on the rate of going public, we used a piece-
wise exponential model because it is extremely flexible with
respect to the form of duration (in this case, age) depen-
dence. The model assumes that the baseline transition rate,
represented by the coefficients on a vector of v time peri-
ods, is constant within each age period but can vary in an
unconstrained manner across periods. Hence, the piecewise
exponential model does not require a strong assumption
about the functional form of age dependence. The model
that we estimate can be written as:
r(t)i = r(t)* exp(o'Mt + 3'Qit + WI3 Zijt + ' (1D)
where r(t)* represents the baseline hazard rate, Mt is a ma-
trix of time-varying (but organization invariant)
measures of
environmental conditions, Qt is a matrix of time-changing
variables representing characteristics of the organizations in
the sample, Zijt are time-changing variables representing
counts of relations between young biotechnology firms and
their business associates, Dit are time-varying measures of
the prominence scores (in a number of different domains) of
the exchange partners of the firms in the sample, and ax, 1,
8, and y are parameters to be estimated.
The variables in the hazard models change over time. To ac-
commodate time-varying covariates, we divided the time pe-
325/ASQ, J une 1999
riod during which each organization is observed into quarter-
year spells (Tuma and Hannan, 1984). All time-changing
covariates were then updated every calendar quarter (i.e., at
three-month intervals). Each of the quarterly spells was
treated as censored on the right, except for the spells that
terminate in an IPO. A very small number of firms experi-
enced a terminal event other than IPO or censoring, which
occurred when the firm was still privately owned at the con-
clusion of our observation window: a handful of firms either
disbanded or were acquired while still private. Because only
a few of the sample firms experienced terminal events other
than an IPO, we chose not to estimate a competing risks
hazard model but, rather, just to censor the final spell for all
firms that were acquired or disbanded.
All 301 firms in our sample were founded between January
1, 1978 and December 31, 1991. Among these, 121 experi-
enced IPOs prior
to the end of calendar year 1991; the re-
maining 180 firms were right censored on December 31,
1991, or on the date they were acquired or disbanded.
Estimating market valuation regressions. We observed a
firm's market capitalization conditional on it undertaking an
IPO (market value and all other nominal figures are con-
verted into real dollars). The market value of a firm at IPO
was defined as:
V* = (puqt - puqj),
where p, is the 1PO
subscription price, qt is the total number
of shares outstanding, and qj is the number of shares of-
fered in the IPO. In other words, we subtracted from the
firm's total market capitalization the dollar amount raised in
the IPO. V* is a measure of the market's assessment of the
value of a biotechnology company at IPO. We used the IPO
subscription price to compute V*; in unreported models, we
observed results similar to the ones we report here when
defining V* on the basis of the first day closing price of a
new issue. Because biotechnology firms at IPO have mini-
mal physical assets, the dollar value of the firm as reflected
in V* is contingent on outsiders' estimates of the value of
the firm's intellectual property, its scientific capabilities, the
quality of its management, its portfolio of strategic alliances,
and its reputation. Because 121 of the firms in the sample
went to IPO during our analysis period, we have public mar-
ket capitalization data on 121 of the 301 companies in the
sample.
We used OLS to estimate the valuation models. The one
complication in these models is that we had to limit the
analysis of market capitalization to the firms in the sample
that underwent IPOs because the dependent variable was
observed for such cases only, but the fact that market capi-
talization was not observed for some firms may result in
sample selection bias in the valuation models. Thus, estimat-
ing these models presented the standard censored sample
problem, which we addressed by using Lee's (1983) general-
ization of Heckman's (1976) two-stage estimator. Similar in
principle to Heckman's use of the reciprocal of the Mills ratio
as an additional regressor to correct for sample selection,
Lee's method allowed us to compute a selection variable by
transforming the cumulative waiting-time distribution (de-
326/ASQ, J une 1999
3
Lee (1983) showed that, conditional on y
being observed (selection is designated
as I = 1), the following censored regres-
sion model can be estimated:
yj = xjP - (lp)+[,P-1F(zjy)y/F(zjy)
+ q,
where E(lIl = 1, x, z) = 0, y designates
market values conditional on IPO, x is a
matrix of independent variables, + is the
standard normal density function, @-1 is
the inverse normal, F(z-y)
designates the
cumulative waiting time distribution (i.e.,
the complement of the survivor function).
Estimates of F(z-y)
are derived from the
hazard rate analysis and up is the (esti-
mated) coefficient on the selection vari-
able (denoted as X in table 4).
Endorsements
rived from the hazard rate models), which was then entered
as a covariate in the OLS models.3
Measures
Independent variables. In operationalizing the prominence
of affiliates, our analysis takes into account the fact that
there is likely to be uncertainty about the quality of a young
technology company in more than one area of its operations.
As a result, outsiders' estimates of the worth of the com-
pany may be affected by endorsements in different domains
of the firm's activities. This has two implications. First, inter-
organizational relationships will be perceived as endorse-
ments only with respect to the operations of a young com-
pany in the area in which the prominent affiliate has
experience or expertise (Goode, 1978). For example, having
its financial statements prepared by a reputable accounting
firm is likely to instill confidence in the accuracy of a new
venture's financial reports, but it is unlikely to affect outsid-
ers' perceptions of the technological capability of the com-
pany. Second, the value of associations with prominent orga-
nizations is likely to be domain-dependent: endorsements of
some components of an organization's operations will be
more meaningful than will others for the perceived value of
the firm. For both of these reasons, it is important to investi-
gate how connections to organizations that are prominent in
different domains affect the performance of young compa-
nies.
Before describing the measures of prominence, we stress
that strategic alliances are very important in the biotechnol-
ogy industry. Young biotechnology companies are well
aware of the significance of intercorporate collaborations as
a means of garnering additional resources, bolstering their
strategic positions, and enhancing their reputations. Perhaps
the most telling indication of this is that biotech firms have
entered many thousands of strategic alliances since the
emergence of the industry, and the identities of partners
and descriptions of alliances figure prominently in biotechnol-
ogy companies' securities registration statements. We re-
viewed a large number of S-1 statements, the SEC docu-
ments that register securities for an initial public stock
offering. In virtually
every case in which a focal biotech firm
had one or more strategic partners, that firm's S-1 statement
profiled its strategic partnerships by the third paragraph of
the document. This signals the importance of alliances to
potential investors.
We measured the prominence of the alliance partners and
equity investors in two areas, which we label the "techno-
logical" and the "commercial." We defined technological
prominence as a measure of the success of an alliance or
equity partner as a biotechnology innovator: technologically
prominent organizations are those that have developed many
influential biotechnology innovations. When a young com-
pany establishes a relationship with a technologically accom-
plished organization, it benefits from an endorsement of its
own technological capability. The effect of this endorsement
is to assuage outside evaluators' concerns about the an-
swers to questions such as, How significant are the prior
technical achievements of a young company? And, What is
327/ASQ, June 1999
the probability that the technological area in which the new
venture participates will become significant? Figuring that
technologically prominent organizations are competent and
selective judges of the technological potential of new ven-
tures in their areas of expertise, outside evaluators upgrade
their opinion of the capabilities of the partners of technologi-
cally prominent organizations.
We defined commercial prominence as a measure of an ex-
change partner's experience in evaluating young biotechnol-
ogy companies: commercially prominent organizations are
those that have a long track record of evaluating and work-
ing with young firms in the industry. When a new venture
forms a relationship with a commercially experienced organi-
zation, it benefits from an endorsement that minimizes out-
side evaluators' concerns about the answers to questions
such as, Is the young company organized effectively to re-
search and develop new biotechnologies? Does the firm
have the connections needed to develop and market a prod-
uct? If the firm is able to develop a product, will it effectively
produce it? And, Is the company run by a strong manage-
ment team? To potential investors, employees, customers,
suppliers, and collaborators, an alliance or equity relationship
with an experienced evaluator of young biotechnology com-
panies conveys an important vote of confidence in the young
venture's organizational integrity and managerial ability.
We measured the commercial and technological prominence
of partners in terms of the pattern of relations that embed
the alliance and equity partners of the biotech firms in our
sample into two different networks: a strategic alliance net-
work and a patent citation network. Following the conven-
tion in the network literature, commercial (or technological)
prominence was defined as the degree to which an organi-
zation's position makes it visible to other actors in the alli-
ance (patent citation) network (Knoke and Burt, 1983).
Among many potential measures of prominence, the most
parsimonious is an actor's degree score, defined as a count
of the number of relations in a network involving that actor.
For the statistical analysis, we operationalized commercial
prominence as an organization's normalized degree score in
the network of strategic alliances, defined broadly to include
all biotechnology-related partnerships. Technological promi-
nence is a strategic partner's degree score in a network con-
figured from the patent citations between the patent portfo-
lios of all biotechnology innovators.
To construct the measure of commercial prominence, we
organized the biotech alliance data as symmetric matrices of
organization-to-organization ties, which were updated four
times a year (quarterly)
to reflect new alliances. The ele-
ments of each matrix at period t are the number of alliances
established prior
to the quarter t between the organization
on the row (i) and on the column (j).
