Strategic
Management Journal, Vol. 18:2, 127-140 (1997)
THE EFFECT OF REPUTATION ON THE DECISION
TO JOINT VENTURE
MARC
J. DOLLINGER
PEGGY
A. GOLDEN
College
of Business,
TODD
SAXTON
This
paper focuses on the impact that reputation has on the decision to proceed with
a strategic alliance. Employing
reputation constructs adapted from the Fortune Corporate Reputation Survey, we
manipulated a target firm's reputation in an experimental design. The subjects were placed in the role of CEO
of the partner firm and asked whether they would engage in the alliance. Findings indicate that (1) reputation is a
multidimensional construct, (2) the personal information-processing
characteristics of the decision-maker mediate the reputation effect and may
suppress the reputation information, (3) subjects may compensate weaker
elements of reputation for stronger ones when making decisions, (4) product and
management reputation are the most important factors, and (5) reputation is a
factor affecting the decision regardless of whether the proposed target is a
supplier or a competitor. @ 1997 by John Wiley & Sons, Ltd.
A
positive reputation indicates that an organization is highly esteemed, worthy
or meritorious; it implies a good name and high regard (Webster's Third New
International Dictionary, 1961). A
firm's reputation is an intangible element of its business strategy. With it the firm may signal its competitive
intentions. For example, a reputation
for retaliation inhibits rivalry (Caves and Porter, 1977). A positive reputation is a strategic factor
that can be employed to earn above-average profit (Bamey,
1986). A firm's reputation influences
trust, and that leads to alliances and other interorganizational relationships
(Oliver, 1988). The reputations of new
firms and their founders, including favorable beliefs, trust, and psychological
commitment, are assets that serve as the foundation of the entrepreneurial
'honeymoon' (Fichman and Levinthal,
1991). Reputation-building activities
are, therefore, strategically important for potential target firms in these
incomplete information settings (Weigelt and Camerer, 1988).
The
purpose of this paper is to build and test a model of alliance formation that
focuses on target firm reputation, and to look at the dimensionality of the
reputation construct. The model has
three components, incorporating the direct effect of reputation, the status of
the alliance partner as a competitor or supplier, and the individual
characteristics of the decision-maker.
The first component is the main effect of reputation on the joint
venture decision. Although it seems
clear that a positive reputation is an unalloyed asset, there are a number of
previously unexamined issu6s here: Which elements of a firm's reputation are
most salient to the alliance partner?
Which are necessary, which sufficient?
And, what are the combinatorial properties of reputational information
regarding the propensity to form a joint venture? For example, a firm might have an excellent
reputation for financial stability, yet its products may be seen as noninnovative and of below-average quality. Can a clean balance sheet offset the weak
product situation in the minds of an alliance partner?
The
second component concerns the position of the target firm in the production
chain. It has been noted that it is
strategically risky for enterprises to engage in alliances directly with
competitors (Bresser and Harl,
1986; Nielsen, 1988; Bresser, 1988). The question of whether a competitor can be
trusted enters into the model. A
competitor firm's reputation for integrity and trust may influence its
attractiveness as a partner. We test the
effect that this has on the decision by manipulating the target firm's
relationship (supplier or competitor) with the partnering firm.
The
third component concerns the individual characteristics of the
decision-maker. The content of a firm's
reputation is information. This
information is often incomplete, and a firm's reputation with a specific
decision-maker may be based on second-hand information. Each decision maker will process this reputational
content according to the individual's proclivities for handling such ambiguous
information. Therefore, if we view a
firm's reputation as a mental map, individual differences in the map reader's
tendencies and biases will influence the decision to enter into an alliance
based upon the target's reputation. We
incorporate the effects that the psychological content of the information has
on the joint venture strategic decision (Bateman and Zeithaml, 1989).
The
model can contribute to research investigating reputation and strategic
alliances and to practitioners considering alliances as part of a strategic
initiative. For researchers, the unidimensionality of corporate reputation found in prior
research (see, for example, Fombrun and Shanley, 1990) has led to a recent wave
of analytical approaches to address a halo effect in corporate reputation as
measured by the Fortune survey (Brown and Perry, 1994; Fryxell
and Wang, 1994). The halo effect and
common method variance based on the data-gathering technique have been
identified, but little has been offered to expand the notion of reputation as a
multidimensional construct. This study
takes an alternative perspective to the post hoc analytical approach. If reputation is multidimensional,
researchers should be able to manipulate dimensions independently in a
decision-making framework. Once validity
for the multidimensionality of reputation has been established through this
manipulation, further field and lab research may be undertaken to measure and evaluate
this important construct.
For
practitioners, the use of alliances has become a common tactic to achieve
strategic objectives (Harrigan, 1986; Fortune, 1992). To engage in these relationships, though, a
firm must provide an attractive combination of assets to a potential
partner. By understanding how potential
partners value intangible assets such as reputation, and how the
decision-maker's characteristics may affect the decision to engage in an
alliance, a company may be able to better position itself as an attractive
partner. Companies such as Coming have
built a core competence based on an ability to successfully engage in alliances
(Fortune, 1992). Thus, research investigating
the role reputation and dimensions of reputation play in alliance decisions is
managerially relevant.
THEORETICAL
DEVELOPMENT The role of reputation
The
reputation construct is an important component in four theoretical models of
management and organization: resource-based theory, game theory, transaction
cost theory, and theories of organizational effectiveness. In each of these models, a positive
reputation can be shown to increase the desirability of the target firm as a
strategic alliance partner.
Resource-based
theory
Within
the resource-based theory of the firm (Penrose, 1959; Wemerfelt,
1984), reputation is seen as one of the key asset bases (Grant, 1991). A reputation can be valuable (Hall, 1992),
rare, hard to duplicate (Mahoney and Pandian, 1992) and nonsubstitutable, thus
providing the firm with a sustainable competitive advantage. A firm's reputation can, therefore, be a
source of rent and profit (Barney, 1991).
