A DISTINCTIVE COMPETENCIES APPROACH TO MULTIMARKET CONTACT: FIRM
AND FIRM-MARKET LEVEL ANALYSES
by
THOMAS E. WILL
(Under the Direction of ANN K. BUCHHOLTZ)
ABSTRACT
This dissertation proposes a theoretical framework for understanding multimarket contact
(MMC) dynamics across levels of analysis, and then tests for nonprice competitive effects at
both firm and the firm-market analytical levels. I theorize that the profit-enhancing effects of
MMC’s short-term collusive outcomes are more than offset in the long-term by the performancedepleting convergence of MMC’s firm and population level consequences. Multimarket contact
leads to rivalry reduction, mimicry, and myopia. These outcomes mediate firm level competency
depletion over time by decreasing problemistic search, increasing path dependence, reducing
competitive experience, inducing mimetic isomorphism, promoting macrocultures, and locking
the firm into foothold commitments. At the population level, multimarket contact destabilizes
concentration levels, generating patterns of ‘punctuated forbearance’ in which flurries of intense
rivalry interrupt extended periods of tacit collusion. Firm level competence depletion and
population level punctuated forbearance converge to undermine long-term firm performance.
Punctuated forbearance ultimately exposes the very competence depletion that it largely drives
and from which it partially derives.
Empirical analyses are set in the U.S. passenger airline industry. At the firm-market level,
findings indicate a negative relationship between multimarket contact and on-time performance,
suggesting that MMC may undermine competence development in service quality. Analysis at
the firm level employs a new measure introduced in this dissertation. Findings indicate strong
positive relationships between firm level multimarket contact and organizational resource
allocation to promotion and sales and to customer service. Contrary to the competence depletion
hypothesis, firm level MMC seems to amplify rather than dampen inter-firm rivalry along certain
nonprice dimensions. Cumulatively, the empirical results suggest that the implications of
multimarket contact extend well beyond the firm-market level pricing outcomes traditionally
emphasized by the mutual forbearance hypothesis.
INDEX WORDS:
Multimarket Contact, Competitive Dynamics
A DISTINCTIVE COMPETENCIES APPROACH TO MULTIMARKET CONTACT: FIRM
AND FIRM-MARKET LEVEL ANALYSES
by
THOMAS E. WILL
B. A., Duke University, 1993
M. A., Clemson University, 1997
A Dissertation Submitted to the Graduate Faculty of The University of Georgia in Partial
Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
ATHENS, GEORGIA
2006
© 2006
Thomas E. Will
All Rights Reserved
A DISTINCTIVE COMPETENCIES APPROACH TO MULTIMARKET CONTACT: FIRM
AND FIRM-MARKET LEVEL ANALYSES
by
THOMAS E. WILL
Electronic Version Approved:
Maureen Grasso
Dean of the Graduate School
The University of Georgia
August 2006
Major Professor:
Ann K. Buchholtz
Committee:
Allen C. Amason
Gideon D. Markman
iv
DEDICATION
To my parents, Gene and Joanne Will, who have unfailingly encouraged my educational
pursuits. Mom and Dad—your devotion to all of your children is as inspiring as it is endless and
selfless. You are two amazing people.
v
ACKNOWLEDGEMENTS
I am exceedingly fortunate to know Ann Buchholtz. Her sage counsel, both on matters
specific to this project and on concerns quite diffuse, has been an invaluable component of my
experience at The University of Georgia. Ann—your scholarship is impressive, but your
understanding of human nature is sublime. Thanks, Ann, for assisting my acclimation to a new
field with patience, understanding, and enthusiasm! Thank you, as well, to Allen Amason and
Gideon Markman for serving on my committee, for providing valuable advice on my
dissertation, and for your general encouragement over the years.
The administrative staff at the UGA Department of Management has been of immense
help for the past five years. Mary Lou Abbott, Ruth Davis, Dana Myers, and Mary Hillier
contribute in a million different ways to make everything run smoothly. I appreciate all that you
have done on my behalf.
Although they will not read this document—nineteenth-century America claims their
time—William Steirer of Clemson University and Bill Blair of The Pennsylvania State
University have profoundly influenced me. Each is, in my mind, a beau ideal of an academic.
They are intellectuals and they are gentlemen, in equal measure. Very few people so successfully
balance depth and grace.
Marcus—you added a ton of fun to my years in Athens, which kept me sane. I will never
think about Athens without recalling (with a smile) dollar bets, Texas Aggies, Jot ‘em Down, and
the countless other (ultimately important) day-to-day minutia that defined my life there. Chris—
we traveled each stage of this process together, and your generosity helped me immensely. I’m
vi
glad you were my “cohort”! Matt, Mark, Bryan—you guys showed me the ropes and let me
know the real story on everything . . . but I stuck around anyway! You three are as solid as they
come, and I was very lucky it was your program that I joined. To all of you guys and to all of the
graduate students: thanks for the innumerable acts of friendship and support that define the
doctoral experience.
Betsy, Dave, Jon, and Gary—though distance dictates that I see you all less than I would
like, I feel your support. Over the years I have come to appreciate more and more that family is
my base. Rick—I love you, brother, and you are with me still and always.
And finally, thanks to my loyal and majestic four-legged buddy, Dantzler, who endured
regularly irregular hours.
vii
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS.............................................................................................................v
LIST OF TABLES......................................................................................................................... ix
LIST OF FIGURES .........................................................................................................................x
CHAPTER
1
INTRODUCTION .........................................................................................................1
2
THE MULTIMARKET CONTACT LITERATURE....................................................7
3
FIRM LEVEL EFFECTS OF MMC: COMPETENCE DEPLETION........................10
Muted Rivalry..........................................................................................................10
Mimicry and Myopia...............................................................................................20
4
POPULATION LEVEL EFFECTS OF MMC: PUNCTUATED FORBEARANCE .25
System Openness.....................................................................................................26
Competitive Vacuums .............................................................................................27
Punctuated Forbearance ..........................................................................................29
5
JOINT FIRM AND POPULATION LEVEL EFFECTS OF MMC............................31
6
IMPLICATIONS OF THE THEORETICAL FRAMEWORK...................................33
7
FIRM-MARKET LEVEL EMPIRICAL EXAMINATION OF U.S. PASSENGER
AIRLINE INDUSTRY............................................................................................37
Theory Development ...............................................................................................39
Methods ...................................................................................................................44
viii
Results .....................................................................................................................48
Discussion ...............................................................................................................55
8
FIRM LEVEL EMPIRICAL EXAMINATION OF U.S. PASSENGER AIRLINE
INDUSTRY.............................................................................................................59
Multimarket Contact Posture: A New Firm Level Construct..................................61
Pervasive Versus Partial Forbearance Effects of MMCP .......................................65
Methods ...................................................................................................................72
Results .....................................................................................................................79
Discussion ...............................................................................................................88
9
CONCLUSION............................................................................................................99
REFERENCES ............................................................................................................................103
ix
LIST OF TABLES
Page
Table 1: Descriptive Statistics and Correlations for Full Sample of all Nonstops ........................49
Table 2: Descriptive Statistics and Correlations for Carrier Delay ≥ 1 Minute.............................50
Table 3: Descriptive Statistics and Correlations for Carrier Delay ≥ 15 Minutes .........................51
Table 4: Results of OLS Regression Analysis for Carrier Delay ..................................................52
Table 5: Results of Logistic Regression Analysis for Carrier Delay.............................................54
Table 6: Raw Un-weighted MMCP Means and Standard Deviations ...........................................63
Table 7: Carriers Ranked by MMCP .............................................................................................81
Table 8: Descriptive Statistics and Correlations for MMCP Analysis ..........................................82
Table 9: Results of OLS Regression Analysis for Allocation of Resources to Promotion and
Sales ................................................................................................................................84
Table 10: Results of OLS Regression Analysis for Allocation of Resources to Customer Service
.........................................................................................................................................85
Table 11: Results of OLS Regression Analysis for Customer Service Quality.............................87
Table 12: Results of OLS Regression Analysis for Return on Assets...........................................89
Table 13: Results of OLS Regression Analysis for Return on Equity...........................................90
x
LIST OF FIGURES
Page
Figure 1: Convergence of MMC’s Firm, Inter-Firm, and Population Level Effects.......................5
Figure 2: Multimarket Contact and Rate of Innovation.................................................................19
Figure 3: Partial Overlap in Extended Oligopolies........................................................................28
Figure 4: Relationship between MMC and Product / Service Quality .........................................42
Figure 5: Negative Relationship between MMCP and Nonprice Competition .............................66
Figure 6: Positive Relationship between MMCP and Nonprice Competition...............................69
Figure 7: MMCP Equation.............................................................................................................80
1
CHAPTER 1
INTRODUCTION
The mutual forbearance hypothesis (Edwards, 1955) holds that there will be a negative
relationship between multimarket contact (MMC) and intensity of inter-firm rivalry. Extending
oligopoly theorists’ work on the relationship between concentration, collusion, and profitability,
MMC researchers theorize that familiarity and deterrence effects emanating from cross-market
contact generate tacit collusion in the form of mutual forbearance. Mutual forbearance, in turn,
results in higher prices and wider margins in the markets involved. With the negative
relationship between rivalry and profitability largely established in the I/O economics literature,
multimarket contact studies concentrate on the mutual forbearance hypothesis, finding evidence
that MMC mutes rivalry in the banking (Barnett, Greve, & Park, 1994; Heggestad & Rhoades,
1978), savings and loan (Haveman & Nonnemaker, 2000), software (Young, Smith, Grimm, &
Simon, 2000), cement (Jans & Rosenbaum, 1997), hotel (Fernandez & Marin, 1998), hospital
(Boeker, Goodstein, Stephan, & Murmann, 1997), cellular telephone (Busse, 2000; Parker &
Roller, 1997) and newspaper (Fu, 2003) industries. No line of business, however, has served as
the focus of more multimarket contact research than the airline industry, where researchers
identify forbearance effects in the form of higher prices (Evans & Kessides, 1994), increased
revenue per passenger seat mile (Gimeno, 1999; Gimeno & Woo, 1996), wider price-cost
margins (Gimeno & Woo, 1999; Singal, 1996), and lower entry and exit rates (Baum & Korn,
1996, 1999).
2
An element of irony surrounds airline studies’ prominence in the MMC literature. After
all, if multimarket contact fosters mutual forbearance, and mutual forbearance promotes higher
prices and fatter margins, one might anticipate high profitability among firms experiencing as
much cross-market contact as airlines. Financial returns in the airline industry, however, are
notoriously poor. More troubling still, the carriers experiencing the greatest financial distress are
precisely those most enmeshed in multimarket relationships. Major airlines with the most interroute contact have struggled mightily for years, while carriers engaged in far fewer multimarket
relationships, such as Southwest and JetBlue, have proven consistently profitable (Rubin & Joy,
2005; Vasigh & Fleming, 2005). This suggests that the mutual forbearance hypothesis is lacking
in accuracy, exclusivity, or importance. If mutual forbearance arguments hold in all three
respects, then carriers with higher levels of multimarket contact should outperform carriers with
lower levels of multimarket contact at the firm level.
Numerous studies support the mutual forbearance hypothesis’ conclusion that
multimarket contact enhances firm-market level profit margins (Gimeno, 1999; Gimeno & Woo,
1996; Gimeno & Woo, 1999; Singal, 1996). Deductively, then, any inverse or insignificant
relationship between multimarket contact and firm profitability—in the airline industry or
elsewhere—must derive from either the mutual forbearance hypothesis’ lack of importance or
lack of exclusivity. On one hand, mutual forbearance may represent MMC’s only substantive
outcome, yet the relationship between inter-firm forbearance and firm level profitability may be
weak and of relatively little importance, given the great variety of factors contributing to firm
profitability. According to this perspective, chance alone accounts for any apparent negative
relationship between MMC and firm profitability. Environmental or firm-specific factors
unrelated to multimarket contact account for firm performance, under this view, such that no
3
systematic relationship ties MMC to firm performance. On the other hand, mutual forbearance
may not constitute MMC’s sole substantive outcome. Other outcomes, currently unexplored,
may negatively affect firm level profitability. According to this perspective, a negative
relationship between MMC and firm profitability is systematic rather than random.
The former perspective is certainly plausible. In the face of evidence that MMC enhances
firm-market level margins, but in the absence of evidence that MMC affects firm level
performance, it is reasonable to suppose that factors unrelated to MMC drown out MMC effects
at the firm level. The extant literature implicitly adopts this assumption. The object of this
dissertation is to investigate whether a compelling argument supports the latter perspective that a
systematic negative relationship exists between multimarket contact and firm performance. The
broad, pressing question is whether environmental attributes and firm characteristics affecting
performance relate systematically to MMC in ways currently not taken into account. In other
words, do unexplored paths mediate the MMC-performance relationship at levels of analysis
above and below the firm-firm level focused on by existing work?
My purpose is to argue that such mediation paths do operate, and to explain how their
convergence undermines firm performance over time. At the firm level, multimarket contact
reduces the quality and variety of competencies in two ways. First, the muted inter-firm rivalry
resulting from MMC curtails problemistic search and competitive experience and heightens path
dependence. Second, interconnectedness between firms fosters isomorphism and macrocultural
myopia. At the population level, multimarket contact attenuates the concentration stability upon
which mutual forbearance depends. Variance in concentration levels leads to variance in
collusion, such that flurries of intense rivalry emerge in—or punctuate—extended periods of
mutual forbearance. Firms sound in competence maintenance and development might deftly
4
weather these patterns of punctuated forbearance. However, for firms exposed at length to the
competence depleting influences of multimarket contact, punctuated forbearance poses an acute
threat to firm performance. Thus long-term firm performance may suffer from the convergence
of MMC-induced firm level competence depletion and MMC-induced population level
punctuated forbearance. Figure 1 illustrates these cross-level relationships, upon which
subsequent chapters elaborate.
Chapter 2 traces multimarket contact theory’s focus on short-term, market-specific
benefits to its roots in oligopoly theory. As accurate as the prevailing perspective may be in its
particulars, its narrow scope threatens to distort MMC’s comprehensive impact. A fuller
appreciation of the phenomenon’s strategic importance calls for a dynamic theory of multimarket
contact. To this end, I broaden the scope of investigation along each of three dimensions: time,
causality, and level of analysis. The subsequent three chapters explore linkages between shortand long-term consequences, between constructs exogenous and endogenous to extant models,
and between constructs occupying firm, firm-market, and population levels of analysis. Chapter
3 focuses on firm level effects of multimarket contact, chapter 4 emphasizes population level
effects, and chapter 5 considers implications of the convergence of the two. I conclude my
development of an MMC meta-framework with chapter 6, where I discuss implications for
theory, research, and practice.
The broad theoretical framework developed in the first six chapters motivates the
empirical analyses in the final two chapters. Chapter 7 examines firm-market level implications
of multimarket contact. Results indicate a statistically significant negative relationship between
MMC and on-time performance in the U.S. passenger airline industry, suggesting that
multimarket contact may endanger competence development in service quality. Chapter 8 moves
5
FIRM LEVEL
INTER-FIRM LEVEL
P1,2,3,4,5
Competence
Depletion
Muted
Rivalry
P9
Sys tem
Openness
MMC
P6,7,8
POPULA TION LEVEL
Mimicry &
Myopia
P10
Punctuated
Forbearance
Competiti ve
Vacuums
P11
Reduced Long-Te rm
Firm Perf ormance
FIGURE 1: Convergence of MMC’s Firm, Inter-Firm, and Population Level Effects
6
analysis of the same industry to the firm level. I introduce a new firm level MMC construct,
propose an associated measure, and test for outcomes in resource allocation patterns and
profitability. While no profitability outcomes emerged, strong effects were found for the
allocation of resources to promotion and sales and to customer service. However, those effects
contradicted the competence depletion hypothesis, suggesting instead that firm level multimarket
contact may amplify rather than dampen inter-firm rivalry. Potential explanations for rivalry
intensification are discussed in chapter 8.
7
CHAPTER 2
THE MULTIMARKET CONTACT LITERATURE
The multimarket contact literature bears traces of the oligopoly research domain from
which it emerged. Oligopoly research examines conditions promoting mutual recognition of
competitive interdependence—and hence tacit collusion—among firms in a single market. The
investigative process flows causally backward, driven by the search for structural antecedents to
collusive outcomes. In single-market contexts, oligopoly research centers on concentration as an
antecedent. The horizontal interdependence hypothesis (Adams, 1974) anchoring oligopoly
theory holds that concentration facilitates mutual recognition of competitive interdependence,
which in turn fosters tacit collusion and dampens rivalry. Multimarket contact theory and
research emerged as scholars began to consider antecedents to collusion beyond single-market
settings. From the start, the research question motivating this literature has been whether interfirm relationships spanning multiple markets promote mutual recognition of extended
interdependence (Areeda & Turner, 1979). The research model’s prevailing outcome of
interest—tacit collusion—is the raison d’etre of the antecedent construct lending the
multimarket contact literature its name. Rather than ask, “what firm, inter-firm, and population
level outcomes derive from multimarket contact?,” scholars began by asking, and for the most
part have continued to ask, the more limited question, “does multimarket contact facilitate tacit
collusion?”
While narrow in scope, the latter question is grounded in theory. Edwards (1955)
specified the theoretical basis for expecting multimarket contact to foster collusion. Edwards
8
noted that firms confronting one another across multiple markets recognize that a competitive
attack can draw a retaliatory response not only in the attacked market, but at other points of
contact as well. Multimarket contact thereby magnifies the expected retaliatory costs of initiating
an attack, providing firms a strong incentive to withhold first-mover competitive actions
(Karnani & Wernerfelt, 1985). As a result, firms recognizing their extended interdependence will
tend to “mutually forbear” (Edwards, 1955), or tacitly collude in the pursuit of rivalry reduction.
Empirical studies generally support Edward’s mutual forbearance hypothesis. Extant
research proceeds along three distinct levels of analysis (Gimeno & Jeong, 2001). First, I/O
economics scholars conceptualize MMC as a market characteristic, measuring the overall degree
of multimarket contact among firms serving a focal market. Feinberg (1985), for example, finds
a positive relationship between industry-wide measures of MMC and industry-wide price-cost
margins. Evans and Kessides (1994) and Singal (1996) conclude that the average number of
external contacts between airlines in a given route positively affects fare levels in that route. Jans
& Rosenbaum (1996) find that cement prices correlate with geographic market MMC levels.
While I/O economics researchers conceptualize and measure MMC as a market
characteristic, management scholars approach MMC directly as a characteristic of the
relationship between firms (Gimeno & Jeong, 2001). Within the management literature, in turn,
multimarket contact is treated at two distinct levels of analysis. Most research approaches MMC
at the firm-market level of analysis, measuring the level of cross-market contact that a firm has
with incumbents in a focal market. For example, Boeker et al. (1997) find a negative relationship
between the extent to which a hospital meets focal market competitors in other markets and that
hospital’s likelihood of exiting the focal market. Gimeno & Woo (1996, 1999) show that an
airline’s MMC with incumbents in a given route tends to increase the prices charged by that
9
airline in that route. Baum and Korn (1996) also find that airline-in-route MMC levels tend to
mute rivalry, but the authors use market entry and exit as dependent variables reflecting levels of
collusion. Other management scholars conceptualize and measure the construct at the dyadic
level of analysis. Rather than measure multimarket contact between all firms in a given market
(market level), or measure MMC between a focal firm and all incumbents in a focal market
(firm-market level), dyadic research seeks to capture an overall level of MMC between two firms
across all of the markets in which the two meet. Baum and Korn (1999), for example, find that
the multimarket contact between two airlines across all of the markets in which they meet bears
an inverted “U-shaped” relationship with market entry and exit. Korn and Baum (1999) examine
antecedents to dyadic multimarket contact.
Cumulatively, empirical work in MMC sheds considerable light on the extent to which
cross-market contact between firms affects their competitive behavior toward one another.
However, existing MMC theory and research skirt the central question pertinent to any topic in
the strategy field – how does the phenomenon under investigation affect firm performance?
Grasping the relationship between MMC and firm performance entails consideration of
interactions between multiple outcomes over an extended period. Existing studies of the MMCrivalry relationship assume fixed firm competencies along with stable and high market
concentration levels. Is it safe to assume, however, that multimarket relationships between firms
do not systematically affect firm and population attributes? A dynamic theory of MMC must
explore how MMC influences competencies and concentration levels over time, while
considering feedback loops crossing levels of analysis (Porter, 1991). The next three chapters lay
the groundwork for such a dynamic theory of MMC.
10
CHAPTER 3
FIRM LEVEL EFFECTS OF MMC: COMPETENCE DEPLETION
This chapter examines firm level effects of multimarket contact. I first analyze ways in
which muted rivalry—an established outcome of multimarket contact—undermines
competencies by reducing problemistic search, entrenching path dependencies, and diminishing
competitive experience. Subsequently, I explore debilitating tendencies toward mimicry and
myopia that derive from the interconnectedness between firms in multimarket contact.
Interconnectedness promotes isomorphism, macrocultures, and foothold commitments that
compromise the quality and variety of firm competencies.