The alliance matrices
include all organizations that had previously formed at least
one alliance with a biotechnology firm. We normalized the
elements in each matrix by the total number of alliances
formed in the industry prior
to t so that commercial promi-
nence scores are proportions. Finally,
we summed over each
row in the alliance matrix to produce a vector of commercial
prominence scores. The interpretation of the commercial
328/ASQ, J une 1999
Endorsements
prominence score for each organization j is the proportion of
all biotechnology alliances in which it had participated.
Substantively, the content of the ties that create high degree
scores in the biotech alliance network should be an appropri-
ate measure of prominence to study the endorsement pro-
cess. First, organizations that are central in the alliance net-
work may be the actors with the strongest incentive to
protect their reputations by exercising care in their selection
of partners. Because they are repeat players, organizations
with high degree scores will be concerned about the reputa-
tion costs that they will incur if they sponsor salient failures.
Perhaps of greater importance, alliance network degree
score is an appropriate measure of prominence because it
equates prominence with the organizations that have the
most experience in evaluating young biotech companies. Be-
cause organizations with high degree scores in the alliance
network have the most familiarity
with the due diligence pro-
cess, they will be perceived as very capable judges of com-
mercial potential. For both reasons, exchange relationships
with central organizations in the biotech alliance network will
be perceived by outside parties as a significant endorsement
of a young company.
We defined a technologically prominent organization as one
with a highly cited portfolio of biotechnology patents. Sub-
stantively, patent citations denote technological building rela-
tionships: to acquire a patent, an inventor must submit an
application that includes a list of citations to all previously
issued patents that made technological claims similar to
those for which the current application seeks intellectual
property protection. In other words, patent applicants must
acknowledge the existing, patented inventions that are near-
est in technical content to the inventions for which they are
pursuing patent protection. Previous research has demon-
strated that highly cited patents have been shown to be the
most important inventions (Trajtenberg, 1990; Albert et al.,
1991).
To compute technological prominence scores, we created an
organization-to-organization patent citation matrix for each
calendar quarter t. The elements in these matrices are
counts of the number of citations from the biotechnology
patents of the organizations on the columns (j)
to the pat-
ents of the organizations on the rows (i). We then normal-
ized the elements of the matrix by the total number of pat-
ent citations made during period t, again so that
technological prominence scores are proportions that can be
compared across time. Finally,
we summed across each
row, yielding a vector of technological prominence scores.
Like the alliance centrality variable, the interpretation of the
technological prominence score for each organization is the
proportion of all biotechnology patent citations flowing to its
patent portfolio. These scores are highest for organizations
that have developed the most influential stocks of biotech-
nology inventions.
After computing the prominence measures, we merged into
the records of each of the biotechnology firms in the sample
the technological and commercial prominence scores of all
of their strategic alliance partners and their (non-venture-capi-
329/ASQ, June 1999
4
In unreported models, we computed
prominence scores by averaging the
prominence scores of all exchange part-
ners and by selecting the maximum
prominence score of any of a focal firm's
exchange partners. The correct specifica-
tion of the affiliative effects rests on
one's theory about how the underlying
social process operates. For example,
using the mean prominence of partners
would be justified if the addition of low-
prestige partners can actually downgrade
perceptions of a focal firm's quality. Not
surprisingly, however, these measures
were all highly correlated, and the results
were generally consistent across the
three measures.
tal) organizational equity investors. In the statistical models
reported below, we chose to define prominence scores for
each biotech venture during each time period by summing
over the prominence scores of that venture's affiliates
(whether we used commercial or technological prominence
scores and whether these scores pertained to a focal com-
pany's equity investors or alliance partners depends on the
model). For example, we defined the commercial promi-
nence of a focal biotech venture as:
di= zijt
djt,
where dit denotes the commercial prominence of the part-
ners of a focal biotech venture i at time t, zijt is coded 1 if
there was an alliance between i and organization j at time
<t, djt is the commercial prominence of organization j, and
the sum is performed over all organizations that participated
in biotechnology.4 In the event history data, we updated
these variables each quarter to reflect the addition of new
affiliates. In the valuation models, each variable assumes its
value at the time that the firm went public.
Prominence of investment banks. In the market valuation
models, we included a measure of the prestige of the in-
vestment banks that served as lead managers for the IPOs
of the firms in the sample, to test the third hypothesis. We
coded from IPO prospectuses the identity of the investment
banks that acted as lead manager in each IPO. Our measure
of investment bank status is the prestige score for the bank
computed by Carter and Manaster (1990). Carter and Manas-
ter constructed prestige scores from the rankings of invest-
ment banks in "tombstone" announcements for IPOs.
Based on investment banks' positions in the tombstone an-
nouncements for new issues, Carter and Manaster con-
structed a scale of investment bank prestige ranging from 0
(lowest prestige) to 9 (highest prestige).
Control variables. The models include two types of control
variables: firm-level controls and measures of time-changing
environmental conditions. For the firm-level controls, we in-
cluded a measure of each firm's intellectual property posi-
tion, a time-updated count of the number of ultimately suc-
cessful biotechnology patent applications assigned to it,
because patents have intrinsic value and play a signaling
role. Because the cost of obtaining a patent is significant and
is inversely correlated with a firm's technological capabilities,
patents conform neatly with Spence's (1974) definition of a
signal. While U.S. patents only become part of the public
record after they are issued by the Patent Office (pending
and rejected patent applications are not released), young bio-
technology companies typically make every effort to publi-
cize their technological accomplishments. These companies
often issue press releases when they receive patents, and
lengthy discussions of their intellectual property positions are
prominently placed in their IPO prospectuses, including de-
scriptions of pending patents. Our expectation is therefore
that firms with more pending patents will go to IPO more
quickly and will be worth more when they do.
Next, for each firm, we entered a time-changing indicator
variable that turns on when a biotechnology company files
330/ASQ, June 1999
5
Twenty-two percent of the firms in the
sample never appeared in the CorpTech
directory. For these firms, we found de-
scriptions of their business focus either
on the Biospace Web Site (www.
biospace.com), in company securities fil-
ings, or in the Bioscan directories. Based
on these descriptions, we coded the in-
dustry segments in which each organiza-
tion participated.
Endorsements
its first investigational new drug application (IND). Before
marketing a new therapeutic, biomedical firms must submit
their products to a rigorous and controlled experimental pro-
cess designed to test the efficacy and safety of the drug as
a precursor to gaining Food and Drug Administration ap-
proval to sell the drug. The human trials testing process is
initiated with the submission of an IND. Hence, the IND
dummy variable is a measure of the progress of the firms in
the sample toward developing (potentially) marketable prod-
ucts.
The firms in our sample all competed in the same broad
market niche-they all focused on the human segment of
the industry, including the application of biotechnology to the
production of human diagnostics or therapeutics. Still, within
this area there are differences in strategic foci. To account
for differences in the strategies of the firms in our sample,
we included in the models a set of indicator variables repre-
senting participation in various segments of biotechnology.
To construct these variables, we looked up each firm in the
sample in the Corporate Technology Directory (CorpTech) in
the first year in which it appeared in the directory. The
CorpTech directories contain a classification system that is
used to characterize each firm according to its presence in
fine-grained technology segments. In the models, we experi-
mented with including dummy variables for as many as
twelve industry segments; for each observation, we included
twelve dummy variables denoting whether or not firm i par-
ticipated in segment k. In the models that we report, we in-
cluded the four largest technology segments, because re-
sults were not materially different from the 12-segment
findings.5
The models include three alliance-based control variables.
First is a time-varying count of the total number of strategic
alliances formed by each of the firms in the sample. Our ar-
gument suggests that interorganizational alliances frequently
serve as channels of resource access for the young compa-
nies in our sample, but because our purpose was to assess
the incremental reputational benefit of interorganizational re-
lationships over and above the benefits of access to intellec-
tual and material capital, we chose to treat the alliance count
as a control variable.
While the companies in our sample established many differ-
ent types of strategic alliances, many of these alliances did
involve significant amounts of funding for start-up firms, typi-
cally payments from a strategic partner to sponsor a re-
search program or to purchase the rights to use an innova-
tion. As a result, alliances can be a direct source of research
funding for biotech firms. More often than not, funded re-
search alliances were structured to include a lump sum, up-
front payment to the biotechnology company, to be followed
with additional payments if technological milestones were
achieved within contractually specified time periods. To re-
flect the fact that strategic alliances are direct funding
sources, we coded the dollar value of each alliance that in-
volved a payment to one of the biotechnology firms in our
sample (alliance partner funding). Although we do not know
if the firms in the sample reached all of the milestones
specified in the alliance
contracts, we effectively included
for
331/ASQ, June 1999
each firm the upper bound on the dollar amount that it could
realize from its strategic partnerships. In the event models,
we cumulated these amounts over time as biotechnology
firms entered new strategic coalitions. Hence, the models
control for the level of research funding supplied by each
new venture's alliance partners.
Corresponding to the alliance count variable, we also in-
cluded in the models a time-changing count of the number
of non-venture-capital equity investors in each of the ven-
tures in our sample. It is relatively common for established
pharmaceutical, agricultural, and chemical companies to ac-
quire equity positions in privately held biotechnology compa-
nies in the human therapeutics and diagnostics segments.
