Business Strategy and Forecasting as Competitive Advantages
ccbtta
RedQueenEffect.pdf
THE RED QUEEN EFFECT: COMPETITIVE ACTIONS AND FIRM PERFORMANCE
PAMELA J. DERFUS
PATRICK G. MAGGITTI Temple University
CURTIS M. GRIMM KEN G. SMITH
University of Maryland
We investigate the Red Queen effect as a contest of competitive moves or actions among rivalrous firms. The results from a multi-industry study of over 4,700 actions confirms the existence of Red Queen competition, whereby a firm’s actions increase perfor- mance but also increase the number and speed of rivals’ actions, which, in turn, negatively affect the initial firm’s performance. We further show that this Red Queen effect depends on industry context and a focal firm’s market position.
The quest to explain performance differences among competing firms is a fundamental issue in strategic management. A number of answers to this complex question have been offered. According to the industry structure viewpoint, positioning firms in industries where they can take advantage of fa- vorable competitive forces, such as barriers to entry or mobility (Caves & Porter, 1977), enhances per- formance. The resource-based view also empha- sizes limiting the behavior of rivals by suggesting that firms acquire or develop unique, valuable, and rare resources that are difficult for rivals to repli- cate (Barney, 1986). Evolutionary theory posits per- formance differences among firms are a function of a competitive race to discover profit opportunities. According to this view, high performance is achieved by speed and innovation that keep firms ahead of rivals (Nelson & Winter, 1982). Our focus in this paper is the latter perspective, perhaps the least understood of the three; more specifically, we explore “Red Queen competition” in the context of actions among rivals.
Evolutionary and ecology theories focusing on Red Queen competition portray how entities dy- namically interact and coevolve with one another. Introduced by the biologist van Valen (1973), the
Red Queen effect is based on the conversation be- tween the Red Queen and Alice in Lewis Carroll’s Through the Looking Glass. In that story, Alice realizes that although she is running as fast as she can, she is not getting anywhere, relative to her surroundings. The Red Queen responds: “Here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!” (Carroll, 1960: 345).1 Van Valen used this analogy to de- scribe the continuous and escalating activity and development of participants trying to maintain rel- ative fitness in a dynamic system. Since then, the- orists have used the notion of the Red Queen to explain behavior in a variety of settings ranging from biology to military arms races (Baumol, 2004; Dawkins & Krebs, 1979).
Applied to a business context, the Red Queen can be seen as a contest in which each firm’s perfor- mance depends on the firm’s matching or exceed- ing the actions of rivals. In these contests, perfor- mance increases gained by one firm as a result of innovative actions tend to lead to a performance decrease in other firms. The only way rival firms in such competitive races can maintain their perfor- mance relative to others is by taking actions of their own. Each firm is forced by the others in an indus- try to participate in continuous and escalating ac- tions and development that are such that all the firms end up racing as fast as they can just to stand still relative to competitors. This self-escalating,
This research was partially supported by the Dingman Center for Entrepreneurship at the Robert H. Smith School of Business, University of Maryland, College Park. The authors would also like to thank our AMJ associate editor, R. Duane Ireland, and the three anony- mous reviewers for their insightful comments and sug- gestions during the revision process.
1 Through the Looking Glass was originally published in 1871.
� Academy of Management Journal 2008, Vol. 51, No. 1, 61–80.
61
Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s express written permission. Users may print, download or email articles for individual use only.
coevolving system of Red Queen competition has been empirically shown to affect founding rates (Barnett & Sorenson, 2002), failure rates (Barnett & Hansen, 1996), and competitiveness (Barnett & McKendrick, 2004). Indeed, Baumol (2004) sug- gested that the Red Queen effect is the most pow- erful mechanism driving economic development in capitalistic society.
Following Barnett and McKendrick (2004), we build a model that captures the Red Queen process of competitive evolution as both a positive and a negative force on focal firm performance in which the gains by one firm must come at the expense of another. As Barnett and McKendrick noted, “A de- fining characteristic of competition is that one or- ganization’s solution becomes its rivals’ problem. The resulting increased constraints, again in turn, are likely to trigger responses among rivals, again intensifying competitive constraints on the first or- ganization, and so on” (2004: 540). The first pur- pose of this study was to explicitly model the Red Queen “running as fast as you can” process by examining the relationships among focal firm ac- tions, rival actions, speed of rival actions, and focal firm performance. In doing this, we illustrate theo- retically and empirically that focal firm actions ver- sus rival actions and speed of rival actions have two opposing effects on focal firm performance. We refer to this formulation as our baseline model. A second goal of our research was to expand the baseline model by developing and testing theory that begins to identify the conditions that moderate the Red Queen effect.
Our examination of firm action and firm perfor- mance and the coevolutionary dependency of firm action and rival action on firm performance is fin- er-grained and more dynamic than prior Red Queen research. Specifically, we contend that firms are prompted to search, undertake new actions, and learn in an effort to improve performance. This use of Red Queen theory is consistent with Schumpet- er’s 1942 argument that a dynamic process of “cre- ative destruction” occurs when firms launch inno- vative actions to gain advantage in the marketplace, which is then eroded by their rivals’ competitive moves (Schumpeter, 1976). Thus, Schumpeter’s perspective figures prominently in the develop- ment of our Red Queen model. Additionally, our study applies Red Queen theory in the context of competitive dynamics, which focuses on the ac- tions and reactions of firms (Smith, Grimm, & Gan- non, 1992). In line with prior competitive dynam- ics research, we define firm actions as “externally directed, specific and observable competitive moves initiated by a firm to enhance its relative competitive position” (Smith, Ferrier, & Ndofor,
2001: 321). Rival actions are defined as the exter- nally directed competitive moves of all rivals in the industry in which the focal firm’s participation is being studied (Young, Smith, & Grimm, 1996).
Although our conceptualization of Red Queen competition is consistent with Schumpeterian and competitive dynamics perspectives, neither Schumpeter nor competitive dynamics research fully explains the motivating factors behind the competitive process. The Red Queen theory out- lined in this article provides a more complete pic- ture of this competitive process by explaining how firms are motivated to search, act, and learn in a desire to improve performance. As a consequence, our Red Queen theory begins to clarify vital aspects of the competitive process—the motivation for ac- tion and reaction—not fully explained in prior literature.
Like Barnett and Hansen (1996), we assume that a firm facing competition is likely to act. Further, we regard the competitive interactions of firms as constituting a mechanism for a simple search, ac- tion, and learning process firms undertake to im- prove performance (March & Simon, 1958). In our model, action allows firms to “learn by doing” (Ar- gote, 1999; Eisenhardt & Tabrizi, 1995; Pisano, 1994). Actions of focal firms that increase their performance may result in a decline in rival perfor- mance, thus prompting those rivals to engage in similar search, action, and learning processes. Our goal was to model this reciprocal system of focal firm actions and rival actions, including the speed at which those rival actions occur, and to show the system’s effect on focal firm performance. We focus on short-run performance, or the effects of focal firm action and rival action on focal firm perfor- mance in a given year. Though we recognize that Red Queen theory also posits long-term conse- quences of action exchanges, such as greater fitness for all competing firms, these long-term effects are beyond the scope of the present work.
As mentioned above, in addition to examining how firm actions, rival actions, and rival action speed impact short-term performance, a second goal of our research was to develop and test theory that begins to identify the conditions that amelio- rate or exacerbate the Red Queen effect. Such work is important as it may offer insights into how firms can successfully adapt or evolve under Red Queen competitive pressures. In this research, we focus on a set of factors that may facilitate or impede how managers learn from action. We argue that a better understanding of these factors can help explain the motivation for search and action and thereby, Red Queen competition. We suggest that industry con- centration, industry growth rate, and a firm’s mar-
62 FebruaryAcademy of Management Journal
ket position affect the search and action process and hence moderate the Red Queen effect.
RED QUEEN COMPETITION: FOCAL FIRM ACTIONS, RIVAL ACTIONS, RIVAL ACTION SPEED, AND FOCAL FIRM PERFORMANCE
In this section, we develop baseline theory that explains the Red Queen effect in terms of focal firm actions, rival actions, and rival action speed, and their combined impact on focal firm performance. We argue that search and firm action are motivated by a desire to learn new ways of improving perfor- mance. However, we assume that the relationship between firm action and performance is uncertain, dynamic, and subject to constant change by the very Red Queen competition that motivates search and action. In our model, firms search to discover opportunities to act. They experiment in taking new actions and learn from the results of action about the relationship between action and perfor- mance. We do not assume that all action is effective or costless, but we do assume that the average per- formance benefits of action outweigh the costs. Otherwise, firms would not have a motivation to act.
Recent theory on the Red Queen effect has spec- ulated about the motivation for firms’ and rivals’ actions. As Barnett and McKendrick (2004) sug- gested, firm aspirations expand, and goals may change quickly as a result of comparisons with others and Red Queen evolution. For instance, a large retailer may increase the frequency of market- ing campaigns on the basis of observing a positive relationship between prior campaigns and perfor- mance. Assume this action results in increased rev- enue and profits for the focal firm at the expense of a rival’s profits. Rivals may then search for and learn of some way to increase their own perfor- mance. In this example, rivals may take new price- cutting actions. These rival actions may adversely affect the performance of the focal firm, thus moti- vating additional search, action, and learning for this retailer. The learning from action in pursuit of profits that drives this action–rival action process captures the incremental and coevolutionary na- ture of Red Queen competition.
The motivation to search, act, and learn elicited by the Red Queen effect extends Schumpeter’s the- ory of creative destruction regarding the relation- ship between action and performance in a compet- itive context. Schumpeter (1934) highlighted the interdependent nature of a competitive market- place, arguing that it was the result of, and the reason for, continuous innovation and firm action. If firms stand still, competitors who introduce new
combinations that appeal to the market erode the inactive firms’ positions. To avoid this erosion, firms must continually strive to introduce new products, methods, and initiatives. Success, or profitable performance, he argued, is more the re- sult of “the new commodity, the new technology, the new source of supply, the new type of organi- zation” than it is the result of control of margins, output, or prices (Schumpeter, 1976: 84 – 85). In Schumpeter’s world, an efficient but lethargic firm would not survive for long.
Schumpeter also pointed out, however, that in- novation and action draw rival action, which he referred to as “creative destruction.” The acting firm “leads in the sense that he draws other pro- ducers in his branch after him” (Schumpeter, 1976: 89). Thus, innovations and the results of new ac- tion are visible to competitors, often spurring rival actions and an ongoing cycle of creation and de- struction. Indeed, if rivals are to survive in the marketplace, they cannot afford to ignore the com- petitive actions of other firms; they must also act creatively. The innovative competitive interaction of firms in pursuit of profits is so fundamental that Schumpeter (1976) argued it was the key source of market expansion and economic growth.