A superior reputation can act as a barrier to imitation. In a recent survey of top executives, several
aspects of reputation (firm and product) were rated the top contributors to
firm performance (Hall. 1992).
Game
theory
A
strategic alliance can be analyzed from a game theory perspective (Dollinger,
1990). Forming an alliance adds value
and transforms a zero-sum game into a positive-sum one. In order for the game to result in the total
maximum utility over the length of the relationship, firms must be able to
predict what the other will do. One cue
as to the target's game behavior is reputation (Weigelt
and Camerer, 1988).
Partners infer from positive reputations that the target is not going to
defect and lower the collective pay-offs.
A positive reputation also encourages future game playing. If an alliance partner were to covet a
short-term gain by defecting, and in an end game scenario be immune to
retaliation, its reputation for future games and alliances would be diminished.
Transaction
cost economics
When
the issue is strategic alliance partner choice, a positive reputation can
reduce transaction costs. Firms with
strong positive and negative reputations are more visible; they are likely to
receive more media coverage than firms with no reputations. Therefore, the searching costs for an
acceptable partner firm are lower. Also,
the implication of a positive reputation is that the target firm can be
monitored and evaluated more easily because it is more visible and its
performance more public.
As
with game theory, the threat of opportunism by the target firm is reduced by
virtue of the target's positive reputation.
Even if opportunistic behavior takes place during the alliance, the
decision-maker is somewhat protected from the negative consequences because a
consensus highreputation firm was originally
picked. The recontracting problem (each
side in the alliance is potentially in the position of a monopolist or monopsonist) is also ameliorated by the desire to maintain
their hard-eamed reputation over a long period of
time.
The
effectiveness literature
In
the previous three models, firm reputation is viewed as an independent
variable, a contributor to firm performance.
But a firm's reputation is also a measure of its effectiveness, a
dependent variable. It may be a function
of financial performance, product quality, management effectiveness or some
combination of factors that appeal to a firm's multiple constituencies (Tsui, 1984).
Accountability to external constituents is viewed as the hallmark of a
positive reputation (Gaertner and Ramnarayan,
1983). Fombrun and Shanley (1990), in
their work on the Fortune Corporate Reputation Survey data base, used
reputation in this way.
This
discussion illustrates the potential benefits of a positive reputation. It is a resource for a firm, and potentially
desirable to partners; a positive reputation reduces the perceived likelihood
of defection in a game-theoretic perspective; it reduces transaction costs for
the partner; and it is an indicator of the firm's overall effectiveness. It is reasonable to expect, then, that a
partner with a positive reputation would be more desirable than one with a poor
reputation. Thus, we propose:
Hypothesis
1: A decision-maker's propensity to engage in a joint venture is increased by
the positive reputation of the target firm.
This
hypothesis is made more interesting by the potential to manipulate dimensions
of the partner firm's reputation. One of
the difficulties with the reputation construct is that although it seems to be
comprised of several fundamental and independent dimensions, there is a halo
effect that masks this multidimensionality (Fombrun and Shanley, 1990; Brown
and Perry, 1994). For example, Rao (1994), in a study of the auto industry, found that
high-visibility events, such as winning product quality certification contests,
improved the company's reputation which subsequently improved effectiveness in
other areas. Similarly, Johnson (1993)
found CEO reputation to be sensitive to stock returns and accounting
earnings. In this case the halo is from companyto-individual reputation instead of from companyto-company element.
Although it appears that reputation consists of multiple dimensions, it
is frequently either measured in a unidimensional
fashion as noted above in the Rao (1994) and Fombrun
and Shanley (1990) examples, or the potential for the halo, spillover and
compensatory effects are ignored by researchers and the market. In this paper, we attempt to 'tease out' the
effects of several dimensions of reputation that have been used or alluded to
in the literature.
Supplier
vs. competitor partners
The
second component of our model concerns the target firm's position in the
production chain, namely the potential differences between a supplier or
competitor firm. Direct contacts and
alliances with noncompeting firms (conjugate
alliances) include long-term purchasing contracts with suppliers and customers
and joint R&D projects. For example,
a joint R&D effort enables a manufacturer to test the operating
characteristics of a supplier's materials (for a fee) and reports back to the
supplier how the material holds up under various real operating conditions (a
gain for the supplier).
Direct
contact and an alliance with a competitor is called a confederate alliance
(Astley and Fombrun, 1983). Partnefing among competitive technology firms offers the
potential to create entry barriers and may prevent their suppliers from flexing
power. In Hagedoom's
study (1993), market access and structure was the most frequently mentioned
reason for engaging in a confederate alliance.
Other often-cited and researched reasons include rapid access to new
technology and markets, organizational leaming, and
improving customer-supplier relationships (Foffest,
1990; Hagedoom, 1993; Hamel, Doz
and Prahalad, 1989). Thus, alliances
between technology firms become a mechanism for cost containment and the
strategic control of competitive drivers.
Engaging
in an alliance, however, is not without risk.
An alliance with a competitor means taking the risk of revealing
proprietary technology (Bresser, 1988; Bresser and Harl, 1986). There is still the risk of defection (from
game theory). And not only technology is
in jeopardy. Organizational skills and
systems can be transferred to the competitor through diffusion and social leaming. Important
human resources can be lured away to work for the competitor. The competitor can attempt to piggyback the
alliance into additional financial resources which could be used to attack the
partner at a later point. So there is a
tradeoff when partnering with a competitor, and it is not clear, a priori,
whether the costs outweigh the benefits.
The competitor-supplier manipulation is included in the experiment, but
without a hypothesis regarding the outcomes in our experiment.
Decision-maker
cognitive characteristics
The
target firm's reputation is received as information. This information possesses elements of
uncertainty, ambiguity and risk. The decisionmaker's ability to process this reputational
information and his tolerance for this uncertainty and risk will vary among
individuals. Information interacts with
cognitive characteristics within the person to produce the frame for the
decision. Some cognitive characteristics
may serve to magnify or enhance the reputational information; others may serve
to suppress the content received.