Hypothesis 2 predicted that when a new venture's non-ven-
ture-capital equity investors are prominent organizations, the
firm benefits from a strong endorsement. Paralleling the way
that we modeled the hypotheses about the prominence of
alliance partners, we controlled for the number of (non-ven-
ture-capital) equity investors to assess the incremental effect
of the prominence of the investors.
We also used two measures of the uncertainty about the
quality of the firms in the sample, firm age and the total
amount of pre-lPO, private equity raised. We included both
of these variables as main effects and as interactions with
the prominence measures to test hypothesis 4, that the
value of endorsements are contingent on the level of uncer-
tainty about new ventures. Both of the variables have been
shown in the finance literature to be proxies for investor un-
certainty about the value of firms.
The age of a firm is indicative of the uncertainty about its
quality because very young companies have limited perfor-
mance histories on which quality can be assessed (Beatty
and Ritter, 1986). Supporting the assertion that uncertainty
declines with organizational age, venture capitalists them-
selves are reluctant to invest heavily in very young compa-
nies because of their inability
to evaluate the quality of such
ventures. This reluctance is apparent in the way that the
venture financing process has become institutionalized: ven-
ture-backed companies receive funding in a sequence of
capital infusions, known as "rounds." Each time a company
undergoes a financing round, it is subjected to a thorough
evaluation by venture capitalists. Typically, the amount of
funding in early rounds is relatively small because venture
capitalists will not make large financial commitments to
young companies about which they lack a sufficient under-
standing of quality (Gompers, 1995). As they learn more
about the company-both from a technological and manage-
rial standpoint-venture capitalists then decide to terminate,
maintain, or increase the level of funding for a project.
Hence, the fact that staged financing has become institution-
alized is, in and of itself, a confirmation of the fact that the
amount of uncertainty about a venture declines as it ages.
The second measure of the uncertainty of quality is the total
dollar amount of prior
venture funding, a cumulating sum of
the cash raised by each firm in private financing rounds. A
robust empirical finding is that there is more uncertainty
about the value of small firms (e.g., Beatty and Ritter, 1986).
332/ASQ, J une 1999
Endorsements
Because all of the firms in the sample are venture-backed
and all participate in the same industry, the level of funding
should be a good approximation of differences in firm size.
When other attributes of a young venture are held constant,
this variable can also be construed as a reflection of venture
capitalists' commitment to a company: the greater the size
of their investment, the higher their expectations for the
young venture and the more certain they are of venture qual-
ity. In the hazard models, we updated this variable quarterly
to reflect any changes. In the market valuation models, we
included the total amount of cash raised in all private financ-
ing rounds.
Environmental conditions. Finally, we controlled for time-
changing environmental conditions. The analyses included an
equity index consisting of biotechnology stocks, to control
for intertemporal differences in the receptivity of the equity
markets to new biotech issues. The index, constructed by
Joshua Lerner (1994), is re-balanced annually and consists of
equal dollar shares of thirteen publicly traded, dedicated bio-
technology firms. It has been observed that there are "hot"
and "cold" markets for IPOs, implying that the ability of a
company to go public may depend on equity market condi-
tions (Ritter, 1984). It is therefore important to control for
the receptivity of the market-ideally, for the particular type
of company under study-in the time-to-IPO and valuation
models. In the hazard rate models, the biotech equity index
was updated quarterly. In the valuation models, the index
assumes its value at the end of the month prior
to when an
IPO occurred.
We also coded and included in the models the population
density of all dedicated biotechnology firms (the density vari-
able includes non-venture-backed and agriculture biotech
firms), which we entered in the models as a monotone and
a quadratic term. We entered these variables to link the em-
pirical analysis to the density dependent model of legitima-
tion and competition, which has become a baseline in organ-
izational analysis. Generalizing from the robust empirical find-
ings that organizational founding rates first increase and then
decrease with growth in population density (Hannan and Car-
roll, 1992), we anticipated that density would have an in-
verted-U-shaped effect on the rate of IPO and on market
valuations. The density dependence model would predict
that rising legitimation will increase the rate of IPOs and the
valuations of new biotechnology issues with initial increases
in population density, while the opposing force of competi-
tion will decrease the rate of IPOs and market valuations as
density increases.
In the valuation models only, we included dummy variables
designating the four years in which biotech IPO activity flour-
ished. In the time period covered by our analyses, the major-
ity of IPOs occurred in four years: 1983, 1986, 1987, and
1991. The primary
factor that singled out these four years
seems to have been the favorable milieu for biotech equities
in the public markets-company valuations were generous in
these years relative to other time periods. In the valuation
models, we included both the year dummies and the level of
the iP0 index to control for market conditions.
333/ASQ, J une 1999
RESULTS
Table 1 reports descriptive statistics. Table 2 presents a cor-
relation matrix, and table 3 contains the results from the
time-to-IPO models.
Time-to-IPO. The baseline model (1) in table 3 includes only
the firm age segments, variables describing focal venture
quality differences and strategic focus, and the measures of
environmental conditions. All models include three age peri-
ods, which allow the baseline hazard to shift at three points.
The piecewise model with three age periods is a statistically
significant improvement over an unreported constant rate
(exponential) model. Model 1 shows that firms with more
patents go to IPO faster. Each additional patent multiplies
the rate of going public by a factor of 1.11 (e
[106 1). The "to-
tal cash raised" variable and the "investigational new drug"
dummy variable are both positive and significant. The coeffi-
cient estimate on the former variable implies that each $10
million in additional venture funding multiplied the baseline
rate by a factor of 1.16. None of the four market-segment
dummy variables achieved statistical significance.
Turning to the controls for environmental conditions, model
1 shows that there is an extremely strong effect of the bio-
tech equity index on the rate of going public (see also
Lerner, 1994). The coefficient suggests that a one-point in-
crease in the equity index almost triples the rate of going
public, demonstrating the great importance of favorable eq-
uity market conditions for the occurrence of biotechnology
IPOs. The density variables in model 1 are not statistically
Table 1
Descriptive Statistics*
Mean S.D. Mean S.D.
Variable (all) (all) (IPO) (IPO) Min. Max.
Firm characteristics
Firm age 3.655 2.451 3.543 2.019 0 12.502
Total cash raised 16.740 20.489 16.741 20.484 0 143.218
BT patents .727 2.377 1.289 3.159 0 19
Investigation new drug dummy .051 .220 .088 .284 0 1
Genetic engineering dummy .197 .391 .272 .447 0 1
Protein engineering dummy .051 .220 .088 .284 0 1
Immunology dummy .151 .357 .211 .409 0 1
Diagnostics dummy .151 .357 .219 .415 0 1
Environmental conditions
BT equity index 3.954 .605 3.484 .762 1.402 4.885
BT density 397.880 36.330 383.577 55.439 114 474
BT density-squared (/1000) 159.627 23.900 150.172 36.447 12.996 224.676
Partner controls
Alliance count 1.597 2.683 2.526 3.135 0 17
Equity investor count .440 .880 .667 1.053 0 5
Funding from partners 4.883 13.649 8.349 3.135 0 112.9
Partner prominence
Com. prominence (alliance) .013 .026 .023 .034 0 .113
Tech. prominence (alliance) .004 .008 .005 .010 0 .057
Com. prominence (equity) .003 .008 .006 .009 0 .095
Tech. prominence (equity) .002 .005 .004 .007 0 .035
Investment bank prominence 5.395 3.109 0 9
* Statistics
from
market valuation
models. Mean
(all)
and S.D. (all)
include
all
301 firms.
Mean
OIPO)
and S.D. OIPO)
include
only
the 121
firms
that had IPOs.
334/ASQ, J une 1 999
Endorsements
Table 2
Correlation Matrix*
Variable 1 2 3 4 5 6 7 8 9 10 1 1 12
1. Firm age
2. Total cash raised .38 -
3. BT patents .27 .45 -
4. Invest. new drug (=1) .36 .25 -.02 -
5. Genetic engineering (=1) .15 .06 .24 .11 -
6. Immunology (=1) -.05 .08 .13 .17 .26 -
7. Protein engineering (=1) .03 .11 .28 -.09 .09 -.08 -
8. Diagnostic (=1) .08 .04 -.03 -.08 .08 .06 .06 -
9. BT equity index .18 .14 .02 -.02 .11 .05 .15 .17 -
10. BT density .33 .18 .06 .02 -.27 -.20 -.02 .07 -.00 -
11. BT density2 (/1000) .33 .18 .05 .01 -.27 -.21 -.04 .06 -.06 .99 -
12. 1983 -.29 -.15 .02 -.02 .23 .19 .05 .04 .39 -.55 -.62 -
13. 1986 .29 .13 .17 -.03 .09 -.02 .13 .03 .21 .11 .10 -.17
14. 1987 -.1 1 -.08 -.13 -.1 1 -.03 -.12 -.12 .01 -.12 .33 .38 -.15
15. 1991 .31 .27 -.02 .14 -.12 -.06 .04 .16 .33 .28 .28 -.28
16. Alliance count .43 .40 .27 .12 .12 -.01 .16 .00 -.04 .19 .19 -.20
17. Funding from partners .14 .57 .36 .04 -.07 .12 .15 .04 .13 .13 .13 -.14
18. Equity investor count .22 .60 .31 .07 .06 .02 .05 .00 .13 .17 .17 -.12
19. Com prom., alliance .28 .30 .26 .16 .29 .14 .00 .02 -.01 -.17 -.17 -.00
20. Tech prom., alliance .35 .36 .09 .08 .22 .11 .07 .29 .13 .18 .18 -.17
21. Com prom., equity .09 .23 -.05 -.08 -.08 .06 -.11 .08 .16 .06 .05 -.05
22. Tech prom., equity .13 .25 .19 .08 .19 .14 .03 -.09 .03 -.21 -.21 .06
23. Investment bank prestige .29 .44 .21 .13 .00 .03 .12 .13 .16 .12 .12 -.1 1
Variable 13 14 15 16 17 18 19 20 21 22
13. 1986 -
14. 1987 -.15 -
15. 1991 -.29 -.26 -
16. Alliance count .22 -.07 .05 -
17. Funding from partners .10 -.10 .20 .28 -
18. Equity investor count .08 -.06 .21 .34 .51 -
19. Com prom., alliance .19 -.07 -.12 .49 .13 .23 -
20. Tech prom., alliance .01 -.11 .30 .58 .29 .32 .31 -
21. Com prom., equity .16 -.07 .10 .08 .15 .35 .21 .22 -
22. Tech prom., equity .09 .05 -.14 .31 .19 .22 .65 .03 .24 -
23. Investment bank prestige .03 .01 .29 .19 .18 .24 .31 .25 .18 .31
* Based on market
value data
set.
