Professional Documents
Culture Documents
Uzzi 1997 Social Structure & Embeddedness
Uzzi 1997 Social Structure & Embeddedness
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
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.
Sage Publications, Inc. and Johnson Graduate School of Management, Cornell University are collaborating
with JSTOR to digitize, preserve and extend access to Administrative Science Quarterly.
http://www.jstor.org
sector at the time of the study. The unit of analysis was the
interfirmrelationship.
My analysis focused on the women's better-dress sector to
controlfor the differences that exist across industrysectors
(other sectors include menswear, fantasywear, etc.). Better
dresswear is a midscale market (retailsfor $80-$180),
comprises off-the-rackdresses, skirts, and jackets, typically
sells in departmentstores and chains, and tends to be price,
quality,and fashion sensitive.
Figure 1 depicts a typicalorganizationalnetwork in this
sector. Productionrevolves aroundmanufacturers(called
"jobbers")that normallyfabricateno partof the garment;
instead, they design and marketit. The first step in the
productionprocess of a garment entails a manufacturer
makinga "collection"of sample garment designs in-house
or with freelance designers and then showing its collection
to retailbuyers, who place orders. The jobberthen "manu-
factures" the designs selected by the retailbuyers by
managinga network of grading,cutting, and sewing con-
tractingfirms that produce in volume the selected designs in
their respective shops. Jobbers also linkto textile mills that
take raw materialssuch as cottons and plant linens and
make them into griege goods-cloth that has no texture,
color, or patterns. Convertersbuy griege goods from textile
mills and transformthem into fabrics (cloththat has color
and patterns).The fabricis then sold to jobbers who use it in
their clothingdesigns.
39/ASQ, March 1997
Places Delivers
order garments
tMakes
Makes / \ \ Sews garment
and delivers to
sample Creates Sends manufacturer
design trim
_ At / Sets Sends
sizes fabric Sewing
(9 Design
Studio 0Contracto
Makes Cuts
pattern fabric
Trust
Respondents viewed trust as an explicitand primaryfeature
of their embedded ties. It was expressed as the belief that
an exchange partnerwould not act in self-interest at anoth-
er's expense and appearedto operate not like calculated risk
but like a heuristic-a predilectionto assume the best when
interpretinganother's motives and actions. This heuristic
qualityis importantbecause it speeds up decision making
and conserves cognitive resources, a point I returnto below.
Typicalstatements about trust were, "Trustis the distin-
guishing characteristicof a personal relationship";"It's a
personalfeeling"; and "Trustmeans he's not going to find a
way to take advantage of me. You are not selfish for your
own self. The partnership[between firms]comes first."
Trustdeveloped when extra effort was voluntarilygiven and
reciprocated.These efforts, often called "favors,"might
entail giving an exchange partnerpreferredtreatment in a
job queue, offering overtime on a last-minuterush job, or
placingan order before it was needed to help a network
partnerthrougha slow period.These exchanges are note-
worthy because no formaldevices were used to enforce
reciprocation(e.g., contracts, fines, overt sanctions), and
there was no clear metric of conversion to the measuring
rod of money. The primaryoutcome of governance by trust
was that it promoted access to privilegedand difficult-to-
price resources that enhance competitiveness but that are
difficultto exchange in arm's-lengthties. One contractor
explained it this way, "Withpeople you trust, you know that
if they have a problemwith a fabricthey're just not going to
say, 'I won't pay' or 'take it back'. If they did then we would
have to pay for the loss. This way maybe the manufacturer
will say, 'OKso I'llmake a dress out of it or I can cut it and
make a short jacket instead of a long jacket'." In contrast,
these types of voluntaryand mutuallybeneficialexchanges
were unlikelyin arm's-lengthrelationships.A production
manager said, "They [arm's-lengthties] go only by the letter
and don't recognize my extra effort. I may come down to
their factory on Saturdayor Sunday if there is a problem ...
I don't mean recognize with money. I mean with working
things out to both our satisfaction."Trust promotedthe
exchange of a range of assets that were difficultto put a
price on but that enriched the organization'sabilityto
compete and overcome problems, especially when firms
cooperativelytraded resources that produced integrative
agreements.