Researchers have empirically found that firms that are more active (i.e., are running faster) than their rivals improve their competitive positions (Ferrier, Smith, & Grimm, 1999) and increase their performance (Young, 1993; Young et al., 1996), while firms that are more sluggish than their rivals experience negative performance consequences (Miller & Chen, 1994). The basic argument of this research has been that more active firms achieve greater performance because they have greater as- piration levels, are more capable at implementing actions, and are perceived by rivals as more aggres- sive competitors than are less active firms (Smith et al., 2001).
Yet neither Schumpeter (1976) nor competitive dynamics researchers have recognized the role that learning from action outcomes may play in fueling competition and evolution. Barnett and McKen- drick (2004) described how learning drives Red Queen competition. When performance falls below aspirations, managers will search, act, and learn until performance reaches expectations. At the heart of this process is a manager seeking to under- stand a dynamic world of action cause and effect. Weick noted that managers cannot “ignore the ac- tion because they are responsible for it” (1995: 134 –135). Thus, Red Queen theory helps to explain how firms incrementally evolve by taking action and learning from the results of action in a desire to
2008 63Derfus, Maggitti, Grimm, and Smith
improve performance. Given these arguments along with prior research, we predict:
Hypothesis 1a. With the number of rival ac- tions held constant, as the number of focal firm actions increases, focal firm performance increases.
However, Red Queen competition can also have negative consequences for an active firm (Barnett & Hansen, 1996; Barnett & McKendrick, 2004). Bar- nett and McKendrick contended that one organiza- tion’s solution to search and action can become another firm’s problem. In this sense, Red Queen competition narrows the options for firms, escalat- ing rivalry and races, sometimes with limited short- term benefits for all. Returning to the above exam- ple, consider a case in which a rival’s response was not to cut prices but to simply imitate an initially acting firm’s behavior by increasing the frequency of its own marketing campaigns, thereby regaining the revenues shifted by the focal firm’s initial action.
Schumpeter argued that all advantages are tem- porary and uncertain because firms interact and the “perennial gale of creative destruction” erodes past accomplishments (1976: 89). Successful action evokes reaction from rivals. Indeed, it is the dy- namic process of firm actions and rival actions that defines the market process. Firms are spurred to engage in a cycle of action as they continually seek to learn more about action-performance relation- ships. When a firm leads with a new product or service, it puts pressure on competitors’ products and services, perhaps to the point of rendering them obsolete. Those competitors must act if they are to stay viable. Some rivals imitate, and others make innovative thrusts of their own. Regardless, the cycle repeats again and again as rivals struggle for profits and market share. It is this competitive interaction of rivals in pursuit of profits that results in market progress and evolution (Schumpeter, 1976).
Recent research supports the coevolutionary na- ture of action and reaction. Specifically, in a variety of different studies, conducted in a variety of dif- ferent industries, researchers have found a positive correlation between firm actions and rival actions. Indeed, response time, often a measure of compet- itive intensity and rival action speed, has ranged from as low as 8 days in the airline industry to 24 days in computer retailing and 124 days in high- technology industries (Grimm & Smith, 1997).
Thus, Red Queen theory recognizes the interde- pendent nature of firms. Specifically, the search, action, and learning process does not end with a focal firm’s actions (Barnett & McKendrick, 2004).
That is, the improved performance of the firm may come at the expense of rivals’ performance, which, in turn, may prompt rivals to search, act, and learn to improve their own performance. In this sense, learning and competition are codependent. To- gether, they explain the incremental and relative process by which firms evolve as they try to im- prove performance.
Thus, we predict:
Hypothesis 1b. As the number of focal firm actions increases, the number of rival firm ac- tions and the speed of rival actions increase.
A number of studies have focused on the perfor- mance consequences of industry rivalry. In a sam- ple of software firms, Young and colleagues (1996) found that as industry rivalry, measured as number of rival actions, increased, focal firm performance decreased. Chen and Miller (1994) found that higher levels of rival responses decreased perfor- mance in the airline industry, and Schomburg, Grimm, and Smith (1994) found a negative relation- ship between rivalry and profitability in the beer, telecommunications, and personal computer in- dustries. Finally, Smith and colleagues (1992) found that increased competitive actions were re- lated to lower profitability in the airline industry.
Evolutionary scholars have also examined the performance consequences of Red Queen competi- tion. Barnett and Hansen (1996) argued that a focal firm’s superior performance leads a rival to search for new opportunities to improve its own perfor- mance. Assuming effective actions are found, the rival’s position improves at the expense of the focal firm. In a study of Illinois banks from 1900 to 1992, Barnett and Hansen (1996) found that a focal firm’s own competitive experience increased its chances of success and survival, whereas its rivals’ aggre- gate relative experience decreased the focal firm’s success. They argued that firms are constrained by their history, falling into competency traps where they respond to new developments with old ac- tions (Ingram, 2002; Levitt & March, 1988).2
In view of this theory and research, we hypothesize:
Hypothesis 1c. With the number of focal firm actions held constant, as the number of rival firm actions and the speed of rival actions in- crease, focal firm performance decreases.
2 As noted, although the long-term effects of competi- tion on performance might be positive (Porter, 1980), we believe the shorter-term effects will be negative.
64 FebruaryAcademy of Management Journal
Hypotheses 1a–1c represent our baseline Red Queen prediction on the positive and negative con- sequences of firm and rival actions for focal firm performance. We next consider how this Red Queen relationship may be affected by industry conditions and market position. Figure 1 is an il- lustration of the hypothesized relationships.
MODERATORS OF THE RED QUEEN EFFECT
We draw from evolutionary theory (Nelson & Winter, 1982), the industry position school of com- petitive advantage (Porter, 1980), and action re- search (Smith et al., 1992) to propose how industry concentration, industry demand conditions, and market position moderate the baseline model re- garding the relationship between focal firm action, rival action, rival action speed, and firm perfor- mance. We theorize that these factors moderate the relationship between focal firm action, rival action, rival action speed, and focal firm performance by affecting the ability of a focal firm and rival firms to learn from search and action.
Industry Concentration
Industry concentration, commonly measured by the percentage of the market share held by the largest firms in an industry (the Herfindahl index), is an important industry characteristic. As Wald- man and Jensen noted, “Seller concentration within a particular market is regarded as a signifi- cant aspect of market structure because of its hy- pothesized relationship to market power and, ulti- mately, to behavior and performance” (2001: 94). Theory from economics suggests that a small num- ber of dominant firms in an industry will recognize their mutual dependence and tacitly coordinate search and action in an effort to limit competition
and rivalry (Scherer & Ross, 1990). The implicit motive for this coordination is that escalation of competition increases costs and hurts performance. Conversely, as the number of firms increases, search, action, and potential learning increase as it becomes more and more difficult for this coordina- tion to occur (Williamson, 1965). Under these con- ditions, it is more challenging for a focal firm to find unique opportunities to act and, as a conse- quence, effective search, action, and learning be- come more costly. One implication of this argu- ment is that the effects of focal firm search, action, and learning on performance are greater in more concentrated industries. In such environments, search and action are less frequent, and so it is easier for firms to learn and to comprehend the consequences of their actions as they receive more attention from market participants. Moreover, when a market consists of just a few large firms, customers have limited choices (fewer competi- tors), and they are therefore more likely to be at- tracted to the new actions of dominant firms.
For the same reasons that a firm’s performance gains from action are likely to be greater in concen- trated industries, because of the high market shares of firms, limited choices of customers, and ease of learning from the effectiveness of search and ac- tion, the effects of rivals’ actions on the focal firm’s performance are also greater. Firm actions in highly concentrated industries are much more likely to garner the attention of competitors because mutual awareness is very high (Bain, 1951). Thus, in highly concentrated industries, rival firms are more apt to learn of the actions of a focal firm and to respond to those actions to stave off the negative consequences of nonresponse. In addition, and perhaps more im- portantly, in highly concentrated industries, rival firms are more inclined to respond, and to respond quickly, to teach their competitors that breaking the
FIGURE 1 Hypothesized Relationships
2008 65Derfus, Maggitti, Grimm, and Smith
unwritten covenant of tacit collusion will be pun- ished severely. On the other hand, in less concen- trated industries, where there are more competitors to keep track of, actions are less likely to provoke responses because rivals will not be aware of the initial behavior. As Scherer and Ross stated, “As the number of sellers increases and the share of industry output supplied by a representative firm decreases, individual producers are increas- ingly apt to ignore the effect of their price and output decisions on rival reactions” (1990: 277). Thus, in concentrated industries, focal firm actions will have a greater impact on performance, evoke a larger number and higher speed of rival ac- tions, and those rival actions and their speed will have a greater impact on the focal firm’s per- formance.
From an evolutionary perspective, Barnett and Hansen (1996) contended that the Red Queen effect would be less constraining when a firm faced a relatively small number of competitively different cohorts, as in the condition of high concentration. These authors described how increases in the num- ber of competitive relationships (i.e., lower concen- tration) constrained effective learning and adapta- tion. They noted that each constraint lowers the likelihood that a focal firm can carry out effective search, action, and learning, arguing that costs of search and action come to outweigh the benefits. Using a sample of Illinois banks, they found that failure rates increased with the number of compet- itive relationships. These findings suggest that de- creased concentration increases the number of competitive relationships to manage and that under low concentration, a firm’s action has less impact on both its own performance and rival action, while rival action has less impact on focal firm performance. Carroll and Hannan’s (1989) density dependence theory also supports this prediction. Specifically, this research showed that firms founded under conditions of many competitors are less likely to survive in the long term because of scarcity of resources (Carroll & Hannan, 2000).
Given the above arguments, we propose:
Hypothesis 2a. Industry concentration posi- tively moderates the relationship between a focal firm’s actions and its performance.
Hypothesis 2b. Industry concentration posi- tively moderates the relationship between a focal firm’s actions and rival actions and the speed of rival actions.
Hypothesis 2c. Industry concentration nega- tively moderates the relationship between rival
actions, the speed of rival actions, and a focal firm’s performance.
Industry Demand
The growth rate of industry demand should also have an impact on Red Queen competition. Studies by Caves (1980) and Bothwell, Cooley, and Hall (1984) showed that firms in high-growth industries are less concerned about competing with rivals be- cause they are able to enhance revenues simply by maintaining their shares of the steadily increasing demand. Therefore, high industry growth leads to a “live-and-let-live” attitude among firms (Bradburd & Caves, 1982; Liebowitz, 1982). A growing market facilitates existing routines, and each firm can in- crease its share of the pie by searching for and carrying out actions that it knows will work with- out affecting rivals. Conversely, a decline in indus- try demand will prompt firms to search for new ways of generating demand, by instituting a new price cut or new marketing campaign, that initiate or escalate warfare (Caves, 1980).