MUTED RIVALRY
Former Quaker Oats CEO Bill Smithburg quipped in 1995, “Competition is a way of life.
If you don’t have a really tough competitor, you ought to invent one.” (Sellers, 1995).
Smithburg’s comment reflects the widespread belief that competition promotes excellence. The
“vague suspicion that competition is the enemy of sloth” (Caves, 1980: 64) is cardinal to
economic theory. Viable competitors motivate firms to reduce costs, improve products, and stay
abreast of technological change (Porter, 1985: 206; Porter, 1991). Porter’s extensive study of
national competitive advantage indicates a strong relationship between the vigor of domestic
rivalry and the persistence of competitive advantage in an industry (Porter, 1990: 117-122).
Numerous I/O economics studies report negative relationships between competitive intensity and
such competency-related outcomes as efficiency, total factor productivity, product/service
quality, marketing spending, and R&D intensity. For example, high levels of concentration
11
correlate with reduced technical efficiency (Caves, 1992; Caves & Barton, 1990), diminished
total factor productivity (Nickell, Wadhwani, & Wall, 1992), lower total factor productivity
growth (Nickell, 1996), decreased service quality (Mazzeo, 2003), and decreased product quality
(Banker, Khosla, & Sinha, 1998).
If rivalry reduction deriving from high concentration depletes competencies, then it
stands to reason that rivalry reduction deriving from multimarket contact depletes competencies
as well. Indeed, evidence suggests that MMC decreases the resources firms devote to marketing
functions such as advertising, promotion, and sales force deployment when introducing new
products (Shankar, 1999). Relationships between MMC and firm decisions to allocate resources
toward competency enhancement in other functional realms remain unexplored, even though
theoretical grounds exist for proposing negative relationships between MMC and a broad range
of competencies related to efficiency, differentiation, learning, and innovation.
Problemistic search, path dependence, and competitive experience. The extent to which a
market is competitive holds powerful implications for firm search, choice, and learning. The
competitive environment affects stimuli to problemistic search (Cyert & Mach, 1963: 169-171);
it shapes interpretations and outcomes of past routines and actions, upon which present actions
are path dependent (Levinthal & March, 1993); and it influences the extent to which firms
undertake competitive actions and thus learn by doing (Levitt & March, 1988). The dampening
of inter-firm rivalry, therefore, affects firm competencies through the mediating processes of
problemistic search, path dependence, and competitive experience.
Problemistic search is search stimulated by a problem and directed toward finding a
solution to the problem (Cyert and March, 1963: 169; March & Simon, 1958: 194). According to
evolutionary economic theory (Nelson & Winter, 1974, 1982: 173) and organizational evolution
12
(Tushman & Romanelli, 1985), firm search tends to be problem-oriented or failure-induced. Kim
(1988) and Winter (2000) show how organizational performance crises, termed “internal
activation triggers” by Zahra and George (2002), intensify firm efforts to achieve and learn new
skills. Inter-firm competition induces performance crises, because the more intensely an
organization competes with others to achieve its objectives, the more likely are results to fall
short of expectations (Barnett, Greve, & Park, 1994). Given that competition stimulates
problemistic search (March, 1988), and problemistic search enhances firm competencies, it
follows that competition enhances firm competencies. Over time, this evolutionary adjustment of
the firm and its referent rivals develops into a self-reinforcing process that has been termed ‘Red
Queen’ evolution (Barnett & Hansen, 1996). Inter-firm competition triggers learning, which
increases an organization’s competitive strength, which in turn triggers learning in its rivals and
thus strengthens competitors, and reciprocally so on.
Muted rivalry, however, blunts the cycle of Red Queen evolution. To the extent that
multimarket contact dampens rivalry, it serves as a positional advantage shielding the firm in the
short-term from performance failure. Muted rivalry conceals problems and thus curtails the
search for answers. Organizational success may make managers complacent with the status quo
(Miller & Friesen, 1984) and blind them to the need for action (Lant, Milliken, & Batra, 1992;
Miller & Chen, 1994). In sum, performance pressures and crises are less prevalent in the absence
of intense competition; problemistic search is less operative in the absence of performance
pressures; and organizational learning and competence enhancement are less prevalent in the
absence of problemistic search. Consequently, I propose the following:
Proposition 1: The more inter-firm rivalry is muted, the less performance
pressures trigger problemistic search and competence-enhancing organizational
13
learning. Thus, a firm’s level of multimarket contact with rivals is negatively
related to that firm’s competence development.
Path dependence, like problemistic search, helps explain how muted rivalry mediates a
negative relationship between multimarket contact and competence development. Whereas the
concept of problemistic search addresses firm search proclivities, the concept of path dependence
emphasizes response behaviors, building on the observation that organizational actions are
history-dependent (Levitt & March, 1988). Path dependence emerges in the extended absence of
problemistic search. Organizations prefer continuation of present programs over alternatives that
represent change, so they do not search for or consider alternatives to the present course of action
unless that present course is deemed unsatisfactory (March & Simon, 1958: 194). In other words,
when an organization meets with success, its managers tend to replicate and perpetuate routines
and actions they perceive as responsible for that success. An organization’s dependence on a
historically successful path becomes evident when its response to challenges is conditioned and
constrained by that path (Arthur, 1989, 1994; Levinthal & March, 1993).
The potentially dysfunctional consequences of path dependence manifest in a number of
‘traps’ jeopardizing competence development. A competency trap can occur when favorable
performance with a procedure leads an organization to accumulate more experience with it, thus
keeping experience with alternate, potentially superior procedures inadequate to make them
rewarding to use (Levitt & March, 1988: 322). The organization’s success with learned
competencies ‘traps’ it into continued reliance on those competencies at the expense of
developing or adopting more optimal competencies. Firms caught in competency traps become
increasingly removed from other bases of experience and knowledge, exacerbating their
vulnerability to environmental change (David, 1985; Levinthal & March, 1993: 102). Ahuja and
14
Lampert (2001) specify several types of competency traps. Familiarity traps occur when the
mutual positive feedback between experience and competence renders the refinement of familiar
technologies and procedures preferable to the exploration of new ones, and propinquity traps
result from the organization’s predisposition to look for new solutions near old solutions when
exploration is pursued (Ahuja & Lampert, 2001).
Environments characterized by muted inter-firm rivalry increase the likelihood of
competency traps. In shielding the firm from performance failure feedback, muted rivalry
promotes path dependence. Managers attribute organizational success to existing routines and
technologies, decreasing their propensity to learn or even consider alternate procedures. If and
when the firm performs poorly, path dependence constrains managerial responses to ‘doing more
of the same.’ Levinthal and March (1993: 102)) link market power to path dependence, noting
the tendency among firms with strong market positions to “impose their policies, products, and
strategies on others, rather than learn to adapt to an exogenous environment.” Environmental
change beyond the firm’s control exposes underdeveloped adaptive skills (Levinthal & March,
1993: 102). Consequently, I propose the following:
Proposition 2: The more inter-firm rivalry is muted, the more likely managers are
to replicate and perpetuate existing routines and technologies, and the less likely
they are to identify and assimilate new routines and technologies. Thus, a firm’s
level of multimarket contact is negatively related to that firm’s breadth of
competence development and adaptive skill.
Muted rivalry affects competitive experience as well as search and response. A critical
element of organizational learning is “learning by doing” (Levitt & March, 1988: 321-22). Firms
develop expertise in those activities that they perform repeatedly over time. The more
15
competitive actions a firm has taken in the past, the wider its knowledge base will be, and the
more skilled, tactile, and efficient it will become at taking future competitive action (Amburgey,
Kelly, & Barnett, 1993; D’Aveni, 1994). A firm’s competitive repertoire is affected by the range
of its own past competitive actions (Miller & Chen, 1994). Young, Smith, and Grimm (1996:
247) explain why the maintenance and enhancement of a productive asset base requires
undertaking activities: “In building on asset strengths, the cost of taking action is lower for the
firm that has efficiencies derived from a rich history of prior activity. Importantly, the firm with
a rich history of activity-derived learning not only has lower costs of supporting superior
performance, but also is capable of undertaking more activities in a given time period.” By
definition, firms engaged in muted rivalry pursue fewer competitive actions than do firms
experiencing intense rivalry. Indeed, studies show that as multimarket contact increases, firms
are less likely to initiate tactical competitive attacks such as price changes (Young et al., 1996;
Young et al., 2000). Dampened rivalry reduces the firm’s competitive experience and skill along
such dimensions as differentiation, cost efficiency, and launching and responding to competitive
assaults. Consequently, I propose the following:
Proposition 3: The more inter-firm rivalry is muted, the lower a firm’s
competitive experience. Thus, a firm’s level of multimarket contact is negatively
related to that firm’s competitive competence.
Rate of innovation and innovative diversity. The preceding discussion of problemistic
search, path dependence, and competitive experience as behavioral dynamics mediating the
relationship between muted rivalry and competence development pertains to a potentially broad
range of competencies, including skills integral to differentiation strategies, cost leadership
strategies, absorptive capacity, and product and process innovation. Innovative competencies
16
merit further elaboration, however. In particular, where distinct pre- and post-innovation markets
exist, competing lines of thought pervade extant theory and bring complexity to the relationship
between muted rivalry and innovative competencies. Propositions are needed to resolve the
countervailing tendencies toward competition-induced innovation on the one hand and
Schumpeterian innovation on the other.
Managerial propensity toward problemistic search informs the general expectation that
weak competition reduces the motivation for innovative activity (Cyert & March, 1963: 188-190;
March & Simon, 1958: 203-207). This view of competition as a stimulant to innovation, argued
over four decades ago by Arrow (1962), finds support in subsequent empirical studies. Geroski
(1990) and Blundell, Griffith, and Van Reenen (1995), for example, find that concentration
dampens innovative activity, while Boone (2001) shows that innovation confers more value to
the firm when competition is intense. Delbono and Denicolo (1990) show that incentives to
introduce cost-reducing innovations are greater where firm decision variables are prices rather
than output levels. Since price competition typically leads to higher output and lower prices than
output-level competition, the former can be thought of as a more intense form of competition
(Bonanno & Haworth, 1998). Thus, the Delbono and Denicolo study (1990) supports the position
that the incentive to innovate is greater under more intense competition.
An alternate line of reasoning, associated with Schumpeter (1942), maintains that
concentration rather than competition serves as a stimulus to innovation. The Schumpeterian
argument rests on two points of logic. On the one hand, ex ante market power possessed by a
firm facing little competition provides the firm the financial wherewithal to invest in risky
innovation. Stable cash flows render the firm capable of pursuing innovation. The behavioralist
approach reflects this point of argument with the notion that ‘slack innovation’ may occur in
17
firms with ample resources (Cyert & March, 1963: 189). On the other hand, ex post market
power provides a firm facing little competition with powerful incentive to make innovative
investments. The firm does not anticipate rents being competed away on the post-innovation
market (Conner, 1991; Schumpeter, 1942: 82-88).
The Schumpeterian perspective on innovation and the view of innovation as competitioninduced represent countervailing logics, but the two are not mutually exclusive. The influence of
the opposing dynamics on the overall relationship between rivalry and rate of innovation is best
captured by disaggregating that relationship into the effect that rivalry in the pre-innovation
market has on innovation and the effect that rivalry in the post-innovation market has on
innovation. Numerous studies linking seller concentration to R&D intensity find an “inverted-U”
shaped relationship (Scott, 1993: 136). This pattern emerges because of the ease with which
price coordination is achieved in post-innovation markets relative to the difficulty with which
coordination for R&D is achieved in pre-innovation markets. Low levels of concentration mute
rivalry in neither market. Thus, competitive incentives in pre-innovation markets are
counterbalanced by the likelihood of post-innovation rents being competed away. As a result,
innovation is relatively low. As concentration levels rise, coordination is more easily obtained in
post-innovation markets than in pre-innovation markets (Scott, 1993). At moderate concentration
levels, therefore, intense competition in pre-innovation markets continues to stimulate innovation
while muted rivalry in post-innovation markets also induces innovation. As a result, innovation
rates are highest at moderate levels of concentration. Finally, as concentration levels further
increase, rivalry is muted in pre-innovations markets as well as post-innovation markets,
removing one innovation stimulant while perpetuating the other. As a result, at high
concentration levels, innovation levels dip downward again.
18
The above logic supports the existence of an “inverted-U” shaped relationship
(represented in Figure 2) between MMC and innovation in extended oligopolies similar to the
one found between concentration and innovation in single-market oligopolies. Consequently, I
propose the following:
Proposition 4: The allocation of resources to innovation is greatest when tacit
collusion is powerful enough to mute rivalry on the post-innovation market, but
not powerful enough to mute rivalry on the pre-innovation market. Thus, a firm’s
level of multimarket contact bears an inverted U-shaped relationship with that
firm’s rate of innovation.
Rivalry affects innovative diversity as well as rate of innovation. Intense competition in
the post-innovation market motivates firms to pursue divergent research strategies. Seeking to
avoid the erosion of post-innovation rents, firms attempt to generate the least imitable
innovations possible. Competition provides an incentive for firms to pursue diverse research
strategies aimed at producing a unique product or process bearing a decisive advantage that is
costly or difficult to imitate (Scott, 1993: 148-166). Muted rivalry, therefore, likely mediates a
negative relationship between multimarket contact and innovative diversity. Therefore, I propose
the following:
Proposition 5: Firms confronting intense rivalry in the post-innovation market
pursue diverse innovation paths according to the logic that inimitability protects
post-innovation rents. The incentive to attain inimitability via innovative diversity
declines as rivalry becomes more muted. Thus, a firm’s level of multimarket
contact is negatively related to that firm’s innovative diversity.
Rate of Innovation
19
Pre-Innovation Mar ket Rivalry
Post-Innovation Market F orbearanc e
Pre-Innovation Mar ket Rivalry
Post-Innovation Market Ri valr y
Low
Pre-Innovation Mar ket Forbearance
Post-Innovation Market F orbearanc e
Moder ate
Multimarket Contact
FIGURE 2: Multimarket Contact and Rate of Innovation
High
20
MIMICRY AND MYOPIA
The relationship between multimarket contact and firm level competence depletion is
partially—not fully—mediated by muted rivalry. This is significant, because rivalry is often
unevenly dampened by MMC across competitive dimensions. For example, multimarket
competitors colluding on price may compete intensely along such other competitive dimensions
as advertising or new product innovation. Even in such contexts, multimarket contact inclines
firms toward various forms of mimicry and myopia entailing their own set of competencedepleting consequences. The interconnectedness and inter-firm familiarity inherent in
multimarket contact encourage homogenizing isomorphism and myopic macrocultures.
Additionally, the extended interdependence of firms in multimarket contact orients deployment
of resources toward foothold commitments that potentially limit resource sharing opportunities.
Mimetic isomorphism and macroculture. The psychological concept of social proof and
the sociological concept of information cascades (Surowiecki, 2004) suggest that individuals
base decisions about their own behavior in part on observed referent group behavior. Mimicry of
this sort reflects not a conformist need for social acceptance, but rather the rationale that ‘if
others are doing it, it must have value.’ The institutional literature captures this dynamic at the
organizational level with the concept of mimetic isomorphism, or firms’ tendency to respond to
uncertainty by modeling referent others (DiMaggio & Powell, 1983: 151). When firms face
problems with ambiguous causes or unclear solutions—which they frequently do—they seek
solutions in the actions and organizing routines of other firms they perceive dealing successfully
with similar problems. Mimetic isomorphic pressures are strongest where the interconnectedness
(Oliver 1991) and interdependence (DiMaggio & Powell, 1983) between firms is highest. Firms
confronting one another across multiple markets share a high degree of interconnectedness,
21
extended interdependence, and familiarity, such that they should be strongly affected by the
homogenizing influence of mimetic isomorphism. While institutional isomorphism can enhance
firm legitimacy and thus, in certain contexts, firm success, mimetic isomorphism nevertheless
works counter to the development of unique organizational strategies, routines, and technologies
(Oliver, 1997). Consequently, I propose the following:
Proposition 6: The greater the interconnectedness and familiarity between firms,
the greater the mimetic isomorphic pressures acting on those firms. Thus a firm’s
level of multimarket contact is negatively related to that firm’s development and
adoption of rare or unique competencies.
The concept of macroculture, like the concept of isomorphism, pertains broadly to the
issue of similarity between firms. Whereas institutional theory examines the sources of
homogenization in structures and activities, however, the macroculture literature explores the
existence and outcomes of managerial perceptions of inter-organizational similarities. Perceived
similarities between firms derive from common values, frames of reference, norms, and
expectations (Cassidy & Loree, 2001). A macroculture consists of “the relatively idiosyncratic,
organization-related beliefs that are shared among top managers across organizations”
(Abrahamson & Fombrun, 1994: 730). A macroculture’s influence reflects the extent to which
organizations’ top managers share a mutual perception of one another as close competitors
(Abrahamson & Fombrun, 1994). Multimarket contact fosters macrocultures (Cassidy & Loree,
2001) because the more firms confront one another across multiple markets, the more likely they
are to perceive one another as close competitors. The familiarity that facilitates mutual
forbearance among MMC cohorts promotes, as well, managerial perceptions of intra-group
similarity in strategic priorities and beliefs about competitive boundaries.
22
Strong macrocultures affect member firms’ strategic vision and decision making. Where
a set of competing firms confront one another repeatedly in multiple domains, top managers tend
to focus on one another to the exclusion of firms, technologies, and modes of operating outside
the group. Thus, macrocultures are inertial in their tendency to blind members to competitive
opportunities and threats originating beyond perceived competitive boundaries (Abrahamson &
Fombrun, 1994). This strategic myopia weakens member firms’ adaptive competencies. The
learning literature strikes a like note in its discussion of spatial myopia (Levinthal & March,
1993), while the strategic groups literature similarly recognizes the tendency of highly
interconnected competitors to overly focus attention on one another and away from outside
competitors (Peteraf & Shanley, 1997).
Macrocultures tend, as well, to blind members to knowledge sources beyond perceived
competitive boundaries (Abrahamson & Fombrun, 1994). Firms defining their competitive
environment narrowly operate in a sterile and homogeneous learning environment (Miller &
Chen, 1994). In curtailing organizational attention to external innovative trends, this strategic
myopia decreases both rate and diversity of innovation. Myopic effects of multimarket contact
with regard to innovation are particularly significant in light of the emerging literature on
knowledge transfer effects of MMC. Several studies find a positive correlation between MMC
and cross-citation of patents, suggesting that MMC promotes knowledge transfer between
participant firms (Cassidy & Loree, 2001; Greve and Mitsuhashi, 2004; Scott, 2001). Myopic
influences on innovative rates and diversity, currently unexplored empirically, may offset or
outweigh within group positive effects on knowledge transfer with between group negative
effects on knowledge exploration and acquisition. Consequently, I propose the following:
23
Proposition 7: The greater the interconnectedness and familiarity between firms,
the more those firms share macrocultural beliefs, such that the narrower is the
range of opportunities, threats, and knowledge sources perceived and acted upon
by those firms. Thus, a firm’s level of multimarket contact is negatively related to
that firm’s adaptive competencies and innovative rate and diversity.
Foothold commitments. While isomorphic and macrocultural processes reflect social
dynamics associated with inter-organizational connectedness, foothold commitments represent a
form of mimicry rooted in economic logic. I define foothold commitment as the presence that a
firm maintains in a particular market at least partly for the purpose of inducing or perpetuating
mutual forbearance. The mutual forbearance hypothesis specifies the rivalry-muting
consequences of the retaliatory potential possessed by firms in multimarket contact. Both logic
and empirical evidence suggest that firms value the short-term benefits of muted rivalry and, as a
result, base market entry and exit decisions at least in part on their desire to establish or preserve
mutual forbearance arrangements. Haveman and Nonnemaker (2000), for example, find that
propensity to enter a given focal market first increases, then decreases, as the firm’s level of
contact in other markets with incumbents in the focal market increases. In other words, firms low
in MMC with incumbents are highly likely to enter the incumbents’ focal market because they
are not yet highly exposed to retaliation, and because they seek to raise their level of MMC to the
point that mutual forbearance can be established with rivals. On the other hand, when a firm
establishes a high level of MMC with incumbents, that firm proves less likely to enter the
incumbents’ focal market because it is highly exposed to retaliation. Additionally, several studies
have found that firms with high levels of MMC in a given market are less likely to exit that
market than are firms with low levels of MMC (Barnett, 1993; Baum & Korn, 1996; Boeker et
24
al., 1997). These results indicate that firms with high levels of MMC value the rivalry-muting
effects of mutual forbearance and consequently hesitate to relinquish those benefits. In sum,
forbearance considerations entice firms to enter some markets they would not otherwise enter,
and to stay in some markets they would otherwise exit (Stephan & Boeker, 2001).