An analysis of the distinctionbetween trust and risk is useful
in explicatingthe natureof trust in embedded ties (William-
son, 1994). I found that trust in embedded ties is unlikethe
calculated riskof arm's-lengthtransactingin two ways. First,
the distributionalinformationneeded to compute the risk
(i.e., the expected value) of an action was not culled by
trusting parties. Rather,in embedded ties, there was an
absence of monitoringdevices designed to catch a thief.
Second, the decision-makingpsychology of trust appearedto
conform more closely to heuristic-basedprocessing than to
the calculativeness that underliesrisk-baseddecision making
(Williamson,1994). Interviewees reportedthat among
embedded ties the informationneeded to make risk-based
43/ASQ, March 1997
PARADOXESOF EMBEDDEDNESS
If a firm becomes too embedded, does adaptationbecome
more difficultas network relationshipsare tuned to specific
tradingpartners,isomorphismwithin the network decreases
diversity,and a concentrated level of exchange with only a
few network partnersreduces nonredundantinformationand
access to new opportunities(Burt,1992)? This question
suggests a paradoxof theoreticalsignificance:The same
processes by which embeddedness creates a requisitefit
with the currentenvironmentcan paradoxicallyreduce an
organization'sabilityto adapt. In this section I explicate three
conditionsthat turn embeddedness into a liability:(1) there is
an unforeseeable exit of a core network player,(2) institu-
tional forces rationalizemarkets, or (3) overembeddedness
characterizesthe network.
The unexpected loss of a network's core organization,or
more generally,a deep and sudden structuralchange in
resource flows can cause embeddedness to shift from an
asset to a liability.Underthese conditions, social processes
that increase integrationcombine with resource dependency
problems to increase the vulnerabilityof networked organiza-
tions. Forexample, a contractormay become highlyskilled
at workingwith a manufacturer'sfabric,productionsched-
ule, and design specifications. If that manufacturercloses
shop or migrates offshore, then the embedded relationship
that had originallybenefited the contractormay now put it at
a higher riskof failurethan if it had diversifiedits ties,
because it is likelyto lackthe resources needed to transition
to a replacement partner(Romo and Schwartz, 1995).
The problemis the opposite of the free-riderproblem:
diligentcommitment, backed by expectations of reciprocity
and social pressure to perform,intensifies an organization's
involvementwith certain network partnerswhile raisingthe
concomitantcosts of keeping ties to extra-networkpartners
that can providea safety net for unexpected or random
fluctuations.Portes and Sensenbrenner (1993: 1340) drew
attentionto this phenomenon in their study of entrepre-
neurs, whose socially embedded relationshipsgave them
access to resources but restrictedtheir actions outside their
network. In the apparelindustry,the unexpected failureof
Leslie Fay, Inc., a manufacturerat the center of a large
network, was most debilitatingfor the primarycontractors
that had benefited from their close tie to Leslie Fay, Inc.
which, before fallingvictim to a few unscrupuloustop
executives, had sheltered them from a glut of low-cost
competitors and downturns in the economy (Uzzi, 1997). The
above arguments suggest the following propositions:
57/ASQ, March 1997
Underembedded
Arm's-length Network Integrated Network Overembedded Network
0 /s0 /u
0~~~~0
0 ~ ~ 0 00
A? o
000~~~0
/ # 0
Key:
..g-- e--*. - = Embedded tie 0=Focal Firm
- = Arm's-length tie a = Manufacturers
.-> ;e t o = Contractors
1st-order 2nd-
Network order
network
DISCUSSION
If economic action is embedded in networks of relations
(Granovetter,1985), then a logicalfirst step is to specify the
dimensions of embedded relationshipsand the mechanisms
by which they influence economic action. This undertaking
builds on the work of others who have launchedthe impor-
tant enterprise of reintroducingsocial structureinto the
analysis of economic phenomena. In tryingto demonstrate
the unique organizationaland marketprocesses that follow
from an understandingof social structureand economic
performance,I analyzedthe propertiesof embedded
relationsand how they create competitive advantages for
firms and networks of firms.