Research exploring the evolution of industries has examined the differing effects that the early, high-demand, stage and the mature, decreasing de- mand, stage of the industry life cycle have on com- petition among firms (Agarwal & Gort, 1996; Agar- wal, Sarkar, & Echambadi, 2002; Carroll & Hannan, 1989). Specifically this work speculates that during periods of high demand growth firms take actions that help create that demand and thereby, benefit all the firms in their industry (Agarwal & Bayus, 2002). Carroll and Hannon (1989) described such early-stage actions as “legitimizing actions,” as op- posed to later-stage “competitive” actions.
In regard to Red Queen competition, high-growth environments provide fertile ground for searching and learning about new opportunities to act, and limit the negative effect such actions have on com- petitors. Increasing industry growth mitigates the Red Queen argument that a firm’s performance gains come at the expense of other firms (Barnett & Hansen, 1996). For example, irrespective of rival actions, a focal firm’s successful new-product launch is going to be even more successful when the number of consumers seeking such products is growing. Accordingly, all the firms in a high- growth industry will be focused on developing suc- cessful “initial” actions and less focused on re- sponding more often or faster to other firms’ actions. Further, when rivals do act, their actions are more likely to increase industry growth overall rather than have a deleterious impact on another firm’s performance. Therefore, we propose:
66 FebruaryAcademy of Management Journal
Hypothesis 3a. Industry demand positively moderates the relationship between a focal firm’s actions and its performance.
Hypothesis 3b. Industry demand negatively moderates the relationship between a focal firm’s actions and rival actions and the speed of rival actions.
Hypothesis 3c. Industry demand positively moderates the relationship between rival ac- tions, speed of rival actions, and focal firm performance.
Market Position
We next theorize about the effect of the relative market position of a focal firm. Research on the Red Queen effect has suggested that market leaders are less affected by Red Queen competition than are other firms. For example, Barnett and McKendrick (2004) found that market share leaders—in their case, large firms—were the most likely to act to develop new products in the disk drive industry. They also found that, when exposed to compe- tition, large firms were less likely to fail. How- ever, they also noted that market share leaders can become isolated from competition and, when this happens, their survival may be threatened as they lose their ability to learn from search and action.
Action research also suggests that market leaders more effectively search and act than their rivals and react more quickly than their rivals (Smith et al., 2001). Presumably, market leaders have the re- sources to engage in more effective search and ac- tion, which facilitates greater learning. In essence, this is how they obtain and defend their market positions. Ferrier, Smith, and Grimm (1999) found that persistent market leaders act more frequently, faster, and with more complexity. The actions of market leaders should be more positively related to performance than the actions of other firms in an industry because they have more experience and enjoy more efficient search and action routines. In a sense, they have more effectively institutional- ized the search, action, and learning process. Spe- cifically, actions of market leaders are more visible to customers and therefore likely to garner more customer attention (Smith et al., 1992). Under these conditions, managers can more effectively learn from their actions. Young et al. (1996) found that market leaders benefit from significant scale effects of action that smaller firms cannot obtain. For mar- ket-leading firms, the cost of search, action, and learning can be spread over a larger customer base.
Although different predictions are possible re-
garding how rivals will behave with regard to mar- ket leaders,3 we believe the more powerful argu- ment is that rivals are unlikely to act or act quickly against leading firms because of fear of retribution (Scherer & Ross, 1990). Specifically, research in industrial-organization (IO) economics has investi- gated the behavior of dominant firms with regard to their rivals on several fronts; it has been found that pricing (Gaskins, 1971; Kamien & Schwartz, 1971), R&D and patenting (Gilbert & Newberry, 1982), product proliferation (Schmalensee, 1976), adver- tising (Comanor & Wilson, 1967; Cubbin & Domberger, 1988), and capacity increase (Masson & Shannan, 1986; Spence, 1977) actions by dominant firms deter rival entry. Ferrier and colleagues (1999) found that market leaders that engaged in more frequent, speedier actions and utilized more complex action repertoires deterred the actions of challengers. Thus, we expect that rivals will be deterred from attacking an industry leader.
Finally, we predict that the frequency and speed of rival actions, if they occur, have less negative impact on leaders’ performance than on nonlead- ers’ performance. Customers of leaders are more likely to remain loyal in the face of rival actions. Leader firms have stronger brand reputation and customers are less likely to defect because the switching costs of moving from a market leader are potentially higher (Scherer & Ross, 1990). In sum- mary, prior work shows that market leaders are more effective than nonleaders, and their actions have a stronger impact on performance than the actions of nonleaders. Because leaders’ actions are more likely to deter, rather than provoke rivals, they evoke fewer and slower rival actions than do nonleaders’ actions. And, because of customer loy- alty and higher switching costs, rivals’ actions and their speed do not detract from leader performance as much as they do nonleaders’ performance. Therefore, we predict:
Hypothesis 4a. Market position positively mod- erates the relationship between a focal firm’s actions and its performance.
Hypothesis 4b. Market position negatively moderates the relationship between a focal firm’s actions and rival actions and the speed of rival actions.
Hypothesis 4c. Market position positively mod- erates the relationship between rival actions,
3 One possible alternative explanation is that rivals are more likely to follow or imitate market leader actions because they are perceived as more legitimate.
2008 67Derfus, Maggitti, Grimm, and Smith
the speed of rival actions, and a focal firm’s performance.
RESEARCH METHODS
Sample
To test the hypotheses, we developed a sample of all the major competitors in 11 different industries across a broad spectrum of the U.S. economy. One requirement for sample inclusion was that firms were competing in the same markets so that their specific actions and firm performance could be di- rectly connected to the competition and perfor- mance in these markets. As a result, we focused solely on the actions and performance of U.S. firms and included only those industries in which 70 percent or more of industry sales was generated by firms that were public, had a distinct single-busi- ness entity competing in the specific U.S. market, and reported performance relative to that U.S. mar- ket. Firms needed to meet these criteria so that we could match the actions of their single-business entities with their performance in only in that market/industry.
Eleven industries met our criteria, including ap- pliance manufacturing, athletic footwear manufac- turing, automobile manufacturing, brewing, gen- eral retailing, book retailing, lumber and hardware retailing, long-distance telephone services, steel manufacturing, and grocery retailing. With repre- sentation from manufacturing, services, and retail- ing, good industry variation was achieved. On average, included firms accounted for 87 percent of U.S. industry sales in their respective market/ industry.
Data Collection: Competitive Actions
Competitive actions are defined as specific and observable moves, such as new marketing cam- paigns or new-product introductions, initiated by a firm to defend or improve its relative competitive position (Chen, 1988; Smith et al., 1992; Young et al., 1996). Actions that are observable to customers, competitors, and other industry watchers are most likely to be reported in the business press (Miller & Chen, 1994) and thereby are available for identifi- cation, data collection, and analysis. We identified and coded observable competitive actions by con- ducting a structured content analysis (Jauch, Os- born, & Martin, 1980) of newspaper and trade mag- azine articles found on the Lexus-Nexus article index. This index allows electronic searching of full-text articles from thousands of newspapers and journals. For each of the industries chosen, at least
one industry trade magazine was searched. Addi- tionally, the New York Times and the Wall Street Journal were searched for all industries. We iden- tified 76,963 article citations via keyword search- ing of the Lexus-Nexus database. Coders then con- tent-analyzed the full texts of articles that potentially contained reports of competitive ac- tions. Only the earliest report of an action was entered into the database. This procedure resulted in a database containing 4,474 actions. To verify the accuracy of the coding, two coders reviewed 10 percent of the article citations for each industry. Action identification and action-type coding agree- ment were obtained for 99.25 percent of the over 7,697 citations they read.
Actions were collected for 58 firms; missing data reduced the sample to 56 firms over a six-year period, 1993 through 1998. The mean number of actions per firm was 12.58. The maximum number of actions per firm per year was 51; the minimum, 0. The most common actions related to pricing, and the least common were geographic actions.
Firm financial data, firm size, and industry con- text variables were collected from Standard & Poor’s Compustat database, which offers financial data on all companies that were publicly traded on North American stock markets in those years. These data are collected from annual reports, Secu- rities and Exchange Commission (SEC) filings, and other publicly available documents. Where neces- sary, we adjusted financial data to remove contri- butions from non-U.S. operations, thereby match- ing the financial figures to the actions accounted for in the study. These adjustments were possible because Compustat offers detailed geographic seg- ment data delineating company operations in vari- ous countries.
Measures
Focal firm total actions and rival total actions. Five types of focal firm and rival actions were mea- sured: pricing, capacity, geographic, marketing, and product introductions. We calculated firm total actions or activity by simply summing the number of all five actions for a focal firm in a given year. We then operationalized rival total actions or compet- itive activity by subtracting a focal firm’s total num- ber of actions in a given year from the total number of actions taken by all competitors in a focal industry.
Rival action speed. As in other competitive dy- namics research (e.g., Ferrier et al., 1999; Young et al., 1996), rival firm action speed quantified the average length of time it took rivals to act after a focal firm acted. To calculate this measure, we de-
68 FebruaryAcademy of Management Journal
termined the number of days between each firm action and the first rival action and then averaged those scores for each focal firm for each year. Fi- nally, we took the reciprocal of this value to aid in interpretation of results. The resulting measure equates high rival action speed values to fast rival action speed and low rival action speed values to slow rival action speed.
Focal firm performance. Focal firm perfor- mance was operationalized with accounting mea- sures of return on sales (ROS) and return on assets (ROA) in the same year as the action measure.
Industry conditions. Industry concentration and industry demand were used to capture the industry context in which firm and rival actions took place. These measures served as independent or control variables in all regressions and were also interacted with firm actions, rival actions, and rival action speed in tests of Hypotheses 2a–2c and 3a–3c. In- dustry concentration was calculated as the Herfin- dahl measure of the market shares of the firms in each industry for each year. Industry demand was measured as industry growth, defined as the per- cent change in sales from the previous year to a focal year.
Relative market position. Relative market posi- tion was measured as rank order based on market share for each firm in each industry for each year. This variable was used as an independent or a control variable in all regressions and was also interacted with firm actions, rival actions, and rival action speed in testing Hypotheses 4a– 4c.
Control variables. To control for unobserved dif- ferences in industry factors that might influence market dynamics, we included industry dummies in all regressions.
In this study, we had two basic regression mod- els. In the first, we regressed our independent ac- tion variables on firm performance. However, since we had two measures of firm performance, return on assets and return on sales, a model is presented for each. In these models, we lagged a focal firm’s prior year return on assets or return on sales per- formance, respectively, to control for the influence the variable might have on performance in the fol- lowing year and also to help control for correlated error terms of our longitudinal data (Young et al., 1996). In these regressions, we also controlled for firm characteristics that have been shown to influ- ence firm actions and performance, including size and slack resources (Smith et al., 1992). Specifi- cally, sales measured size, and the quick ratio mea- sured slack resources (e.g., Ferrier et al., 1999). The logic was that firms with more assets and resources are able to undertake more actions (Smith et al.,
2001). We then repeated these regressions replac- ing rival action speed with rival total actions.