significant, and their signs are the reverse of those predicted
by density dependence theory. Because the industry was
still very young during the course of our analyses and be-
cause our analyses did not begin until five or so years after
the birth of the industry, we do not view the results as
strong evidence against density dependence theory in this
population.
Model 2 adds the control variables for the number of alli-
ances, number of organizational equity investors, and the
amount of resources committed to a biotech venture by its
alliance and equity partners. The alliance count variable is
positive and statistically significant: according to the esti-
mates, each new alliance led to a 12-percent increase in the
rate of going public. The number of organizational equity in-
vestors is also positive and significant. The amount of fund-
ing purveyed by alliance partners is positive, but it is not sta-
tistically significant.
The next four models add the prominence-of-affiliate vari-
ables. Model 3 includes the commercial prominence of alli-
ance partners; model 4 the technological prominence of alli-
ance partners; model 5 the commercial prominence of equity
335/ASQ, J une 1 999
Table 3
Maximum Likelihood Estimates of Time-to-lPO*
Variable Model 1 Model 2 Model 3 Model 4 Model 5
Age: <3 years -4.591 - -4.227- -4.346- -4.210- -6.280-
(2.250) (2.131) (2.241) (2.231) (2.569)
Age: 3-7 years -3.916- -3.682 -3.831 - -3.669 -5.740-
(2.270) (2.258) (2.285) (2.252) (2.591)
Age: >7 years -4.460 -4.349- -4.449- -4.344- -6.333-
(2.287) (2.271) (2.300) (2.240) (2.588)
Total cash raised .015- .003 .005 .003 .002
(.006) (.005) (.005) (.005) (.005)
BT patents .106- .078- .075- .083- .068-
(.039) (.041) (.040) (.042) (.041)
Invest. new drug (=1) .574 .663- .601 .692- .555
(.354) (.354) (.361) (.364) (.359)
Genetic engineering (=1) .080 -.088 -.117 -.094 -.121
(.232) (.237) (.241) (.239) (.239)
Immunology (=1) .227 .209 .271 .299 .320
(.262) (.260) (.261) (.258) (.255)
Protein engineering (=1) .431 .420 .423 .371 .399
(.401) (.404) (.410) (.410) (.408)
Diagnostic (=1) .092 .151 .187 .138 .212
(.243) (.242) (.242) (.246) (.241)
BT equity index 1.071-* 1.072- 1.060- 1.049- 1.059-
(.113) (.116) (.118) (.118) (.118)
BT density -.011 -.012 -.012 -.012 -.007
(.015) (.015) (.015) (.014) (.016)
BT density2 (/1000) .016 .016 .016 .016 .016
(.023) (.023) (.023) (.023) (.023)
Alliance count .114- .066 .1070 .1000
(.032) (.051) (.037) (.033)
Funding from partners .003 .005 .003 .004
(.006) (.006) (.006) (.006)
Equity investor count .238' .235- .231- .209-
(.118) (.118) (.121) (.122)
Com. prom., alliance 6.554
(4.820)
Tech. prom., alliance 4.148
(12.059)
Com. prom., equity 25.833-
(11.150)
Tech. prom., equity
Tech. prom., alliance x Total
cash raised
Com. prom., equity x Total
cash raised
Com. prom., alliance x Age <3
Com. prom., alliance x Age
3-7
Com. prom., alliance x Age >7
Tech. prom., equity x Age <3
Tech. prom., equity x Age 3-7
Tech. prom., equity x Age >7
Log-likelihood -313.5 -304.67 -303.85 -304.61 -302.26
p < .05, one-sided tests.
* Standard
errors are in parentheses.
Total of 4739 spells and 121 events (IPOs).
336/ASQ, June 1999
Endorsements
Table 3 (continued)
Model 6 Model 7 Model 8 Model 9 Model 10
-4Q037- -5.097 -5.5555 -5.023- -4.122-
(2.280) (2.248) (2.495) (2.287) (2.240)
-3.595 -4.735- -5.185- -4.346' -3.574
(2.243) (2.508) (2.517) (2.229) (2.260)
-4.154- -5.337- -5.818- -5.044- -4.253-
(2.259) (2.511) (2.522) (2.294) (2.280)
.003 .010 .015- .003 .003
(.006) (.007) (.007) (.006) (.006)
.095- .104- .097- .081 - .094'
(.041) (.041) (.040) (.041) (.042)
.860- .660- .935- .622- .856-
(.361) (.364) (.374) (.366) (.359)
-.057 -.154 -.134 -.139 -.077
(.238) (.243) (.240) (.245) (.241)
.224 .286 .340 .309 .201
(.257) (.257) (.252) (.261) (.261)
.422 .340 .342 .382 .422
(.410) (.410) (.410) (.419) (.410)
.065 .137 .112 .174 .117
(.240) (.247) (.249) (.242) (.243)
1.056' 1.068- 1.048- 1.075- 1.060'
(.117) (.118) (.117) (.118) (.118)
-.014 -.012 -.012 -.008 -.013
(.015) (.015) (.016) (.015) (.015)
.016 .016 .016 .016 .016
(.023) (.023) (.024) (.023) (.023)
.118 .097- .107- .078 .118-
(.033) (.053) (.052) (.054) (.033)
.007 .008 .011 ' .007 .007
(.006) (.006) (.006) (.006) (.006)
.085 .043 .058 .042 .065
(.133) (.136) (.129) (.138) (.136)
.334 2.822 2.938
(5.218) (5.658) (5.023)
26.492 7.914 1.905 4.700
(20.893) (13.809) (13.347) (13.084)
11.561 29.102' 7.888 10.130
(12.445) (13.165) (12.913) (12.829)
56.289- 46.004- 51.051' 50.724-
(12.871) (14.538) (14.265) (14.418)
-.770
(.491) -1.173-
(.489) 16.917-
(8.086)
7.113
(5.251)
.868
(12.752) 74.630-
(25.544)
41.7710
(16.000)
18.896
(137.22)
-299.361 -296.05 293.56 -295.53 -295.33
337/ASQ, June 1999
6
One might suspect a high correlation be-
tween the commercial and technological
prominence of alliance partners as we
have operationalized the concepts. This
would occur if firms with the most devel-
oped biotech capabilities were also the
most active participants in strategic alli-
ances. But the actual correlation between
the commercial and technological promi-
nence of alliance partners is only .20. The
low correlation indicates that affiliates
with high technological prominence (well-
honed, in-house biotech operations) are
not much more active in biotech alliances
than other organizations. While firms do
not appear to treat in-house research and
alliances as substitutes (the two variables
are positively correlated, albeit weakly
so), commercial and technological promi-
nence do appear to capture different di-
mensions of affiliates' activities in bio-
technology.
7
Because of the weak effect of the alli-
ance partner prominence variables in
models 3 and 4, we wondered if there
was an additional contingency that we
had failed to consider. In particular, we
considered whether technological promi-
nence exerts an effect only in the con-
text of R&D alliances and commercial
prominence only in the context of license
and marketing alliances. When we reesti-
mated the models to allow for this con-
tingency, we still found weak effects of
the alliance partner prominence, parallel-
ing the results that we have reported.
investors; and model 6 the technological prominence of eq-
uity investors. Although all four prominence variables have
positive coefficients, only the two measures of equity inves-
tor prominence are statistically significant (models 5 and 6).6
According to the parameter estimates, a standard deviation
increase in the commercial prominence of the equity inves-
tors in a focal biotech firm multiplies the rate of going public
by a factor of 1.23. A standard deviation increase in the tech-
nological prominence of equity investors raises the IPO rate
by 33 percent.7
The remaining models in table 3 include interaction effects
between the four affiliate prominence variables and the two
measures of uncertainty in focal venture quality, the amount
of venture capital funding and the age of the firm. We esti-
mated these models to decouple evidence of an endorse-
ment process from the effects of the more tangible value of
an alliance (e.g., access to a partner's resources). Because
the arguments for a contingency in the endorsement pro-
cess are least subject to alternative explanations, hypothesis
4 offers the strongest test of our theory. Our hypothesis is
that the amount of uncertainty surrounding the quality of a
company determines the degree to which prominent ex-
change partners affect perceptions of new venture quality.
