The notion of intra-industry groups has been developed within the strategic-management field to address the fundamental questions of competitor definition and strategic behaviors among competitors (Daems & Thomas, 1994; Hunt, 1972; Ketchen, Snow, & Hoover, 2004; McGee & Thomas, 1986; Short, Ketchen Jr., Palmer, & Hult, 2007; Thomas & Pollock, 1999). The term "strategic group" has been used traditionally in the strategy literature to refer to intra-industry groups comprising firms that "follow the same or similar strategy along strategic dimensions" (Porter, 1980, p. 20).
An alternative conceptualization of intra-industry groups can be found within the cognitive school of strategy. Scholars working in this tradition have shown that when managers' subjective interpretations of the competitive landscape in which they operate are shared, a "cognitive community" arises (Porac & Thomas, 1990, 1995; Porac et al., 1989). For example, cognitive communities have been found in the case of Scottish knitwear manufacturers (Porac, Thomas, & Baden-Fuller, 1989), Chicago banks (Reger, 1990), US hospitals (Nath & Gruca, 1997) and commercial banks (Deephouse, 1999), and the US pharmaceutical industry (Osborne et al., 2001).
Another conceptualization of intra-industry groups is that of "legitimacy-based group" (Barreto & Baden-Fuller, 2006; Deephouse, 1996). Similar to cognitive communities, legitimacy-based groups are also related to managers' shared cognition. Unlike cognitive communities, however, the formation of shared cognition is not based on managers' interactions with other firms or among themselves but on the categorization by an external observer with the authority to assess whether firms behave according to the required standards. The notion of legitimacy-based groups posits that the strong normative isomorphic influence provided by a powerful authority renders a certain grouping of firms legitimate. Once built, legitimacy-based groups are used by firms as a reference. Legitimacy-based groups are, in this sense, socially constructed.
The commensurability of these intra-industry groups has recently attracted the attention of scholars; a few studies have reported much overlap between strategic groups and cognitive communities (G. Hodgkinson, 2005; G. P. Hodgkinson, 1997b; Osborne, Stubbart, & Ramaprasad, 2001). However, although earlier implied by Barreto and Baden-Fuller (2006), whether a legitimacy-based group would function as a cognitive community for the managers of firms has not been tested empirically. If this is the case, the role of a legitimacy provider, or an external categorizer, in the strategy formation should be emphasized. Furthermore, the similarity of legitimacy-based groups and cognitive communities justifies deductive cluster analysis in the identification of cognitive communities.
This article tries to compare legitimacy-based groups and cognitive communities, which are identified separately. Taking into account such factors as data availability and comparability to previous studies (Barreto & Baden-Fuller, 2006; Deephouse, 1996), we chose the Korean asset-management industry during the 2001-2007 period as a suitable context for the purposes of our research.
Intra-industry structure and the strategic group
The notion of the strategic group links back to the early works of industrial organization (IO) researchers examining the intra-industry heterogeneity of firms. The original version of SC-P (Structure-Conduct-Performance) theory posits the existence, within any industry, of a single oligopolistic group of firms with an optimal strategy. The only difference among firms that could be allowed for was the effect of the scale. However, this explanation could not account for the variety of strategies that firms adopt in any industry. Subsequently, some scholars coined the term strategic group, and this term has been used traditionally in the strategy literature to refer to intra-industry groups comprising firms that "follow the same or similar strategy along strategic dimensions" (Porter, 1980, p. 20).
With the aid of the strategic group construct, the long-run heterogeneity of firms in an industry becomes compatible with the S-C-P paradigm, for the strategy and resulting profitability can vary across strategic groups. According to the traditional S-C-P paradigm, firms within different strategic groups will display different strategies, which lead to differences in performance across strategic groups; whereas, the performance of firms within a strategic group will be similar due to the similarity of their strategies, as is implied by the definition of a strategic group. To date, however, empirical research has cast doubts on the validity of such an argument.