When we tested the effect of focal firm actions on rival actions and rival action speed, we controlled for prior year aggregate performance utilizing lagged rival firm prior year return on sales. This lagged composite was calculated as the aggregated net income of all rival firms in an industry divided by their aggregated sales. As in our regressions on firm performance, we also controlled for rival size and rival slack resources. Rival’s size was calcu- lated as an aggregate average measure of the relative size of each firm’s pool of rivals. It is an annual sum of the sales of the unique group of rivals pertaining to each focal firm. Rival quick ratio is a composite average of the quick ratios of those rival firms.
We followed the practice of prior competitive dynamics researchers by investigating the impact of actions on the same year’s performance (e.g., Ferrier et al., 1999; Young et al., 1996). During the years of our study, for these 11 industries, the av- erage number of days between a focal firm action and a rival action is 12.3 (s.d. � 19.3) and the maximum number of days between actions is 149. This relatively short time frame provided further support for our same-year analysis of actions and performance.
In calculating the measures in this study, it was important to precisely define the focal and rival firms. For example, in the U.S. brewing industry there were three sampled firms: Anheuser-Busch, Miller Brewing, and Adolph Coors. When An- heuser-Busch was the focal firm, Miller Brewing and Adolph Coors were the rival firms. Similarly, when Miller Brewing was focal, Anheuser-Busch and Adolph Coors were the rivals. We calculated firm total actions and performance measures for each firm individually and the rival total actions measure for each firm’s unique set of rivals. Since the unit of observation was the firm-year, changes in the rival set from year to year must be accounted for. Therefore, the rival set was officially defined as all the other firms competing in the focal firm’s industry for the year under consideration.
RESULTS
Table 1 reports the means and correlations among all variables in this study. We tested hy- potheses with random-effects regression models to ensure that error due to serial correlation in our panel data set was specified and analyzed (Erez, Bloom, & Wells, 1996). In addition, we used nega- tive binomial regression analysis for regression models in which the dependent variable was rival total actions, a count-type variable. This type of
2008 69Derfus, Maggitti, Grimm, and Smith
T A
B L
E 1
D es
cr ip
ti v e
S ta
ti st
ic sa
V a ri
a b le
b M
ea n
s. d
. 1
2 3
4 5
6 7
8 9
1 0
1 1
1 2
1 3
1 4
1 5
1 6
1 7
1 8
1 9
2 0
2 1
2 2
2 3
2 4
2 5
2 6
2 7
2 8
2 9
3 0
3 1
3 2
3 3
1 .
R O
A 3 .7
4 5 .5
1 2 .
R O
S 4 .5
3 5 .3
3 .4
5 3 .
F ir
m to
ta l
ac ti
o n
s 1 3 .1
8 1 0 .6
5 .2
2 .1
4 4 .
R iv
al to
ta l
ac ti
o n
s 6 3 .5
7 4 4 .6
9 �
.1 3
� .3
9 .3
1 5 .
R iv
al ac
ti o n
sp ee
d 0 .1
8 0 .1
4 �
.1 1
� .3
2 .4
6 .9
5 6 .
F ir
m m
ar k et
sh ar
e ra
n k
0 .3
8 0 .2
7 .0
4 �
.0 7
.1 5
.3 4
.3 2
7 .
In d
u st
ry H
er fi
n d
ah l
0 .2
2 0 .1
8 .0
8 .3
8 .0
9 �
.5 4
� .4
9 �
.5 5
8 .
In d
u st
ry gr
o w
th 0 .0
9 0 .1
1 .0
3 �
.1 5
� .1
5 �
.1 9
� .2
2 �
.0 8
.1 4
9 .
M ar
k et
sh ar
e ra
n k
� fi
rm ac
ti o n
s 5 .5
5 6 .1
9 .1
9 .0
1 .7
2 .3
5 .4
2 .6
7 �
.2 8
� .0
8
1 0 .
M ar
k et
sh ar
e ra
n k
� ri
v al
ac ti
o n
s 2 8 .6
6 3 2 .2
0 �
.0 7
� .2
5 .2
0 .6
9 .6
3 .8
5 �
.6 2
� .0
9 .6
4
1 1 .
H er
fi n
d ah
l �
fi rm
ac ti
o n
s 3 .4
7 4 .4
2 .1
6 .4
0 .7
0 �
.1 5
� .0
1 �
.2 1
.6 2
� .1
1 .2
1 �
.2 6
1 2 .
H er
fi n
d ah
l �
ri v al
ac ti
o n
s 1 1 .9
4 1 0 .0
1 �
.0 7
.0 2
.3 8
.3 5
.4 1
� .3
4 .4
6 �
.1 3
� .0
4 �
.1 3
.4 6
1 3 .
In d
u st
ry gr
o w
th �
fi rm
ac ti
o n
s 1 .0
9 1 .5
5 .2
0 �
.0 6
.3 8
.0 8
.1 3
.1 2
.0 0
.5 4
.3 6
.1 1
.1 3
.0 0
1 4 .
In d
u st
ry gr
o w
th �
ri v al
ac ti
o n
s 5 .2
2 5 .7
9 �
.0 8
� .3
3 .1
1 .6
0 .5
3 .2
6 �
.3 2
.4 2
.2 3
.5 0
� .2
0 .1
3 .5
3
1 5 .
M ar
k et
sh ar
e ra
n k
� ri
v al
sp ee
d 0 .0
8 0 .0
9 �
.0 6
� .2
2 .2
8 .6
7 .6
6 .8
7 �
.6 1
� .1
2 .7
1 .9
8 �
.2 1
� .1
0 .1
3 .4
6
1 6 .
H er
fi n
d ah
l �
ri v al
sp ee
d 0 .0
4 0 .0
3 �
.0 5
.1 0
.5 6
.3 4
.4 8
� .2
8 .4
1 �
.1 6
.0 7
� .1
1 .6
3 .9
4 .0
6 .1
1 �
.0 5
1 7 .
In d
u st
ry gr
o w
th �
ri v al
sp ee
d 0 .0
1 0 .0
2 �
.0 6
� .3
0 .2
2 .5
9 .5
8 .2
4 �
.2 9
.4 0
.2 8
.4 6
� .1
2 .1
8 .6
4 .9
7 .4
5 .2
0
1 8 .
L ag
ge d
R O
A 0 .0
4 0 .0
6 .5
2 .2
2 .0
6 �
.0 5
� .0
7 .0
0 .0
3 �
.0 2
.0 7
� .0
6 �
.0 2
� .0
1 .0
8 �
.0 9
� .0
7 �
.0 5
� .0
9 1 9 .
L ag
ge d
R O
S 0 .0
4 0 .0
5 .3
1 .6
4 .0
1 �
.3 3
� .2
9 �
.0 9
.2 6
� .1
7 �
.0 7
� .2
4 .1
8 .0
1 �
.0 9
� .3
5 �
.2 2
.0 4
� .3
5 .5
4 2 0 .
L ag
ge d
ri v al
R O
S 0 .0
5 0 .0
4 .0
5 �
.1 6
� .0
5 �
.0 4
� .1
5 �
.0 8
.0 5
.0 9
� .0
4 �
.0 8
� .0
6 �
.0 6
� .0
1 �
.0 9
.1 3
� .1
5 �
.0 9
.1 3
.0 3
2 1 .
F ir
m sa
le s
1 5 ,6
6 3
2 5 ,1
9 0
.1 5
.1 0
.4 6
.0 5
.1 4
.0 2
.1 1
� .1
1 .2
8 .0
0 .3
9 .1
6 .1
4 �
.0 4
.0 5
.2 9
� .0
4 .0
6 .0
3 �
.1 0
2 2 .
F ir
m q u
ic k
ra ti
o 0 .6
0 0 .4
3 .1
1 .3
0 �
.0 7
� .1
4 �
.0 8
� .0
6 .1
0 �
.1 2
� .1
5 �
.1 1
.1 4
.0 9
� .0
6 �
.1 0
� .0
9 .1
1 �
.1 0
.1 1
.2 1
� .2
5 �
.1 6
2 3 .
R iv
al sa
le s
8 2 ,2
0 2
9 1 ,6
3 0
� .0
9 �
.2 2
.3 1
.6 1
.6 6
� .0
5 �
.2 6
� .2
0 .1
4 .2
0 �
.0 4
.3 9
.0 6
.2 4
.2 2
.4 4
.2 4
� .0
8 �
.1 9
� .0
3 .4
3 �
.2 2
2 4 .
In d
u st
ry q u
ic k
ra ti
o 0 .6
1 0 .3
1 .0
3 .3
2 �
.0 6
� .1
5 �
.0 8
.0 2
.0 8
� .1
9 �
.1 0
� .0
6 .1
5 .1
0 �
.1 2
� .1
2 �
.0 3
.1 2
� .1
2 .0
0 .2
1 �
.3 4
� .1
4 .7
4 .2
9 2 5 .
In d
u st
ry 1
0 .0
6 0 .2
4 .0
9 .2
7 .0
0 �
.2 3
� .2
0 �
.1 6
.2 9
� .1
8 �
.1 0
� .1
8 .1
5 �
.0 3
� .1
2 �
.1 9
� .1
8 .0
0 �
.1 9
.0 6
.2 1
� .2
0 �
.0 7
� .0
1 �
.1 8
� .0
1 2 6 .
In d
u st
ry 2
0 .0
7 0 .2
6 .2
8 .0
7 .1
4 �
.0 6
� .0
8 �
.1 2
.2 4
.0 3
.0 0
� .1
3 .2
4 .3
1 .1
7 .0
1 �
.1 4
.2 2
.0 1
.2 6
.1 1
.1 2
� .1
5 .3
6 �
.2 3
.4 5
� .0
6 2 7 .
In d
u st
ry 3
0 .1
9 0 .3
9 �
.2 0
� .1
5 �
.2 1
.2 1
.2 1
.5 8
� .5
3 �
.1 4
.1 4
.4 7
� .3
2 �
.3 5
� .1
0 .1
7 .4
9 �
.3 1
.1 7
� .1
4 �
.1 2
� .2
9 �
.2 6
.2 8
� .2
2 .4
1 �
.1 2
� .1
4 2 8 .
In d
u st
ry 4
0 .0
6 0 .2
4 �
.0 4
.2 3
� .2
4 �
.3 2
� .2
9 �
.1 9
� .0
8 �
.1 3
� .2
0 �
.2 1
� .1
5 �
.2 5
� .1
5 �
.2 1
� .2
2 �
.2 2
� .2
1 �
.0 6
.1 7
� .2
0 �
.1 1
.2 2
� .1
9 .3
1 �
.0 6
� .0
7 .1
3 2 9 .