With two measures of alliance partner prominence, two
measures of equity investor prominence, and two measures
of uncertainty, there are a total of eight possible promi-
nence-by-uncertainty interactions. We report four of the
eight possible interactions (one model for each of the four
affiliate prominence variables), but the results in the unre-
ported models are similar to those in table 3.
Models 7 through 10 generally provide strong support for
the hypothesized endorsement process. Model 7 contains an
interaction of total cash raised and the technological promi-
nence of alliance partners. In this model, the technological
prominence variable becomes more significant, although it is
still shy of the 5-percent level. The interaction between
prominence and the measure of uncertainty is negative, as
we had predicted, although it too is slightly beneath the con-
ventional significance level. Model 8 includes an interaction
of total cash raised and the commercial prominence of eq-
uity investors. Here, the commercial prominence main effect
is positive and statistically significant, and the coefficient on
the interaction between prominence and uncertainty is nega-
tive and significant. Therefore, the effect of equity investor
prominence does appear to be contingent on the amount of
uncertainty about the focal venture's quality. Moreover, the
commercial prominence of alliance partners becomes a sig-
nificant determinant of the IPO rate when it is entered in a
specification that permits the variable to depend on the
amount of uncertainty about the focal venture's quality.
Models 9 and 10 report interactions of partner prominence
with focal firm age. Because firm age is the duration clock in
the time-to-IPO models, age dependence in the rate is cap-
tured by the coefficients on the three age periods in the
piecewise model. Therefore, we interacted the prominence-
of-affiliate variables with the age segments to determine
how their effects vary across organizational ages. Model 9
includes the commercial prominence of equity investors in-
338/ASQ, June 1999
Endorsements
teracted with the age segments, and model 10 reports an
interaction of age and the technological prominence of alli-
ance partners. Supporting our theory, in both models the af-
filiate prominence variable has the greatest effect on the IPO
rate for young companies. To illustrate the magnitude of the
decline in coefficient magnitudes across age periods, figure
1 plots the IPO rate multiplier as the equity investor promi-
nence variable increases within each age period. Because
we are primarily interested in the differential impact of tech-
nological prominence across the three age levels, we have
not adjusted the intercepts in figure 1 by the estimates of
the baseline hazard. As the figure demonstrates, having
prominent equity investors had a much greater impact on
the IPO rate for very young organizations. This is evidenced
by the fact that the effect of technological prominence on
the rate multiplier rises most precipitously over the variable's
range for young firms (organizations in the 0 to 3 age range).
Market valuation results. The findings from the market
valuation models are reported in table 4. In the baseline
model (1), older firms and those that had raised larger
amounts of cash from their venture capitalists garnered sig-
nificantly higher valuations. The "protein engineering" seg-
ment dummy variable is also positive and statistically signifi-
cant. The coefficients on the number of biotechnology
patents is positive but slightly shy of statistical significance
in the baseline model. Among the controls for environmental
conditions, new ventures that had IPOs in 1987 and 1991
received more generous valuations than otherwise compa-
rable firms that went public in other years. The direction of
the effects on firm density and density-squared is consistent
with density dependence theory, but neither variable is sta-
tistically significant. The biotechnology equity index is also
not significant; the level of the index appears to have a
much greater impact on the timing of IPOs than on market
Figure 1. Effect of technological prominence of equity investors on IPO
rate.
3
2.5 -
Age0to3yrs
2 4 /
~~~ i.~~~~~~ ge~~~3
to 7yrs
1.5.S
Age > 7 yrs
00l I I I I I I I I I I I I I I I I I 1 10
.0000 .0011 .0022 .0033 .0044 .0055 .0066 .0077 .0088 .0099 .0110
Technological Prominence of Equity Investors
339/ASQ, J une 1 999
Table 4
OLS Estimates of the Log of Market Value of Biotech Firms at 1PO*
Variable Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
Firm age .150- .180- .159- .179- .139- .176-
(.052) (.085) (.085) (.048) (.049) (.049)
Total cash raised .039- .039- .033- .038- .034- .038-
(.005) (.005) (.005) (.005) (.005) (.005)
BT patents .035 .051- .042 .055- .031 .052-
(.028) (.027) (.027) (.027) (.027) (.027)
Invest. new drug (IND=1) -.158 -.104 -.237 -.100 -.147 -.023
(.168) (.279) (.265) (.284) (.253) (.283)
Genetic engineering (=1) -.028 -.104 -.159 -.061 -.066 -.032
(.174) (.178) (.169) (.187) (.161) (.177)
Immunology (=1) .112 .121 .070 .119 .103 .087
(.179) (.182) (.172) (.185) (.165) (.182)
Protein engineering (=1) .525- .477- .464- .479- .479- .486-
(.245) (.229) (.227) (.229) (.221) (.230)
Diagnostics (=1) .106 .116 .229 .112 .284 .144
(.294) (.298) (.282) (.302) (.273) (.296)
BT equity index .056 .085 .058 .081 .001 .08
(.168) (.168) (.156) (.157) (.154) (.157)
BT density .018 .018 .013 .018 .009 .017
(.013) (.013) (.013) (.014) (.012) (.014)
BT density2(/1000) -.032 -.032 -.023 -.032 -.014 -.031
(.023) (.023) (.022) (.023) (.021) (.023)
1983 (=1) .282 .273 .325 .274 .580 .274
(.396) (.399) (.376) (.401) (.368) (.401)
1986 (=1) .384 .378 .345 .382 .386 .325
(.320) (.321) (.303) (.328) (.292) (.322)
1987 (=1) 1.128- 1.118- 1.059- 1.1860 .813- 1.181 -
(.362) (.366) (.344) (.370) (.340) (.362)
1991 (=1) .936- .958- .983- .9570 .998' .938-
(.310) (.311) (.292) (.312) (.282) (.310)
Alliance count .107- .055 .099- .068- .106-
(.026) (.043) (.029) (.030) (.026)
Funding from partners .007 .007 .007 .005 .007
(.004) (.005) (.005) (.005) (.004)
Equity investor count .024 .023 .037 .019 .011
(.084) (.076) (.076) (.074) (.079)
Com. prom., alliance 8.851 -
(3.276)
Tech. prom., alliance 5.285
(8.769)
Comi. prom., equity 27.431 -
(8.811)
Tech. prom., equity 1.957
(1.510)
I-Bank
prom.
Com. prom., equity x Firm age
Tech. prom., equity x Firm age
Com. prom., alliance x Tot. cash raised
Tech. prom., alliance x Tot. cash raised
I-Bank
prom. x Tot. cash raised
Lambda -1.375- -1.898- -1.606- -1.920- - 1.430- -1.862-
(.316) (.356) (.402) (.359) (.383) (.366)
Constant 1.081 .652 .855 .661 1.101 .688
(1.782) (1.584) (1.578) (1.589) (1.539) (1.592)
R-squared .626 .687 .706 .689 .711 .689
p < .05, one-sided tests.
* Standard errors in parentheses; 121 IPOs.