Some researchers, mainly those from the resource-based view (RBV) vein, have focused more on the similarity of firms in terms of the composition of resources, than on strategic similarity, when clustering firms in an industry. For example, Hatten and Hatten (1987, p. 329) define a strategic group as "a grouping of organizations which pursue similar strategies with similar resources." Similarly, Cool and Schendel (1987) consider a strategic group as "a set of firms competing within an industry for the similar combination of scope and resource commitment" (p. 1106).
While the IO approach to the identification of strategic groups is mainly the classification of an industry into groups (Bogner, Pandian, & Thomas, 1994; Bogner & Thomas, 1993), the RBV approach starts at the level of the firm, thus aggregating firms with similar strategies and resources into groups. The emphasis upon the resource similarity of firms within a strategic group implied by the RBV perspective suggests the possibility for substantial intragroup performance differences caused by strong competition among firms with similar resources and intergroup performance differences.
With alternative theoretical perspectives and the mixed empirical results, some scholars have questioned the validity of the strategic group concept as a theoretical construct, while others have argued for the use of more sophisticated methodologies in empirical studies.
Cognitive perspective on the strategic group: cognitive community
Scholars in the cognitive school believe that managers have a very narrow, asymmetric cognition of competition, which in turn affects the development of competitive strategy. Managers pay attention to the activities of firms whose action repertoire includes the types of actions they perceive to be important (from experience and learning), and of course, they pay little attention to firms that use fewer of these actions, according to the managers' perception and valued types of action (Marcel, Barr, & Duhaime, 2005). Therefore, the selection of competitors' action repertoires to be responded to is conducted with reference to managers' mental models (Miller, 1993). Similarly, the selection of competitors to be paid attention to also depends on managers' mental models.
In the seminal study carried out by Porac, Thomas, and Baden-Fuller (1989), it was found that Scottish knitwear manufacturers only consider other Scottish manufacturers as their competitors even though Scottish knitwear manufacturers only account for 3% of the global market share. This narrow focus on competitor identification was certainly one good reason for the late, and poor, response by the American automobile manufactures toward their Japanese counterparts in the 1980s (Yates, 1983). Similar to the organizational identity that members of a firm have developed and shared about their firm, managers may have another type of identity in mind, group identity, so to speak (Peteraf & Shanley, 1997), about who their competitors are and are not. The categorization of competitors according to managers' mental models may be inevitable, but it can often lead to blind spots (Zajac & Bazerman, 1991).
Motivated by studies and practical examples, cognitive school researchers argue that managers in a firm focus on a small number of competitors whom they believe are similar to their firm. In other words, managers set the reference group for firms and do not search whole firms in their industry when developing strategy (Daniels & Henry, 1998; Fahey & Narayanan, 1989; Marcel, Barr, & Duhaime, 2005; Reger & Huff, 1993). Therefore, group structure becomes a reality as a social construction by industry participants and firms sharing a group identity. Peteraf and Shanley (1997) thus rephrased the definition of the strategic group with this sociocognitive group in mind:
We used the term "strategic group " to denote a meaningful substructure of firms within an industry--one that is acknowledged by industry participants and has significance for them. In contrast to some other definitions, we neither define strategic groups on the basis of similar strategies nor distinguish them on this basis. While members of the same strategic group may employ similar strategies, this is not a necessary condition for the groups to have significance. (p. 166) These sociocognitive groups have been investigated since the early 1990s (Churchman & Woodard, 2004; G. P. Hodgkinson, 1997b; Osborne, Stubbart, & Ramaprasad, 2001; Porac, Thomas, & Baden-Fuller, 1994; Wry, Deephouse, & McNamara, 2006). They are often called "cognitive communities" instead of strategic groups, for firms within each group do not necessarily show similarities in strategies. This cognitive perspective is distinct from economic perspectives on strategic groups in the sense that it regards group structure as resulting not solely from economic reality but from an interweaving of objective reality and subjective beliefs.
How are these groups formed and maintained? It has been argued that there is a process of competitive enactment among firms within a cognitive community....