In d
u st
ry 5
0 .1
2 0 .3
2 �
.1 3
� .1
2 .3
0 .4
2 .5
8 �
.1 5
� .0
1 �
.0 5
.0 3
.0 1
.1 5
.5 9
.1 6
.2 6
.0 7
.6 8
.2 6
� .1
1 .1
2 �
.3 0
.2 9
.0 0
.7 2
.0 0
� .0
7 �
.0 8
� .1
5 �
.0 7
3 0 .
In d
u st
ry 6
0 .0
6 0 .2
3 �
.0 1
.4 0
.1 8
� .1
7 �
.1 3
� .1
9 .6
2 �
.1 3
� .0
5 �
.1 9
.5 3
.3 5
� .1
6 �
.2 4
� .1
8 .3
5 �
.2 4
� .0
6 .2
5 .0
3 .1
8 .1
2 �
.0 9
.1 6
� .0
6 �
.0 7
� .1
4 �
.0 7
� .0
8 3 1 .
In d
u st
ry 7
0 .0
8 0 .2
8 .0
3 �
.0 2
� .1
4 �
.2 4
� .2
6 �
.0 1
.0 7
.1 6
� .0
9 �
.1 8
� .0
6 �
.1 .0
4 �
.0 3
� .1
9 �
.1 2
� .0
3 .0
3 �
.0 1
.2 9
� .0
8 �
.2 2
� .1
9 �
.3 5
� .0
7 �
.0 8
� .1
6 �
.0 8
� .0
9 3 2 .
In d
u st
ry 8
0 .0
8 0 .2
7 .0
2 �
.0 3
.0 1
� .1
7 �
.1 8
� .1
5 .0
2 �
.0 5
� .0
7 �
.1 8
.0 1
� .0
4 .0
0 �
.1 1
� .1
9 �
.0 6
� .1
1 �
.0 1
� .0
3 �
.0 2
.5 4
� .0
6 .5
1 �
.0 8
� .0
7 �
.0 8
� .1
5 �
.0 7
� .0
8 �
.0 8
� .0
9 3 3 .
In d
u st
ry 9
0 .1
5 0 .3
6 .1
3 �
.2 1
.2 1
.5 6
.4 4
.2 2
� .4
3 �
.1 4
.3 3
.4 5
� .2
1 �
.1 4
.0 0
.1 7
.4 0
� .1
6 .1
7 .1
2 �
.1 7
.3 5
.0 2
� .4
1 .3
5 �
.5 2
� .1
1 �
.1 2
� .2
4 �
.1 1
� .1
3 �
.1 3
� .1
4 �
.1 3
3 4 .
In d
u st
ry 1 0
0 .0
7 0 .2
6 �
.1 1
� .1
3 �
.0 5
� .1
5 �
.1 8
� .1
5 .2
2 .1
4 �
.0 8
� .1
7 .0
6 .1
1 .0
8 .0
3 �
.1 8
.0 2
.0 3
� .0
7 �
.0 4
.0 8
� .1
6 �
.1 4
� .2
2 �
.2 0
� .0
6 �
.0 7
� .1
3 �
.0 6
� .0
7 �
.0 7
� .0
8 �
.0 7
� .1
2
a n
� 2 8 1 .
C o rr
el at
io n
s ab
o v e
.1 1
ar e
si gn
if ic
an t
at p
� .0
5 .
b R
O A
an d
R O
S ar
e p
er ce
n ta
ge s.
S al
es ar
e in
m il
li o n
s o f
d o ll
ar s.
regression was used in these models for two rea- sons. First, these count data are not normally dis- tributed, violating a key assumption of generalized least squares (GLS) regression analysis (Greene, 1993). Secondly, as is typically the case, our count data are overdispersed, meaning the variance of the event counts exceeds their means (Cameron & Tra- vendi, 1986). A likelihood-ratio test of overdisper- sion also indicated that negative binomial regres- sion was an appropriate choice. Negative binomial regression overcomes distribution problems and es- timates an additional parameter that corrects for overdispersion (Frome, Kutner, & Beauchamp, 1973). Tables 2 and 3 report the regression results.
Table 2 reports the results for regressions relating firm actions, firm performance, rival actions, and speed of rival actions. Table 3 reports the regres- sion results that examine the impact of industry environment and market leadership on the rela- tionships between firm actions, firm performance, rival actions, and speed of rival actions.
Hypothesis 1a states that as firm total action in- creases, firm performance increases. This hypothe- sis is fully supported. As seen in models 1 and 3 in Table 2, firm actions have a positive, significant coefficient for both return on assets (� � 0.08, p � .01) and return on sales (� � 0.08, p � .01). These results are repeated in models 2 and 4 of Table 2, in which rival action speed replaces rival actions in the models. That is, firm actions again have a pos-
itive, significant effect on both return on assets (� � 0.12, p � .01) and return on sales (� � 0.11, p � .01).
Hypothesis 1b states that as firm actions increase, rival actions and rival action speed also increase. This hypothesis is supported. Specifically, models 5 and 6 in Table 2 report a significant, positive coefficient for the relationship between firm ac- tions and rival actions (� � 0.01, p � .05) and firm actions and rival speed (� � 0.00, p � .01), respectively.
Hypothesis 1c states that, as rival actions and rival action speed increase, focal firm performance decreases. This hypothesis is also fully supported. As reported in Table 2, rival actions is significantly and negatively related to focal firm ROA in model 1 (� � �0.05, p � .01) and ROS in model 3 (� � �0.03, p � .01). Similarly, as shown in models 2 and 4, rival action speed is significantly and nega- tively related to focal firm’s ROA in model 1 (� � �17.02, p � .01) and ROS in model 2 (� � �12.93, p � .01).
Hypotheses 2a–2c, 3a–3c, and 4a– 4c explore the boundary conditions of the first hypothesis set, Hy- potheses 1a–1c. Specifically, in the second and third sets of hypotheses we examine how various industry situations condition the relationship be- tween firm actions, rival actions, rival action speed, and firm performance. In the fourth set of hypoth- eses, we look at how these relationships differ on
TABLE 2 Results of Random-Effects Regression Analyses of the Main Relationshipsa
Variables Model 1:
ROA Model 2:
ROA Model 3:
ROS Model 4:
ROS
Model 5: Rival Total
Actions
Model 6: Rival Speed of Actions
Lagged ROA 35.74** (4.81) 35.73** (4.82) Lagged ROS 38.78** (4.34) 39.93** (4.31) Lagged rival ROS �2.14* (1.03) �0.30** (0.10) Firm sales 0.29† (0.21) 0.30† (0.24) �0.14 (0.17) �0.14 (0.17) Quick ratio 1.84* (0.75) 1.89* (0.75) 2.03** (0.60) 2.05** (0.60) Rival sales �0.00* (0.00) �0.00* (0.00) Industry quick ratio 0.22 (0.16) 0.03 (0.02) Market share rank 0.93 (1.27) 0.84 (1.27) 2.95** (1.04) 2.81** (1.03) �0.13* (0.18) �0.04** (0.01) Herfindahl index �5.52 (5.57) �5.78 (5.59) �1.19 (4.61) �1.17 (4.63) �0.34 (0.31) �0.12* (0.07) Industry growth 6.12* (3.11) 5.06* (3.06) 2.83 (2.51) 2.29 (2.49) 1.19** (0.29) 0.06* (0.04) Firm total actions 0.08** (0.03) 0.12** (0.04) 0.08** (0.03) 0.11** (0.03) 0.01* (0.00) 0.00** (0.00) Rival total actions �0.05** (0.01) �0.03** (0.01) Rival speed of actions �17.02** (5.21) �12.93** (4.23)
Constant 1.66 (2.54) 1.99 (2.55) 0.58 (2.13) 0.66 (2.14) 1.97** (0.58) 0.05 (0.04) Wald chi-square 187.20** 185.18** 394.58** 406.50** 248.93** 2,117.46**
a Standard errors are in parentheses. Industry dummy variables were included in all regression models. Results are available upon request. n � 281.
† p � .10 * p � .05
** p � .01
2008 71Derfus, Maggitti, Grimm, and Smith
the basis of the market share of a focal firm relative to the other firms in the industry. Table 3 reports the results of our tests of Hypotheses 2a–2c, 3a–3c, and 4a– 4c.
Hypothesis 2a predicts industry concentration positively moderates the relationship between fo- cal firm actions and focal firm performance. This hypothesis is partially supported. Specifically, and as predicted, the effect of the interaction of indus- try concentration on the relationship between focal firm actions and firm performance is positive and significantly related to firm ROS in models 3 and 4 of Table 3 (� � 0.59, p � .01; � � 0.50, p � .01). There was no significant finding with respect to focal firm ROA. This result suggests that the posi-
tive effect of focal firm actions on performance, in terms of return on sales, is higher in concentrated industries than in nonconcentrated industries.
In Hypothesis 2b, we predict that industry con- centration positively moderates the relationship between focal firm actions and both rival actions and speed of rival action. Although there were no significant findings with respect to rival actions, the results shown in model 6 of Table 3 run counter to this hypothesis for speed of rival action (� � �0.01, p � .01). That is, in concentrated industries the relationship between firm actions and speed of rival action is weaker than it is in less concentrated industries. No significant findings were found to support or refute our Hypothesis 2c, that industry
TABLE 3 Results of Random-Effects Regression Analyses of Interactionsa
Variables Model 1:
ROA Model 2:
ROA Model 3:
ROS Model 4:
ROS Model 5: Rival Total Actions
Model 6: Rival Speed of Actions
Lagged ROA 34.48** (5.04) 34.12** (5.09) Lagged ROS 42.02** (4.30) 42.23** (4.34) Lagged rival ROS �2.18** (0.65) �0.30** (0.10) Sales 0.24 (0.23) 0.25 (0.23) �0.31* (0.18) �0.33* (0.18) Rival sales �0.00 (0.00) �0.00 (0.00) Quick ratio 1.89* (0.76) 1.91* (0.77) 1.83** (0.59) 1.85** (0.59) Industry quick ratio 0.06 (0.17) 0.02 (0.02) Market share rank 5.40* (3.01) 5.53* (3.09) 5.78* (2.35) 5.75* (2.41) �0.42* (0.23) 0.01 (0.02) Herfindahl index �4.89 (6.08) �5.42 (5.90) �5.11 (4.84) �3.90 (4.71) 1.53* (0.88) �0.05 (0.07) Industry growth 0.36 (4.91) 1.20 (4.63) 2.34 (3.84) 2.97 (3.62) 3.61** (0.46) 0.14** (0.05) Firm total actions 0.03 (0.11) 0.04 (0.12) �0.18* (0.09) �0.14† (0.09) 0.02** (0.01) 0.01** (0.00) Rival total actions �0.02 (0.04) �0.04† (0.03) Rival speed of actions �0.06 (13.65) �11.49 (10.65) Herfindahl � firm total
actions 0.05 (0.18) 0.08 (0.18) 0.59** (0.14) 0.50** (0.14) 0.01 (0.02) �0.01** (0.00)
Herfindahl � rival total actions
�0.04 (0.09) 0.10† (0.07)
Herfindahl � rival speed of actions
�28.20 (36.17) 19.54 (28.36)
Industry growth � firm total actions
0.11 (0.28) 0.15 (0.31) �0.11 (0.22) �0.17 (0.25) �0.14** (0.02) �0.01** (0.00)
Industry growth � rival total actions
0.11 (0.10) 0.15* (0.08)
Industry growth � rival speed of actions
13.43 (34.07) 34.13 (26.78)
Market share rank � firm total actions
0.07 (0.14) 0.11 (0.15) 0.20* (0.11) 0.23* (0.12) �0.01* (0.01) �0.00** (0.00)
Market share rank � rival total actions
�0.06* (0.03) �0.05* (0.03)
Market share rank � rival speed of actions
�24.84* (13.5) �19.82* (10.44)
Constant 2.24 (3.01) 2.14 (3.00) 1.21 (2.40) 0.81 (2.39) �0.50 (0.68) �0.00 (0.047) Wald chi-square 193.19** 189.51** 459.54** 453.06** 1,244.63** 2,242.03**
a Standard errors are in parentheses. Industry dummy variables were included in all regression models. Results are available upon request. n � 281.
