Intended rationality in exit decisions: The organizational field of rivals as a source of signal in segment exit decisions by Korean SI ventures,2000

Abstract

This paper examines how the aggregated actions of rivals belonging to the same organizational field affect a firm’s exits from existing market segments. We argue that while a firm’s exit decisions are strongly influenced by the aggregated actions by other firms which may serve as an uncertainty-reducing signal, the focal firm’s strategic responses to signals from the organizational field of rivals are more complicated, informed, and rational, than is usually suggested. For instance, in its evaluation of signals about segment attractiveness based on the aggregated actions of rivals, the focal firm will give different weights to rivals’ actions according to the strength of competitive threat, will make refined inferences about the attractiveness of a segment based on the performances of rivals, and will give a greater weight to recent actions by rivals than to older ones. The results of our event history analysis of all segment exits by Korean SI ventures support all our hypotheses.

Keywords

Intended rationality segment exit signal SI ventures uncertainty

1. Introduction

This paper attempts to complement the mimetic isomorphism argument of institutional theory by examining the dynamics of segment exit in which a firm changes its niche composition by exiting an existing market segment, from the perspective of intended rationality (DiMaggio and Powell, 1983, 1991; March and Simon, 1958). We argue that while a firm’s exit decision under uncertainty is strongly influenced by the aggregated actions of rival firms belonging to the same organizational field as suggested in the mimetic isomorphism argument of institutional theory (DiMaggio and Powell, 1983), the focal firm’s evaluation of such uncertainty-reducing signal from the field of rivals about segment attractiveness is far more complicated, informed, and calculative at least at the intention level, than is usually suggested in the existing institutional theory literature.

Since the late 1970s, institutional theory has radically enriched the field of management studies by proposing an alternative model of the firm as a social actor, which may significantly substitute and/or complement the previously dominant model of the firm as a rational, economic, and strategic actor (Cabantous et al., 2010; DiMaggio and Powell, 1983; Hough and White, 2003; Meyer and Rowan, 1977). Although it is undisputable that the emergence of institutional theory was one of the most important theoretical developments in the history of management studies, this new perspective nonetheless has been constantly criticized for its image of the firm as an overly non-rational social actor which conforms to social pressures (DiMaggio, 1988, 1991; Tolbert and Zucker, 1983). To integrate these two competing conceptions of the firm, this paper returns to the initial formulation of the bounded rationality argument which assumed that an organization is intendedly rational, but only limitedly so (March and Simon, 1958).

The perspective of intended rationality in this paper posits that, although choices of a firm under uncertainty are often socially influenced by other firms belonging to the same organizational field in a bandwagon manner as suggested in institutional theory rather than based solely on its rational and maximizing calculation due to various limits of rationality, the firm still intends and attempts to evaluate social influences from the organizational field of rival firms as a type of uncertainty-reducing in a far more rational and calculative way than usually assumed in the existing institutional theory literature. From this perspective, this paper examines how the aggregated actions of rivals belonging to the same organizational field affect the focal firm as an intendedly rational actor when it is faced with the uncertainty surrounding exit decisions, such as uncertainty about the future attractiveness or risk of a market segment.

Although exits from existing market segments have been seldom examined in strategic management literature, which has been interested primarily in entries to new market segments or industries (Bowman and Singh, 1993; Brauer, 2006; Haveman, 1993; Karakaya, 2000; Tong and Reuer, 2007), the small number of existing studies on this topic have attributed exit decisions to economic calculation of various factors at the levels of firm, segment, and industry (Bergh and Lawless, 1998; Bruyaka and Durand, 2011; Hamilton and Chow, 1993). Unlike these studies, we focus on the aggregated actions of rivals as an uncertainty-reducing signal in exit decisions following the mimetic isomorphism argument of institutional theory (Denrell, 2003; DiMaggio and Powell, 1983; Gimeno and Woo, 1996; Greve, 2003). That is, we argue that, faced with the uncertainty regarding the future attractiveness or risk of a market segment, the firm as a boundedly rational actor will monitor and learn from the aggregated actions of rivals in and around the segment to find a clue that may guide its exit decision (Cohen and Levinthal, 1990; Denrell, 2003; DiMaggio and Powell, 1983).