The tendency of firms to base market entry and exit decisions at least partly on their
desire to establish and preserve forbearance arrangements may detract from their capacity to
locate and pursue internal resource sharing opportunities. Devoting resources and managerial
attention to foothold commitments diverts resources and attention away from potential resourcesharing opportunities in other markets. This is not to suggest that foothold commitments entirely
replace or preclude firm pursuit of scope economies or other synergies; rather, I suggest that at
the margins, forbearance objectives may dilute resource-sharing objectives. In short, forbearance
considerations may crowd out resource-sharing considerations in scope decisions. A substantive
body of literature, in turn, supports the importance of resource-sharing to competence
development (Lengnick-Hall & Wolff, 1999; Prahalad & Hamel, 1990). Consequently, I propose
the following:
Proposition 8: The more managerial attention is focused on establishing and
preserving forbearance benefits deriving from MMC, the less it will be focused on
entering and remaining in markets with strong resource-sharing opportunities.
Thus, a firm’s level of multimarket contact is negatively related to that firm’s
competence development.
25
CHAPTER 4
POPULATION LEVEL EFFECTS OF MMC: PUNCTUATED FORBEARANCE
Multimarket contact affects characteristics of firm populations as well as attributes of
member organizations. In particular, the extent of multimarket contact within a population
influences concentration stability. Extended oligopolies are more open systems and thus more
accessible to new entrants than are single-market oligopolies. Additionally, firm level variety
reduction and competence depletion deriving from MMC aggregates to the population level,
creating competitive vacuums that induce new entry. The confluence of system openness and
competitive vacuums de-stabilizes concentration levels in MMC contexts, generating patterns of
punctuated forbearance in which convergent periods of muted rivalry are periodically
interrupted by rivalrous reorientations.
Semi-stability in concentration levels has powerful implications for the preservation of
mutual forbearance. The MMC literature’s focus on the MMC-forbearance relationship should
not obscure the importance of concentration to the development of tacit collusion, in multimarket
contexts just as in single-market contexts. While oligopoly studies explicitly operationalize
market concentration far more often than do MMC studies, Scott (1982, 1991, 1993) has shown
that multimarket theory collapses in the absence of high concentration levels. In fact, in markets
with low levels of concentration, multimarket contact is negatively associated with firm
profitability. The greater the number of firms confronting each other across multiple markets, the
more familiarity declines, the more unwieldy coordination becomes, the more difficult and costly
monitoring defection becomes, and the more mutual recognition of competitive interdependence
26
fades. Absent mutual recognition of interdependence, tacit collusion collapses and rivalry ensues.
In sum, persistently high concentration levels represent the foundation upon which the entire
edifice of current multimarket contact theory rests. A potential causal relationship between
MMC and concentration destabilization at the population level, therefore, is a matter of critical
importance to multimarket theory.
SYSTEM OPENNESS
The persistence of high concentration in a single-market oligopoly is contingent upon
sustained barriers to entry in that market alone. Extended oligopolies, on the other hand, are
more open to entry by virtue of their greater structural complexity. Each market occupied by a
cohort of firms engaged in multimarket competition must remain highly concentrated in order for
incumbents to mutually recognize interdependence and monitor/enforce coordination. Contact in
one market affects intensity of rivalry in another market (and vice versa) if and only if
concentration is—and remains—high in both markets. A simple example illustrates the greater
system openness in extended oligopolies relative to single-market oligopolies. Consider two
scenarios: a five-firm single-market oligopoly, and a five-firm extended oligopoly in which each
of the five firms confronts each of the other four firms in each of three markets. All other things
being equal, the probability of new entry is three times higher in the MMC scenario than in the
single-market scenario, simply because three markets are vulnerable to entry as opposed to one.
The extent of MMC system openness is often even greater that suggested by the simple
five-firm, three-market extended oligopoly example introduced above. A firm in multimarket
contact typically meets one competitor in a certain set of markets, another competitor in another
set of markets that partially but not fully overlap with the first, a third firm in another set of
markets that partially but not fully overlap with the first and second, and so forth. Partial overlap
27
of market contact destabilizes concentration levels even further than full overlap. Under
conditions of partial overlap, entry into a market in which a firm does not even operate can
initiate a series of rivalrous actions and reactions that reverberate to affect collusive
arrangements in markets in which the firm does operate. Consider, for example, five firms that
operate in three markets each, confronting each other across five markets such that each firm is
in multimarket contact with at least two others (see Figure 3). Entry and subsequent decreases in
concentration in one market can undermine forbearance arrangements in a chain reaction down
the line, potentially engulfing all five firms. In this scenario, fully five markets are vulnerable to
entry, or five times the amount in a single-market oligopoly. Consequently, I propose the
following:
Proposition 9: The large number of entry points in extended oligopolies relative
to single-market oligopolies entails a higher probability of new entry in the former
than the latter. Thus, mutual forbearance deriving from multimarket contact is less
stable and enduring than tacit collusion in single-market oligopolies.
COMPETITIVE VACUUMS
Firm level effects of multimarket contact express themselves aggregately at the
population level. As previously discussed, muted rivalry, mimicry, and myopia mediate
reductions in the quality and variety of organizational competencies. Over time, the affected
population of firms comprising a given MMC cohort will begin to suffer from a fixed and narrow
capacity to meet customer wants. This group condition might remain unnoticed in the absence of
new entrants. Ironically, however, homogeneity and atrophy characterizing the population’s
competencies induce the very new entrants that expose the state of those competencies. While
multimarket contact may mute rivalry between involved firms, competition and competency
28
Firm 1
Mkt 1
Firm 3
Mkt 2
Firm 2
Mkt 3
Mkt 4
Firm 4
FIGURE 3: Partial Overlap in Extended Oligopolies
Mkt 5
Firm 5
29
development are certain to proceed outside the bounds of the MMC cohort. A portion of
competency development outside the MMC cohort is likely, eventually, to pertain to the needs of
customers of the firms in multimarket contact. Profitable discrepancies between the knowledge
and competencies possessed by those within MMC networks and those outside MMC networks
are likely to be seized upon by entrepreneurs. Competitive vacuums do not remain unfilled in
perpetuity. Consequently, I propose the following:
Proposition 10: Multimarket contact reduces the quality and variety of
competencies possessed by a population of firms, creating competitive vacuums
that ultimately induce market entry by new firms.
PUNCTUATED FORBEARANCE
System openness and competitive vacuums combine to generate patterns of forbearance
and rivalry I term punctuated forbearance, borrowing from the concepts of convergent periods
and reorientations enumerated in Tushman and Romanelli’s (1985) punctuated equilibrium
model of organizational evolution. The punctuated equilibrium model maintains that
“organizations progress through convergent periods punctuated by reorientations which demark
and set bearings for the next convergent period” (Tushman & Romanelli, 1985: 173).
Convergent periods consist of relatively long time spans of incremental change and adaptation
during which organizational structures, systems, controls, and resources are increasingly coaligned. Reorientations are relatively short periods of discontinuous change during which
organizational attributes are fundamentally transformed toward a new basis of alignment
(Tushman & Romanelli, 1985). The framework embodied by the punctuated equilibrium model
of organizational evolution captures the dynamics I propose among populations of firms in
multimarket contact. Prolonged convergent periods witness increasingly collusive and myopic
30
co-alignment between firms, as repeated interaction over time entrenches familiarity,
interdependence, and forbearance norms. Co-alignment within an MMC cohort slowly builds
competency gaps between the cohort and external firms/entrepreneurs. Competitive vacuums
operate in conjunction with system openness to induce reorientations—the ‘punctuation’ in
punctuated forbearance. While convergent periods represent the forestalling of competition,
reorientations reflect its inexorability. Brief and infrequent, reorientations consist of new entry of
sufficient scale to undermine existing forbearance arrangements, thus transforming the
competitive landscape. The frequency and magnitude of reorientations are likely to be inversely
related.
31
CHAPTER 5
JOINT FIRM AND POPULATION LEVEL EFFECTS OF MMC
The framework advanced thus far encompasses and reconciles, in the context of
multimarket contact, two competing views on the source of above average firm performance.
One view, rooted in the structure-conduct-performance paradigm of industrial organization
(Bain, 1959), emphasizes the actions firms take to create defensible positions against competitive
forces (Caves & Porter, 1977; Porter, 1980; Teece, Pisano, & Shuen, 1997). This ‘positional
advantage’ view is countered by a second major approach that regards superior performance as
the result of idiosyncratic organizational attributes (Barnett et al., 1994). This ‘distinctive
competencies’ or ‘resource-based’ perspective emphasizes the development of rare, valuable,
non-substitutable, and inimitable firm capabilities and assets (Wernerfelt, 1984; Barney, 1991).
Both perspectives, I contend, are necessary to a dynamic theory of multimarket contact. The
positional advantage view accounts for performance outcomes in the short-term, while the
distinctive competencies view is critical to understanding long-term firm performance.
The mutual forbearance hypothesis captures the positional advantage conferred upon
firms by multimarket contact. Mutually recognized interdependence fosters tacit collusion, which
shields member firms from competitive pressures. Empirical studies of multimarket contact
reflect this positional advantage in findings of higher prices (Evans & Kessides, 1994; Fernandez
& Marin, 1998; Jans & Rosenbaum, 1996) and fatter price-cost margins (Gimeno & Woo, 1999;
Hughes & Oughton, 1993; Parker & Roller, 1997; Singal, 1996) at the firm-market level. My
framework recognizes short-term positional advantages with propositions one and two.
32
I depart from extant theory by integrating the distinctive competencies perspective into
consideration of performance outcomes. The distinctive competencies view inherently orients
attention toward the long term. The processes mediating reduction in competence quality and
variety do not generate immediate effects. Organizational decision making tendencies reflected
in the concepts of problemistic search, path dependence, experience, isomorphism, macroculture,
and footholds are gradual, cumulative affairs. Additionally, in the MMC context, positional
advantage disguises competency effects. Indeed, competence depletion largely derives from the
blunting of competitive forces. Firm performance does not reflect competence depletion as long
as mutual forbearance persists. I have argued, however, that positional advantage affects
competencies, which recursively affect positional advantage. The dissolution of mutual
forbearance exposes the competence depletion that it in part drove and from which it in part died.
The full cycle is not evident in the short-term. Competitive vacuums result from slowlydeveloping competence gaps between firms internal and external to the MMC network, and
entrepreneurial gap recognition and subsequent entry entail additional time. In sum, positional
advantage in MMC contexts evidences the Icarus Paradox (Miller, 1990) in that it contains the
seed of its own destruction. However, time is necessary for the seed to bear fruit. Consequently, I
propose the following:
Proposition 11: A firm’s level of multimarket contact is positively related to that
firm’s short-term financial performance and negatively related to its long-term
financial performance.
33
CHAPTER 6
IMPLICATIONS OF THE THEORETICAL FRAMEWORK
Implications for Theory. The dynamic perspective advocated in this dissertation has
important implications for multimarket contact theory. By extending consideration of
multimarket issues outward in time and across levels of analysis, my approach invites
reevaluation and expansion of the bounds of endogeneity in MMC theory. Excessively narrow
formulations dominate current research. Short-term focus goes hand-in-hand with defining as
fixed and exogenous certain variables that might well belong within MMC models. Future
theoretical efforts might build on my framework to explore additional causal pathways involved
in MMC dynamics. One theoretical avenue in need of further exploration is the integration of
positional advantage and distinctive competencies perspectives. Nuanced analyses informed by
both views promise to substantially enhance our understanding of cross-level, long-term
recursive relationships between market position and competencies. In a more detailed sense, the
potential exists for theoretical refinement of particular relationships introduced in this article. For
example, network theory or NK modeling might shed additional light on the ‘system openness’
concept I discuss. Additional insights from behavioral or economic theory, or from elsewhere,
might be applied to analysis of the relationship between multimarket contact and firm level
competencies.
Implications for Research. A dynamic theory of multimarket contact has numerous
research implications. Moving research forward entails developing a firm level multimarket
contact construct against which firm performance may be examined. Therefore, the need exists
34
for operationalizing a construct reflecting a firm’s overall orientation toward multimarket contact
throughout its portfolio. Measurement might be indexed, such that a market-level MMC value is
assessed for the firm in each market in which it operates, and then aggregated to the firm level by
scaling each market-level MMC value according to the proportion of total firm revenues (or
profits) derived from the market in question. I measure and test firm level MMC in the U.S.
passenger airline industry in chapter 8.
Many other research opportunities arise from the framework I have proposed. For
example, empirical work could test for relationships between firm level MMC and competence
quality/variety along such dimensions as rate/diversity of innovation, product/service quality,
cost efficiency and absorptive capacity. Another research opportunity pertains to the relationship
between firm level MMC and the existence of particular mediators I propose. For example, does
evidence support the relationships I suggest between firm level MMC and isomorphism, path
dependence, macrocultures, or resource-sharing? At the population level, the need exists for
empirical studies pertaining to the relationship between MMC, concentration, and mutual
forbearance. Is high concentration really necessary, in every market occupied by an MMC
cohort, for mutual forbearance to hold? Are there discernable patterns in concentration and
collusion over time approximating the convergent periods and reorientations comprising
punctuated forbearance? Finally, the role of moderators remains conspicuously unaddressed.
What firm attributes (such as age or size) and what industry/environmental characteristics (such
as dynamism or complexity) accelerate or decelerate competence depletion, and how do
moderators affect ease of entry and thus patterns of forbearance punctuation?
Implications for Practice. Multimarket contact is framed one-dimensionally by the
current literature as a positional advantage positively associated with profit margins. The
35
theoretical framework introduced in this dissertation paints a very different picture. The principal
implication for top executives is that multimarket contact should be approached warily, as
something of a poisoned fruit. Multimarket contact does confer market power, but it should not
be regarded as a factor contributing to sustainable competitive advantage. Managers of firms
with postures high in multimarket contact must be on guard against the numerous tendencies
toward competence depletion associated with MMC. When making scope decisions, executives
should consciously avoid pursuing forbearance benefits at the expense of resource-sharing
opportunities. Firms with a multimarket orientation must stay attuned to the potential for
strategic myopia, and should scan their environments beyond the bounds of their MMC cohort
for competitors, knowledge sources, novel routines and emerging technologies. Finally,
managers must avoid falling prey to the path dependent tendencies arising from a multimarket
orientation. Where problemistic search is curtailed, search should be routinized.
Conclusion. I have sought to expand multimarket contact theory beyond the relatively
narrow temporal and causal focus it currently possesses. To do so, I have drawn insights from
behavioral and economic theory to elaborate an approach integrating the positional advantage
with the distinctive competencies perspective on competitive advantage. The framework
introduced here does not contradict the mutual forbearance hypothesis, but offers instead that
collusive arrangements represent but one stage in a broader, dynamic MMC cycle. Multimarket
contact forestalls rivalry, but competition is inexorable. Causal chains and feedback loops
transverse levels of analysis, such that market power influences competence development and
competence development recursively affects market power.
My attentiveness to part-whole relationships in the form of firm-population dynamics,
along with my argument that MMC market power contains the seeds of its own destruction,
36
evinces the dialectical sensibilities motivating my approach. The framework I advance should
not be interpreted as deterministic, however. While I identify behaviorally- and economicallygrounded tendencies that, if unaddressed, are likely to surface in firms with strong MMC
orientations, my analysis remains strategic in spirit. Managerial choice plays a critical role in
shaping the relationships between variables that I propose. Indeed, my ultimate purpose is to
inform theory and research that better equips managers to navigate multimarket contexts.
Considerable room remains for theoretical and empirical contributions to dynamic MMC theory.
The propositions forwarded here provide a foundation on which future endeavors may build.
The remaining two chapters relate empirical analyses of firm-market and firm level
MMC outcomes along nonprice competitive dimensions. Focusing on customer service
effectiveness and resource allocation to customer service and marketing, I consider whether
MMC evinces implications for competence development.
37
CHAPTER 7
FIRM-MARKET LEVEL EMPIRICAL EXAMINATION OF U.S. PASSENGER AIRLINE
INDUSTRY
Multimarket contact (MMC) research is motivated by the mutual forbearance hypothesis,
which maintains that inter-firm contact across multiple markets promotes tacit collusion
(Edwards, 1955). Multimarket contact dampens inter-firm rivalry by establishing inter-firm
familiarity, engendering mutually recognized interdependence, and investing constituent firms
with the potential to retaliate disproportionately against rivals’ competitive moves. Thus,
multimarket contact represents a positional advantage protecting constituent firms against
competitive pressures. Evidence of the MMC positional advantage has been found in the form of
higher prices (Alexander, 1985; Busse, 2000; Evans & Kessides, 1994; Fernandez & Martin,
1998; Fu, 2003; Jans and Rosenbaum, 1996; Singal, 1996) and wider price-cost margins
(Feinberg, 1985; Gimeno, 1999; Gimeno & Woo, 1996, 1999; Hughes & Oughton, 1993; Parker
& Roller, 1997; Scott, 1982). Consistent with the structure-conduct-performance paradigm of
industrial organization (Bain, 1959), which emphasizes the actions firms take to create defensible
positions against competitive forces (Caves & Porter, 1977; Porter, 1980), studies indicate that
firms actively seek and maintain the positional advantage conferred by multimarket contact
(Baum & Korn, 1996; Boeker, Goodstein, Stephan, & Murmann, 1997).
The positional advantage perspective on multimarket contact, while theoretically
grounded and empirically supported, may capture but one facet of the relationship between
MMC and firm-in-focal-market outcomes. In the strategic management literature, the positional
38
advantage view on the source of competitive advantage is countered by an alternate approach
that regards superior performance as the result of idiosyncratic organizational characteristics and
competencies (Barnett, Greve, & Park, 1994). This ‘resource-based’ or ‘distinctive
competencies’ perspective attributes variance in firm performance primarily to inter-firm
differences in the development of rare, valuable, non-substitutable, and inimitable capabilities
and assets (Wernerfelt, 1984; Barney, 1991). Multimarket contact has not been examined
through the distinctive competencies lens. If MMC affects firm capability, asset, and competence
development, existing theory and research may overlook important performance ramifications of
multimarket contact.
In this chapter, I approach multimarket contact from a distinctive competencies
perspective. Specifically, I consider the implications that MMC has for firm-market level
competence in delivering service quality. I propose a two-step sequence by which multimarket
contact engenders competence depletion. First, MMC establishes both incentives and means for
firms to reduce their commitment to service quality. Second, inertial pressures associated with
MMC entrench diminished quality levels such that, over time, quality capabilities atrophy.
Empirically, I examine the linchpin in this two-step sequence—service quality. In the context of
my argument, diminished service quality represents an outcome of MMC, a potential antecedent
to competence depletion, and a potential indicator of competence depletion. I test the
relationship between multimarket contact and firm-market level service quality in the U.S.
passenger airline industry. Results indicate a statistically significant relationship between
multimarket contact and lower on-time performance. The arguments and findings advanced in
this article suggest that the pricing benefits highlighted by the positional advantage view
represent an incomplete accounting of multimarket contact’s effects on firm performance. The
39
ledger has another side. This chapter’s distinctive competencies approach to multimarket contact
discloses quality effects with the potential to counteract and subsume positional advantages.
THEORY DEVELOPMENT
Multimarket Contact and Service Quality. The central theme running through
multimarket contact theory and research is that MMC mutes the intensity of inter-firm rivalry.
Edwards (1955) observed that firms confronting one another across numerous markets recognize
the potential for a competitive move to induce retaliation not only in the assaulted market, but at
other points of contact as well. The threat of disproportionate retaliation magnifies the expected
costs of initiating a competitive attack (Karnani & Wernerfelt, 1985). Furthermore, inter-firm
familiarity fostered by MMC enhances constituent firms’ aptitude in communicating and
apprehending one another’s intentions (Jayachandran, Gimeno, & Varadarajan, 1999). Therefore,
firms engaged in high levels of multimarket contact have not only the incentive to tacitly collude,
but the communicative capabilities to monitor and enforce collusive arrangements. Multimarket
contact’s tendency to de-intensify rivalry has been established solidly, with findings deriving
from such varied industries as banking (Barnett et al., 1994; Heggestad & Rhoades, 1978),
software (Young, Smith, Grimm, & Simon, 2000), cement (Jans & Rosenbaum, 1997), hotel
(Fernandez & Marin, 1998), hospital (Boeker, Goodstein, Stephan, & Murmann, 1997), cellular
telephone (Busse, 2000) newspaper (Fu, 2003), and airlines (Evans & Kessides, 1994; Gimeno,
1999; Gimeno & Woo, 1996, 1999; Singal, 1996). Evidence that MMC de-intensifies rivalry
derives largely from analyses of pricing behavior (Alexander, 1985; Busse, 2000; Evans &
Kessides, 1994; Feinberg, 1985; Fernandez & Martin, 1998; Fu, 2003; Gimeno, 1999; Gimeno &
Woo, 1996, 1999; Hughes & Oughton, 1993; Jans and Rosenbaum, 1996; Scott, 1982; Singal,
40
1996; Parker & Roller, 1997) and market entry and exit behavior (Baum & Korn, 1996; Boeker,
Goodstein, Stephan, & Murmann, 1997).