† p � .10 * p � .05
** p � .01
72 FebruaryAcademy of Management Journal
concentration negatively moderates the relation- ship between rival firm actions and focal firm performance.
Although we found no support for Hypothesis 3a, predicting that industry demand conditions positively moderate the relationship between focal firm actions and focal firm performance, our Hy- pothesis 3b, in which we predict that industry de- mand negatively moderates the relationship be- tween focal firm actions and rival actions and rival action speed, was supported. That is, models 5 and 6 of Table 3 indicate that the interaction between focal firm actions and industry growth was nega- tive and significantly related to both rival actions (� � �0.14, p � .01) and the speed of rival actions (� � �0.01, p � .01). Thus, and as predicted, as industry demand increases, the effect of firm ac- tions on rival actions and their speed declines.
Hypothesis 3c predicts that industry demand positively moderates the relationship between fo- cal firm performance and both rival actions and the speed of rival actions. We found some support for this hypothesis, as shown in model 3 of Table 3. Specifically, the significant and positive effect of the interaction between rival actions and industry growth on focal firm return on sales (� � 0.15, p � .05) is consistent with our hypothesis.
The influence that focal firm market position has on rival actions and firm performance was explored in Hypotheses 4a– 4c. In Hypothesis 4a, we predict that market position positively moderates the rela- tionship between focal firm actions and focal firm performance. This hypothesis was partially sup- ported, as indicated in models 3 and 4 of Table 3, in which the interaction of market position and firm actions is positive and significantly related to re- turn on sales (� � 0.20, p � .05; � � 0.23, p � .05). This result suggests that the positive impact of a firm’s actions on performance is greater for firms that have higher market shares in an industry.
Hypothesis 4b predicts that market position neg- atively moderates the relationship between focal firm actions and both rival actions and their speed. This hypothesis is supported. That is, the interac- tion of firm actions with higher market position is negatively and significantly related to both rival firm actions in model 5 of Table 3 (� � �0.01, p � .05) and the speed of rival actions in model 6 of the same table (� � �0.004, p � .01). Thus, the actions of firms with higher market shares than their rivals tend to not increase rival actions and rival action speed as much as do the actions of firms with lower market shares.
Hypothesis 4c predicts market position posi- tively moderates the relationship between both ri- val firm actions and rival firm action speed and
focal firm performance. Thus, we expected that rival actions and rival action speed would not af- fect market leaders in the same way as they would affect the performance of nonleaders. Results did not support this hypothesis and, in fact, were con- trary to our expectation. Specifically, Table 3 shows that the interaction between market position and rival actions is negatively and significantly related to both focal firm return on assets in models 1 and 2 (� � �0.06, p � .05; � � �24.84, p � .05) and focal firm return on sales in models 3 and 4 (� � �0.05, p � .05; � � �19.82, p � .05). Thus, rival actions have a greater negative impact on firms with larger shares of an industry market than on firms with smaller market shares.
DISCUSSION
This study has shed light on Red Queen compe- tition by investigating the relationships between focal firm actions, rival firm actions, and focal firm performance in a variety of industries. In line with Red Queen theory, all the relationships in our base- line model were supported and showed that even though a focal firm’s actions do increase its perfor- mance, they also increase the number and speed of rivals’ actions which, at least partially, negatively impact the focal firms’ performance. Indeed, to paraphrase the Red Queen in Lewis Carroll’s Through the Looking Glass (1960), it is true that the firms studied here have “to run as fast as they can to stay in place, and twice as fast as that” to get ahead. Although portions of this baseline model have been tested elsewhere, we are aware of no previous study that has examined both the positive and negative effects of actions, as we do in the present study. Studying these effects together en- ables us to elucidate the positive and negative as- pects of action, and to clarify the relative importance of these aspects with regard to firm performance. Below we offer Figures 2 and 3 to illustrate this.
Plotting the data from the baseline models we used to test Hypotheses 1a–1c, models 1– 4 in Table 2, we graphically present the Red Queen effect in Figure 2. This graph illustrates the counter-balanc- ing effects of firm and rival actions on firm perfor- mance as measured by return on assets (ROA) and return on sales (ROS). Specifically, we see the ac- tual average number of firm actions and rival ac- tions associated with various levels of perfor- mance. As expected, performance gains from action are maximized when firm actions are high and rival actions are low. Perhaps less expected, the number of rival actions necessary to seriously, negatively impact firm performance is surprisingly high rela-
2008 73Derfus, Maggitti, Grimm, and Smith
tive to the number of firm actions necessary to positively increase performance.
Figure 3 illustrates the strength of the Red Queen effect by presenting the net incremental effect that firm actions have on firm performance directly, and indirectly, through rival actions and rival action speed. The line with the steepest positive slope in both the ROA and ROS plots represents the direct effects of firm actions on performance. For ROA, this slope is calculated as the mean of the coeffi- cients for firm actions in models 1 and 2 of Table 1 (0.08 and 0.12, respectively).
The three lines below the “firm actions line” in Figure 3 reveal the strength of the Red Queen effect
in our research. Specifically, line 2 shows how much the direct positive effect of actions on perfor- mance is reduced by the indirect negative impact focal firm actions have when they stimulate rival action. We calculated this negative impact by tak- ing the derivative of rival total actions with respect to firm total actions in model 5 multiplied by the coefficient on rival total actions in model 1 (– 0.05). Similarly, line 3 shows the direct positive effect of actions on performance along with the indirect neg- ative impact focal firm actions have by stimulating rival speed of actions. This negative impact was calculated as the derivative of rival speed of actions with respect to firm total actions in model 6 mul-
FIGURE 2 Effects of Firm Actions and Rival Actions on Firm Performance
FIGURE 3 Incremental Effects of Firm Actions on Firm Performance
74 FebruaryAcademy of Management Journal
tiplied by the coefficient on rival speed of actions in model 2 (–17.02). Comparing lines 2 and 3, we see that rival action speed has a greater negative effect than rival actions. Line 4 shows the cumula- tive negative effects of rival action and rival action speed.
Importantly, even net of the negative effects that rival actions and rival action speed have on perfor- mance, the relationship between firm action and firm performance is still positively sloped. Even though a Red Queen effect is present in our re- search, the benefits of focal firm action outweigh the potentially negative consequences of rival ac- tion in this competitive contest overall. The same analysis using ROS to measure performance is also presented in Figure 3; the results are similar to those for ROA.
The equations used to generate the plots in Fig- ure 3 can also be used to calculate the incremental impact that firm actions, rival actions, and rival action speed can have on firm performance. In the case of firm ROA, each additional firm action has an incrementally positive effect of increasing ROA by .104 percent while also causing a negative effect through rival actions and rival action speed that decreases ROA by .048 percent. Therefore. the net incremental increase in firm ROA from one firm action is .056 percent.
Furthermore, using the results from these equa- tions, it is possible to explore the impact that being more or less active can have on firm performance. For example, if we define active firms as those taking a total number of actions one standard devi- ation above the mean (23.83 actions) and less active firms as those taking a total number of actions one standard deviation below the mean (2.53 actions), we can make comparisons based on the differential number of total actions between the two categories (21.3 actions). The positive effect of those 21.3 ac- tions is 2.2 or 59.2 percent of the average ROA (3.74%) in our sample. When we include the neg- ative effect of those actions that occurs through rival actions and rival action speed, the results are still substantial: the net effect on ROA is 1.2 per- cent, 31.9 percent of the average ROA. Similar re- sults are found with respect to ROS.
This research contributes by advancing under- standing of Red Queen competition and the coevo- lutionary nature of firm search and action, and the relationship between focal firm action and rival action on focal firm performance, a key question in strategy. Specifically, conceiving competition as a contest of actions, we found support for the Red Queen effect by theoretically specifying, and em- pirically detailing, how firm actions and rival ac- tions have opposing effects on focal firm perfor-
mance. As competitive dynamics is a fairly new stream of research (Smith et al., 1992), it has lacked theoretical roots that could give traction to future research agendas (Smith et al., 2001). The present study, with its focus on the Red Queen effect, sug- gests that evolutionary theory may offer important insights that can advance understanding of the dy- namics of competition.
In an effort to better understand the boundaries of the Red Queen hypothesis, we also developed theory on how Red Queen evolution might depend upon different industry conditions and market po- sitions. Importantly, our results indicate that these factors significantly moderate the effects of firm actions on rival actions and their joint influence on performance. We speculated that these moderating factors block, or facilitate, the learning associated with search and action for both a focal firm and its rivals. Our study of these contextual moderating factors went well beyond antecedent research from IO economics. That is, despite the central role of conduct in the structure-conduct-performance par- adigm, IO researchers testing this theory have fo- cused mainly on the relationship between structure and performance and often assumed or not mea- sured the role of conduct. Additionally, their re- search has examined the direct effect that industry context and market position may have on firm per- formance, to the exclusion of actions or conduct; in contrast, our study explores the relationship be- tween actions and performance in the context of varying concentration, industry demand, and mar- ket position.