Tolerance
of ambiguity refers to the tendency of an individual to view situations which
are uncertain, potentially without solutions, and novel as desirable (Budner, 1962).
Scenarios in which information is conflicting, and interconnections
complex and novel are avoided by persons with low tolerance of ambiguity. A mix of positive and negative reputation
factors has the potential to create problems that do not have apparent
solutions; mixed reputation increases the ambiguity of the decision
environment. In the application proposed
here, tolerance of ambiguity functions as a mediator between the reputation of
the target firm and the final decision to form an alliance. The cognitive characteristic is hypothesized
to act as a suppressor of the negative information. Reputation is the main effect and the
decision to engage in the alliance is the criterion. Therefore the following hypothesis is
proposed:
Hypothesis
2: Tolerance of ambiguity will interact with finn
reputation to predict alliance approval.
High tolerance of ambiguity subjects will be more likely to suppress
mixed and negative information regarding target firin
reputations, and therefore more likely to approve the alliance.
While
an individual with high tolerance of ambiguity might not feel discomfort in
processing discordant reputational information, that in ivi
ual might feel the lack of 'latitude of managerial
discretion' needed to make the joint venture decision (Hambrick and Finklestein, 1987).
There are a number of forces that have been argued to influence
managerial discretion. Among these are
the aspiration level of the manager, the commitment level, tolerance of
ambiguity and locus of control (Hambrick and Finklestein,
1987: 373). Locus of control refers to
an individual's beliefs that outcomes stem from internal or external factors
(Rotter, 1966). If a decision-maker is
internal in his locus of control, he might be inclined to increase the domains
and possible courses of action available to him. Conversely, if an individual is external in
his locus of control, he perceives that his discretion to make decisions is
limited by factors, entities, or other individuals (Hambrick and Finklestein, 1987).
Miller, Kets de Vries
and
This
leads us to the third hypothesis:
Hypothesis
3: Locus of control will interact with firm reputation to predict alliance
approval. Internals (high locus of
control) subjects will be more likely to suppress mixed and negative
information regarding target firm reputations, and therefore more likely to
approve the alliance.
To
summarize, we are testing a model of alliance behavior that incorporates three
major components: target firm reputation, decision-maker characteristics, and
target firm position in the production chain.
In this model, reputation is construed as the primary factor affecting
the decision to partner in a joint venture or similar alliance. We anticipate that there may be a difference
in the propensity to engage in the joint venture depending upon whether the
target firm is a competitor of or supplier to the partnering firm. Lastly, because a firm's reputation might
contain discordant and conflicting information, we hypothesize an interaction
between the cognitive characteristics of the decision-maker and the content of
the reputation information.
METHODS
The subjects for this experiment were 170 MBA
and Executive MBA students at a large public
university. The subjects' mean age was 28 years
old. Nearly half of the subjects were
finance/accounting
majors, with the balance of
the
majority made up by marketing, management,
and
operations. The mean number of years of
industry
experience was 4.3; only 22 subjects had
less
than 2 years of industry experience.
This
limited experience was in a wide variety of
industries,
spanning
banking and financial services, con-
health care, government, and other indus-
tries. Functional responsibilities closely
paralleled
the
chosen majors, with the majority of subjects
in
finance and accounting (31%), sales and mar-
keting (18%), production
(13%), and general
management
(8%). Nearly 90 percent of the sub-
jects were bom in the U.S.A. Seventy-three per-
the cent of the subjects were male.
Task
familiarity
In
using student subjects, the issues of task familiarity and generalizability
come into question (Sears, 1986; Gordon, Slade, and Schmitt, 1986; Fromkin and Streufert,
1983). In this experiment, the concern
is tempered by the use of MBA and Executive MBA students. As indicated above, the mean age and
experience reflect a reasonable degree of familiarity with organizational
processes. Although the subjects
certainly are not seasoned in the choice of strategic alliance partners, they
are very familiar with the process of selecting suppliers or dealing with
customers, as suggested by their range of functional responsibilities. Moreover, their work experience levels and
presence in an advance education program should make them aware of the criteria
for making such decisions as alliance partners.
Additionally, the use of student subjects to capture the effect of
individual characteristics on decision-making is well founded in decision-making
research.
In
a postexperiment set of questions, we asked subjects
to rank order their preferences for alliance types. The joint venture was the most preferred
type, followed by acquisition/merger, licensing agreements and comarketing arrangements.
The rank order was significant (
Preference
may be inferred to suggest task familiarity.
Subjects would be likely to choose as their top preference an alliance
type that they were familiar with and understood. Thus, we suggest that task familiarity, a
critical boundary variable, is moderate-to-high from a cognitive standpoint,
moderate from a practical standpoint, and sufficient for this experiment.
Task
The
experiment involved a joint venture decision based on a joint venture reported
in the Wall Street Joumal. The facts presented in the newspaper article
were rewritten, with actual company names disguised, and a scenario developed
to indicate that the joint venture had not yet occurred, but was just about to
occur. Subjects were asked to assume the
role of CEO, and were given the scenario describing a potential joint
venture. The target was described as the
only company remaining under consideration following a screening of potential
partners. Substantial benefits to the
subject's company were indicated. The
scenario stated that the partner firm projected it would save 'hundreds of
millions of dollars over the next 5 years' by implementing the joint venture
rather than assuming the cost of new facilities. In addition, the scenario indicated that the
agreement would enable the partner firm 'to decrease new product development
time.' The scenario description can be found in Appendix 1. The rationale for
the alliance encompasses literature-derived reasons for technologically
oriented alliances and joint ventures (Hagedoom,
1993; Geringer, 1988; Harrigan, 1986).
The explicit benefits were held constant regardless of the target's
reputation; thus, the pay-off in financial terms was a constant, with the
partner reputation and status as supplier or competitor the only sources of
variance for subjects.