If multimarket contact affects firm pricing, entry, and exit decisions, might it also affect
product and service quality? A body of theory supports a negative relationship between
multimarket contact and quality. Multimarket contact may induce tacit collusion in quality
competition for the same reasons that it promotes collusion in pricing. Firms compete for
customer dollars along quality dimensions as well as price dimensions. Firm decisions along
both dimensions are contingent upon historical and anticipated decisions and actions of
competing firms. Therefore, the intensity of inter-firm quality rivalry—like price rivalry—
depends sensitively on the structure of inter-firm relationships. Inter-firm communication,
mutual apprehension of intent, and leverage shape the intensity with which rivals attempt to best
one another in terms of quality. Where the structure of inter-firm relationships fosters mutual
recognition of competitive interdependence and enhances firms’ capability to monitor and
enforce tacit collusion, product/service quality rivalry tends to soften. The single-market
oligopoly literature finds empirical support in this regard. High concentration levels in single
market contexts have been shown to decrease service quality in the airline industry (Mazzeo,
2003), decrease product offerings in the banking industry (Heggestad & Mingo, 1976), and
decrease product quality when the fixed costs of quality improvement display economies of scale
(Banker, Khosla, & Sinha, 1998). Given the demonstrated tendency of MMC to dampen price
rivalry and the tendency of high concentration levels to dampen quality rivalry, there is reason to
anticipate that MMC dampens quality rivalry. Multimarket contact builds familiarity between
firms, enhancing their skill at signaling and interpreting competitive intent. Multimarket contact
promotes firms’ capacity to enforce as well as nurture tacit collusion, because defectors in a
41
single market find themselves vulnerable to disproportionate retaliation ranging across multiple
markets. In sum, MMC fosters the mutual understanding requisite for tacit collusion, provides
strong incentives for constituent firm to toe the forbearance line, and affords the means of
enforcing collusive arrangements. Consequently, I propose the following hypothesis, represented
in Figure 4:
Hypothesis 1: Multimarket contact will be negatively related to customer service
quality
Why quality matters: Inertia and competence depletion under MMC. Disadvantageous
financial performance implications of a negative relationship between multimarket contact and
service quality are not immediately apparent. After all, in instances where attributes of MMC
promote quality collusion, they are likely to promote price collusion and higher margins
concomitantly. In the event that MMC-related collusion does disintegrate or fail to continue to
protect margins, one might expect constituent firms to simply adopt a more competitive posture
along quality dimensions. I propose, however, that multimarket contact is peculiarly inertial,
entrenching low service quality to the point that competence in delivering quality service is
compromised. Constituent firms may find themselves lacking the assets and attributes to adopt a
more competitive posture when changes in external conditions so warrant. The potential for
inertia to calcify quality effects of MMC into competence depletion has important implications
for financial performance.
Inertia is particularly likely to entrench diminished service quality in multimarket
contexts because of the high degree of interconnectedness between firms. Firms confronting one
another frequently across multiple markets are likely to focus on each other to the exclusion of
42
MMC
H1
FIGURE 4: Relationship between MMC and Product / Service Quality
Product/
Service
Quality
43
firms, technologies, and modes of operating outside the group. Multimarket contact thus fosters
macrocultures that blind members to opportunities and threats originating beyond perceived
competitive boundaries (Abrahamson & Fombrun, 1994; Cassidy and Loree, 2001; Peteraf &
Shanley, 1997). The spatial myopia (Levinthal & March, 1993) of MMC macrocultures may not
become evident until a threat presents itself. If changes in external market conditions demand
heightened competitive acumen in, for example, product or service quality, the
interconnectedness of MMC macrocultures may undermine the capacity of member firms to
respond. The institutional literature notes that mimetic isomorphism is strongest where
interconnectedness (Oliver, 1991) and interdependence (DiMaggio & Powell, 1983) between
firms is highest. Thus, members of a multimarket contact cohort are particularly likely to respond
to uncertainty or problems by modeling one another (DiMaggio & Powell, 1983). Such a
proximal focus may be particularly dysfunctional if the ‘problem’ at hand is the need for
increased product/service quality, and the cohort has a recent history of collusion along quality
dimensions. Where constituent firms perceive historical collusion to be the wellspring of success,
the dynamics of path dependence may interact with the dynamics of interconnectedness to
engender myopic tendencies. Firms with a history of success deriving from collusion may
become dependent on collusion, such that their responses to challenges is conditioned and
constrained by that path (Arthur, 1989; Levinthal & March, 1993). In a variant of the
‘propinquity trap’ described by Ahuja and Lampert (2001), firms with collusive histories may
seek new solutions near old ones, responding to newly competitive market conditions by seeking
to reconstitute collusive arrangements. Strategic myopia, homogenization, and path dependence,
in sum, feed inertial momentum that potentially turns firms opting not to deliver high quality
service into firms lacking the competence to deliver high quality service.
44
The more inelastic are quality levels in multimarket contexts, the greater are their
implications for firm performance. The peculiarly inertial nature of multimarket contact
heightens the importance and consequences of the negative relationship I propose between MMC
and service quality. I now describe my test of the hypothesized relationship.
METHODS
Sample. I conducted my analysis in the U.S. passenger airline industry for several
reasons. Most importantly, numerous studies set in this industry have found relationships
between multimarket contact and mutual forbearance in the form of higher prices (Evans &
Kessides, 1994), increased revenue per passenger seat mile (Gimeno, 1999; Gimeno & Woo,
1996), and wider price-cost margins (Gimeno & Woo, 1999, Singal, 1996). Examining
forbearance along quality dimensions in an industry where forbearance along pricing dimensions
has been well established was attractive, because it enabled me to ask whether the former
accompanied the latter. I considered it particularly interesting to investigate whether potentially
performance-damaging quality effects existed side-by-side with margin-enhancing pricing
effects. Additionally, the airline industry was attractive because it had previously unused data
reflecting service quality at the transaction level.
To test my hypothesis, I collected data on nonstop and one-stop flights conducted by the
17 largest U.S. airlines to and from the 68 largest airports in the continental United States. All
airports included in the analysis are categorized by the Department of Transportation (DOT) as
either medium- or large-sized, while the airlines chosen were those accounting for at least one
percent of total market share in 2003. Following previous studies, I defined a market as a citypair (Gimeno & Woo, 1996, 1999). Airports separated by 30 miles or less were treated as
occupying the same city. As a result, 17 of the airports were collapsed into 8 cites, as follows:
45
Midway and O’Hare (Chicago); Hobby and Intercontinental (Houston); Love Field and
Dallas/Ft. Worth International (Dallas); Dulles and Reagan (D.C.); San Francisco International
and Metro Oakland International (S.F.); Burbank and Los Angeles International (L.A.); Fort
Lauderdale and Miami International (Miami); and La Guardia, J.F.K., and Newark (N.Y.). The
59 cities examined yielded 1740 potential city-pair markets, with the 17 focal carriers actually
flying nonstops and/or one-stops in 1704 city-pair markets.
I used U.S. Department of Transportation data. Information pertaining to one-stop flights
was derived from the database DB1B Market, which represents a ten percent sampling of all U.S.
passenger airline flights. A total of 1,428,861 one-stop flights were flown by the 17 focal airlines
between 59 focal cities during the third quarter of 2003. For nonstop flights, I used the database
On-Time Performance, which contains information for all nonstop U.S. passenger airline flights
(as opposed to a ten percent sampling). A total of 398,833 nonstops were flown by the focal
airlines between the focal cities in July 2003. The dependent variable in my analysis was
reported for nonstop flights only, limiting my ultimate sample size to 398,833. Information from
the DB1B Market database of 1,428,861 pertinent one-stop flights was, however, used to
calculate multimarket contact values.
Dependent variable. Measuring service quality at the flight level poses a challenge.
While customers value on-time performance, and data on flight delays has been long available,
until recently the data did not indicate whether a given delay was attributable to the carrier’s
decisions and actions or to causes beyond the carrier’s control. In June 2003, however, the BTS
began collecting data on the causes of flight delays. Carriers accounting for at least one percent
of domestic passenger airline revenues were charged by the BTS with apportioning their delay
minutes to one of five sources: Air Carrier, Extreme Weather, National Aviation System, Late-
46
Arriving Aircraft, and Security. An Air Carrier delay, according to the BTS, occurs “due to
circumstances within the airline’s control (e.g. maintenance or crew problems, aircraft cleaning,
baggage loading, fueling, etc.).” The variable Carrier Delay, therefore, excludes departure
delays clearly attributable to causes beyond the airline’s control, such as weather or security
problems. The variable’s sensitivity to carrier decisions and actions suits it to this study’s
purpose of examining relationships between MMC and service quality outcomes. The Bureau of
Transportation Services collects measures of Air Carrier Delay only for nonstop flights.
Therefore, Carrier Delay reflects the number of departure delay minutes attributed to Air Carrier
Delay for each nonstop flight flown by the focal airlines in the focal city-pair markets during
July 2003.
Independent variable. The independent variable Multimarket Contact was calculated at
the firm-market level. In line with prior research, I used a count measure that yielded a
multimarket contact value for each market served by a focal firm (Baum & Korn, 1996; Evans
and Kessides, 1994; Gimeno & Woo, 1996, 1999). For each carrier in each city-pair market, I
counted the number of city-pair markets, other than the focal market, in which the carrier met
focal market rivals. Multimarket Contact was computed by dividing the total number of
multimarket contacts by the number of rivals. Thus, a value was determined for each firm in each
market served by that firm. For instance, if firm a met two rivals in market m, and firm a met one
of those rivals in 200 other markets and the other rival in 400 other markets, then the
Multimarket Contact value for firm a in market m was 300 ([200 + 400] / 2).
Control variables. I used eight control variables in my analysis. To control for firm size, I
used Firm Flights, which consisted of the total number of flights flown by the focal carrier
during 2003. I used Hour to control for the effect that time of departure had on the dependent
47
variable. The control variable Hour reflects the average Carrier Delay for all nonstop U.S.
passenger flights departing during the hour in question in July 2003. Just as mean Carrier Delay
values vary according of the time of day that planes depart, so too do they vary between airports.
To control for airport effects, I employed the variables Origin and Destination. The former
variable represents the average Carrier Delay for all planes departing the origin airport during
July 2003, while the latter represents the average Carrier Delay for all planes arriving at the
destination airport during July 2003. The final four control variables reflect market attributes,
market structure, and the firm’s dominance within the market. The variable Firm Passengers per
Market measures the number of passengers flown by the focal airline in the focal city-pair
market from the third quarter of 2002 through the second quarter of 2003. Market Share reflects
the percentage of total focal market passengers flown by the focal carrier during the same time
frame. To control for market concentration, I calculated the Herfindahl-Hirshman index (HHI),
or the sum of squares of the market shares of firms operating in a city-pair from the third quarter
of 2002 through the second quarter of 2003. City Share represents the average of a firm’s share
of total enplanements at both end-cities of a city pair.
Statistical Analysis. I first used OLS regression to test my hypothesis. However, Carrier
Delays are relatively infrequent occurrences. Of the 398,833 nonstop flights in my sample, only
29,436 (7.4%) experienced a Carrier Delay of one or more minutes. This extremely high nonoccurrence rate threatened to conceal any relationship between Multimarket Contact and Carrier
Delay. To address this problem, I created two different matched samples for binary logistic
regression analysis. In the first sample, I matched the 29,436 flights experiencing a Carrier
Delay of one or more minutes with an equal number of flights experiencing no Carrier Delay,
yielding a total sample size of 58,872. In the second sample, I matched the 18,750 flights
48
experiencing a Carrier Delay of 15 or more minutes with an equal number of flights
experiencing no Carrier Delay, generating a sample of 37,570 flights. The rationale behind the
second sample is that there is a meaningful difference, in terms of customer utility, between a
carrier delay of a few minutes and a more burdensome delay of twenty minutes, a half-hour, or
more. I chose 15 minutes as a cutoff point, borrowing from the Department of Transportation’s
definition of “late departures” as flights leaving 15 or more minutes after their scheduled
departure time. For both samples, I matched according to carrier. For example, 3,849 American
Airlines flights had Carrier Delays of 15 or more minutes. I matched by randomly selecting
3,849 flights from among the American Airlines flights that had Carrier Delays of 14 minutes or
less. I pursued this matching strategy for each carrier.
RESULTS
Tables 1, 2, and 3 provide descriptive statistics and Pearson correlations for, respectively,
the full sample of 398,833 flights, the matched sample of 58,872 flights (half of which have
Carrier Delays of greater than one minute), and the matched sample of 37,570 flights (half of
which have Carrier Delays of 15 or more minutes). The regression equation for Carrier Delay in
the full sample has a variance inflation factor of 1.661, which is well below 10, the level at
which multicollinearity becomes a concern (Ryan, 1997).
Results for the OLS regression on the full sample of 398,833 flights appear in Table 4. A
positive, statistically significant (p < .001) relationship was found between MMC and Carrier
Delay, lending support to my hypothesis that multimarket contact is negatively related to
customer service quality. However, effect sizes were not strong (∆R2 = .000). Overall Model F,
while significant both before and after the introduction of MMC, was reduced by MMC’s
introduction.
49
TABLE 1
Descriptive Statistics and Correlations for Full Sample of all Nonstopsa
Variables
1. MMC
2. Firm Flights
Mean
s.d.
470.71
246.68
562642.60 253786.38
1
2
3
4
5
6
7
8
9
.40***
3. Hour
3.11
0.71
-.02***
.02***
4. Origin
2.26
0.76
.10***
-.14***
.05***
5. Destination
2.12
0.42
.11***
-.01***
.02***
.11***
6. Firm Psgrs/Mktb
8.68
1.46
.26***
.31***
-.00
.04***
7. Market Share
0.46
0.26
.34***
.41***
.02***
-.18*** -.02***
8. HHI
0.45
0.18
-.06***
.20***
.03***
-.23*** -.03*** -.04***
.58***
9. City Share
0.23
0.13
.52***
.59***
.01***
-.07*** -.01***
.44***
.78***
10. Carrier Delay
2.61
16.80
.01***
-.02***
.02***
.05***
-.01*** -.02*** -.02*** -.01***
n = 398,833
a
Pearson Correlations (2-tailed)
b
Natural Logarithmic transformation used to correct for adverse skew
***
p < .001
-.04***
.03***
.42***
.41***
50
TABLE 2
Descriptive Statistics and Correlations for Carrier Delay ≥ 1 Minutea
Variables
Mean
s.d.
472.59
247.24
561763.90
252185.45
.42***
3. Hour
3.18
0.71
-.01
4. Origin
2.37
0.76
.10***
-.13*** .08***
5. Destination
2.15
0.45
.10***
-.02***
.01**
.12***
6. Firm Psgrs/Mktb
8.69
1.45
.28***
.32***
.00
.01**
7. Market Share
0.46
0.26
.34***
.40***
.03*** -.17***
8. HHI
0.45
0.18
-.05***
.19***
.04*** -.21*** -.03*** -.03*** .59***
9. City Share
0.23
0.13
.53***
.57***
.03*** -.06***
-.01†
.45***
10. Delay ≥ 1 Min.
0.50
0.50
.00
-.01**
.12***
.06***
.00
1. MMC
2. Firm Flights
1
2
3
4
5
6
7
8
9
.02***
n = 58,872
a
Pearson Correlations (2-tailed)
b
Natural Logarithmic transformation used to correct for adverse skew
†
p < .10
**
p < .01
***
p < .001
.17***
-.04***
-.01†
.44***
.79*** .43***
.01
-.01** .02***
51
TABLE 3
Descriptive Statistics and Correlations for Carrier Delay ≥ 15 Minutesa
Variables
Mean
s.d.
471.75
258.49
544022.40
242739.08
.50***
3. Hour
3.18
0.72
-.01
.02**
4. Origin
2.42
0.76
.08***
-.09***
.10***
5. Destination
2.17
0.45
.12***
.03***
.02***
.11***
6. Firm Psgrs/Mktb
8.57
1.55
.34***
.32***
-.01†
.03***
-.02***
7. Market Share
0.44
0.26
.41***
.38***
.02***
-.13***
.02**
.46***
8. HHI
0.44
0.17
-.05***
.12***
.03***
-.19***
-.03***
-.08***
.52***
9. City Share
0.23
0.13
.59***
.56***
.03***
-.03***
.02***
.49***
.80***
.36***
10. Delay ≥ 15 Mins.
.050
0.50
.01
.00
.10
.16***
.06***
.00
.00
-.02***
1. MMC
2. Firm Flights
1
2
3
n = 37,570
a
Pearson Correlations (2-tailed)
b
Natural Logarithmic transformation used to correct for adverse skew
†
p < .10
**
p < .01
***
p < .001
4
5
6
7
8
9
.01**
52
TABLE 4
Results of OLS Regression Analysis for Carrier Delaya
Variables
Intercept
Firm Flights
Hour
Origin
Destination
Firm Psgrs / Mktb
Market Share
HHI
City Share
Step 1
-1.431***
(.282)
p = .000
.000***
(.000)
-.008
p = .000
.513***
(.038)
.022
p = .000
.939***
(.037)
.042
p = .000
.877***
(.064)
.022
p = .000
-.115***
(.022)
-.010
p = .000
.181
(.194)
.003
p = .352
-1.036***
(.202)
-.011
p = .000
.533
(.383)
.004
p = .165
Step 2
-1.623***
(.285)
p = .000
.000***
(.000)
-.009
p = .000
.518***
(.038)
.022
p = .000
.919***
(.037)
.041
p = .000
.843***
(.064)
.021
p = .000
-.105***
(.022)
-.009
p = .000
.108
(.195)
.002
p = .579
-.718**
(.213)
-.008
p = .001
-.124
(.409)
-.001
p = .761
.001***
(.000)
.009
p = .000
.003
.003
.004
.003
MMC
R2
Adjusted R2
53
∆R2
Model F
∆F
.003
.000
***
173.474
156.543***
p = .000
p = .000
***
173.474
21.025***
p = .000
p = .000
a
n = 398,833. Unstandardized regression coefficients, standard errors (in parentheses),
and standardized coefficients (in italics) are presented, in that order.
b
Natural Logarithmic transformation used to correct for adverse skew
†
p < .10
**
p < .01
***
p < .001
I conducted binary logistic regression on the two matched samples in order to correct for
the infrequent occurrence of Carrier Delay. Results appear in Table 5. For Carrier Delays of 1
minute or greater, MMC manifested a positive, statistically significant (p < .001) effect. The
MMC Wald Statistic of 84.60 was higher than the Wald Statistics for five of the eight control
variables. Model chi-square was significant at p < .001 both before and after the introduction of
MMC, as was block chi-square. For Carrier Delays of 15 or more minutes, MMC also showed a
positive, statistically significant (p < .001) effect. The MMC Wald Statistic of 28.85 was again
larger than the Wald Statistics for five of the eight control variables. Both model chi-squares and
the block chi-square were significant at p < .001. While statistically significant, effects were not
large. For Carrier Delays of 1 or more minutes, the inclusion of MMC in the regression equation
increased the percent of correct predictions from 58.3 to 58.4, increased the Cox & Snell R2 from
.041 to .043, and increased the Nagelkerke R2 from .055 to .057. For Carrier Delays of 15 or
more minutes, MMC had no effect on the percent of correct predictions, increased the Cox &
Snell R2 from .033 to .034, and increased the Nagelkerke R2 from .044 to .045.
54
TABLE 5
Results of Logistic Regression Analysis for Carrier Delaya
Variables
Constant
Firm Flights
Hour
Origin
Destination
Firm Psgrs / Mktb
Market Share
HHI
City Share
MMC
Delay ≥ 1 Minute
n = 58,872
Block 1
Block 2
-2.218***
-2.105***
(.091)
(.092)
593.12
524.29
p = .000
p = .000
.000†
.000
(.000)
(.000)
2.73
.000
p = .098
p = .996
.301***
.298***
(.012)
(.012)
627.86
613.13
p = .000
p = .000
.428***
.441***
(.012)
(.012)
1316.12
1374.36
p = .000
p = .000
.182***
.202***
(.019)
(.019)
88.55
108.03
p = .000
p = .000
-.030***
-.035***
(.007)
(.007)
16.81
23.50
p = .000
p = .000
.223***
.268***
(.063)
(.064)
12.47
17.82
p = .000
p = .000
-.104
-.309***
(.065)
(.069)
2.59
20.24
p = .108
p = .000
.397**
.809***
(.123)
(.131)
10.39
38.07
p = .001
p = .000
.000***
(.000)
84.60
p = .000
Delay ≥ 15 Minutes
n = 37,570
Block 1
Block 2
-2.043***
-1.977***
(.112)
(.113)
332.03
307.46
p = .000
p = .000
.000
.000*
(.000)
(.000)
.480
3.93
p = .025
p = .048
.235***
.232***
(.015)
(.015)
249.12
242.81
p = .000
p = .000
.389***
.397***
(.015)
(.015)
719.05
739.40
p = .000
p = .000
.191***
.206***
(.024)
(.024)
63.45
72.67
p = .000
p = .000
-.015†
-.018*
(.009)
(.009)
3.10
4.42
p = .078
p = .035
.115
.145†
(.077)
(.078)
2.20
3.47
p = .138
p = .062
-.071
-.209*
(.079)
(.083)
.81
6.30
p = .368
p = .012
.147
.458**
(.155)
(.166)
.89
7.64
p = .344
p = .006
.000***
(.000)
28.85
p = .000
55
Model Chi-Square
2471.70***
p = .000
Block Chi-Square
% Correct Predictions
Cox & Snell R2
Nagelkerke R2
58.3
.041
.055
2556.55***
p = .000
84.85***
p = .000
58.4
.043
.057
1249.90***
p = .000
57.4
.033
.044
1278.79***
p = .000
28.89***
p = .000
57.4
.034
.045
a
Unstandardized regression coefficients, standard errors (in parentheses), and Wald statistics (in
italics) are presented, in that order.
b
Natural Logarithmic transformation used to correct for adverse skew
†
p < .10
*
p < .05
**
p < .01
***
p < .001
In sum, logistic regression analyses on the two matched samples lend statistically
significant support to my hypothesis that MMC has a negative relationship with customer service
quality. Relatively small effect sizes, however, raise doubt concerning the importance of the
relationship.