340/ASQ, June 1999
Endorsements
Table 4 (continued)
Model 7 Model 8 Model 9 Model 10 Model 11 Model 12
.142- .149- .203- .071 .045 .085-
(.051) (.053) (.0828) (.052) (.054) (.048)
.032- .030- .029- .037- .038- .061 -
(.005) (.006) (.005) (.006) (.007) (.009)
.039 .036 .039 .015 .011 .016
(.027) (.026) (.027) (.026) (.026) (.024)
-.088 -.083 -.712 -.087 -.004 -.289
(.268) (.262) (.260) (.284) (.283) (.288)
.052 -.053 -.196 -.074 -.064 -.155
(.169) (.151) (.166) (.187) (.179) (.165)
.025 .027 .117 .145 .065 .082
(.169) (.169) (.169) (.187) (.182) (.163)
.396- .404- .504- .564- .4950 .566-
(.230) (.215) (.225) (.216) (.232) (.222)
.050 .192 .223 .233 .122 .018
(.277) (.158) (.275) (.268) (.297) (.266)
.094 .006 -.017 -.006 .099 .022
(.155) (.156) (.160) (.147) (.160) (.152)
.013 .004 .011 .005 .019 .018
(.013) (.009) (.012) (.012) (.012) (.013)
-.022 -.005 -.020 -.007 -.028 -.035-
(.022) (.014) (.021) (.021) (.023) (.020)
.261 .444 .561 .663- .280 .168
(.395) (.372) (.379) (.360) (.396) (.353)
.209 .185 .527- .459 .346 .384
(.314) (.301) (.305) (.287) (.322) (.288)
.761 - .460 1.107- .771 - 1.115- 1.113-
(.346) (.327) (.337) (.333) (.362) (.325)
.642- .767- 1.0620 .980- 1.0410 .677-
(.301) (.281) (.287) (.287) (.307) (.279)
.104- .024 .048 .017 .011 .047
(.026) (.044) (.043) (.044) (.044) (.040)
.007 .001 .005 .003 .005 .005
(.004) (.005) (.004) (.004) (.004) (.004)
.004 -.014 .002 .009 .016 .030
(.072) (.076) (.072) (.072) (.069) (.067)
1.101 2.745 11.9060 .370 2.445
(3.309) (3.451) (4.982) (3.252) (3.088)
1.552 2.375 5.127 51.2190 .643
(8.993) (9.109) (9.272) (18.782) (8.380)
57.408- 2.631 - 15.174 25.636- 21.676-
(18.321) (9.781) (9.755) (9.200) (8.773)
1.552 34.744 -1.206 -2.101 - .961
(8.994) (27.256) (8.878) (8.875) (8.443)
.119- .0632- .0735- .068- .073- .126-
(.024) (.023) (.023) (.022) (.022) (.025)
-9.501 -
(4.457) -9.147-
(5.197) -.180-
(.065) -1.631-
(.559) -.004-
(.001)
-1.677- - 1.283- -1.231 - -.997- -.575 -1.036-
(.371) (.423) (.429) (.427) (.471) (.396)
1.087 1.955 1.729 1.636 1.596 1.510
(1.539) (1.499) (1.513) (1.456) (1.448) (1.381)
.742 .777 .772 .785 .788 .810
341/ASQ, June 1999
valuations, although the effect of the index is dampened in
the valuation models because of the inclusion of the year
dummy variables. The selection parameter, lambda, is signifi-
cant in all of the models.
Model 2 adds to the baseline the alliance-based control vari-
ables. The alliance count variable has a positive and signifi-
cant coefficient, the amount of funding from alliance part-
ners is positive and nearly significant, and the number of
non-venture-capital, organizational equity investors is positive
but insignificant.
Models 3 through 7 add the five affiliate prominence vari-
ables. The coefficient on the commercial prominence of alli-
ance partners is positive and significant (model 3): a standard
deviation increase in alliance partner prominence raises pre-
dicted market value by 35 percent. Both the technological
prominence of a focal biotech firm's alliance partners (model
4) and its equity investors (model 6) are positive, but neither
has a statistically significant impact on valuations. Model 5
shows that the commercial prominence of equity investors
is positive and statistically significant; a standard deviation
increase in that variable raises the predicted value by 28 per-
cent. Similarly, model 7 shows that the prestige of the in-
vestment banks that lead new securities offerings has a
strong effect on value: a standard deviation increase in in-
vestment bank prestige increases the predicted value by 45
percent.
Models 8 through 12 report the interactions between the
affiliate prominence variables and the two measures of un-
certainty about focal firm quality, the total amount of venture
capital funding and organizational age. These interactions are
included to test the prediction in hypothesis 4 that endorse-
ments have contingent effects on market valuations. With
two measures of focal firm uncertainty and five prominence-
of-affiliate variables, there are ten possible prominence-by-
uncertainty interactions. We present five of the ten interac-
tions, but the results from unreported models confirm a
pattern of relationships similar to those reported. In all five
of the reported models, the coefficient on the main effect of
affiliate prominence is positive, and the coefficient on the
interaction of affiliate prominence with firm age or with the
amount of venture funding is negative. In all of the models
except model 9, the parameter estimates are statistically sig-
nificant. Paralleling
the findings from the time-to-IPO analy-
ses, the models in table 4 offer strong support for the hy-
pothesis that affiliate prominence matters most when there
is high uncertainty about the quality of a focal venture. Over-
all, the results show a precipitous decline in the exchange
partner prominence effects when focal ventures are older
and when they have raised larger amounts of private fund-
ing.
We focus on the results in model 8 to demonstrate the mag-
nitude of the interaction effects. The predicted market val-
ues for two firms in the sample, Cytogen and Invitron,
offer
an interesting comparison. At IPO, both firms had organiza-
tional equity investors with almost identical commercial
prominence scores: American Cyanamid had purchased a
stake in Cytogen, and Monsanto held an equity position in
342/ASQ, J une 1 999
8
These estimates are based on the equa-
tion: exp(57.41 *prominence -
9.50*prominence*age). For both Invitron
and Cytogen, prominence was 0.019 (the
commercial prominence score for their
equity investors, Monsanto and American
Cyanamid). For Cytogen (age 6 years),
the prominence market value multiplier is
1.01. For Invitron (age 2.83 years), it is
1.78. The reported valuation multipliers
(1.01, 1.78) do not incorporate the posi-
tive "main effect" of age in model 8.
When we add in the positive age term,
the predicted multiplier for Invitron is
2.72 and the multiplier for the older Cyto-
gen is 2.47.
Endorsements
Figure 2. Effect of technological prominence of equity investors on market
value at different ages.
5.
4.5 Mean
Age-S.D.
4
MeanAge
3.5
3
2.5
2.
Mean
Age+S
D.I
1.5
0.5
.000 .003 .006 .009 .012 .015 .018 .021 .024 .027 .030
Technological
Prominence
of Equity
Investors
Invitron. Cytogen was six years old at IPO, however, while
Invitron
underwent an IPO when it was less than three years
old. According to our theory and results, the benefit of hav-
ing a commercially prominent equity investor should be
greater for Invitron than for Cytogen, because Invitron went
public at a young age. Based on the coefficients in model 8,
the effect of having Monsanto as an equity investor was a
91-percent increase in market value for Invitron. By contrast,
having Cyanamid as an equity investor would produce a neg-
ligible increase (less than 1 percent) in the predicted market
value of Cytogen. This difference is created by the decline in
importance of affiliate prominence as firms age.8
Figure 2 offers a more systematic illustration of the joint ef-
fect of age and prominence at different levels of both vari-
ables, using the coefficients in model 8. Figure 2 demon-
strates the effect of the commercial prominence of equity
investors over the variable's range at three different levels of
organizational age: the mean of the age at IPO for the firms
in the sample, the mean age plus a standard deviation, and
mean age minus a standard deviation. The vertical axis of
the figure represents the multiplier of market value resulting
from the combined effects of age and equity investor promi-
nence, while other variables are held constant. With respect
to the determination of market value, the figure demon-
strates the powerful effects of prominence and the strong
dependence of the prominence-of-affiliate effect on the age
of the biotech firm. In particular,
the figure shows that the
predicted market value increases steeply over the range of
technological prominence for young firms, but the increase
343/ASQ, June 1999
for firms a standard deviation above the mean age is rela-
tively modest.
The results thus show strong evidence that the advantage of
having prominent affiliates is contingent on the level of un-
certainty about the quality of a biotech venture. The greater
the uncertainty, the more that outside evaluators rely on the
prominence of affiliates to draw inferences about quality.
DISCUSSION AND CONCLUSION
In uncertain contexts, patterns of affiliation become bases
for evaluations. Because uncertainty pervades attempts to
evaluate new and unproven companies, the social structure
of business relationships is a primary consideration in the
market's assessment of the quality of new ventures. Fore-
most among the factors that affect attributions of value are
the characteristics of those sponsoring and affiliated with
new companies. In this study, we focused on the attributes
of three kinds of exchange partners of young biotech compa-
nies-alliance partners, equity investors, and investment
banks-as determinants of the ability of new ventures to
attract resources. Overall, our empirical analyses demon-
strated consistent and cogent effects of the prominence of
exchange partners on the performance of entrepreneurial
ventures. Moreover, the affiliative effects held up in models
that included a comprehensive set of control variables cap-
turing firm differences and environmental conditions. Build-
ing a new company is a highly competitive endeavor, and it
is difficult to overstate the advantages of having the right
intercorporate relationships when competing to create a vi-
able organization.
The results raise several points worth emphasizing about the
impact of interorganizational relationships on the perceived
value of new ventures. First, the analyses clearly demon-
strate that sponsorship has the capacity to substitute for ac-
complishment and experience as a basis for the success of
young companies. In both the valuation and the rate models,
the exchange partner's prominence had much stronger ef-
fects on the performance of companies about which there
was high uncertainty. This contingency in the impact of affili-
ations is exhibited by the precipitously declining effects of
the partner prominence variables as firms age and as they
garner greater amounts of venture capital funding. By the
same token, experience and accomplishments take on
added significance for firms that lack notable sponsors. For
instance, models 8 and 9 in table 4 show that each addi-
tional year of age increases the predicted market value by
approximately 20 percent for firms that have no strategic
alliance partners and no organizational equity investors (the
effect of a one-year change in organizational age evaluated
at zero affiliate prominence).
The implications of the positive correlations among the affili-
ate-prominence variables (table 2) also merit emphasis.