Manipulation
The
experimental design created two groups: one group was cast in the role of joint
venturing with a competitor, and one group was cast in the role of joint
venturing with a supplier. Within each
group, the product quality, management quality, and financial reputations of
the target firm of the joint venture were manipulated into positive or negative
conditions, based on a one-paragraph description of each dimension. Each subject received treatments
simultaneously on each dimension of the three reputation factors; the positive
or negative treatments were varied and randomized such that approximately 20
subjects received each of the eight possible combinations of treatments. The manipulation of the reputation variables
is reproduced in Appendix 2. The experiment was administered in a mandatory second
year MBA class and Executive MBA management class. Treatments were completely randomized between
and within each class. All subjects were
read the same instructions regarding the assignment, which took approximately
30 minutes to complete.
Measures
Following
the scenario description, subjects were asked to answer a number of questions
about the venture possibility. The first
question and principal dependent variable for analysis was a decision as to
whether to form the venture or not. The
question read: 'Based on the description, would you proceed with the strategic
alliance? (Circle One) Yes No.' Fifty-five respondents indicated they would
proceed with the alliance and the remaining 115 respondents indicated that they
would not proceed. Analysis of the
demographics indicated no systematic difference within the sample. Contingency analysis yielded X' values of
3.67 by major; 6.90 by functional area of work experience; 29.70 for industry;
6.90 for national origin; 0.003 for gender; and 1.64 for marital status. ANOVA Fs for mean differences were
0.015
for GPA, 1.77 for years in industry, and 0.24 for age. None of these results were statistically
significant.
Tolerance
of ambiguity was measured using the 16-item scale developed and validated by Budner (1962). Scale
reliability for this study was 0.7062. Locus of control was measured using the
16-item scale developed and validated by Spector
(1988). This scale is different from the
Rotter (1966) scale in that its items are directly relevant for work-related
behavior and is therefore a more specific scale than the Rotter instrument. The reliability for this study was 0.8061.
Both scale reliabilities are sufficiently high to accept the instruments a
priori. The simple correlation
coefficient between tolerance of ambiguity (TA) and locus of control (LOC) was
0.34 (p < 0.01).
Manipulation
check RESULTS
A
manipulation check was performed on the multidimensional reputation variable
with ANOVA using a scale assessing each dimension of reputation (I = strongly
disagree, 7 = strongly agree that reputation is 'excellent'). This confirmed that the reputation dimensions
were appropriately understood by the subjects.
Results are shown in Table 1. In each case, the mean response of the validity
scale was significantly higher when the manipulation on that dimension was
positive. For example, respondents
indicated that the quality of management was excellent only when the
manipulation was positive (F = 55.787, p = 0.000).
An
intriguing result of the experimental manipulation is the suggestion of
spillover effects between reputation dimensions. Table I captures these spillover effects
among the reputation descriptions. This
occurred even though the reputations were manipulated independently, and the
descriptions of the target firm's reputations did not refer to any of the other
descriptions. For example, when all of
the target firm's reputational descriptions were presented in a positive
manner, the subjects rated them above 6 on the 7-point scale of excellence
(Table 1, column 1). However, when
financial reputation was presented as negative (column 2), product reputation
went up and management reputation went down.
Similar patterns are found in this table and additional remarks about
this will appear in the Discussion section.
Loglinear models were
developed to test the relationship between the manipulations and the decision
to engage in the joint venture. As
reported earlier, the dependent variable was a YesINo
decision. When the response variable is
dichotomous with multiple predictors, one appropriate technique is logit modeling. In
this type of statistical analysis, the frequency of the response is used to
create multiple crosstab tables and the significance of the model is dependent
on whether the model fits well as determined by maximum likelihood estimators
to compute a chi-square statistic. A
good model should show no significant difference between the response
categories overall. Any significant difference
will then be evident in certain predictor variables. This modeling approach is suitable to experimental
situations in which there are only two choices: go or no go.
Reputation
effects
In
loglinear models, average cell size must be at least
five observations (Demaris, 1992). Therefore, our average cell size of II is
sufficient to produce reliable and stable results. The manipulations were coded into one
variable with eight states that represented all of the combinations of the
three variables in two possible states.
This new variable was then used to create an unsaturated logit model to assess the role of reputation in the
decision to joint venture. As Table 2
indicates, the model showed significant differences between the groups
(likeliness ratio X = 39.659,
Table
1. Reputation ratings for the eight scenarios
Manipulation
checks
(1) (2) (3) (4) (5) (6) (7) (8)
All Management/Management/ Product/ Management
Product Finance
All
positive product finance finance positive positive positive
negative
positive positive positive
Quality
of 6.21 5.68 5.85 2.41 4.86 1.48 1.71 1.30 55.787***
management
Financial 6.11 1.95 6.30 5.95 1.43 1.48 5.81 1.65 103.872***
reputation
Quality
of 6.05 6.41 1.70 6.14 1.48 6.38 1.52 1.52 157.804***
product
***p
< 0.001.
Table
2. Propensity irrespective of integration form
Manipulation Adj.
Response (management/product/finance) Observed
Expected residual
No
group +/+/+ 9.00 14.37 -2.643
5.00 14.37
12.00 14.37
15.00 14.37
20.00 14.37
17.00 14.37
18.00 14.37
19.00 14.37
Yes
Group +/+/+ 10.00 6.88
17.00 6.88 4.128
8.00 6.88
7.00 6.88
1.00 6.88 -2.395
4.00 6.88
4.00 6.88
4.00 6.88
Likeliness
ratio X' 39.659;
P=0.000
p
= 0.000). In logit, adjusted residuals above the
absolute value of 1.96 indicate a significant (p < 0.05) contribution to the
grouping variable. The sign indicates
the direction of the relationship with respect to the grouping variable (Demaris, 1992; Knoke and Burke,
1980). The adjusted residuals for each
manipulation indicated that three manipulations were significantly out of the
expected range. In the 'no' group, the
'all positive' reputation caused the observed rejections of the alliance to be
smaller than expected. In the 'yes'
group, the manipulation that was positive for management and product and
negative for finance had more acceptances than expected. And the positive management reputation
coupled with negative product and financial reputation was associated with
fewer acceptances than expected. Within
the 'no' group, rejections decreased as the manipulated reputation became more
positive. Also, within the 'yes' group,
acceptances decreased as the manipulated reputation was more negative. Thus, Hypothesis I regarding the overall
effect of reputation on the joint venture decision is accepted.