DISCUSSION
This chapter approached multimarket contact from a distinctive competencies
perspective. My motivation in pursuing this analysis was that, while extant theory and research
emphasize positional advantages deriving from MMC, there exist sound theoretical grounds for
anticipating a negative relationship between multimarket contact and organizational competence
in delivering product/service quality. I proposed a two-step sequence through which MMC first
decreases and then ossifies product/service quality. I tested the relationship between MMC and
service quality in the U.S. passenger airline industry, finding statistically significant support for
negative effects.
56
While statistically significant, effects were of minimal size. The most likely explanation
for small effect sizes is that variance in Carrier Delay proved very difficult to explain. For
example, adjusted-R2 for the full regression equation in the full sample of 398,833 flights was an
exceedingly small .003. Cumulatively, therefore, all of the controls and MMC explained less
than one percent of variance in Carrier Delay. In the matched samples, Cox & Snell R2 never
exceeded .043 and Nagelkerke R2 never exceeded .057. Thus, although Carrier Delay seemed
attractive because it represented delay minutes under a carrier’s control, very little variance in
the dependent variable was ultimately accounted for by variance in firm attributes, market
attributes and structure, and inter-firm relationships.
In spite of the small effect sizes, the fact that the results were statistically
significant in the predicted direction is of theoretical importance. This is the first study to
propose and test a relationship between multimarket contact and service quality. While
the findings fail to validate conclusively the importance of the relationship in this
particular context, they do suggest that a negative relationship exists. The theoretical
implications of this finding are substantive because extant MMC theory and research are
wholly oriented toward explaining and confirming outcomes beneficial to the firm. This
study provides the first indication of an MMC outcome with potentially damaging
consequences from the firm’s perspective. Diminished service quality, even if closely
associated with margin-enhancing collusion, portends poorly for long-term firm
performance should competitive conditions change.
Limitations and future research. The study’s cross-sectional design represents its
primary limitation. It is difficult to distinguish, under any design, between a firm opting
to deliver low service quality and a firm lacking the competence to deliver anything but
57
low service quality. With a cross-sectional design, it is impossible to distinguish between
the two. This study measured service quality. I was unable, however, to determine with
confidence whether quality levels were outcomes of recent/current MMC levels or
indicators of current competence depletion deriving from MMC effects in the more
distant past. The former relationship seems most likely, but longitudinal research is
needed to explore the nature of the causal flow from MMC to quality delivered to quality
competence.
As with any study set in a single industry, generalizability is a concern. Future
studies of MMC-quality relationships in other industries are needed to determine whether
the arguments advanced in this work hold beyond the U.S. passenger airline industry.
Another limitation pertains to the breadth and quality of the dependent variable.
Assessing product or service quality at the transaction level or at the firm-market level, in
an industry that lends itself to calculating multimarket contact, poses a significant
challenge. However, a quality measure with greater levels of explainable variance will be
essential to better apprehend the importance of multimarket contact. Especially valuable
would be a dependent variable reflecting firm resource allocation to quality at the firmmarket level.
Conclusion. This chapter represents an important counterpoint to the extant
literature’s positional advantage perspective on multimarket contact. The primary
contribution of this work is in shedding light on multimarket contact’s latent dark side. I
introduce, for the first time, both a theoretical explanation for—and empirical evidence
of—a negative relationship between MMC and service quality. The arguments and
findings advanced here warrant a re-orientation in the underlying MMC research question
58
from, “How much of a positional advantage does MMC confer?” to, “Does the
positional advantage conferred by MMC outweigh the competence depletion induced by
MMC?” It is my hope that this study will promote a more balanced accounting of
MMC’s performance implications.
59
CHAPTER 8
FIRM LEVEL EMPIRICAL EXAMINATION OF U.S. PASSENGER AIRLINE INDUSTRY
Multimarket contact (MMC) research examines how inter-firm relationships outside a
focal market affect inter-firm behavior within the market. Multimarket theory concentrates on
the mutual forbearance hypothesis (Edwards, 1955), which holds that MMC engenders sufficient
inter-firm familiarity and retaliatory potential to mute focal market rivalry. Numerous empirical
studies lend support to the mutual forbearance hypothesis, finding evidence of muted rivalry in
higher prices (Alexander, 1985; Busse, 2000; Evans & Kessides, 1994, Fu, 2003; Jans &
Rosenbaum, 1996; Singal, 1996), higher margins (Feinberg, 1985; Gimeno, 1999; Gimeno &
Woo, 1996, 1999; Hughes & Oughton, 1993; Scott, 1982), and lower entry and exit rates (Baum
& Korn, 1996; Boeker, Goodstein, Stephan, & Murmann, 1997). Extant research, therefore,
suggests that firms derive certain benefits from meeting focal market rivals in multiple other
markets. From this perspective, a high level of MMC has salutorious effects on firm-market
prices, margins, and entry rates.
The compelling body of research on firm-market level forbearance outcomes, however,
does not broach the issue of firm level implications of multimarket contact. Neither extant theory
nor empirical evidence explicitly supports any relationship between MMC and firm behavior or
outcomes. Absent demonstrated links to the firm level, the strategic import of multimarket
contact remains uncertain. Is MMC a peripheral consideration, drowned out by more powerful
concerns and of little consequence to the way a firm behaves and performs? Or do forbearance
dynamics rooted in a firm’s individual markets percolate upward one level of analysis, shaping
60
organization-wide actions and outcomes? If multimarket contact does find expression at the firm
level, what is the nature of its influence? Are forbearance effects pervasive across competitive
dimensions, or does multimarket contact influence nonprice competition differently than price
competition?
My purpose is to more firmly establish the nature and significance of MMC’s strategic
impact. To this end, I employ two fresh investigative tactics. First, I introduce a firm level
construct capturing the overall extent to which a firm experiences multimarket contact. I term the
new construct Multimarket Contact Posture (MMCP). A firm level MMC construct is
theoretically justified, for firm level managerial processes are deeply implicated in the prevailing
firm-market level construal of MMC. Measurement of firm level MMC is empirically valuable,
as well, because it facilitates examination of firm level strategic outcomes. Second, I utilize a
competing hypotheses framework that highlights and helps resolve opposing perspectives on the
manner in which multimarket contact may influence competitive intensity. The traditional
emphasis on prices and margins as dependent variables in MMC research underscores the need
to explore how far the bounds of mutual forbearance extend. Firms compete along nonprice as
well as price dimensions, but it remains unclear how—or even whether—forbearance pertains to
the former as well as the latter. While a body of theory supports a negative relationship between
MMCP and nonprice competition mirroring that found between multimarket contact and price
competition, viable opposing arguments anticipate MMCP amplifying nonprice competition. If
MMCP relates positively to such nonprice competitive dimensions as advertising, promotions,
and product/service quality, then the mutual forbearance hypothesis must be reconceptualized as
permeating but a limited rivalrous space. I test competing pervasive and partial forbearance
61
hypotheses in the U.S. passenger airline industry, where extant MMC research most
compellingly demonstrates forbearance in pricing.
MULTIMARKET CONTACT POSTURE: A NEW FIRM LEVEL CONSTRUCT
The multimarket contact construct has been conceptualized and measured at three distinct
levels of analysis (Gimeno & Jeong, 2001). Early empirical work conducted by economists
treated MMC as a market attribute. Studies in this tradition measure the construct as the overall
degree of multimarket contact among firms serving a focal market (Evans & Kessides, 1994;
Feinberg, 1985; Jans & Rosenbaum, 1996; Singal, 1996). Management scholars, on the other
hand, typically regard multimarket contact as a characteristic of the relationship between firms
(Gimeno & Jeong, 2001). Within the management literature, MMC has been measured at both
the dyadic and the firm-market levels of analysis. The dyadic approach measures the overall
level of multimarket contact between two firms across all of the markets in which the two meet
(Baum & Korn, 1999; Korn & Baum, 1999). The more prevalent firm-market approach measures
the level of cross-market contact that a firm has with incumbents in a focal market. In the airline
industry, for example, Gimeno & Woo (1996, 1999) find that a carrier’s multimarket contact
with focal route rivals tends to increase the prices charged by that carrier in that route. Baum &
Korn (1996) find that high airline-in-route MMC levels result in tacitly collusive market entry
and exit patterns.
No existing construct reflects the extent to which a firm confronts multimarket
competitors across the breadth of its corporate portfolio. Conceptualizing and measuring MMC
at the firm level has theoretical and empirical value, however, because the prevailing firmmarket level construal does not relate MMC to all types of competitive activity that it may in fact
affect. Firms undertake competitive actions specific to individual markets, and they undertake
62
competitive actions common to the entire firm. The firm-market level approach, which is the
focus of existing MMC studies, is incapable of examining relationships between MMC and
competitive actions at the firm level. Only a construct reflecting a firm’s overall multimarket
contact posture (MMCP) can be linked comprehensively to competitive activity, both marketspecific and common to the entire firm.
A firm’s market-level MMC values, like the competitive activities and outcomes they
influence, inform a firm level perspective when they are considered cumulatively and averaged
across the firm. For instance, market-specific competitive moves such as promotions or pricing
decisions, or market-specific competitive outcomes such as profit margins, can be absorbed into
firm level constructs by averaging across the firm’s markets. Market-level values can, where
appropriate, be weighted by proportion of total firm revenues accounted for by the market in
question. The straightforward averaging procedure lends itself comfortably to MMC values as
well. Standard deviations for any given firm’s market-level MMC values are relatively small, as
indicated in Table 6. The MMC mean of all of a firm’s markets closely resembles the value of
nearly any individual market. In short, a firm’s numerous market-level MMC values
cumulatively define its overall MMC posture, with most markets adhering closely to the overall
definition. Averaging both antecedents and outcomes across markets translates the relationship
between market-level MMC and market-specific competitive activity to the firm level.
Firm level MMCP can be related, as well, to competitive activity pursued by the firm as a
single entity. Competitive moves such as firm-wide advertising campaigns, firm-wide price cuts,
or firm-wide product/service quality initiatives may be sensitive to MMCP. A firm level vantage
on multimarket contact with a rival appears nearly identical to a firm-market level vantage. Delta
and Northwest, for example, confront one another in non-stop and one-stop service in 1066 city-
63
TABLE 6
Raw Un-weighted MMCP Means and Standard Deviations
Carrier
American Eagle Airlines
United Airlines
American Airlines
Delta Airlines
Southwest Airlines
Skywest Airlines
Independence Air
Continental Airlines
Expressjet Airlines
America West Airlines
Northwest Airlines
US Airways
Atlantic Southeast Airlines
ATA Airlines
Alaska Airlines
JetBlue Airways
Airtran Airways
Raw Un-Weighted
MMCP Mean
Standard
Deviation
54
595
707
719
449
103
69
607
226
358
705
374
94
125
71
42
168
11
132
139
189
96
20
11
125
44
56
118
86
16
20
10
4
26
pair markets, which constitutes 78% of Delta’s business in those service segments. From Delta’s
perspective in, say, the Atlanta-Boston market, Northwest is a rival is 1065 other markets. At
either the firm- or firm-market level, Northwest represents a highly salient rival from Delta’s
perspective. Much as Northwest’s retaliatory potential may deter Delta from initiating a
competitive move in the Atlanta-Boston market, it may deter Delta from initiating a firm-wide
competitive move. I use the dyadic example of Delta and Northwest for the sake of simplicity,
but the principle that deterrence may cross levels of analysis can be extended to multiple
competitors.
64
A final point demonstrates the mutual implication of competitive perspectives at different
levels of analysis, and supports the case for a firm level MMC construct. Firm level vision and
decision-making are critical components in multimarket theory, which posits that a firm’s
behavior in market A is affected by the competitive landscape in markets B, C, D, and so forth.
The mutual forbearance hypothesis assumes a central decision-making body possessing the
access and authority to integrate information from multiple corporate units and to coordinate
actions across those units. The choice between forbearing and not forbearing resides at the firm
level. The mutual forbearance hypothesis has firms recognizing that MMC magnifies rivals’
retaliatory potential, firms understanding the incentive to withhold first-mover competitive
actions, and firms tacitly colluding in pursuit of rivalry reduction (Karnani & Wernerfelt, 1985).
Thus, firm level decision processes link the traditional firm-market level MMC construct to its
firm-market level outcomes.
Given that little variance distinguishes market-level MMC values from one another
within a given firm, that multimarket rivals present as much retaliatory potential when viewed
from a firm-wide perspective as from a market-level perspective, and that theory anticipating
forbearance at the market level assumes a large measure of firm-wide strategic vision, a sound
basis exists for thinking in terms of a firm-wide multimarket contact posture. I address MMCP
measurement in greater detail in the methods section; it entails calculating a revenue-weighted
average of the firm’s MMC in its various markets, then controlling for the aggregate number of
markets served by the firm. The construct promises to be useful and informative as well as
theoretically grounded. Examination of relationships between MMCP and firm level activities
and outcomes—including both averaged market-specific activities and outcomes and those
65
attributable to the firm as a single entity—will enhance our understanding of multimarket
contact’s strategic import.
PERVASIVE VERSUS PARTIAL FORBEARANCE EFFECTS OF MMCP
Even at the firm-market and dyadic levels of analysis, where MMC research has
concentrated, the breadth and boundaries of forbearance effects remain uncertain. Initially
focused during the 1970s and 1980s on whether MMC induces tacit collusion, research has only
in the past decade begun to shift toward examination of how firms engaged in multimarket
contact collude. Prices and margins—readily accessible standbys of collusion studies in
oligopoly contexts—represent the most studied outcomes in the MMC literature. Forbearance
effects along nonprice competitive dimensions have begun to receive attention in recent years.
Scholars have found that MMC mutes rivalry in innovation (Greve & Mitsuhashi, 2004),
advertising (Shankar, 1999), and focal market entry and exit (Baum & Korn, 1999; Boeker et al.,
1997; Korn and Baum, 1999). Still, the nature of relationships between MMC and various types
of nonprice competition is far from settled. In this section, I approach the issue at the firm level. I
introduce competing rationales and testable hypotheses addressing relationships between MMCP
and nonprice competition. One perspective anticipates widespread forbearance effects in service
quality and marketing similar to those associated with price competition, while the counter-view
anticipates MMCP amplifying rather than muting nonprice competition. Implications for firm
financial performance are hypothesized.
Parallel price and nonprice effects of MMCP. A body of theory anticipates MMCP
engendering pervasive forbearance effects along nonprice as well as price dimensions of
competition. Muted rivalry in advertising or product/service quality may derive from either or
both of two distinct sets of processes. Figure 5 represents direct and mediated paths from MMCP
66
Price
Competition
-
+
MMCP
-
Nonprice
Competition
FIGURE 5: Negative Relationship between MMCP and Nonprice Competition
to nonprice competition. First, MMCP may reduce nonprice competition as it reduces price
competition—directly, by promoting tacit collusion. Collusion research in oligopoly contexts
suggests that the boundaries of muted rivalry can indeed be broad. High concentration—a wellestablished antecedent to collusion in pricing—has been linked as well to decreased service
quality in the airline industry (Mazzeo, 2003), decreased product offerings in the banking
industry (Heggestad & Mingo, 1976), and decreased product quality when the fixed costs of
quality improvement display economies of scale (Banker, Khosla, & Sinha, 1998). The effects of
multimarket contact posture may be similarly broad. Firms confronting rivals across many
markets risk inducing retaliation along several fronts should they pursue promotional campaigns
or quality improvement initiatives. A promotional campaign in one market, for instance, may be
67
met by rivals with counter-campaigns in numerous markets. The threat of retaliation decreases
the expected value of such initiatives, therefore dampening the motivation to pursue them. Interfirm familiarity resulting from cross-market contact promotes the mutual recognition of
interdependence upon which collusion depends. This view, in sum, interprets the mutual
forbearance hypothesis as broadly applicable to nonprice as well as to price competition.
An alternate, distinct set of processes may link MMCP indirectly to muted nonprice
rivalry. According to this view, price competition mediates outcomes in nonprice competition.
Tacit collusion in pricing shields the firm from market forces, artificially propping up margins.
Beneath the price forbearance shield, firms experience less pressing incentive to allocate
resources to advertising, promotions, or product/service quality. Firm search tends to be
problem-oriented or failure-induced (Nelson & Winter, 1982: 173; Tushman & Romanelli,
1985), meaning that it is stimulated by a problem and directed toward finding a solution to the
problem (Cyert & March, 1963: 169; March & Simon, 1958: 194). Forbearance pricing and
margins reduce firm perception of performance problems, thus curtailing search for solutions in
advertising or quality initiatives. Muted rivalry along nonprice dimensions, according to this
view, does not result directly from collusion, but rather indirectly from contextual munificence
associated with forbearance in pricing.
Direct and indirect forbearance effects along nonprice dimensions are not mutually
exclusive; either or both may operate in a given context. Singly or in conjunction with one
another, rivalry-muting processes induced by MMCP may reduce the extent to which a firm
allocates resources to promotions and sales, product quality, or customer service functions.
Muted rivalry resulting from MMCP may reduce, as well, the quality of products and services a
firm delivers to its customers. Accordingly:
68
Hypothesis 1a: Multimarket Contact Posture will be negatively related to firm
allocation of resources to promotion and sales.
Hypothesis 2a: Multimarket Contact Posture will be negatively related to firm
allocation of resources to customer service
Hypothesis 3a: Multimarket Contact Posture will be negatively related to
customer service quality
Divergent price and nonprice effects of MMCP. In contrast to the theoretical arguments
supporting a pervasive negative relationship between MMCP and both price and nonprice
competitive intensity, an alternate body of logic anticipates MMCP affecting price rivalry
negatively and nonprice rivalry positively. Oligopoly studies find inverse relationships between
price competition and competition in arenas such as advertising and R & D (Symeonidis, 2000a,
2000b), suggesting that similar relationships may hold in multimarket contact scenarios. Three
lines of reasoning inform the position that MMCP mutes price rivalry while increasing nonprice
competition. First, price coordination and collusion may be easier to establish and maintain than
nonprice (Symeonidis, 2000b). Firm pricing decisions have immediate, public, standardized
outcomes. Firm resource allocation decisions in the realms of marketing and, especially,
product/service quality generate less immediate, public, and standard outcomes. Marketing and
quality outcomes are therefore more difficult to interpret, rendering firm intentions more difficult
for rivals to divine. Rivals’ mutual recognition and apprehension of competitive intentions are
necessary preconditions to tacit collusion. In their absence, inter-firm contact in multiple markets
fuels rather than mutes intensity of rivalry. These considerations support a direct, positive
relationship between MMCP and nonprice competition, as represented in Figure 6.
69
Price
Competition
-
-
MMCP
+
Nonprice
Competition
FIGURE 6: Positive Relationship between MMCP and Nonprice Competition
A second, “supply side” dynamic may contribute to a positive relationship between
MMCP and nonprice rivalry. Price forbearance confers excess resources upon the firm.
Therefore, firms high in multimarket contact posture may allocate more resources to marketing
and product/service quality by virtue of the fact that they have more resources available for
allocation. According to this logic, price forbearance mediates the relationship between MMCP
and nonprice outcomes.
Third and finally, firms with high multimarket contact orientations may recognize that
they and their rivals, as a group, stand to benefit from nonprice rivalry. Increased nonprice
competition in high-MMC contexts may represent a rational group effort to grow overall market
size. This outcome is uniquely possible along nonprice dimensions because price and nonprice
70
competition have fundamentally different implications for the alignment of individual and group
interests. Tacit collusion emerges when “prisoner’s dilemmas” are mitigated by a combination of
inter-firm communication and the mutual capacity to punish defection. Price competition
represents a prisoner’s dilemma. While a unilateral price cut may serve a firm’s short-term
interests by increasing market share, revenues, and profits, the move undermines rivals’ interests.
The dilemma is that while each firm thereby has an incentive to cut prices, all will be worse off if
all cut prices. Familiarity and deterrence associated with multimarket contact circumvent the
dilemma by aligning group and individual interests, assuring each firm that a unilateral price cut
will be punished. What hurts the group hurts the first-mover even more. Individual and group
interests are less clearly opposed in nonprice competition, however. Unilateral improvements in
actual or perceived quality may steal some market share from rivals, but it may also grow market
size without undermining rivals’ margins. More importantly, if all rivals pursue advertising
campaigns and/or product/service quality initiatives, all stand to gain by increasing overall
market size and by differentiating in ways that justify sustained high prices and margins.