Based on the bivariate correlations and the fact that the mag-
nitude of the coefficients on each of the prominence vari-
ables declined when all of the prominence variables were
entered in the same models, we suspect that new ventures
that obtained prominent exchange partners tended to do so
344/ASQ, J une 1 999
Endorsements
across the board: firms that had well-known equity partners
were more likely to have prominent alliance partners and to
use prestigious investment banks. The tendency for firms
that obtain a highly regarded exchange partner to gain others
is not surprising. When a new venture secures a prominent
exchange partner, that partner often serves as an ally, intro-
ducing the company to its associates. For example, Bygrave
and Timmons (1992) noted that high-status venture capital
firms maintained close relationships with leading investment
banks. Thus, start-ups funded by leading venture capital
firms tended to secure prestigious investment banks to syn-
dicate their IPOs. Moreover, our theory suggests that in ad-
dition to gaining access to the contacts of a prominent affili-
ate, young companies gain attention and recognition when
they first obtain a prominent associate. This recognition may
facilitate the acquisition of additional, well-known exchange
partners, creating the possibility of a cumulating cycle of ad-
vantage accruing to young firms that gain prominent organi-
zational associates.
It is important to stress, however, that credentialing is often
an ancillary consideration in the decisions of young biotech
firms to form strategic alliances and to solicit equity invest-
ments. Alliances in the biotechnology industry are instrumen-
tal ties aimed at acquiring resources to fund speculative de-
velopment projects and to unite complementary assets, such
as technical and marketing capabilities, held by different
firms. Both parties in the transaction enter it with clearly
specified objectives, and agreements are consummated only
when extensive contracts detailing the legal terms of the
relationship are signed. The instrumental objectives of such
agreements notwithstanding, our results show that these
interfirm relations carry reputational consequences. Bearing
this out, the positive effects of the count of the number of
alliance partners and of the number of equity investors in the
baseline models in table 3 and table 4 often dropped out
once we entered the complete vector of exchange-partner
prominence scores, a trend most evident in the valuation
models. Overall, the results suggest that the impact of inter-
organizational relations is driven more by who a company is
associated with than by the volume of its relations.
One possible explanation for this pattern of results is that
the instrumental value of many interorganizational ties is
highly uncertain. For instance, the contractual terms of alli-
ances between young biotech companies and established
corporations often permit the latter to terminate the agree-
ments on relatively short notice and without cause. Hence,
even if an alliance entails milestone or other payments to
the biotech firm, the existence of the contract does not in
itself guarantee a future revenue stream. Furthermore, stra-
tegic alliances carry claims against a young biotech firm's
future outputs. Therefore, even when alliances convey sig-
nificant resources, the future costs associated with ceding
downstream development and commercialization rights often
exceed the present financial benefits of the relationships. In
contrast, the endorsement benefit derived from obtaining a
prominent strategic partner is unequivocally positive: it
stems from the mere fact that the biotech company has sur-
345/ASQ, June 1999
vived the due diligence of a capable or highly motivated
evaluator.
This study raises two issues that are worth discussing. First
is causality. Can we be sure that our effects are driven by
the hypothesized endorsement process rather than unob-
served differences between firms that happen to correlate
with affiliate prominence? To rule out this possibility, we
made every effort to control for differences in quality be-
tween the firms in our empirical analysis. We also restricted
our sample to a narrow segment of the market to reduce
the effect of differences in firm strategy. While one can
never completely rule out unobserved heterogeneity, the
contingency in the affiliate effects greatly increases our con-
fidence in the theory. Other than the hypothesized endorse-
ment process, it is difficult to construct a cogent explanation
for the strong and consistent pattern of interaction effects
observed in tables 3 and 4. The most compelling alternative
explanation for the interaction effects is that alliances (inde-
pendent of partner characteristics) are more valuable to
young and small firms, perhaps because these are the orga-
nizations that face the greatest resource constraints. Under
this scenario, the interaction effects between partner promi-
nence and focal firm age and size could be significant and
negative because of the positive correlation between affiliate
prominence and the count of alliances. To rule out this possi-
bility, we estimated models that included interactions be-
tween the alliance count and firm age along with the promi-
nence-by-age interactions. In about half of the models, the
prominence-by-age interactions lost significance-presum-
ably due to collinearity among the predictors-but in no
case, however, was the alliance count-by-age variable statis-
tically significant.
The second issue is generalizability. While caution must al-
ways be exercised when generalizing from a single industry
study, we believe that the processes we have observed in
the biotech industry do operate in other contexts. Ours is a
study of how relationships affect evaluations under condi-
tions of uncertainty. We believe that the social process un-
derlying the endorsement effects operates widely in organi-
zational domains. Consistent with this belief, others have
demonstrated that affiliations resolve uncertainty in adoption
decisions (e.g., Burt, 1987; Davis, 1991; Podolny and Stuart,
1995). With the notable exception of Baum and Oliver's
(1991, 1992) research on the advantages conferred by ties to
certain types of institutions, however, there has not been
much work on the status-enhancing properties of interorgani-
zational relations.
The results of this study open some avenues for future re-
search. First, evaluations of new ventures are affected by
other types of associations than those that we have scruti-
nized here. In particular,
the characteristics of affiliated orga-
nizations, such as law firms, accounting firms, and especially
venture capital firms themselves, are likely to affect percep-
tions of a new venture's quality. Moreover, young compa-
nies can acquire status by solidifying relationships with pres-
tigious individuals (Burton, Sorensen, and Beckman, 1999).
For example, Gilead Sciences is a biotech firm that has
gained attention because of the members of its board of di-
346/ASQ, June 1999
Endorsements
rectors. Gilead's board included a former secretary of de-
fense (Donald Rumsfeld), a former secretary of state
(George Schultz), and the chairman of Intel (Gordon Moore).
Similarly, technology companies often establish scientific ad-
visory boards, and staffing these panels with prestigious re-
searchers is another tactic for enhancing the visibility of the
venture. As Werth (1994: 22) succinctly stated, "Most [sci-
entific advisory boards] are ballast for the letterhead." Added
to the potential access and information benefits of relation-
ships with prominent individuals, relationships with them
confer status on new ventures. It would be useful to know
how other types of affiliations affect evaluations of start-ups
and the conditions under which they are most beneficial.
Second, we have for the most part ignored the strategic in-
tent and the incentives of the exchange partners of the firms
in our sample, yet, in their formal dealings with biotechnol-
ogy companies, pharmaceutical companies and other types
of alliance partners have pursued a wide variety of strate-
gies. Some exchange partners are noted for having estab-
lished alliances with a large number of biotech firms, many
of which have overlapping and competing ambitions. Other
firms make more exclusive commitments to their biotech
partners, acquiring only one strategic partner in any particular
area of development. It is almost certain that the cooperative
strategies of the partners of biotechnology firms affect the
reputational value of alliances with them. It is therefore im-
portant to develop more fine-grained measures of partner
characteristics to better understand the boundary conditions
around status transfer.
Third, the obvious implication of our findings for new venture
strategy is that young technology companies should actively
seek prominent exchange partners. The interesting question
is the conditions under which they will succeed at recruiting
such partners. It would greatly improve our knowledge of
the organization-building process to understand the relation-
ship between a focal firm's accomplishments and interorga-
nizational endorsements. Do the achievements of early-stage
ventures attract high-status exchange partners, implying that
gaining a well-regarded affiliate is an accurate reflection of
merit? Do ties to prominent actors facilitate the acquisition
of resources and thus the ability to fund expensive projects?
If so, this raises the possibility that obtaining a prominent
partner invokes a cycle of accumulating advantage for young
companies in which the addition of a well-known affiliate ex-
pedites the acquisition of the resources that enable future
accomplishments. Or are there only weak ex ante and ex
post relationships between actual quality and the promi-
nence of an organization's exchange partners?
One possibility is that the short-term advantage of a connec-
tion to a prominent actor is followed by sub-par perfor-
mance. We might expect to find this if the true quality of a
biotech firm is orthogonal to the prominence of its affiliates.
A research design that could be used to test this hypothesis
would be to compute the post-IPO stock market perfor-
mance of a sample of new ventures and then to investigate
the effect of the prestige of pre-IPO affiliates on post-IPO
performance. If the prestige of associates is unrelated to a
focal firm's quality (and if the "ntreatment
effect" of having
347/ASQ, June 1999
high-status partners is not too large), then the performance
of new issues will suffer in the years following IPOs, be-
cause the market will reassess the values of companies as it
learns more about firms' actual quality. In other words, we
might observe poor post-IPO performance among firms that
have prominent associates because of the "halo" effect
stemming from connections to prominent affiliates. Investi-
gating possibilities such as this one will enlighten the pro-
cesses through which reputations are created and the con-
sequences of intercorporate relations for the performance of
the producers in a market.
REFERENCES
Albert, M. B., D. Avery, F. Narin,
and P. McAllister
1991 "Direct validation of citation
counts as indicators of indus-
trially important patents." Re-
search Policy, 20: 251-259.
Aldrich, Howard, and Ellen R.
Auster
1986 "Even dwarfs started small:
Liabilities
of age and size and
their strategic implications." In
L. L. Cummings and B. M.
Staw (eds.), Research in Orga-
nizational Behavior, 8:165-
198. Greenwich, CT:
JAI
Press.
Aldrich, Howard E., and C.
Marlene Fiol
1993 "Fools rush in? The institu-
tional context of industry cre-
ation." Academy of Manage-
ment Review, 19: 645-670.