Supplier-competitor
effects
Since
the literature suggests that joint ventures may be used either to moderate
industry drivers or control costs, we tested a 'no difference'
hypothesis
involving the supplier and competitor treatment groups. An unsaturated logit
model was tested. As Table 3 indicates,
there was no significant difference between the observed and the expected
occurrence of the competitor and supplier treatments (likelihood ratio X2 = 0.
15683, p = 0.692).
TA
and LOC effects
The
roles of LOC and TA were tested using logistic regression with an interaction
between the cognitive variable and the reputation factor. Logistic regression can be used when the
dependent variable is dichotomous and the predictors are continuous or interval
in nature. If one of the
Table
3. Supplier-competitor impact on joint venture
decision
Adj.
Response
Relationship Observed Expected residual
No Competitor 56.00 54.79 0.396
Supplier 59.00 60.21
Yes Competitor 25.00 26.21 -0.396
Supplier 30.00 28.79
Likelihood
ratio X2 = 0. 15683; p = 0.692
Small
Firm Reputation and Decision Making
Table
4. Logistic regression interaction of tolerance of
ambiguity
and reputation on the joint venture decision
Reputation
effect
(Management/
product/finance) Parameter (B) Wald
statistic
+/+/+ -0.0122 3.5691 ***
-0.0334 19.2296***
-0.0048 0.5709
+0.0019 0.0858
+0.0289 5.1296*
+0.0057 0.6318
-0.0095 1.6620
+0.8673 16.6786***
***p
< 0.001; *p < 0.05
predictors
is a categorical variable, the model can be built in a manner equivalent to a
dummy regression. This was the approach
used here. The eight possible
combinations of manipulations were constructed as an interaction with the
interval LOC and TA scales. In this
analysis, the 'all negative' combination was the reference manipulation. As Tables 4 and 5 indicate, both tolerance of
ambiguity (-2 log likelihood = 169.47, p = 0.2179; goodness of fit X2 = 164.38,
p = 0.3 1) and locus of control (-2 log likelihood = 171.359, p = 0.24;
goodness of fit X2 = 164.86, p = 0.3588) have an interaction with the overall
reputation effects on the joint venture decision.
For
the tolerance of ambiguity interaction (Table 4), the parameters were
significant in four
cases: (1) when all elements of reputation were
Table
5. Logistic regression interaction of locus of
control
and reputation on the joint venture decision
Reputation
effect
(management/
product/finance) Parameter (B) Wald
statistic
+/+/+ -0.0162 6.1815**
-0.0291 17.7551
-0.0078 1.4671
+0.002] 0.1136
+0.0305 4.9165*
+0.0062 0.7501
-0.0088 1.4254
+0.9245 17.9972***
***p
< 0.00 1; **p < 0.0 1; *p < 0.05
positive
(Wald = 3.569 1, p < 0.00 I), (2) when management/product was positive (Wald
= 19.2296, p < 0.001), (3) when management only was positive (Wald = 5.2196,
p < 0.05), and (4) when all reputation elements were negative (Wald =
16.6786, p < 0.001). In order to interpret these findings, one needs to
examine the sign on the Parameter.
The
first two significant conditions show that when TA was high and the reputation
treatments either all positive or mostly positive, the subjects were actually
more likely to reject the alliance. That
is, the high TA suppressed the positive information. The last two significant conditions show that
when TA was high and the reputation treatments either all negative or mostly
negative, the subjects were more likely to accept the alliance. These last two findings are in accordance
with Hypothesis 2 which anticipated that negative information would be
suppressed, but the first two are not.
So although we can accept the hypothesis that an interaction occurs, the
nature of the interaction is for the cognitive characteristic to suppress both
the most positive and the most negative information.
The
pattern of interaction between LOC and reputation (Table 5) parallels the TA
ones. An interaction between LOC and the
reputation treatment occurs in four conditions: (1) when reputation is all
positive (Wald = 6.1815, p < 0.01), (2) when management and product are
positive but finance is negative (Wald = 17.7551, p < 0.001), (3) when
management is positive but product and finance are negative (Wald = 4.9165, p
< 0.05), and (4) when all reputation treatments are negative (Wald =
17.9972, p < 0.001).
In
the first two significant interactions, the high-LOC subjects (intemals) were more likely to reject the alliance when
faced with the positive reputation treatments.
Again the cognitive characteristic suppressed the most positive
information. In the last two significant
interactions, the intemals were more likely to accept
the alliance when the reputation was all negative or mostly negative. Here the cognitive characteristic suppressed
the negative information, as expected.
Therefore, we cannot accept Hypotheses 2 and 3 in their current
form. A more accurate hypothesis would
have stated that high-TA individuals and intemals
were more likely to suppress both the most positive and most negative
reputation information to decide the alliance issue.
136 M. J. Dollinger, P. A. Golden and T Saxton
DISCUSSION
The
objective of this study was to examine the impact that reputation has on the
formation of a strategic alliance. We
modified the reputation constructs adapted from the Fortune Corporate
Reputation Survey, and manipulated a target firm's reputation in an
experimental design. The design had
three major components: the effect of reputation, the effect of the target's
position in the production chain, and the information-processing
characteristics of the decision-maker.
There
were five major findings in this experimental study of the effect of reputation
on alliance partner choice. First, firm
reputation counts. It is an important
resource and is capable of attracting other resources in the form of an
alliance partner. The better a firm's
reputation, the more likely it is to be targeted for joint venture activity
and, we believe, similar interorganizational relationships. In fact, Michelet (1992) found that in
cross-border alliances reputation contributed to success in the local markets.