Nonprice competition, therefore, does not necessarily represent a prisoners’ dilemma. A positive
relationship between MMCP and nonprice competition may derive not so much from difficulties
in communicating, interpreting, and enforcing collusion, or from excess resources, as from
rivals’ mutual recognition that they stand to gain more than they lose by rigorous multilateral
attention to advertising and quality.
The three strains of thought supporting a positive relationship between MMCP and
nonprice competition may operate singly or in tandem. The arguments considered suggest that
MMCP heightens the allocation of resources to marketing and product/service quality—whether
71
directly, or mediated through price forbearance, or both—and enhances the quality of products
and services actually delivered. Accordingly:
Hypothesis 1b: Multimarket Contact Posture will be positively related to firm
allocation of resources to promotion and sales.
Hypothesis 2b: Multimarket Contact Posture will be positively related to firm
allocation of resources to customer service
Hypothesis 3b: Multimarket Contact Posture will be positively related to customer
service quality
Profitability effects of MMCP. A potentially complicated relationship exists
between MMCP and firm profitability. On one hand, MMCP may increase profit
margins. Research establishing the positive relationship between firm-market MMC and
both prices and margins (Evans & Kessides, 1994; Feinberg, 1985; Gimeno & Woo,
1996) may be interpreted as suggesting that these outcomes percolate to the firm level in
the form of higher profitability. According to this view, MMCP has a direct, positive
effect on profit margins. Indeed, where the pervasive forbearance perspective sees
MMCP influencing nonprice competition principally through the path mediated by muted
price rivalry, high margins may represent a necessary precondition to low levels of
nonprice competition. The partial forbearance perspective, as well, anticipates a positive
relationship between MMCP and firm profitability. Nonprice competition hones the
firm’s competitive acumen; differentiates the firm’s products or services, adding marginjustifying value in the eyes of customers while increasing market share; and grows the
aggregate market. From this perspective, amplification of nonprice competition couples
72
with price forbearance to enhance profitability. Thus, according to both pervasive and
partial forbearance rationales, MMCP may have a positive effect on firm profit margins:
Hypothesis 4a: Multimarket Contact Posture will be positively related to
profitability.
On the other hand, MMCP may decrease profit margins. If forbearance proves
pervasive, muted rivalry along nonprice dimensions may, over time, deplete firm
competencies to the detriment of profitability (Will, 2006). Firms pay a potentially steep
price for withholding competitive activity. A firm’s competitive repertoire and acumen
are affected by the range of its own past competitive actions (Miller & Chen, 1994).
Reduced competitive activity may constrain the firm’s knowledge base, making it less
skilled, tactile, and efficient at taking future competitive action (Amburgey, Kelly, &
Barnett, 1993; D’Aveni, 1994). Path-dependent on the market power embedded in
multimarket contact, firms may prove incapable of adapting to environmental change
beyond their control (Levinthal & March, 1993: 102). The fruit of market power, in short,
has potentially poisonous consequences for firm competencies and, in time, firm
profitability. Strong competence depletion effects of muted nonprice competition may,
according to the pervasive forbearance perspective, counteract and outweigh the profitenhancing effects of muted price competition. Thus:
Hypothesis 4b: Multimarket Contact Posture will be negatively related to
profitability.
METHODS
Sample. The U.S. passenger airline industry was a particularly attractive setting for my
empirical analysis of firm level MMCP, given the prominence of airline studies in the MMC
73
literature. The mutual forbearance hypothesis has met with repeated support at the firm-market
level in airline studies. Researchers have found that multimarket contact generates higher prices
(Evans & Kessides, 1994), increased revenue per passenger seat mile (Gimeno, 1999; Gimeno &
Woo, 1996), wider price-cost margins (Gimeno & Woo, 1999; Singal, 1996), and lower entry
and exit rates (Baum & Korn, 1996, 1999). The numerous indications of carriers’ firm-market
level forbearance make the airline industry an ideal context for examining whether tacit collusion
translates to the firm level as well. Additionally, the demonstrated presence of price forbearance
in the industry is appropriate to my purpose of determining whether nonprice effects accompany
price effects of multimarket contact. The questions motivating this article build on and extend
existing MMC work, so the heavily-studied airline industry represents an ideal context for my
analysis.
I collected flight, financial, and customer service data for each of the 17 U.S. passenger
airlines that accounted for at least 1 percent market share in 2003. For these target airlines, I
analyzed nonstop and one-stop flights conducted in July 2003 to and from the 68 airports in the
continental United States categorized by the Department of Transportation (DOT) as either largeor medium-sized. It was necessary to calculate firm-market level MMC in order to obtain values
for the firm level dependent variable MMCP, because the former is an input in the equation
yielding the latter. Calculation of MMC entails defining market parameters. Consistent with
previous multimarket contact studies in the airline industry (Gimeno and Woo, 1996, 1999), I
defined a market as a city pair. Airports were considered in the same city if the distance
separating them was 30 miles or less. The following airports were treated as occupying the same
city: Midway and O’Hare (Chicago); Hobby and Intercontinental (Houston); Love Field and
Dallas/Ft. Worth International (Dallas); Dulles and Reagan (D.C.); San Francisco International
74
and Metro Oakland International (S.F.); Burbank and Los Angeles International (L.A.); Fort
Lauderdale and Miami International (Miami); and La Guardia, J.F.K., and Newark (N.Y.). The
68 airports in 59 cities generate 1740 potential distinct city-pair markets, and the 17 target
airlines actually flew nonstop and/or one-stop flights in 1704 city-pair markets.
Data were obtained from U.S. Department of Transportation databases. Nonstop flights
were determined from the database On-Time Performance, while one-stop flights were
determined from the database DB1B Market. For the purposes of coding a carrier’s presence in
or absence from a particular market, I utilized two data sets: one consisting of 398,833 flights, or
all nonstops flown in July 2003; and a second consisting of 1,428,861 flights, which was a ten
percent random sampling of all one-stops flown in the third quarter of 2003. Balance sheet and
income statement data were collected from the Schedule B-1 and Schedule P-12 components of
the Form 41 Financial Schedule. Service quality data were obtained from the Air Travel
Consumer Report.
Dependent variables. I used two variables to measure firm allocation of resources to
promotion and sales. Promotion and Sales Expenses describes a carrier’s allocation of resources
to promotion and sales relative to its allocation of resources to other functions. The variable was
calculated by dividing each carrier’s 2003 promotion and sales expenses by its total operating
expenses. The Bureau of Transportation Services reports as promotion and sales expenses those
costs “incurred in promoting the use of air transportation generally and creating a public
preference for the services of particular air carriers.” The BTS includes in this category the
functions of selling, advertising, publicity, and developing flight schedules for publication.
Promotion and Sales Share describes a carrier’s resource commitment to promotion and sales
relative to rivals’ resource commitments to the same functions. The variable’s numerator was
75
calculated by dividing the focal carrier’s promotion and sales expenses by the total promotion
and sales expenses of all 17 firms. The variable’s denominator, intended to adjust the variable to
market share differentials between firms, was calculated by dividing the focal carrier’s operating
revenues by the total operating revenues of all 17 firms. Promotion and Sales Share, therefore, is
a measure of a firm’s promotion and sales expenses market share relative to its revenue market
share.
I used two variables to measure firm allocation of resources to customer service.
Passenger Service Expenses describes a carrier’s allocation of resources to passenger service
relative to its allocation of resources to other functions. The variable was calculated by dividing
each carrier’s 2003 passenger service expenses by its total operating expenses. The Bureau of
Transportations Services defines passenger service expenses as the costs of “activities
contributing to the comfort, safety, and convenience of passengers while in flight or when flights
are interrupted.” The category includes flight attendant salaries and passenger food expenses.
Passenger Service Share describes a carrier’s resource commitment to passenger service relative
to rivals’ resource commitments to the same function. The variable’s numerator was calculated
by dividing the focal carrier’s passenger service expenses by the total passenger service expenses
of all 17 firms. The variable’s denominator, intended to adjust the variable to market share
differentials between firms, was calculated by dividing the focal carrier’s operating revenues by
the total operating revenues of all 17 firms. Thus, Passenger Service Share is a measure of a
firm’s passenger service expenses market share relative to its revenue market share.
I used three variables to measure customer service quality. The variables Mishandled
Baggage and Consumer Complaints reflect rates of reports and complaints filed by passengers
with the U.S Department of Transportation in 2003. Airline passengers may file DOT reports and
76
complaints against carriers by telephone, e-mail, and in person. Mishandled Baggage indicates
the number of reports lodged against a carrier, per 1000 passengers, concerning lost, damaged,
delayed, or pilfered baggage. I used the natural logarithm of this variable to correct for adverse
skew. Consumer Complaints is a measure of the number of complaints lodged against an airline,
per 100,000 passengers. The DOT recognizes a wide range of complaint categories
encompassing customer service broadly: complaints about rude or unhelpful employees;
inadequate meals or cabin service; treatment of delayed passengers; cancellations; delays;
oversales; airline mistakes in reservations and ticketing; problems making reservations or
obtaining tickets due to busy phones, waiting in line, or delays in mailing tickets; overcharges;
incorrect or incomplete information about fares or fare conditions and availability; problems in
obtaining refunds for unused or lost tickets or for fare adjustments; claims for lost, damaged, or
delayed baggage; charges for excess baggage; and carry-on problems. Therefore, Consumer
Complaints is a multidimensional reflection of the general service quality a carrier delivers to it
passengers. Submitting a compliant to the DOT is a rather unusual step for dissatisfied customer
to take. Most passengers likely take their complaints to the airline before they take them to the
U.S. Department of Transportation. Complaints filed with the DOT, therefore, indicate a strong
customer mindset and a twofold failure in customer service on the carrier’s behalf. Not only did
some initial occurrence engender passenger dissatisfaction, but the carrier’s failure to assuage the
dissatisfaction motivated the customer to invest the time and effort into seeking out the DOT.
The Consumer Complaints variable, therefore, is highly appropriate to my purpose of capturing
the quality of service an airline delivers to its customers. Additionally, the measure possesses the
virtue of objectivity, in that it is recorded and reported in a standardized manner by a third party.
77
The third variable used to measure customer service quality, Late and Cancelled Flights,
indicates the 2003 percentage of a carrier’s flights that did not arrive or that arrived 15 or more
minutes later than the scheduled time shown in the carrier’s computerized reservation system.
Cancelled and diverted operations, as well as flights arriving late for any reason, were counted in
the Late and Cancelled Flights variable. Customers value dependability, and carriers have some
discretion over their on-time performance. Airline decisions concerning preemptive
maintenance, number of reserve planes kept on-site, plane turn-around processes, and scheduled
turn-around times all factor in to on-time performance. Thus, Late and Cancelled Flights is a
valid measure of the service quality a carrier delivers to its customers.
I used four variables to measure profitability. For the fiscal year ending 2003, I used both
return on assets and return on equity, designated ROA 2003 and ROE 2003. To mitigate the
influence of short-term variance in profitability, I also used three-year averages of both return on
assets and return on equity, designated ROA 2002-2004 and ROE 2002-2004.
Independent variable. The independent variable multimarket contact posture (MMCP)
was obtained by a three-step process. First, firm-market level multimarket contact (MMC) values
were determined for each market served by a focal firm. Following prior research, I used a count
measure of multimarket contact (Baum & Korn, 1996; Evans & Kessides, 1994; Gimeno &
Woo, 1996, 1999). For each airline in each city-pair market, I counted the number of markets in
which the carrier in question met a specific focal market rival outside the focal market. I
computed MMC by summing all multimarket contacts that the focal firm had with focal market
rivals, then dividing by the number of rivals. In this way, a multimarket contact value was
obtained for each firm in each market. For example, if firm a encountered two rivals in market
m, and one rival shared 200 multimarket contacts with firm a while the other rival shared 400
78
multimarket contacts with firm a, then the MMC value for firm a in market m was 300 ([200 +
400] / 2).
The first step yielded an MMC value for each market served by each firm. Because my
objective was to examine the extent to which multimarket contact posture affects firm level
behavior and outcomes, it was necessary to create an independent variable that captured not only
the level of MMC in each market served by the firm, but also the relative impact that the firm’s
participation in a market was likely to have on firm level strategic orientation. Some markets are
more salient to organizational decision-makers and more important to firm level outcomes than
are others. Competitive position in the Albuquerque-Omaha market is likely to be of less
consequence to most firms than is competitive position in the Chicago-New York market.
Therefore, the second step in calculating MMCP involved weighting each market according to
the proportion of total firm operating revenues for which it accounted. I divided firm operating
revenues derived from the focal market by total firm operating revenues, then multiplied this
proportion by the firm’s MMC value in the focal market. This calculation yielded a revenueweighted MMC value for each firm in each market. I summed these revenue-weighted MMC
values from all markets served by a given firm, obtaining a raw revenue-weighted MMCP value
for each firm.
The third and final step in obtaining MMCP involved indexing the raw MMCP values
according to number of markets served by the focal firm. Carriers participating in a great many
markets are more likely to meet rivals in multiple markets than are carriers that serve fewer
markets. Raw MMCP, therefore, strongly reflects the breadth of a firm’s presence in its industry.
My interest, however, was in the firm’s relative exposure to multimarket relationships, given the
breadth of its presence in its industry. To capture this relative exposure to multimarket
79
relationships, I divided each firm’s raw MMCP value by the number of city-pair markets in the
sample served by the firm. Finally, I multiplied resulting values by 1000 for rescaling. The entire
equation used to calculate multimarket contact posture is represented in Figure 7, and the values
for markets served, raw revenue-weighted MMCP, and multimarket contact posture are listed in
Table 7.
Control variables. The denominator of the equation for MMCP embeds within the
independent variable a control for one dimension of organizational size. The number of markets
served by a firm reflects the breadth of the firm’s resource deployments, but it does not fully
capture the size of the firm’s resource endowments. Resource allocation to promotion, sales, and
passenger service functions, the ability to deliver quality service, and profitability are all likely
influenced by economies of scope and economies or diseconomies of scale. To account for
effects attributable to the latter size dimension, I included the control variable Total Assets at
fiscal year ending 2003. I used the natural logarithm of this variable to correct for skew.
Dependent variables relating to service quality and resource allocations are likely as well to be
sensitive to differences in profitability. More profitable firms may be better equipped than less
profitable firms to devote resources to promotion, sales and passenger service functions, and to
deliver quality service. For this reason, I included the control variable ROE 2002-2004 in
regression analyses for hypotheses 1a, 1b, 2a, 2b, 3a, and 3b. I used ROE averaged across the
three years 2002, 2003, and 2004 to limit the effects of short-term variance in profitability.
RESULTS
Table 8 provides descriptive statistics and Spearman correlations for the study variables. I
tested for multicollinearity by calculating the variance inflation factors (VIFs) in each regression
equation. Multimarket contact posture had VIFs of 1.361 in regression equations for all resource
80
(Revsam)
(Revsan)
…
(MMCam)
+ (MMCan)
(RevsmTot)
(RevsnTot)
X 1000
MMCP =
markets serveda
Where:
MMC = firm-in market level multimarket contact
a = firm a
m, n, . . . = market m, market n, . . .
FIGURE 7: MMCP Equation
81
TABLE 7
Carriers Ranked by MMCP
Carrier
American Eagle Airlines
United Airlines
American Airlines
Delta Airlines
Southwest Airlines
Skywest Airlines
Independence Air
Continental Airlines
Expressjet Airlines
America West Airlines
Northwest Airlines
US Airways
Atlantic Southeast Airlines
ATA Airlines
Alaska Airlines
JetBlue Airways
Airtran Airways
MMCP
449
490
502
511
517
540
546
548
573
575
579
604
634
688
695
714
752
Raw RevenueWeighted
MMCP
53
570
641
699
417
104
63
579
227
335
680
352
96
118
75
42
168
Markets Served
(non-stops &
one-stops)
118
1165
1277
1369
807
192
116
1057
395
583
1174
583
151
171
108
59
223
82
TABLE 8
Descriptive Statistics and Correlations for MMCP Analysisa
Variables
Mean
s.d.
1. MMCP
583.420
86.594
2. Raw Revenue-wghted MMCP
307.048
244.793
-.35
3. Promotion and Sales Expenses
0.067
0.040
.27
.29
4. Promotion and Sales Share
0.752
0.456
.09
.46†
.92***
5. Passenger Service Exps.
0.078
0.028
-.11
.52*
.57*
.65**
6. Passenger Service Share
0.753
0.316
-.14
.63**
.54**
.70**
.94***
7. Mishandled Baggageb
1.502
0.500
-.46†
-.20
-.77***
-.66**
-.36
-.32
8. Consumer Complaints
0.645
0.269
.03
.59*
.22
.36
.42†
.50*
-.27
9. Late and Cncld Flights
0.189
0.034
.24
-.37
-.42†
-.41
-.64**
-.54*
.31
.09
10. ROA 2003
0.038
0.083
.43†
-.42†
-.22
-.43†
-.71**
-.71**
.06
-.35
.31
11. ROA 02-04
-0.042
0.138
-.18
-.38
.03
-.20
-.33
-.48*
.14
-.52*
-.05
.56*
12. ROE 2003
0.034
0.096
.44†
-.50*
-.15
-.38
-.66**
-.66**
.05
-.30
.29
.95***
.60*
13. ROA 02-04
-0.045
0.097
.15
-.39
-.05
-.17
-.28
-.42†
-.22
-.51*
-.20
.56*
.96***
.59*
14. Total Assetsb
8.233
1.328
-.49*
.67**
.43†
.58*
.72**
.79***
-.20
.44†
-.47†
-.74**
-.41
-.64**
a
1
2
3
4
Spearman Correlations (2-tailed)
Natural logarithmic transformation used to correct for adverse skew
*
**
***
†p < .10
p < .05
p < .01
p < .001
b
5
6
7
8
9
10
11
12
13
-.34
83
allocation and service quality variables, and VIFS of 1.360 in regression equations for
profitability variables. Both VIF values are well below the level of 10 that is regarded as
problematic (Ryan, 1997).
I used OLS regression analysis to test my hypotheses. Hypothesis 1a predicts that MMCP
will be negatively related to firm allocation of resource to promotion and sales functions.
Hypothesis 1b, on the other hand, anticipates a positive relationship between MMCP and the
allocation of resources to promotion and sales. Results appear in Table 9. After controlling for
total assets and return on equity, I found that MMCP contributed significantly to prediction of
the two dependent variables tested. Multimarket contact posture predicted statistically significant
variance in promotion and sales expenses (∆R2 = .43, ∆F = 16.06, p < .01) and promotion and
sales share (∆R2 = .36, ∆F = 14.32, p < .01). Overall model F was nonsignificant for promotion
and sales expenses before the introduction of MMCP, but with the inclusion of that single
variable, the model explained a sizable proportion of variance at a high level of statistical
significance (Adjusted R2 = .58, F = 8.22, p < .01). For promotions and sales share, as well,
overall model significance and explained variance increased appreciably with the introduction of
MMCP (Adjusted R2 = .23, F = 3.33, p < .10 without MMCP; Adjusted R2 = .60, F = 9.12, p <
.01 with MMCP). Signs were positive for both regression analyses, supporting Hypothesis 1b.