Barley, Stephen R., John Freeman,
and Ralph C. Hybels
1992 "Strategic alliances in com-
mercial biotechnology." In
Nitin Nohria and Robert Eccles
(eds.), Networks and Organi-
zations: Structure, Form and
Action: 31 1-347. Boston: Har-
vard Business School Press.
Baum, Joel A. C., and Christine
Oliver
1991 "Institutional linkages and or-
ganizational mortality." Admin-
istrative Science Quarterly, 36:
187-218.
1992 "Institutional embeddedness
and the dynamics of organiza-
tional populations." American
Sociological Review, 57: 540-
559.
Beatty, Randolph, and Jay Ritter
1986 "Investment banking, reputa-
tion and the underpricing of
initial public offerings." Jour-
nal of Financial Economics,
15: 213-232.
Blau, Peter M.
1964 Exchange and Power in Social
Life. New York:
Wiley.
Burrill, G. Steven, and Kenneth B.
Lee, Jr.
1993 Biotech 93: Accelerating Com-
mercialization. San Francisco:
Ernst & Young.
Burt, Ronald S.
1987 "Social contagion and innova-
tion: Cohesion versus struc-
tural equivalence." American
Journal of Sociology, 92:
1287-1 335.
1992 Structural Holes: The Social
Structure of Competition.
Cambridge, MA: Harvard Uni-
versity Press.
Burton, Diane, Jesper Sorensen,
and Christine Beckman
1999 "Coming from good stock:
Career histories and new ven-
ture formation." Working pa-
per, Harvard Business School.
Bygrave, William D., and Jeffry A.
Timmons
1992 Venture Capital at the Cross-
roads. Boston: Harvard
Busi-
ness School Press.
Cahill, Scott, Richard Caligaris,
and David Williams
1992 "Have pharmaceutical compa-
nies missed the boat on bio-
technology?" Medical Market-
ing and Media, January: 28-
38.
Carter, Richard, and Steven
Manaster
1990 "Initial
public offerings and
underwriter reputation." Jour-
nal of Finance, 45:
1045-1067.
Cohen, Wesley M., and Daniel A.
Levinthal
1990 "Absorptive capacity: A new
perspective on learning and
innovation." Administrative
Science Quarterly, 35: 128-
152.
Coombs, J.
1986 The Biotechnology Directory:
1986. New York: Stockton
Press.
Coombs, J., and Y. R. Alston
1989 The Biotechnology Directory:
1989. New York: Stockton
Press.
Davis, Gerald S.
1991 "Agents without principles?
The spread of the poison pill
through the intercorporate
network." Administrative Sci-
ence Quarterly, 36: 583-613.
Dibner, Mark D.
1988 Biotechnology Guide: U.S.A.
New York: Stockton Press.
1991 Biotechnology Guide: U.S.A.
New York: Stockton Press.
Freeman, John
1997 "Venture capital as an
economy of time." Unpub-
lished manuscript, Haas
School of Business, University
of California, Berkeley.
Freeman, John, Glenn R. Carroll,
and Michael T. Hannan
1983 "The liability
of newness and
age dependence in organiza-
tional death rates." American
Sociological Review, 48: 692-
710.
Gompers, Paul A.
1995 "Optimal investment, monitor-
ing, and the staging of ven-
ture capital." Journal of Fi-
nance, 50: 1461-1489.
Goode, William J.
1978 The Celebration of Heroes:
Prestige as a Social Control
System. Berkeley, CA: Univer-
sity of California Press.
Granovetter, Mark S.
1973 "The strength of weak ties."
American Journal of Sociol-
ogy, 78: 1360-1380.
Hannan, Michael T., and Glenn R.
Carroll
1992 Dynamics of Organizational
Populations. New York: Ox-
ford University Press.
Hannan, Michael T., and John
Freeman
1984 "Structural inertia and organi-
zational change." American
Journal of Sociology, 49: 149-
164.
348/ASQ, June 1999
Endorsements
Heckman, James J.
1976 "The common structure of
statistical models of trunca-
tion, sample selection and lim-
ited dependent variables and a
simple estimator for such
models." Annals of Economic
and Social Measurement, 5:
475-492.
Henderson, Rebecca, and lain
Cockburn
1994 "Measuring competence? Ex-
ploring firm effects in pharma-
ceutical research." Strategic
Management Journal, 1
5 (spe-
cial issue): 63-84.
Kenney, Martin
1986 Bio-technology: The Univer-
sity-Industrial Complex. New
Haven, CT: Yale University
Press.
Kogut, Bruce, and Udo Zander
1996 "What firms do? Coordination,
identity, and learning." Organi-
zation Science, 7: 502-518.
Knoke, David, and Ronald S. Burt
1983 "Prominence." In Ronald S.
Burt and Michael J. Minor
(eds.), Applied Network Analy-
sis: 195-222. Beverly Hills,
CA: Sage.
Latour, Bruno
1987 Science in Action. Cambridge,
MA: Harvard
University Press.
Lee, Lung-Fei
1983 "Generalized econometric
models with selectivity."
Econometrica, 51: 507-512.
Lerner, Joshua
1994 "Venture capitalists and the
decision to go public." Journal
of Financial Economics, 35:
293-316.
Levinthal, Daniel A.
1991 "Organizational
adaptation,
environmental selection, and
random walks." Administra-
tive Science Quarterly, 36:
397-420.
Lin, Nan, Walter M. Ensel, and
John C. Vaughn
1981 "Social resources and
strength of ties: Structural fac-
tors in occupational status at-
tainment." American Socio-
logical Review, 46: 393-405.
Marsden, Peter V.
1983 "Restricted access in net-
works and models of power."
American Journal of Sociol-
ogy, 88: 686-717.
Megginson, William L., and
Kathleen A. Weiss
1991 "Venture capitalist certification
in initial public offerings."
Journal of Finance, 46: 879-
903.
Merton, Robert C.
1973 "The Matthew Effect in sci-
ence." (First published in
1968.) In Norman W. Storer
(ed.), The Sociology of Sci-
ence: 439-459. Chicago: Uni-
versity of Chicago Press.
Podolny, Joel M.
1993 "A status-based model of
market competition." Ameri-
can Journal of Sociology, 98:
829-872.
1994 "Market uncertainty and the
social character of economic
exchange." Administrative
Science Quarterly, 39: 458-
483.
Podolny, Joel M., and Damon J.
Phillips
1996 "The dynamics of organiza-
tional status." Journal of In-
dustrial and Corporate
Change, 5: 453-471.
Podolny, Joel M., and Toby E.
Stuart
1995 "A role-based ecology of tech-
nological change." American
Journal of Sociology, 100:
1224-1260.
Powell, Walter W., Kenneth W.
Koput, and Laurel Smith-Doerr
1996 "Interorganizational
collabora-
tion and the locus of innova-
tion: Networks of learning in
biotechnology." Administrative
Science Quarterly, 41: 116-
145.
Rao, Hayagreeva
1994 "The social construction of
reputation: Certification con-
tests, legitimization, and the
survival of organizations in the
American automobile industry:
1895-1912." Strategic Man-
agement Journal, 15: 29-44.
Ritter, Jay R.
1984 "The 'hot issue' market of
1980." Journal of Business,
57: 215-240.
Singh, Jitendra V., Robert J.
House, and David J. Tucker
1986 "Organizational change and
organizational mortality." Ad-
ministrative Science Quarterly,
31: 587-611.
Sorensen, Jesper, and Toby E.
Stuart
1999 "Aging, obsolescence and or-
ganizational innovation."
Working paper, Graduate
School of Business, University
of Chicago.
Spence, A. Michael
1974 Market Signaling. Cambridge,
MA: Harvard University Press.
Stinchcombe, Arthur L.
1965 "Social structure and organiza-
tions." In James G. March
(ed.), Handbook of Organiza-
tions: 142-193. Chicago: Rand
McNally.
Stuart, Toby E.
1998 "Network positions and pro-
pensities to collaborate: An
investigation of strategic alli-
ance formation in a high-tech-
nology industry." Administra-
tive Science Quarterly, 43:
668-698.
Trajtenberg, Manuel
1990 "A penny for your quotes:
Patent citations and the value
of information." Rand Journal
of Economics, 21: 172-187.
Tuma, Nancy Brandon, and
Michael T. Hannan
1984 Social Dynamics: Models and
Methods. Orlando, FL: Aca-
demic Press.
Tushman, Michael L., and Philip
Anderson
1986 "Technological discontinuities
and organizational environ-
ments." Administrative Sci-
ence Quarterly, 31: 439-465.
Tushman, Michael L., and Lori
Rosenkopf
1992 "Organizational
determinants
of technological change: To-
ward a sociology of techno-
logical evolution." In L. L.
Cummings and B. M. Staw
(eds.), Research in Organiza-
tional Behavior, 14: 311-347.
Greenwich, CT:
JAI Press.
Werth, Barry
1994 The Billion Dollar Molecule.
New York: Simon & Schuster.
White, Harrison
1985 "Agency as control." In J.
Pratt and R. Zeckhauser
(eds.), Principals and Agents:
The Structure of Business:
187-212. Cambridge, MA:
Harvard Business School
Press.
349/ASQ, June 1999