However,
from the managerial point of view, no benefit is without cost. So the reputationbuilding
activities of a firm also come with a cost.
These costs can take three forms.
First, there are the out-of-pocket costs of maintaining public
relations, corporate affairs, and communications offices. For a large firm these may be trivial
compared to total turnover or profits, but for a smaller company, looking to
attract a partner, this may represent a substantial amount.
Second,
there is the extra cost of 'doing good' itself.
Improved quality has a cost, in labor (e.g., training and compensation)
and capital (e.g., investment and inspections).
Companies have gone overboard in the past by focusing only on the
quality variable. A number of Baldrige Award winners have had difficult financial
problems. Third, there is the
opportunity cost of not engaging in activities that have the potential to go
wrong. Managers may be risk averse
enough without the extra burden of maintaining a pristine reputation in all
activities. If companies focus too much
on what the rest of the world thinks of it, it may miss legitimate and highly
profitable opportunities because of the risk of damaging its reputation. So although the results here indicate that a
good reputation is a valuable asset, we recognize that it is not the only asset
in the company's portfolio.
The
study showed that the cognitive characteristics of the decision-makers
influenced the result of the experiment.
People who were more tolerant of ambiguity were more likely to discount
and suppress negative information regarding the target's reputation and proceed
with the alliance. And surprisingly,
they were more likely to suppress the most positive reputation information as
well. The latter result requires
explanation. We may speculate that the
high-TA subjects, comfortable with uncertain and risky situations, imputed risk
where none was explicitly stated to exist.
The positive target reputation combined with the description of high
profitability if the alliance was approved seemed too good to be true to these
subjects. They suppressed the positive
information and the lack of ambiguity made them uncomfortable. Thus the higher tendency to reject the alliance.
Similarly,
we may speculate that the high intemals were
suspicious of something that sounded too good to be true. These individuals are used to attributing
success to personal factors, not the environmental conditions as received in
the experiment. Thus, when faced with
positive information, they had a higher tendency to reject the alliance,
imposing their own will and framework on the situation. The parallels between the TA and LOC results
may be due to the moderate positive correlation (0.34) between the two
variables.
Next,
we found that a firm's reputation is perceived as multidimensional. In this experiment we differentiated between
three components of reputation: product quality and innovation, management
integrity, and financial soundness.
Although a factor analysis of the Fortune survey data indicated a unidimensional variable (Fombrun and Shanley, 1990), we
were able to induce combinations of three dimensions in the mental maps of the
subjects. The evidence for this
multidimensionality (beyond the forcing that we did by the design) is found in
the manipulation checks reported in Table 1. Here, we found that by 'damaging'
one part of the firm's reputation, we by no means damaged the entire
reputation. When one component of the
organization's reputation was 'ruined' by the manipulation, the others were not
equally affected. This indicates that
the subjects were able to separate the different components of reputation.
In
fact, in certain cases a decrease in the
Small
Firm Reputation and Decision Making
reputation
of one component led to an attributed increase in the reputation of another
component. For example, in column 2 of
Table 1, we see that when we 'ruined' the financial reputation of the target
firm, the management reputation also went down, but not very much. This is a simple spillover and to be
expected. The presence of the spillover
in no way detracts from the validity of the manipulations, which are still
strong and consistent. But the same
manipulation caused the product reputation to rise, from its baseline in column
I of 6.05 to its high of 6.41. The subjects have introduced a notion of
compensating reputation. The mental maps
of the target firm created by the manipulations indicate that, in this example,
a firm with a poor financial reputation that still produces high-quality
products is being given even more credit for that quality than when the
financial activities of the firm are in order.
We
can speculate that, in this case, the subjects perceive that the problems that
the firm is having in the financial areas are associated with its commitment to
higher product quality. Of course the
experimental manipulation made no such claim.
But subjects, just like customers, try to make sense of the world. And it seems to make sense that if the
company is experiencing financial difficulties and its products are of high
quality, that the two might be directly related.
The
same effect can be seen from the other end of Table 1. The baseline for
negative managerial reputation (column 8) is 1.30. But when both product
quality and financial soundness are high, the managerial reputation rating goes
up to 2.41. Again, the map created by the manipulations indicates that the
subject compensates for the strong product and financial reputations by raising
the perceived rating of the 'ruined' managerial one. These combinations of perceptual movement,
spillover effects and compensating reputations deserve additional study.
The
important implication of this result is that managers can focus attention on
different parts of their reputation and even compensate in one area for
deficiencies in another. The exception
seems to be with the product reputation.
When the product reputation is negative, there seems to be less ability
to compensate for other reputation attributes.
This is explained in the next section.
We can integrate this conclusion with the previous one concerning the
costs of maintaining and communicating reputation and see that the
firm
faces an objective function of maximizing its reputation, constrained by cost
factors and combinatorial effects.
Again, this deserves additional inquiry.
Fourth,
we found that the 'product by management' interaction was the most powerful
effect. When both the product and
management reputation are positive (regardless of the financial reputation) the
target firm's probability of being chosen for the alliance was greatly
increased. Thus, it seemed to be
somewhat less important that the financial reputation was positive. This has been home out in the IBM-Lotus
situation as well as in other alliances and ventures: if the product quality is
beyond question, the earnings and stock market shortfalls are less
important. Of course, these results and
this particular ordering are heavily influenced by the demand characteristics
of the task: choice of a manufacturing partner for a joint venture. In a situation where the task is more
financially driven, the order might change.
However, for the small and medium-size manufacturer whose fortunes
depend on joint venture and alliance activity, the implication is clear. It is more important to spend and invest in
product quality and innovation than it is to keep a 'clean and tidy' balance
sheet. For all firms there is often a
trade-off between maintaining product momentum and financial soundness, at
least in the short run. This study
provides evidence that when the strategy entails attracting an alliance
partner, a firm should choose investment in product reputation.
The
last finding was the relative unimportance to the decision-maker of the
position of the target in the production chain.
There were no differences in the chances of being chosen as partner
between the supplier and competitor groups.