According to Hypothesis 2a, MMCP will be negatively related to firm resource allocation
to customer service, while Hypothesis 2b predicts a positive relationship between MMCP and the
allocation of resources to customer service. The results, shown in Table 10, indicated that
MMCP contributed significantly to prediction of the two variables tested, controlling for total
assets and profitability. Multimarket contact posture predicted statistically significant variance in
passenger service expenses (∆R2 = .16, ∆F = 5.89, p < .05) and passenger service share (∆R2 =
84
TABLE 9
Results of OLS Regression Analysis for Allocation of Resources to
Promotion and Sales a
Variables
Intercept
Total Assets b
ROE 2002-04 Avg
MMCP
R2
Adjusted R2
∆ R2
Overall Model F
∆F
a
Promotion and Sales Expenses
Step 1
Step 2
-.055
-.358**
(.061)
(.087)
p = .381
p = .001
.015†
.027**
(.008)
(.006)
.501
.898
p = .065
p = .001
.036
.042
(.103)
(.071)
.087
.101
p = .732
p = .570
.000***
(.000)
.762
p = .001
.228
.118
.228
2.071
p = .163
2.071
p = .163
.655
.575
.426
8.219**
p = .003
16.060**
p = .001
Promotion and Sales Share
Step 1
Step 2
-.858
-4.018**
(.656)
(.958)
p = .212
p = .001
.196*
.320***
(.081)
(.066)
.569
.931
p = .029
p = .000
.019
.079
(1.102)
(.789)
.004
.017
p = .987
p = .922
.004**
(.001)
.695
p = .002
.322
.225
.322
3.329†
p = .066
3.329†
p = .066
.678
.603
.355
9.106**
p = .002
14.324**
p = .002
n = 17. Unstandardized regression coefficients, standard errors (in parentheses), and
standardized coefficients (in italics) are presented, in that order.
b
Natural logarithmic transformation used to correct for adverse skew
†
p < .10
*
p < .05
**
p < .01
***
p < .001
85
TABLE 10
Results of OLS Regression Analysis for Allocation of Resources
to Customer Service a
Variables
Intercept
Total Assets b
ROE 2002-04 Avg
MMCP
R2
Adjusted R2
∆ R2
Model F
Passenger Service Expenses
Step 1
Step 2
-.041
-.173*
(.036)
(.063)
p = .267
p = .016
.014**
.020**
(.004)
(.004)
.671
.914
p = .005
p = .001
-.021
-.018
(.060)
(.052)
-.070
-.062
p = .736
p = .732
.000*
(.000)
.466
p = .030
Passenger Service Share
Step 1
Step 2
-.677†
-1.878**
(.336)
(.600)
p = .063
p = .008
.171**
.218***
(.041)
(.042)
.716
.914
p = .001
p = .000
-.573
-.550
(.564)
(.494)
-.176
-.169
p = .327
p = .286
.001*
(.001)
.381
p = .039
.488
.647
.631
.737
.415
.566
.578
.677
.488
.160
.631
.107
6.665**
7.959**
11.954**
12.163***
p = .009
p = .003
p = .001
p = .000
∆F
6.665**
5.891*
11.954**
5.277*
p = .009
p = .030
p = .001
p = .039
a
n = 17. Unstandardized regression coefficients, standard errors (in parentheses), and
standardized coefficients (in italics) are presented, in that order.
b
Natural logarithmic transformation used to correct for adverse skew
†
p < .10
*
p < .05
**
p < .01
***
p < .001
86
.11, ∆F = 5.28, p < .05). Explained variance in passenger service expenses was sizable and
significant for the model without MMCP (Adjusted R2 = .42, F = 6.67, p < .01), but grew
appreciably with the addition of MMCP (Adjusted R2 = .57, F = 7.96, p < .01). Similarly, the
model was effective in predicting passenger service share without MMCP (Adjusted R2 = .58, F
= 11.95, p < .01), but was even more effective with MMCP (Adjusted R2 = .68, F = 12.16, p <
.001). Signs were positive for analyses on both dependent variables, lending support to
Hypothesis 2b.
Hypothesis 3a predicts that MMCP will be negatively related to customer service quality,
while Hypothesis 3b predicts that MMCP will be positively related to customer service quality.
Results for the three measures of customer service quality appear in Table 11. Controlling for
total assets and return on equity in all regression analyses, I found that MMCP contributed
significantly to the prediction of mishandled baggage (∆R2 = .28, ∆F = 6.25, p < .05). The sign
was negative, lending support to Hypothesis 3b, for lower rates of mishandled baggage reflect
better customer service. Overall model F was nonsignificant for mishandled baggage before the
introduction of MMCP, but with MMCP the model explained a statistically significant
proportion of variance (Adjusted R2 = .27, F = 3.01, p < .10). However, MMCP did not
contribute significantly to the prediction of consumer complaints or late and cancelled flights,
nor did models with or without MMCP account for variance in these two dependent variables. In
all, Hypothesis 3b received partial support.
Hypothesis 4a predicts that MMCP will be positively related to profitability; Hypothesis
4b counters that a negative relationship exists between MMCP and profitability. I conducted
regression analyses for four profitability measures: return on assets for the fiscal year 2003;
return on assets averaged over the period 2002 to 2004; return on equity for the fiscal year 2003;
87
TABLE 11
Results of OLS Regression Analysis for Customer Service Quality a
Variables
Intercept
Total
Assets b
ROE 2002-04
Avg
MMCP
R2
Adjusted R2
∆ R2
Model F
∆F
a
Mishandled
Baggage b
Step 1
Step 2
**
2.552
5.646**
(.816)
(1.419)
p = .007
p = .002
-.134
-.256*
(.098)
(.100)
-.681
-.357
p = .022
p = .202
-1.243
-1.302
(1.371)
(1.169)
-.242
-.253
p = .380
p = .286
-.004*
(.001)
-.622
p = .027
Consumer
Complaints
Step 1
Step 2
.212
-.348
(.370)
(.762)
p = .576
p = .655
.045
.067
(.045)
(.053)
.222
.331
p = .339
p = .663
-1.392*
-1.382*
(.621)
(.628)
-.503
-.499
p = .042
p = .046
.001
(.001)
.209
p = .414
Late and
Cancelled Flights
Step 1
Step 2
***
.302
.308**
(.047)
(.099)
p = .000
p = .008
-.015*
-.015*
(.006)
(.007)
-.576
-.585
p = .023
p = .049
-.172*
.172†
(.079)
(.082)
-.492
-.493
p = .046
p = .055
.000
(.000)
-.017
p = .949
.126
.001
.126
1.006
p = .391
1.006
p = .391
.381
.292
.381
4.300*
p = .035
4.300*
p = .035
.377
.287
.377
4.227*
p = .037
4.227*
p = .037
.410
.273
.284
3.007†
p = .069
6.253*
p = .027
.413
.277
.032
3.044†
p = .067
.711
p = .414
.377
.233
.000
2.619†
p = .095
.004
p = .949
n = 17. Unstandardized regression coefficients, standard errors (in parentheses), and
standardized coefficients (in italics) are presented, in that order.
b
Natural logarithmic transformation used to correct for adverse skew
†
p < .10
*
p < .05
**
p < .01
***
p < .001
88
and return on equity averaged over the period 2002 to 2004. I controlled for total assets. The
results, which appear in Tables 12 and 13, did not indicate that multimarket contact posture
contributed significantly to explained variance for any of the four measures. The model with
MMCP did predict variance in ROA 2003 (Adjusted R2 = .32, F = 4.75, p < .05), but with less
power and statistical significance than the model without MMCP (Adjusted R2 = .36, F = 10.16,
p < .01). None of the models predicted statistically significant variance in the other three
profitability measures. Therefore, neither Hypothesis 4a not Hypothesis 4b received support.
DISCUSSION
This chapter considered firm level implications of multimarket contact. I introduced
MMCP—a construct and associated measure reflecting firm level multimarket contact—and
proposed competing hypotheses on the reach of mutual forbearance effects. One set of
hypotheses represented the pervasive forbearance argument that MMCP mutes nonprice
competitive intensity. An opposing set of hypotheses represented the partial forbearance
argument that MMCP amplifies nonprice competition. Profitability implications for each set of
hypotheses were proposed as well. I tested my hypotheses in the airline industry, where
numerous existing studies report evidence of mutual forbearance in pricing. On balance, results
backed the partial forbearance perspective. In support of partial forbearance hypotheses 1b and
2b, findings indicate a positive relationship between MMCP and the allocation of resources to
promotion and sales and to customer service. Partial forbearance hypothesis 3b received mixed
support. A positive relationship was found between MMCP and baggage-handling service
quality, but no relationship appeared between MMCP and either rate of consumer complaints or
on-time performance. Finally, no link was found between MMCP and profitability, indicating
89
TABLE 12
Results of OLS Regression Analysis for Return on Assets a
Variables
Intercept
Total Assets b
ROA 2003
Step 1
.365**
(.104)
p = .003
-.040**
(.012)
-.635
p = .006
Step 2
.340
(.225)
p = .154
-.039*
(.015)
-.620
p = .022
.000
(.000)
.031
p = .901
ROA 2002-04 Avg
Step 1
Step 2
.059
.367
(.223)
(.475)
p = .794
p = .453
-.012
-.024
(.027)
(.032)
-.118
-.233
p = .652
p = .456
.000
(.000)
-.224
p = .474
.404
.364
.404
10.158**
p = .006
10.158**
p = .006
.404
.319
.001
4.754*
p = .027
.016
p = .901
.014
-.052
.014
.212
p = .652
.212
p = .652
MMCP
R2
Adj. R2
∆R2
Model F
∆F
a
n = 17.Unstandardized regression coefficients, standard errors (in parentheses), and
standardized coefficients (in italics) are presented, in that order.
b
Natural logarithmic transformation used to correct for adverse skew
*
p < .05
**
p < .01
.051
-.085
.037
.374
p = .695
.542
p = .474
90
TABLE 13
Results of OLS Regression Analysis for Return on Equity a
Variables
Intercept
Total Assets b
ROE 2003
Step 1
.346*
(.132)
p = .020
-.038*
(.016)
-.524
p = .031
Step 2
.177
(.283)
p = .541
-.031
(.019)
-.433
p = .119
.000
(.000)
.177
p = .509
ROE 2002-04 Avg
Step 1
Step 2
.165
.186
(.148)
(.321)
p = .281
p = .571
-.026
-.026
(.018)
(.021)
-.349
-.360
p = .170
p = .238
.000
(.000)
-.022
p = .941
.275
.227
.275
5.688*
p = .031
5.688*
p = .031
.298
.198
.023
2.971†
p = .084
.459
p = .509
.122
.063
.122
2.076
p = .170
2.076
p = .170
MMCP
R2
Adj. R2
∆R2
Model F
∆F
a
.122
-.003
.000
.972
p = .402
.006
p = .941
n = 17. Unstandardized regression coefficients, standard errors (in parentheses), and
standardized coefficients (in italics) are presented, in that order.
b
Natural logarithmic transformation used to correct for adverse skew
†
p < .10
*
p < .05
91
that neither pervasive forbearance hypothesis 4a nor partial forbearance hypothesis 4b received
support.
It is not surprising that findings were stronger for resource allocation variables than for
customer service variables. Resource allocation reflects the depth of a firm’s commitment to
particular objectives. I interpret relative resource allocation to promotion, sales, and passenger
service as indicative of a firm’s intent to compete or not compete vigorously along those
dimensions. Commitment and intent, however, do not necessarily translate to success. The
quality of service a carrier actually delivers is conditioned not only by the carrier’s intent, but as
well by the carrier’s effectiveness in actualizing its intent. While the hypothesized relationship
between MMCP and intent to deliver quality service is direct, the relationship between MMCP
and quality of service delivered is indirect. The latter relationship is mediated by intent and
moderated by managerial effectiveness. Thus, one might expect stronger relationships between
MMCP and resource allocation than between MMCP and passenger service delivered. This may
explain why strong results were obtained for the former while the latter met with mixed results.
Measurement issues represent an alternate explanation for mixed customer service
results. The lack of significant findings for consumer complaints and late and cancelled flights
may derive from problems pertaining to those dependent variables specifically. Reliability of the
variable consumer complaints may be compromised by extremely low report rates for all airlines.
The mean report rate for all carriers for 2003 was .645 per 100,000 passengers, or just 1
complaint reported per 155,111 passengers. No airline had a consumer report rate greater than
.95 per 100,000 passengers. Therefore, the variable was exceedingly sensitive to small
variations, and may not constitute a reliable indicator of consumer dissatisfaction with service
quality. The variable Late and Cancelled Flights may measure a service dimension over which
92
carriers exercise too little discretion for the purposes of this study. While airlines can take certain
actions to mitigate delays and cancellations, such as keeping extra planes on hand in case of
maintenance problems or building slack into schedules, multiple other factors may drown out
those that carriers can influence. Flights are delayed or cancelled for many reasons beyond
carriers’ control, including weather, airport security issues, and random passenger behavior.
Minimal carrier discretion over variance in outcomes measured by Late and Cancelled Flights,
therefore, may account for the nonsignificant relationship between the variable and MMCP.
The absence of significant effects for any of the profitability variables is interesting, in
light of previous studies’ support of a positive relationship between MMC and margins at the
firm-market level (Gimeno & Woo, 1999; Singal, 1996). A possible explanation is that increases
in firm-market level profit margins are consumed at the firm level by larger marketing and
customer service budgets. Alternately, MMC’s effect on firm-market level margins may simply
be too minor to manifest at the firm level. A multitude of factors shape firm financial
performance, with environmental determinants such as oil prices and terrorist threats especially
salient in the airline industry. While the airline industry is conducive to examining links between
MMCP and competitive intent/behavior, the link between competitive behavior and firm
financial performance in this industry may be unusually tenuous. Less regulated industries that
are less exposed and sensitive to environmental shocks may constitute better arenas for
examining profitability effects.
The robust effect sizes found for promotion and sales expenses, promotion and sales
share, passenger service expenses, passenger service share, and mishandled baggage ground
two major deductions with important implications for multimarket contact theory and research.
First, this study establishes that multimarket contact contributes to the explanation of firm level
93
behaviors and outcomes. The MMCP construct I propose and measure bridges the gap between
levels of analysis, for it is compiled of firm-market level data and it accounts for sizable variance
in firm level outcomes. Results represent the first empirical confirmation of cross-level
relationships deriving from multimarket contact. The theoretical implication is that part-whole
relationships merit deeper consideration in the multimarket contact literature. Traditionally,
MMC theory cleaves organizations, addressing how contact between segments of rival
organizations affects segmented behavior. The effects found for MMCP underscore the
inextricability of organizational segment and organizational whole. Micro- and macro-level
decisions and behaviors are mutually constitutive. Decisions and behaviors at the periphery
invoke firm-wide vision—at least if they are to bear an MMC imprint—while, reciprocally,
organizational strategic orientation is produced by contacts and actions experienced at the
periphery. Research implications arising from this study are substantial. Firm level measurement
of multimarket contact activates a broad range of potential new dependent variables reflecting
strategic impact.
A second major point emerging from this study is that multimarket contact amplifies
rather than mutes competitive intensity along certain dimensions. Carriers with high multimarket
contact postures were more likely than those with lower MMCPs to channel resources toward
promotion, sales, and customer service. Prevailing MMC theory, which emphasizes the mutual
forbearance hypothesis, neither predicts nor explains these findings. Evidence of a positive
relationship between MMCP and nonprice competitive intensity signifies the need for
reconceptualizing the mutual forbearance hypothesis as bounded and contingent. Specifically, it
appears that multimarket contact may engender forbearance only with regard to forms of rivalry
that are readily monitored or particularly damaging. The research implication is that greater
94
emphasis is due nonprice competitive dimensions in the multimarket contact literature. Whereas
research traditionally asks whether MMC engenders mutual forbearance in the form of higher
prices and margins in a given context—assuming a yes or no answer—this study’s results
support a re-orientation toward asking how MMCP affects rivalry, where it mutes rivalry, and
where it amplifies rivalry.
An alternate explanation for nonprice rivalry amplification merits consideration. The
most influential studies linking multimarket contact to price and entry/exit forbearance in the
airline industry have drawn on data from the years 1979 to 1988, or within the first decade
following the 1978 federal deregulation. Gimeno (1999) and Gimeno and Woo (1996, 1999)
examined data from 1984 to 1988; Singal (1996) examined data from 1985 to 1988; and Baum
and Korn (1996, 1999) and Korn and Baum (1999) studied data from 1979 to 1984. This study,
on the other hand, examines data from 2003. Thus, my assumption that multimarket contact
amplifies marketing and customer service competition while muting price competition may be
inaccurate. Existing evidence for price forbearance may be outdated. It is possible that in 2003,
multimarket contact amplified price as well as nonprice competition.
Chapter four’s discussion of system openness, competitive vacuums, and semi-stable
concentration in MMC cohorts affords a theoretical basis for questioning whether price
forbearance has persisted since the mid-1980s. Chapter 4 elaborated a rationale for concentration
level decrease in MMC contexts over time, and as Scott (1982, 1991, 1993) demonstrates, the
mutual forbearance hypothesis collapses in the absence of high concentration. Under low
concentration conditions, MMC is negatively associated with profitability. As concentration
declines, familiarity erodes, coordination becomes unwieldy, monitoring becomes more costly,
and mutual recognition of competitive interdependence fades. As a result, tacit collusion
95
disintegrates and rivalry ensues. Therefore, this study’s findings that MMC amplifies nonprice
rivalry may reflect low industry concentration levels rather than divergent price and nonprice
effects.
The spate of new entrants into the airline industry since the 1978 deregulation lends
support to the low concentration argument. Tacit collusion borne of—and nurtured by—
governmental regulation may have persisted into the early- to mid-1980s time period comprising
the focus of influential MMC studies. However, during the past two decades, competitive
vacuums engendered by competence depletion may have combined with system openness to
induce new entry, consistent with the theoretical framework advanced in chapters 3 through 6.
Future studies might extend analysis of MMC price effects past the 1980s and up to the present,
in order to determine whether they parallel or diverge from nonprice amplification.
Limitations and future research. Although this study sheds light on firm level
implications of multimarket contact, it has certain limitations. The direction of causal pathways
cannot be assessed confidently under the study’s cross-sectional design. While it is more
plausible that MMCP affects marketing and service orientation than vice versa, it cannot be ruled
out that firms with higher marketing and service orientations are more likely to enter into
multimarket contact. Longitudinal research is needed to confidently assess the causal direction
between specified relationships. The study’s relatively small sample size de-sensitizes OLS
regression analyses to all big very large effect sizes. Larger sample sizes will better equip future
researchers to detect relationships between MMCP and nonprice competitive behavior.
Generalizability is a concern for any study focused on a single industry. Examination of other
industries is needed to determine whether hypothesized relationships hold beyond the U.S.
passenger airline industry. A final limitation pertains to the range of dependent variables used to
96
measure nonprice competition. On-time performance, baggage-handling complaints, general
consumer complaints, and resource allocation to promotion and sales and to passenger service
represent important indicators of marketing and customer service competition, but they do not
capture the full dimensionality of nonprice competitive orientation. Other industries may afford
better measures of customer service delivered, as well as the opportunity to examine additional
nonprice competitive dimensions such as innovation.
The two lines of inquiry introduced in this study might be advanced in several directions.
Research exploring both firm level MMC implications and forbearance parameters should
concentrate on resolving three fundamental issues. First, what attributes of a competitive
dimension promote forbearance and what attributes intensify rivalry under multimarket contact?
For example, does firm proclivity toward forbearance or rivalry along a given dimension ride on
the difficulty of communicating and interpreting competitive behavior, or does the potential for a
form of competition to erode profit margins or grow overall market size appear more
instrumental? Specification of the properties of competitive dimensions driving forbearance and
the properties driving rivalry entails further examination of the nonprice dimensions featured in
this study, as well as examination of alternate nonprice dimensions.
Second, future research should identify contextual moderators influencing relationships
between MMCP and competitive behavior. For example, do industry concentration levels
moderate the MMCP-rivalry relationship? Analysis of this issue will entail either a multiindustry approach where concentration levels vary between industries or a longitudinal singleindustry approach where concentration varies within the same industry over time. Similarly,
future research might explore moderators of the relationship between MMCP and firm financial
performance. For instance, if MMCP is found to mute nonprice rivalry along marketing or
97
customer service dimensions in some contexts, or along alternate nonprice dimensions such as
innovation, might muted rivalry along a given dimension enhance firm financial performance in
certain contexts while undermining performance in other contexts? In particular, might variance
in concentration levels resulting from industry entry by new competitors be necessary to expose
competence depletion among industry incumbents to the detriment of financial performance?
Analysis of this issue will entail longitudinal examination of an industry with both variance in
concentration levels and a link between firm competencies and firm financial performance that is
not heavily diluted by the influence of external environmental factors.
Third, future research should seek a deeper understanding of the process by which
MMCP mutes or amplifies firm level rivalry. Do the behaviors and outcomes of market-level
decisions percolate upward to the firm level, or do market-level structural relationships
aggregately affect firm level decisions? For instance, does market-level MMC affect marketlevel promotional activity, which firm level measurement merely reflects? Or does market-level
MMC define firm level MMCP, which in turn affects firm level promotional activity? While my
study found evidence of an overall relationship between MMCP and firm level behaviors and
outcomes, I was unable to specify the level at which decision processes were actually affected by
multimarket contact because the data were not conducive to examining nonprice competitive
behaviors and decisions at the firm-market level. Future research in other industries should seek
dependent variables at the firm-market level reflecting resource allocation to marketing, resource
allocation to customer service, customer service delivered, and so forth. Comparisons between
MMCP effects at the firm-market level and the firm level will enable researchers to parse out the
relative influence of decisions at each level, and thus to more precisely explain the process by
which MMCP mutes or amplifies rivalry.
98
Conclusion. This chapter has highlighted a latent tension between rivalry reduction and
rivalry amplification under multimarket contact. The evidence provided here compels a broader,
more complex conceptualization of multimarket contact than currently pervades the literature.
The well-established tendency to reduce rivalry represents but one facet of multimarket contact’s
contradictory nature. A janus-faced MMCP emerges from this study, with rivalry amplification
representing the understudied countenance. The ramifications of multimarket contact extend well
beyond the bounds of the mutual forbearance hypothesis, and so too should future research.
Subsequent work in this domain will benefit from substituting a partial forbearance perspective
for pervasive forbearance assumptions. Further validation of the mutual forbearance hypothesis
represents a less pressing need than does resolution of the parameters of rivalry reduction and
amplification under multimarket contact. It is my hope that, by divulging another side of
multimarket contact, this study will promote a more comprehensive accounting of MMC’s
strategic impact.