While these processes and outcomes are not quickly
summarized,Figure3 diagramsthe propositionsand logic
presented in the preceding sections into their antecedents
and consequences. The figure lays out, from left to right,the
social structuralantecedents of embedded ties, the compo-
nents of embedded ties, the positive outcomes of an
integratednetwork structure,and the negative outcomes of
overembedded networks. The plausibilityof these proposi-
tions has been partlyestablished in the empiricalresearch
analyzed here. Testing and refinement await future research.
The programmaticimplicationof this work is that embedded-
ness is a unique logic of exchange that results from the
distinct social structureof organizationnetworks and the
microbehavioraldecision-makingprocesses they promote. In
an embedded logic of exchange, trust acts as the primary
governance structure.Calculativeriskand monitoring
systems play a secondary role. Informationtransfer is more
fine-grained,tacit, and holistic than the typical price data of
pure marketexchanges, and joint problem-solvingarrange-
ments promote voice ratherthan exit. On a microbehavioral
level, actors follow heuristicand qualitativedecision rules,
ratherthan intensely calculativeones, and they cultivate
long-termcooperativeties ratherthan narrowlypursue
self-interest. These factors furnishan alternativemechanism
for coordinatingadaptation,speeding productsto market,
and matchingconsumer demand to production.
This paper offers an explanationof the links between social
structure,microbehavioraldecision-makingprocesses, and
economic outcomes within the context of organizational
networks. At the same time, it adds complexityto models of
exchange such as game theory, agency theory, and transac-
tion costs economics. Typicalclassificationschemes differ-
entiate these theories accordingto an actor's motivation
(selfish or cooperative)and rationality(pureor bounded)
(Williamson,1985: 50). But I found that in embedded ties, an
actor's motivationand rationalityresist characterization
within these distinctions,thereby demonstratingthe unique
logic of embeddedness. My findings suggest that in net-
works of close ties, motivationis neither purelyselfish or
cooperative but an emergent propertyof the social structure
within which actors are embedded and that rationalityis
neither purelyrationalor boundedlyrational,but expert. Thus
embeddedness broadens the typicaltypology of exchange
theories to include a new categorizationfor motives (emer-
61/ASQ, March1997
C >
D lw
Lu
0~~~~~~
2 U) W
~~~~ )
y 0 > '
OWa IL L o0 Z + C? C
+~~~~~~~~~~~~~~~~~~~M40L
-
U ~
0 U)0
~ 0
+ . ::
U
*. .
E00Uo0 4
LU~~~~~~~~
> O'6 *. *~ + C ~~~~~0
0
; >x * =E E_ E
H **> ~ * * x * *) *->z * E
' : Co : CD
LC
Cn# Cn Cn .
o@~
OZ~
~~~..J 0
W
+. C,) +U) ' 0 +*.
4-0~~~~~~~~~~~~~~~~4 +
'-'LJ~~~co . ..m * +
~~
0
Zz ~C=W~~~
0
C
D
~+CM
CDUW 4-0.
(DU C + + U) 0 C0D
2 U)U)~~~~
(X >V x EU
LU ..
CC o
-' o.
D4
*E L-
(D
U . *_
EC 4*_U)U 4o0 E(
U s
0
0
Z .
.~~~~~~~>
. .C .
.=
W
0>>&
. a. Ua
0 0):f** 2
0
o
0)
0 m +-,
CU*-s-
CO
0 0U
E-og,
NU)U)
o c-
C0 C
0~~~~~~~~~~~~~~~
0 -~~~~~~~~~~~~~~
U. 4- )0 C>C
B-
0 L-CU C ) C
~~o U) Cl 0 * :
{ H uj
CC
_C ia
.3aL CUU) )
U=Um
U) @0'a
a
o~~~~WCJ)L oi Co a)'
0 - 0
-'H 0 0 a m- C
U 0U)X-CD l
0 ~ ~ U)~US ~ 4-'oU)0 U)L.
C
E >
-LLJ
H0
>0
4~~~C A LCO
(D0CU3, AWO E
<U
0~~~~(
0~~~~~0 ~ ~
U))
REFERENCES