But, even though our data suggest that the propensity to form an
alliance was not affected by position in the production chain, under different
circumstances, with a different scenario, the motivational characteristics
might prove salient.
CONCLUSION
This
study provides confirmation for research which has suggested that reputation is
a multidimensional construct. While the
dimensions are interrelated, they may be manipulated independently in an
experimental setting to influence a strategic decision. In fact, the data even suggest a hierarchy of
relationships between the dimensions of target reputation and the decision to
proceed with an alliance, with product quality most important, management
second, and financial reputation the least important dimension. This finding is contrary to studies based on
the Fortune data, which suggest that the reputation halo effect is based
primarily on financial performance.
Thus, while the overall importance and effect of a positive reputation
are supported, the more interesting finding relates to the complexities and
interrelationships between the dimensions and the importance of product
quality. Individual decision-maker
characteristics further complicate these relationships. These issues deserve more study in a field
setting.
For
managers, the results affirm the importance of a positive reputation. 'Being good' may not be good enough if it is
not communicated and if the communication does not adhere to the firm's
reputation. A firm needs a proactive
strategy to build and promote its visibility and reputation. The ideas of compensating reputation and
spillover effects, though, would suggest that there are trade-offs involved in
reputation-building with potential consequences for being selected as an
alliance partner. Second, understanding
the cognitive characteristics of the decision-maker may yield insight into the
likelihood of an alliance taking place or being expanded in an uncertain
environment. Finally, decision-makers
may be less reluctant to partner with a competitor than executives
believe. Aligning with competitors
should not be dismissed a priori due to concerns about a potential negative
reaction from the partner. The results
suggest that while concerns about loss of trade secrets and other related
elements may be of greater concern in these confederate relationships,
decision-makers remain willing to engage despite these concerns.
REFERENCES
Astley,
G. and C. Fombrun (1983). 'Collective
strategy: Social ecology of organizational environments', Academy of Management
Review, 8, pp. 576-587. Barney, J.
(1986). 'Strategic factor markets:
Expectations, luck and business strategy', Management Science, 32, pp.
1231-1241.
Barney,
J. (1991). 'Firm resources and sustained
competitive advantage', Journal of Management, 17, pp. 99-120.
Bateman,
T. and C. Zeithami (1989). 'The psychological context of strategic
decisions: A model and convergent experimental findings', Strategic Management
Journal, 10( I), pp. 59-74.
Bresser, R. K.
(1988). 'Matching collective and
competitive strategies', Strategic Management Journal,
9(4), pp. 375-385.
Bresser, R. K. and J. E. Hari
(1986). 'Collective strat-
egy: Vice or virtue?', Academy of Management
Review,
11, pp. 408-427.
Brown,
B. and S. Perry (1994). 'Removing the
financial performance halo from Fortune's "most admired" companies',
Academy of Management Journal, 37, pp.1347-1359.
Budner, S. (1962). 'Intolerance of ambiguity', Journal of
Personality, 30, pp. 29-52.
Caves,
R. and M. Porter (1977). 'From entry
barriers to mobility barriers', Quarterly Journal of Economics, 91, pp.
421-434.
Demaris, A.
(1992). Logit
Modeling: Practical Applications. Sage,
Newbury Park, CA.
Dollinger,
M. (1990). 'The evolution of collective
strategies in fragmented industries', Academy of Management Review, 15, pp.
266-285.
Fichman, M. and D. Levinthal (1991).
'Honeymoons and the liability of adolescence', Academy of Management
Review, 16, pp. 442-468.
Fombrun,
C. and M. Shanley (1990). 'What's in a
name? Reputation building and corporate
strategy', Academy of Management Journal, 33, pp. 233-258. Forrest, J. (July 1990). 'Strategic alliances and the small
technology-based firm', Journal of Small Business Management, pp. 37-45.
Fortune
(21 September 1992). 'Are strategic
alliances working?', pp. 77-78.
Fromkin, H. L. and S. Streufert (1983).
'Laboratory experimentation'. In
M. D. Dunnette (ed.), Handbook of Industrial and
Organizational Psychology. Wiley, New
York, pp. 415-466.
Fryxell, G. E. and J.
Wang (1994). 'The Fortune corporate
reputation index: Reputation for what?', Journal of Management, 20, pp. 1-14.
Gaertner, G. and S. Ramnarayan (1983).
'Organizational effectiveness: An alternative perspective', Academy of
Management Review, 8, pp. 97-107.
Geringer, M. (1988). Joint
Venture Partner Selection.
Quorum,
New York.
Gordon,
M. E., L. A. Slade and N. Schmitt (1986).
'The "science of the sophomore" revisited: From conjecture to
empiricism', Academy of Management Review, 11, pp. 191-207.
Grant,
R. (Spring 1991). 'The resource-based
theory of competitive advantage: Implications for strategy formulation',
California Management Review, pp.114-135.
Hagedoom, J.
(1993). 'Understanding the rationale of
strategic technology partnering: Interorganizational modes of cooperation and sectoral differences', Strategic Management Journal, 14(5),
pp. 371-385.
Hall,
R. (1992). 'The strategic analysis of
intangible resources', Strategic Management Journal, 13(2), pp.135-144.
Hambrick,
D.. and S. Finklestein (1987). 'Managerial discretion: A bfidge
between polar views of organizational outcomes'. In L. L. Cummings and B. M. Staw (eds.), Research in Organizational Behavior, Vol. 9.
JAI Press, Greenwich, CT, pp. 369-406.
Hamel, G., Y. Doz and C. K. Prahalad
(1989). 'Collaborate with your
competitors-and win', Harvard Business Review, 67(l), pp. 133-139.
Harfigan, K. ( 1
986). Managing for Joint Venture Suc-
cess. Lexington Books, Lexington, MA.
Johnson,
W. B. (1993). 'Managerial reputation and
the informativeness of accounting and marketing
measures of performance', Contemporary Accounting Research, 10(i), pp. 305-332.