With regard to focal firm performance exhibiting a more positive relationship to firm performance in highly concentrated or high-growth industries, our findings run somewhat counter to our predictions. We speculate that, in the case of high concentra- tion, the fact that firm action was only more posi- tively related to firm performance for one measure reflects the extent to which firms closely monitor each other’s actions, are very familiar with each other’s capabilities and developments, and are so highly interdependent that they create an environ- ment in which “surprise” actions are rare, and ri- vals are more likely to counter actions quickly and efficiently, wiping out excessive gains. In high- growth industries, it may be that we didn’t find a more positive relationship between firm action and firm performance because firms often act ineffi- ciently in an effort to keep up with the demands of the market. Ample demand opportunities may cre- ate an atmosphere in which firms’ actions are stop- gaps carried out to meet rising demand without consideration of their costs. That is, high-demand environments may create a situation in which firms
2008 75Derfus, Maggitti, Grimm, and Smith
do not have the time to investigate the least costly way to take action. In this way, firms may waste resources undertaking actions in haste or perhaps when they are unnecessary.
With respect to the influence of high concentra- tion and demand on Red Queen competition, we found support for the proposition that the relation- ship between focal firm actions and rival actions is more intense in highly concentrated industries and less intense in high-growth industries. These re- sults support our contention that firms in concen- trated industries are much more interdependent than firms in less concentrated industries, while firms in high-growth industries are less interdepen- dent than those in low-growth industries. Both re- sults show how the industry context in which com- petition takes place moderates the Red Queen effect on firm evolution and performance.
Market position also appears to have an influ- ence on Red Queen competition, or the relationship between focal firm actions, rival actions, and focal firm performance. Specifically, we predicted that the positive relationship between a firm’s actions and its performance would be stronger when the firm was a market leader, and our findings partially support this notion. Results for ROS indicate that firms in stronger leadership positions do receive greater performance benefits from action. While it has been suggested that large firms may become insulated from competitive forces and unrespon- sive to Red Queen competition (Barnett & McKen- drick, 2004), our results suggest that large firms with greater market share can become better com- petitors and enhance performance by being aggres- sive with their actions.
Our predictions regarding the influence that mar- ket leaders’ actions have on rival actions were also supported. We found that the positive relationships between focal firm actions and rival actions, and rival action speed, were weaker when the focal firm was more of a market leader. We speculate that either rivals are less likely to act against leading firms out of fear of retribution, or market leaders take actions to which it is more difficult for rivals to respond.
Contrary to our hypothesis, we found that rival actions have a greater negative impact on the per- formance of market leaders than on the perfor- mance of non–market leaders; in essence we found that “the larger they are, the harder they fall.” This result may be related to the concept of “judo strat- egy” as developed by Yoffie and Kwak (2001). With judo strategy, small rivals can effectively hurt mar- ket leaders, by eliciting responses that hurt the market leader more than the rival. When the market leader’s response affects all customers, it can be
more costly for the leader than the nonleader with its lower market share. Interestingly, while our findings did not replicate Barnett and McKen- drick’s (2004) observation that smaller organiza- tions were more responsive to Red Queen compe- tition than larger organizations, we did observe that smaller firms can be more effective against their rivals by being aggressive with their actions.
Overall, our findings highlight the intricacies of the relationship between competition and perfor- mance and the complexity of studying Red Queen competition. Though the tests of our baseline hy- pothesis yielded results that are completely consis- tent with Red Queen theory, our moderation find- ings revealed that the effects of search, action, and learning on firm performance and rival interdepen- dence largely depend on industry and competitive context. Further, while context definitely impacts whether Red Queen competition constrains or en- hances learning and performance, it does not al- ways do so in the ways that antecedent research would suggest. To better understand these relation- ships, more research needs to be done. As the focus of this research was on short-term performance, future research could fruitfully explore longer-term performance consequences of Red Queen competi- tion. Are the most active and aggressive firms the best performers in the long run? How does industry context influence Red Queen competition over the long term?
Another potentially fruitful avenue for future re- search would be to focus on varying action types. To demonstrate how future research might evolve, we examined one possible characterization of ac- tion type, positive sum actions, which may allow firms to mitigate or reduce the negative aspects of Red Queen competition. Following Porter’s (1985) argument that competitors can provide strategic benefits by helping to develop markets and in- crease industry demand, some action types, namely geographic expansions, new marketing campaigns, and new-product introductions, may represent a “positive sum competition.” An illustration of this win-win dynamic can be seen when Starbucks en- ters a new geographic market in the retail coffee industry. Rather than negatively affecting the exist- ing competition in the new market, these competi- tors often witness increases in their business, as consumers become more comfortable with spe- cialty coffee and overall demand increases with the presence of the new Starbucks (Helliker & Leung, 2002). Similarly, new-product introductions can positively increase or create demand for all com- petitors in an industry. For example, Sony’s intro- duction of the Walkman and Apple’s introduction of the iPod opened the door for a multitude of
76 FebruaryAcademy of Management Journal
imitations from competitors who garnered reve- nues from the expanded market.
At least three types of moves are acknowledged in the literature to have the potential for positive, demand-expanding effects: geographic expansion, promotional campaigns, and new-product intro- ductions.4 To explore the potential of positive sum actions for future research, we created a measure of focal firm and rival positive sum actions by sum- ming actions categorized as geographic, marketing, and/or product introduction. We calculated firm positive sum actions by simply summing the num- ber of instances of all three actions for a focal firm in a given year. Rival positive sum actions were operationalized by subtracting a focal firm’s total number of positive sum actions in a given year from the total number of positive sum actions taken by all competitors in the industry.
Post hoc regression of these types of actions yielded interesting results.5 Specifically, it appears that focal firm positive sum actions are signifi- cantly and positively related to firm performance in the case of return on sales. This result is consistent with our baseline model. Unlike in our baseline results, however, here there is a negative and sig- nificant relationship between firm positive sum ac- tions and both rival actions and rival action speed. In addition, rival positive sum actions and firm performance exhibited no significant relationship.
Taken together, these findings suggest that firm positive sum actions incite less rivalrous action and slower rival action speed. Further, when rivals take positive sum actions, focal firm performance does not suffer significantly. This finding is again consistent with the speculation that firms take ac- tions that build legitimacy and benefit all players in an industry during the early, high-growth period of the industry life cycle (Agarwal & Bayus, 2002). These results also indicate that future research in- corporating action type could enhance understand- ing of Red Queen competition. One avenue for fu- ture research would be to combine action type effects with an examination of longer-term conse- quences of Red Queen competition. To clarify, the above analysis revealed that positive sum actions may mitigate the negative effects of Red Queen competition by reducing both the number and speed of rival actions. However, one could also conjecture that when a focal firm takes positive sum actions, rivals are more likely to take positive sum actions in response, a sequence that may lead to a rivalry-reducing “loop” with positive perfor- mance consequences over the longer term. Al- though a complete analysis is beyond the scope of our current study, it is interesting to note that there is a relatively high correlation between firm posi- tive sum actions and rival positive sum actions within our data set (r � .41), providing an indica- tion that firm positive sum actions beget rival pos- itive sum actions.
This study has implications for practice. Specif- ically, the study of Red Queen competition pro- vides a number of insights regarding what actions managers can take and under what conditions they should take them in an effort to increase perfor- mance. Managers of competing firms in highly con- centrated industries or low-growth industries should be acutely aware of their mutual interde- pendence and be cautious of taking actions for fear of competitive reprisal. Our research also suggests that challenger firms can effectively hurt industry leaders by taking action, which is somewhat con- trary to conventional wisdom. Still, leading firms received a bigger payoff for acting than did non- leading firms. Although more research is required, the post hoc results suggest that managers can po- tentially avoid the negative consequences of rivalry by emphasizing positive sum actions such as geo- graphic actions and product introductions.
Like most research, our study has limitations. First, although we studied a minimum of 70 per- cent of the business activity in each of 11 different industries, our sample favors large, public, single U.S. business firms that perhaps are in the later stages of the organizational life cycle. In particular,
4 Geographic expansion, extending a firm’s reach to customers not previously served, can allow a firm to avoid head-to-head competition with existing rivals. Geographic expansion can be targeted where competition is weak or nonexistent, perhaps in the process filling a previously underserved geographic segment (Porter, 1980). Promotional campaigns that generate new custom- ers and new demand are also consistent with a positive sum notion. As both parties increase their marketing efforts, the actual or potential customer base can be ex- panded, creating a situation in which the marketing ac- tions of the focal firm and the rival firm will both have positive benefits for the focal firm (Warren, 2002). Mar- keting campaigns can also aid firms in differentiating their products. Barney pointed out that product differen- tiation “reduces the threat of rivalry, because each firm in an industry attempts to carve out its own unique product niche” (Barney, 1997: 237). New-product introductions can also have positive impacts on demand. Each new- product introduction sets expectations and challenges rivals to act creatively with their own products in order to catch up (Schumpeter, 1934). The consequence of such innovation is that overall demand may increase as more and better new products are introduced to the mar- ket over time.
5 Regression results are available upon request from the authors.
2008 77Derfus, Maggitti, Grimm, and Smith
the competitive dynamics between firms that are young, small, diversified, and private are not cap- tured by this study. We speculate that the compet- itive dynamics among firms in fragmented indus- tries and in early stages of the life cycle, where mutual interdependence is lower, would be quite different. Future research should continue investi- gation of these issues in these other industry con- texts. As is the case in competitive dynamics re- search, our study also captures only observable moves reported in the press and publications we examined. In addition, our use of product-market- based industries to define the competitive land- scape may not address the potentially changing nature of competition. For example, it is likely that focal firms in our sample face competitors from outside their industry or outside the United States that offer substitute products, have similar resource positions, or exist in geographic markets that are new to the focal firms (Chen, 1996; Peteraf & Ber- gen, 2003). Future research can overcome this lim- itation by identifying competitors on the basis of criteria other than product-market commonality.
Additionally, although we theorize that search and learning take place in this Red Queen com- petitive context, we do not directly specify and measure these constructs. Future research could further elucidate Red Queen competition by ex- amining more specifically the search and learn- ing that take place. Future research could also consider additional links beyond the scope of the current study, such as whether and how firm competitive experience and performance shape future firm action. We can speculate that a firm’s prior learning and experience with action affect its future action and action speed. Moreover, ad- ditional research could examine the extent to which a focal firm learns vicariously from other firms (Baum, Li, & Usher, 2000).
In conclusion, the most important contribution of this research is its theoretical and empirical examination of Red Queen competition, con- ceived of as the positive and negative conse- quences of firm actions on performance. Specifi- cally, the research shows that firm actions can play out as a Red Queen race among rivals: Firm actions are related to rival actions and rival ac- tion speed, and all can impact focal firm perfor- mance. However, we also demonstrated that the effects of firm action on performance are complex and dependent on the industry context and the market positions of competitors. Additional the- ory and research are needed to improve under- standing of Red Queen competition.
REFERENCES
Agarwal, R., & Bayus, B. 2002. The market evolution and take-off of new product innovations. Management Science, 48:1024 –1041.
Agarwal, R., & Gort, M. 1996. The evolution of markets and entry, exit and survival of firms. Review of Economics and Statistics, 78: 489 – 498.