99
CHAPTER 9
CONCLUSION
This dissertation has extended multimarket contact theory beyond its current focus on
inter-firm price collusion. The principal theme informing my view has been that MMC has
implications for competence development. Chapters 3 through 6 developed a broad theoretical
framework tying together multimarket contact, competence depletion at the firm level,
concentration variance at the population level, and long-term firm performance. In those
chapters, I argued that multimarket contact has latent adverse strategic consequences.
Multimarket contact engenders competence depletion, I theorized, which in turn both induces
and converges with punctuated forbearance to undermine long-term form performance.
The empirical analyses comprising chapters 7 and 8 addressed particular relationships
embedded in the broader theoretical framework advanced in the earlier chapters. Chapter 7
examined the relationship between MMC and service quality at the firm-market level. The
rationale guiding this analysis was that a negative relationship between MMC and service quality
would provide initial evidence of a relationship between multimarket contact and competence
depletion along the service quality dimension. While results were statistically significant in the
predicted direction, effect sizes were too minimal to be interpreted as convincing evidence of
competence depletion. Chapter 8 introduced a new measure of firm level MMCP, and then
examined relationships between MMCP and resource allocation to promotion and sales, resource
allocation to customer service, customer service delivered, and profitability. Effect sizes were
large and statistically significant for the resource allocation variables and for one dimension of
100
customer service quality. However, relationships were positive, suggesting that MMCP
amplified nonprice competition rather than depleted competencies by dampening rivalry.
Thus, chapter 7 yielded results consistent with the broader theoretical framework but of a
size too minimal to lend convincing support, while chapter 8 generated results inconsistent with
the broader theoretical framework. How should results from the two chapters be interpreted in
light of one another, and what implications do the findings have for the broad theoretical
framework developed in chapters 3 through 6? Differences between the firm and firm-market
level results may be attributable to measurement differences. While customer service was
measured at both levels, it was operationalized differently. Carrier delay (firm-market level) and
mishandled baggage (firm level) represent distinct customer service dimensions reflecting
distinct firm competences. Furthermore, both variables involve report mechanisms that
complicate the relationship between the variable measured and the underlying construct. Finally,
the firm level analysis included variables for which there were no comparables at the firmmarket level, such as resource allocation to promotion and sales and to customer service. Thus,
caution is in order when interpreting results from chapters 7 and 8 in light of one another and
with an eye toward implications for the broad theoretical framework developed at the front of
this dissertation.
Nevertheless, the results do lend themselves to theoretically substantive interpretation.
One possible interpretation is that inconsistencies between the firm and firm-market level results
express real differences in the way multimarket contact affects nonprice competitive activity at
the two levels. It is possible that firm-market level MMC has little effect on firm-market level
service quality, while at the same time positively affecting service quality and intent to deliver
service quality (as well as intent to compete along marketing dimensions) at the firm level.
101
Customer service (and marketing ) decisions emanating from the corporate center may reflect a
strategic posture or mindset shaped in part by the aggregate of the firm’s pattern of contacts with
rivals across the breadth of its markets. Those centralized decisions may affect overall firm
competitive intent, orientation, and outcomes without affecting outcomes at the firm-market
level in proportion to MMC values at the firm-market level. In the passenger airline industry,
this interpretation stands to reason. Decisions pertaining to baggage handling systems,
turnaround time, advertising campaigns, and other quality or marketing initiatives tend to be
under the purview of the corporate center and tend to affect firm-wide outcomes. If overall firm
multimarket contact posture does affect firm competitive posture, therefore, we might expect
variance in competitive outcomes at the firm level but not at the firm-market level.
While the above interpretation of results highlights implications for our understanding of
the level at which processes occur, the single most important implication of the empirical
analyses pertains to the positive/negative sign of the relationship between multimarket contact
and both competitive activity and competence development. On balance, empirical results favor
the arguments for partial forbearance and nonprice rivalry amplification over the arguments for
pervasive forbearance and nonprice rivalry reduction. This, of course, contradicts the argument
for competence depletion advanced in chapters 3 through 6, and undermines the conclusion that
multimarket contact threatens long-term firm performance. Future research in other industry
contexts may provide support for the competence depletion argument, but in this dissertation
support was limited to small effect sizes at only the firm-market level.
The firm level indications of rivalry amplification are important in and of themselves.
This dissertation marks the first attempt to conceptualize multimarket contact at the firm level,
and the first to find and explain a link between any measure of multimarket contact and
102
heightened competitive activity. In mapping a debate between partial and pervasive forbearance
arguments—and between competence depletion and rivalry amplification hypotheses—this
dissertation establishes the guideposts for future research. The competing hypotheses introduced
here underscore the need to more seriously consider the attributes of the competitive dimensions
examined in the MMC research domain. The turn toward rivalry amplification on the one hand
or rivalry reduction on the other hand may hinge, for example, on the degree of ease in signaling
and interpreting competitive intent along a given competitive dimension. If future research
confirms—as this study suggests—that multimarket contact amplifies rivalry along nonprice
dimensions such as marketing and service quality, then a fundamental shift in the prevailing
understanding of MMC’s effect on firm performance will be warranted. Re-conceptualized as an
antecedent both to rivalry reduction in pricing and to rivalry amplification along nonprice
dimensions, MMC appears to confer an ideal combination of positional advantage and spur to
competence enhancement. Ironically, in contrast to the framework developed in chapters 3
through 6, this view would suggest that multimarket contact enhances firm performance even
more strongly than the prevailing mutual forbearance hypothesis indicates.
This dissertation has mustered a substantial body of theory to inform fresh and competing
approaches to multimarket contact. The future of MMC theory lies in the hands of empiricists.
While empirical results in this dissertation lean toward the partial forbearance and nonprice
rivalry amplification perspectives, theory supporting the opposing perspectives remains too
strong to discard absent additional empirical work.
103
REFERENCES
Abrahamson, E., & Fombrun, C. 1994. Macrocultures: Determinants and Consequences.
Academy of Management Review, 19: 728-755.
Adams, W. J. 1974. Market structure and corporate power: The horizontal dominance hypothesis
reconsidered. Columbia Law Review, 74: 1276-1297.
Alexander, D. L. 1985. An empirical test of the mutual forbearance hypothesis: The case of bank
holding companies. Southern Economic Journal. 51: 122-140.
Amburgey, T. L., Kelly, D., & Barnett, W. P. 1993. Resetting the clock: The dynamics of
organizational change and failure. Administrative Science Quarterly, 38: 51-73.
Ahuja, G., & Lampert, C. M. 2001. Entrepreneurship in the large corporation: A longitudinal
study of how established firms create breakthrough inventions. Strategic Management
Journal, 22: 521-543.
Areeda, P. & Turner, D. 1979. Conglomerate mergers: Extended interdependence and effects on
interindustry competition as grounds for condemnation. University of Pennsylvania Law
Review, 127: 1082-1103.
Arrow, K. 1962. Economic welfare and the allocation of resources for inventions. In Nelson, R.,
(Ed.), The rate and direction of inventive activity. Princeton, NJ: Princeton University
Press.
Arthur, W. B. 1989. Competing technologies, increasing returns, and lock in by historical events.
Economic Journal, 99: 116-131.
104
Arthur, W. B. 1994. Increasing returns and path dependence in the economy. Ann Arbor, MI:
University of Michigan Press.
Bain, J. S. 1959. Industrial Organization. New York: Wiley.
Banker, R. D., Khosla, I., & Sinha, K. K. 1998. Quality and competition. Management Science,
44: 1179-1192.
Barnett, W. P. 1993. Strategic deterrence among multipoint competitors. Industrial and
Corporate Change, 2: 249-278.
Barnett, W. P., Greve, H. R., & Park, D. Y. 1994. An evolutionary model of organizational
performance. Strategic Management Journal, 15: 11-28.
Barnett, W. P., & Hansen, M. T. 1996. The red queen in organizational evolution. Strategic
Management Journal, 17: 139-157.
Barney, J. B. 1991. Firm resources and sustained competitive advantage. Journal of
Management, 17: 99-120.
Baum, J. A. C., & Korn, H. J. 1996. Competitive dynamics of inter-firm rivalry. Academy of
Management Journal, 39: 255-291.
Baum, J. A. C., & Korn, H. J. 1999. Dynamics of dyadic competitive interaction. Strategic
Management Journal, 20: 251-278.
Bernheim, D., & Whinston, M. D. 1990. Multimarket contact and collusive behavior. RAND
Journal of Economics, 21: 1-26.
Blundell, R., Griffith, R., & Van Reenen, J. 1995. Dynamic count data models of technological
innovation. The Economic Journal, 105: 333-344.
Boeker, W., Goodstein, J., Stephan, J., & Murmann, J. P. 1997. Competition in a multimarket
environment: The case of market exit. Organization Science, 8: 126-142.
105
Bonanno, G., & Haworth, B. 1998. Intensity of competition and the choice between product and
process innovation. International Journal of Industrial Organization, 16: 495-510.
Boone, J. 2001. Intensity of competition and the incentive to innovate. International Journal of
Industrial Organization, 19: 705-726.
Busse, M. R. 2000. Multimarket contact and price coordination in the cellular telephone industry.
Journal of Economics & Management Strategy, 9: 287-320.
Cassidy, C. M., & Loree, D. 2001. Dynamics of knowledge transfer among multimarket
competitors. In J. A. C. Baum & H. R. Greve, (Eds.), Advances in strategic
management: Multiunit organization and multimarket strategy, vol. 18: 141-174.
Stamford, CT: JAI Press.
Caves, R. E. 1980. Industrial organization, corporate strategy, and structure. Journal of
Economic Literature, 18: 64-92.
Caves, R. E. (Ed.) 1992. Industrial efficiency in six nations. Cambridge, MA: MIT Press.
Caves, R. E. & Barton, D. R. 1990. Efficiency in U.S. manufacturing industries. Cambridge,
MA: MIT Press.
Caves, R. E. & Porter, M. E. 1977. From entry barriers to mobility barriers. Quarterly Journal
of Economics, 241-261.
Conner, K. R. 1991. A historical comparison of resource-based theory and five schools of
thought within industrial organization economics: Do we have a new theory of the firm?
Journal of Management, 17: 121-154.
Cyert, R. M. & March, J. G. 1963. A behavioral theory of the firm. (2nd ed.). Malden, MA:
Blackwell Publishers.
106
D’Aveni, R. A. 1994. Hypercompetition: Managing the dynamics of strategic maneuvering.
New York: Free Press.
David, P. A. 1985. Clio and the economics of QUERTY. American Economic Review, 75: 332337.
Delbono, F. & Denicolo, V. 1990. R&D investment in a symmetric and homogeneous oligopoly.
International Journal of Industrial Organization, 8: 297-313.
DiMaggio, P. J. & Powell, W. W. 1983. The iron cage revisited: Institutional isomorphism and
collective rationality in organizational fields. American Sociological Review, 48: 147160.
Edwards, C. D. 1955. Conglomerate bigness as a source of power. In G. Stigler, (Ed.), Business
concentration and price policy (pp. 331-359). Princeton, NJ: Princeton University Press.
Evans, W. N., & Kessides, I. N. 1994. Living by the “golden rule”: Multimarket contact in the
U.S. airline industry. Quarterly Journal of Economics, 109: 341-366.
Feinberg, R. M. 1985. Sales-at-risk: A test of the mutual forbearance theory of conglomerate
behavior. Journal of Business, 58: 225-241.
Fernandez, N., & Marin, P. L. 1998. Market power and multimarket contact: Some evidence
from the Spanish hotel industry. The Journal of Industrial Economics, 46: 301-315.
Fu, W. W. 2003. Multimarket contact of U.S. newspaper chains: Circulation competition and
market coordination. Information Economics and Policy, 15: 501-519.
Geroski, P. A. 1990. Innovation, technological opportunity, and market structure. Oxford
Economic Papers, 42: 586-602.
Gimeno, J. 1999. Reciprocal threats in multimarket rivalry: Staking out ‘spheres of influence’ in
the U.S. airline industry. Strategic Management Journal, 20: 101-128.
107
Gimeno, J. & Jeong, E. 2001. Multimarket contact: Meaning and measurement at multiple levels
of analysis. In J. A. C. Baum & H. R. Greve, (Eds.), Advances in strategic management:
Multiunit organization and multimarket strategy, vol. 18: 357-408. Stamford, CT: JAI
Press.
Gimeno, J., & Woo, C. Y. 1996. Hypercompetition in a multimarket environment: The role of
strategic similarity and multimarket contact in competitive de-escalation. Organization
Science, 7: 322-341.
Gimeno, J., & Woo, C. Y. 1999. Multimarket contact, economies of scope, and firm
performance. Academy of Management Journal, 42: 239-259.
Greve, H. R., & Mitsuhashi, H. 2004. Multiunit organization and multimarket strategy: The
dynamics of market entry and commitment. Scandinavian Journal of Management, 20:
9-30.
Haveman, H. A., & Nonnemaker, L. 2000. Competition in multiple geographic markets: The
impact on growth and market entry. Administrative Science Quarterly, 45: 232-267.
Heggestad, A. A. & Mingo, J.J. 1976. Prices, nonprices, and concentration in commercial
banking. Journal of Money, Credit, and Banking, 8: 107-117.
Heggestad, A. & Rhoades, S. 1978. Multi-market interdependence and local market competition
in banking. The Review of Economics and Statistics, 4: 523-532.
Hughes, K. & Oughton, C. 1993. Diversification, multimarket contact, and profitability.
Economica, 60: 203-224.
Jans, I., & Rosenbaum, D. I. 1996. Multimarket contact and pricing: Evidence from the U.S.
cement industry. International Journal of Industrial Organization, 15: 391-412.
108
Jayachandran, S., Gimeno, J., & Varadarajan, R. 1999. The theory of multimarket competition:
A synthesis and implications for marketing strategy. Journal of Marketing, 63: 49- 66.
Karnani, A., & Wernerfelt, B. 1985. Multiple point competition. Strategic Management
Journal, 6: 87-96.
Kim, L. 1998. Crisis construction and organizational learning: Capability building in catching-up
at Hyundai Motor. Organization Science, 9: 506-521.
Korn, H. J., & Baum, J. A. 1999. Chance, imitative, and strategic antecedents to multimarket
contact. Academy of Management Journal, 42: 171-193.
Lant, T. K., Milliken, F. J., & Batra, B. 1992. The role of managerial learning and interpretation
in strategic persistence and reorientation. Strategic Management Journal, 13: 585-608.
Lengnick-Hall, C. A., & Wolff, J. A. 1999. Similarities and contradictions in the core logic of
three strategy research streams. Strategic Management Journal, 20: 1109-1132.
Levinthal, D. A., & March, J. G. 1993. The myopia of learning. Strategic Management Journal.
14 (Winter Special Issue): 95-112.
Levitt, B. & March, J. G. 1988. Organizational learning. Annual Review of Sociology, 14: 319340.
March, J. G. 1988. Decisions and organizations. Cambridge, MA: Basil Blackwell.
March, J. G., & Simon, H. A. 1958. Organizations. (2nd ed.). Cambridge, MA: Blackwell.
Mazzeo, M. J. 2003. Competition and service quality in the U.S. Airline Industry. Review of
Industrial Organization, 22: 275-296.
Miller, D. 1990. The Icarus paradox: How exceptional companies bring about their own
downfall. New York, NY: Harper Business.
109
Miller, D., & Chen, M. 1994. Sources and consequences of competitive inertia: A study of the
U.S. airline industry. Administrative Science Quarterly, 39: 1-23.
Miller, D. & Friesen, P. 1984. Organizations: A quantum view. Englewood Cliffs, NJ: Prentice
Hall.
Nelson, R. R., & Winter, S. G. 1974. Neoclassical vs. evolutionary theories of economic growth:
Critique and prospectus. Economic Journal, 84: 886-905.
Nelson, R. R., & Winter, S. G. 1982. An evolutionary theory of economic change. Cambridge,
MA: Harvard University Press.
Nickell, S. J. 1996. Competition and corporate performance. Journal of Political Economy, 104:
724-746.
Nickell, S. J., Wadhwani, S. B., & Wall, M. 1992. Productivity growth in U.K. companies, 19751986. European Economic Review, 36: 1055-85.
Oliver, C. 1991. Strategic responses to institutional processes. Academy of Management
Review, 16: 145-179.
Oliver, C. 1997. Sustainable competitive advantage: Combining institutional and resource-based
views. Strategic Management Journal, 18: 697-713.
Parker, P. M., & Roller, L. H. 1997. Collusive conduct in duopolies: Multimarket contact and
cross-ownership in the mobile telephone industry. RAND Journal of Economics, 28:
304-322.
Peteraf, M. & Shanley, M. 1997. Getting to know you: A theory of strategic group identity.
Strategic Management Journal, 18 (summer special issue): 165-186.
Porter, M. E. 1980. Competitive Strategy. New York: Free Press.
110
Porter, M. E. 1985. Competitive advantage: Creating and sustaining superior performance.
New York: The Free Press.
Porter, M. E. 1990. The competitive advantage of nations. London: MacMillan.
Porter, M. E. 1991. Towards a dynamic theory of strategy. Strategic Management Journal, 12
(Special Issue): 95-117.
Prahalad, C. K., & Hamel, G. 1990. The core competence of the corporation. Harvard Business
Review, 68(3): 79-93.
Rubin, R. & Joy, J. 2005. Where are the airlines headed? Implications of airline industry
structure and change for consumers. The Journal of Consumer Affairs, 39: 215-228.
Ryan, T. Y. 1997. Modern Regression Analysis. New York: Wiley.
Schumpeter, J. 1942. Capitalism, socialism, and democracy. (3rd ed.). New York: Harper and
Row.
Scott, J. T. 1982. Multimarket contact and economic performance. Review of Economics and
Statistics, 64: 368-375.
Scott, J. T. 1991. Multimarket contact among diversified oligopolists. International Journal of
Industrial Organization, 9: 225-238.
Scott, J. T. 1993. Purposive diversification and economic performance. Cambridge, UK:
Cambridge University Press.
Scott, J. T. 2001. Designing multimarket-contact hypothesis tests: Patent citations and
multimarket contact in the product and innovation markets of the chemicals industry. In
J. A. C. Baum & H. R. Greve, (Eds.), Advances in strategic management: Multiunit
organization and multimarket strategy, vol. 18: 175-202. Stamford, CT: JAI Press.
111
Sellers, P. 1995. Can Coke and Pepsi make Quaker sweat? Fortune, 132 (issue 1, July 10, 1995):
20.
Shankar, V. 1999. New product introduction and incumbent response strategies: Their
interrelationship and the role of multimarket contact. Journal of Marketing Research,
36: 327-344.
Singal, V. 1996. Airline mergers and multimarket contact. Managerial and Decision
Economics, 17: 559-574.
Stephan, J., & Boeker, W. 2001. Getting to multimarket competition: How multimarket contact
affects firms’ market entry decisions. In J. A. C. Baum & H. R. Greve, (Eds.), Advances
in strategic management: Multiunit organization and multimarket strategy, vol. 18:
229-261. Stamford, CT: JAI Press.
Surowiecki, J. 2004. The wisdom of crowds. New York: Doubleday.
Symeonidis, G. 2000a. Price competition, non-price competition, and market structure: Theory
and evidence from the U.K. Economica, 67: 437-456.
Symeonidis, G. 2000b. Price and nonprice competition with endogenous market structure.
Journal of Economics & Management Strategy, 9: 53-83.
Teece, D. J., Pisano, G., & Shuen, A. 1997. Dynamic capabilities and strategic management.
Strategic Management Journal, 18: 509-533.
Tushman, M., & Romanelli, E. 1985. Organizational evolution: A metamorphosis model of
convergence and reorientation,” in Cummings, L., & Staw, B., (Eds.), Research in
Organizational Behavior, 7: 171-222.
112
Vasigh, B. & Fleming, K. 2005. A total factor productivity based structure for tactical cluster
assessment: Empirical investigation in the airline industry. Journal of Air
Transportation, 10: 3-19.
Wernerfelt, B. 1984. A resource-based view of the firm. Strategic Management Journal, 5: 171180.
Will, T. E. 2006. Dynamic theory of multimarket contact: Competence depletion, punctuated
forbearance, and performance. In K. Mark Weaver (Ed.), Proceedings of the Sixty-fifth
Annual Meeting of the Academy of Management (CD), ISSN 1543-8643.
Winter, N. 2000. The satisficing principle in capability learning. Strategic Management
Journal, 21: 981-996.
Young, G., Smith, K. G., & Grimm, C. M. 1996. “Austrian” and industrial organization
perspectives on firm level competitive activity and performance. Organization Science,
7: 243-254.
Young, G., Smith, K. G., Grimm, C. M., & Simon, D. 2000. Multimarket contact and resource
dissimilarity: A competitive dynamics perspective. Journal of Management, 26: 12171236.
Zahra, S. & George, G. 2002. Absorptive capacity: A review, reconceptualization, and extension.
Academy of Management Review, 27: 185-203.