Knoke, D. and P. Burke
(1980). Log-linear Models.
Sage,
Newbury Park, CA.
Mahoney,
J. and J. Pandian (1992). 'The resourcebased view within the conversation of strategic
management', Strategic Management Joumal, 13(5),
pp.363-380.
Michelet,
R. (1992). 'Forming successful strategic
marketing alliances in Europe', Joumal of European
Business, 4(l), pp. 11-15.
Miller,
D., M. Kets de Vries and J.
Toulouse (1982).
'Top
executive locus of control and its relationship to strategy-making, structure
and environments',
Academy
of Management Joumal, 25, pp. 237-253. Nielsen, R. P. (1988). 'Cooperative strategy', Strategic Management Joumal, 9(5), pp. 475-492.
Oliver,
C. (1988). 'The collective strategy
framework:
An
application to competing predictions of isomorphism', Administrative Science
Quarterly, 33, pp. 543-561.
Penrose,
E. (1959). The Theory of the Growth of
the Firm. John Wiley, New York.
Rao, H. (1994). 'The social construction of reputation:
Certification contests, legitimation, and the
survival of organizations in the automobile industry: 18951912', Strategic
Management Joumal, Winter Special Issue, 15, pp.
29-44.
Rotter,
J. B. (1966). 'Generalized expectancies
for internal vs. extemal locus of reinforcement',
Psychological Monographs, 91, pp. 482-497.
Sears,
D. 0. (1986). 'College sophomores in the
laboratory: Influences of a narrow data base on social psychology's view of
human nature', Journal of Personality and Social Psychology, 3, pp. 515-530. Spector, P.
(1988). 'Development of the work locus
of control scale', Journal of Occupational Psychology, 61, pp. 335-340.
Tsui, A. (1984). 'A role set analysis of managerial
reputation', Organizational Behavior and Human Perforrnance,
34, pp. 64-96.
Webster's
Third New Intemational Dictionary (1961). G. & C. Merriam, Springfield, MA.
Weigelt, K. and C. Camerer (1988).
'Reputation and corporate strategy: A review of recent theory and
application', Strategic Management Joumal, 9(5),
pp.443-454.
Wemerfelt, B.
(1984). 'A resource-based view of the
firm', Strategic Management Journal, 5(2), pp.171-180.
APPENDIX
I
Scenario
outline
You
are the CEO of the Acme Computer Company.
Your firm has been evaluating other firms for a possible joint venture,
and your staff has screened out all the potential partners, save one, Niche
Manufacturing. Niche is a much smaller
company than yours and Niche management has already agreed, in principle, to
the pact. The immediate issue facing you
is to make the final decision and to establish some of the parameters of the
alliance.
A
brief description of the situation is provided below. After reading the scenario, please answer the
questions that follow.
The
Acme Company is considering a technology sharing agreement with the Niche
Manufacturing Company that will enable Acme to cut its costs by tuming over to Niche responsibility for its semiconductor
manufacturing. Over the course of the
agreement it is proposed that Acme will spend between $10 and $50 million to
help Niche implement the new arrangement.
Niche
has in the past been a strong
competitor/supplier
of Acme's in various
markets.*
INSERT
NICHE REPUTATION HERE (see
Appendix
2)
Chip
manufacturing is expensive and is growing more so all the time. As electronic devices on the chips shrink, it
becomes more difficult to keep the manufacturing facilities clean enough so
that specks of dust don't render the chips useless. Acme, which has been cutting costs whenever
possible, says the agreement will save it hundreds of millions of dollars over
the next 5 years by avoiding the cost of the new fabrication facilities. Acme will also be cutting between 100 and 250
jobs as part of the proposed agreement.
You
believe that the agreement will enable your firm, Acme, to decrease new product
development time. Acme will continue to
work on semiconductor techniques, such as chip packaging and design, which it
will contribute to the arrangement with Niche.
The agreement also gives Niche access to new semiconductor manu-
*
Competitor/supplier manipulation.
factoring
technology much sooner than it might ordinarily obtain. It is estimated that you will need to commit
between 30 and 70 highly skilled technical people to collaborate with Niche on
the project.
Such
cost-cutting efforts as this one proposed with Niche have begun to pay off for
Acme. After three poor years, Acme
reported a surpris-
Financial
The
Niche Manufacturing Company is also known for its financial soundness. It has consistently provided investors with
excellent returns and has proved to be a very valuable long-term investment. It has a reputation for efficient and
effective use of corporate assets
facturing technology
much sooner than it might ordinarily obtain.
It is estimated that you will need to commit between 30 and 70 highly skilled
technical people to collaborate with Niche on the project.
Such
cost-cutting efforts as this one proposed with Niche have begun to pay off for
Acme. After three poor years, Acme
reported a surprisingly strong fourth quarter profit of $80.5 million, or 31o
per share; even though revenue dropped 16 percent to $2.46 billion.
APPENDIX
2
Management
The
Niche Manufacturing Company is known for the high quality of its top managers
and their integrity. The executives of
the firm display concem for their community and are
known as responsive to environmental concerns.
They have a reputation for being able to attract, develop and keep
talented people,
The
quality of the top managers of Niche Manufacturing and their integrity is
suspect. The executives display little
concern for their community and are known to be unresponsive to environmental
issues. Niche does not have a reputation
for being able to attract, develop and keep talented people.
The
financial soundness of Niche Manufacturing is suspect. It has consistently provided investors with
below-average returns and has not proved to be a very valuable long-term
investment. It does not have a
reputation for efficient and effective use of corporate assets.
Product/service
quality
The
Niche Manufacturing Company is also respected for the high quality of its
products and services. Customers believe
that the firm's reputation for value, quality products at a reasonable price,
is among the best in the industry. It
has a reputation for developing innovative products.
The
Niche Manufacturing Company's products and services are reputed to be below
industry standards. Customers believe
that the firm's reputation for value, quality products at a reasonable price,
is among the worst in the industry. It
does not have a reputation for developing innovative products.