Agarwal, R., Sarkar, M. B., & Echambadi, R. 2002. The conditional effect of time on firm survival. Academy of Management Journal, 45: 971–994.
Argote, L. 1999. Organizational learning: Creating, re- taining and transferring knowledge. Boston: Klu- wer.
Bain, J. S. 1951. Relation of profit rate to industry con- centration: American manufacturing, 1936 –1940. Quarterly Journal of Economics, 65: 293–324.
Barnett, W. P., & Hansen, M. T. 1996. The Red Queen in organizational evolution. Strategic Management Journal, 17: 139 –157.
Barnett, W. P., & McKendrick, D. 2004. Why are some organizations more competitive than others? Evi- dence from a changing global market. Administra- tive Science Quarterly, 49: 535–571.
Barnett, W. P., & Sorenson, O. 2002. The Red Queen in organizational creation and development. Industrial and Corporate Change, 11: 289 –325.
Barney, J. 1986. The types of competition and the theory of strategy: Toward an integrative framework. Acad- emy of Management Review, 11: 791– 800.
Barney, J. 1997. Gaining and sustaining competitive advantage. Reading, MA: Addison-Wesley.
Baum, J. A. C., Li, S. X., & Usher, J. M. 2000. Making the next move: How experiential and vicarious learning shape the locations of chains’ acquisitions. Admin- istrative Science Quarterly, 45: 766 – 801.
Baumol, W. J. 2004. Red-queen games: Arms races, rule of law and market economies. Journal of Evolution- ary Economics, 14: 237–247.
Bothwell, J. T., Cooley, T., & Hall, T. 1984. A new view of the market structure-performance debate. Journal of Industrial Economics, 32: 397– 417.
Bradburd, R. M., & Caves, R. 1982. A closer look at the effect of market growth on industries profits. Re- view of Economics and Statistics, 64: 635– 645.
Cameron, A., & Travendi, P. 1986. Economic models based on count data: Comparisons and applications of some estimators and tests. Journal of Applied Econometrics, 1: 29 –53.
Carroll, G., & Hannan, M. 1989. Density dependence in the evolution of newspaper organizations. American Sociological Review, 54: 524 –541.
Carroll, G., & Hannan, M. 2000. The demography of corporations and industries. Princeton, NJ: Prince- ton University Press.
78 FebruaryAcademy of Management Journal
Carroll, L. 1960. The annotated Alice: Alice’s adven- tures in wonderland and through the looking- glass. New York: New American Library.
Caves, R. 1980. American industry: Structure, conduct, performance (6th ed.). Englewood Cliffs, NJ: Pren- tice-Hall.
Caves, R., & Porter, M. E. 1977. From entry barriers to mobility barriers. Quarterly Journal of Economics, 91: 241–261.
Chen, M. J. 1988. Competitive strategic interaction: A study of competitive actions and responses. Un- published manuscript, University of Maryland, Col- lege Park.
Chen, M. J. 1996. Competitor analysis and interfirm ri- valry: Toward a theoretical integration. Academy of Management Review, 21: 100 –135.
Chen, M. J., & Miller, D. 1994. Competitive attack, retal- iation, and performance: An expectancy-valence framework. Strategic Management Journal, 15: 85– 102.
Comanor, W., & Wilson, T. 1967. Advertising, market structure, and performance. Review of Economics and Statistics, 49: 423– 440.
Cubbin, J., & Domberger, S. 1988. Advertising and post- entry oligopoly behaviour. Journal of Industrial Economics, 37: 123–140.
Dawkins, R. & Krebs, J. R. 1979. Arms races between and within species. Proceedings of the Royal Society of London, B205: 489 –511.
Eisenhardt, K. M., & Tabrizi, B. N. 1995. Accelerating adaptive processes: Product innovation in the global computer industry. Administrative Science Quar- terly, 40: 84 –110.
Erez, A., Bloom, M. C., & Wells, M. T. 1996. Using ran- dom rather than fixed effects models in meta-analy- sis: Implications for situational specificity and valid- ity generalization. Personnel Psychology, 49: 275.
Ferrier, W. J., Smith, K. G., & Grimm, C. M. 1999. The role of competitive action in market share erosion and industry dethronement: A study of industry leaders and challengers. Academy of Management Journal, 42: 372–383.
Frome, E. L., Kutner, M. H., & Beauchamp, J. J. 1973. Regression analysis of Poisson-distributed data. Journal of American Statistical Association, 68: 935–940.
Gaskins, D. W. 1971. Dynamic limit pricing: Optimal limit pricing under threat of entry. Journal of Eco- nomic Theory, 3: 306 –322.
Gilbert, R. J., & Newberry, D. M. 1982. Preemptive pat- enting and the persistence of monopoly. American Economic Review, 72: 514 –526.
Graham, E. M. 1979. Technological innovation and the dynamics of the U.S. comparative advantage in in- ternational trade. In C. T. Hill & J. M. Utterback
(Eds.), Technological innovation for a dynamic economy. New York: Pergamon.
Greene, W. H. 1993. Econometric analysis (2nd ed.). New York: MacMillan.
Grimm, C. M., & Smith, K. G. 1997. Strategy as action: Industry rivalry and coordination. Cincinnati: South-Western College Publishing.
Helliker, K., & Leung, S. 2002. Counting beans: Despite the jitters, most coffeehouses survive Starbucks. Wall Street Journal, September 24, 2002: A1.
Ingram, P. 2002. Interorganizational learning. In J. A. C. Baum (Ed.), Companion to organizations: 642– 664. Oxford, UK: Blackwell.
Jauch, L. R., Osborn, R. N., & Martin, T. N. 1980. Struc- tured content analysis of cases: A complimentary method for organizational research. Academy of Management Journal, 5: 517–526.
Kamien, M., & Schwartz, N. 1971. Limit pricing and uncertain entry. Econometrica, 39: 441– 454.
Levitt, B., & March, J. G. 1988. Organizational learning. In W. R. Scott (Ed.), Annual review of sociology, vol. 14: 319 –340. Palo Alto, CA: Annual Reviews.
Liebowitz, S. J. 1982. Measuring industrial disequilibria. Southern Economic Journal, 49: 119 –136.
March, J., & Simon, H. 1958. Organizations (2nd ed.). Cambridge, MA: Blackwell.
Masson, R. T., & Shannan, J. 1986. Excess capacity and limit pricing: An empirical test. Economica, 53: 365–378.
Miller, D., & Chen, M. J. 1994. Sources and consequence of competitive inertia: A study of the U.S. airline industry. Administrative Science Quarterly, 39: 1–23.
Nelson, R. R., & Winter, S. G. 1982. An evolutionary theory of economic change. Cambridge, MA: Belknap.
Peteraf, M. A., & Bergen, M. E. 2003. Scanning dynamic competitive landscapes: A market-based and re- source-based framework. Strategic Management Journal, 24: 1027–1041.
Pisano, G. P. 1994. Knowledge, integration and the locus of learning: An empirical analysis of process devel- opment. Strategic Management Journal, 15: 85– 100.
Porter, M. E. 1980. Competitive strategy. New York: Free Press.
Porter, M. E. 1985. Competitive advantage: Creating and sustaining superior performance. New York: Free Press.
Scherer, F. M., & Ross, D. 1990. Industrial market struc- ture and economic performance. Boston: Hough- ton-Miller.
Schmalensee, R. 1976. Advertising and profitability: Fur-
2008 79Derfus, Maggitti, Grimm, and Smith
ther implications of the null hypothesis. Journal of Industrial Economics, 25: 45–54.
Schomburg, A. J., Grimm, C. M., & Smith, K. G. 1994. Avoiding new product warfare: The role of industry structure. In P. Shrivastava, A. Huff, & J. Dutton (Eds.), Advances in strategic management, vol. 10: 145–174. Greenwich, CT: JAI.
Schumpeter, J. 1934. The theory of economic develop- ment. Cambridge, MA: Harvard University Press.
Schumpeter, J. A. 1976. Capitalism, socialism, and de- mocracy (5th ed.). London: George Allen & Unwin.
Smith, K. G., Ferrier, W. J., & Ndofor, H. 2001. Competi- tive dynamics research: Critique and future direc- tions. In M. A. Hitt, R. E. Freeman, & J. S. Harrison (Eds.), Handbook of strategic management: 315– 361. Oxford, UK: Blackwell.
Smith, K. G., Grimm, C. M., & Gannon, M. J. 1992. Dy- namics of competitive strategy. London: Sage.
Spence, A. M. 1977. Entry, capacity, investment and oligopolistic pricing. Bell Journal of Economics, 8: 531–544.
van Valen, L. 1973. A new evolutionary law. Evolution- ary Theory, 1: 1–30.
Waldman, D. E., & Jensen, E. J. 2001. Industrial organi- zation. Boston: Addison-Wesley.
Warren, K. 2002. Competitive strategy dynamics. New York: Wiley.
Weick, K. 1995. Sensemaking in organizations. Thou- sand Oaks, CA: Sage.
Williamson, O. E. 1965. A dynamic theory of interfirm behavior. Quarterly Journal of Economics, 79: 579 – 607.
Yoffie, D. B., & Kwak, M. 2001. Judo strategy. Boston: Harvard Business School Publishing.
Young, G. 1993. Engaging a rival: Industry and firm specific predictors of rivalrous firm conduct in the U.S. software industry. Unpublished manuscript, University of Maryland, College Park.
Young, G., Smith, K. G., & Grimm, C. M. 1996. “Austrian” and industrial organization perspectives on firm- level competitive activity and performance. Organi- zation Science, 7: 243–254.
Pamela J. Derfus ([email protected]) received her Ph.D. in strategic management from the Robert H. Smith School of Business at the University of Maryland. Her research interests lie in the areas of competitive dynam- ics, strategy implementation, and cooperation between firms.
Patrick G. Maggitti ([email protected]) is an assistant professor of management and entrepreneurship in the Fox School of Business at Temple University. His re- search focuses on the dynamics of competition and the decision making of executives, entrepreneurs, and inves- tors. He received his Ph.D. in strategic management from the Robert H. Smith School of Business at the University of Maryland.
Curtis M. Grimm ([email protected]) is the Dean’s Professor of Supply Chain and Strategy at the Robert H. Smith School of Business, University of Mary- land. He received his Ph.D. in economics from the Uni- versity of California, Berkeley, with primary focus on industrial organization. Professor Grimm’s research has focused on the interface of business and public policy with strategic management, with a particular emphasis on competition and competition policy.
Ken G. Smith ([email protected]) is the Dean’s Chair and a professor of business strategy at the Robert H. Smith School of Business, University of Maryland. He earned a Ph.D. in business policy from the University of Washington. His research interests include strategic de- cision making, competitive dynamics, and the manage- ment of knowledge and knowledge creation.
80 FebruaryAcademy of Management Journal