Literature from the two fields of management studies – i.e. strategic management and organization theory – propose different theories about the types of signals from the organizational field of rivals. Whereas most strategy studies on interfirm interaction treat rivals as a threat by focusing on competitive rivalry (Grimm et al., 2006; MacMillan et al., 1985; Smith et al., 1991), most organization studies view rivals as a means of judging the attractiveness (i.e. legitimacy) of an action which the focal firm may imitate (Baum et al., 2000; Denrell, 2003; Haveman, 1993). We argue that under circumstances of high uncertainty, the aggregated actions of rivals belonging to the same organizational field in fact play both roles simultaneously. That is, the aggregated actions of rivals operating in a market segment provide the focal firm with mixed signals about both the attractiveness and competitive threat of the segment.

However, by underlining the ‘intended’ rationality aspect of the bounded rationality thesis (March and Simon, 1958), we argue that the focal firm’s strategic responses to uncertainty-reducing signals from the aggregated action of rivals when making exit decisions are considerably more complicated, informed, and rational (in terms of intention), than is usually suggested in organization theory literature (DiMaggio, 1988). For instance, the focal firm evaluate the competitiveness of the segment with a considerable degree of accuracy by giving different weights to rivals’ actions according to the strength of competitive threat that each rival exerts, will make inferences about the attractiveness of the segment based on the performances of rivals operating in that segment, will give a greater weight to recent actions by rivals than to older ones, and will adjust its sensitivity to influences from rivals depending on its own performance, in aggregating actions of rivals belonging to the same organizational field.

To study the strategic dynamics of segment exits empirically, we analyzed entrepreneurial ventures, which usually execute niche change far more frequently than large established firms (Katila and Shane, 2005; Kuratko et al., 2001; Naman and Slevin, 1993). According to the results of our event history analysis of segment exits from 199 Korean systems integration (SI) ventures over the seven-year period from 2000 to 2006, all of our hypotheses were strongly supported. General implications from these findings are discussed.

2. Theory

2.1. Segment exit and uncertainty

The key question to consider in this paper is how a firm assesses the future attractiveness or competitiveness of a market segment when it has to make a decision of whether or not to exit it under uncertainty. Compared to entry into a new market segment or industry, exit decision has received only scant attention in existing strategy literature, due to a common perception that an exit is a mirror image of an entry (Brauer, 2006). Besides this, strategy scholars have been interested mainly in entries into new markets or industries (Haveman, 1993; Karakaya, 2000; Tong and Reuer, 2007). Although limited in quantity, existing studies on exit decisions have focused on the following two groups of factors that may facilitate or obstruct exit from an industry: industry-level factors and firm-level factors (Porter, 1980; Tirole, 1988).

Studies within the former of these groups have paid particular attention to the industry-level conditions that may facilitate or deter exits, such as uncertainty and turbulence (Bergh and Lawless, 1998; Chatterjee et al., 2003), stagnated market growth (Lovejoy, 1971), and intensified competition (Ilmakunnas and Topi, 1999; Van Kranenburg et al., 2002). In contrast, studies within the latter group have focused on firm-level factors, such as financial situation (Hamilton and Chow, 1993; Markides, 1992; Montgomery and Thomas, 1988), performance problems (Moliterno and Wiersema, 2007; Montgomery and Thomas, 1988), excessive diversification (Chang and Singh, 1999; Hoskisson and Hitt, 1994), and ownership structure (Chatterjee et al., 2003).

However, most of these studies have depicted exit decision as being an economic calculation of various factors at the levels of firm, segment, and industry, based on the assumption that the firm is a rational economic actor. Thus, they have been limited in that neither high uncertainty surrounding exit decisions nor the bounded rationality of the firm was taken into serious consideration. It has been argued that uncertainty seriously hampers the rationality of decision making by economic actors (Loomes and Sugden, 1982; March and Olsen, 1976). Moreover, if we assume there is bounded rationality, the economic rationality of decision making may be even further hampered.

Moreover, uncertainty poses far more serious risks in exit decisions than it does in entry decisions. Unlike entry decisions, exit decisions often involve the potential loss of sunk costs invested in the focal segment (Bergh, 1997; Hoskisson and Hitt, 1994). Following from this, it has been suggested that exit decisions are risky, especially in knowledge-intensive high-tech industries such as the systems integration industry, because firms will incur large sunk costs from R&D expenditures (O’Brien and Folta, 2009). It is also suggested that market exit often causes damage to quality, variety, customer service, or innovation, and thus in the long run to consumer welfare (Guiltinan and Gundlach, 1996; Karakaya, 2000). Therefore, exit decisions require more accurate prediction about the future outcomes of strategic options than do entry decisions. Moreover, the real-option type experimentations that are often employed in entry decisions under uncertainty are difficult to apply to exit decisions (McGrath and Nerkar, 2004; Miller and Arikan, 2004).

Above all, to make a rational decision about segment exit based on economic calculation, a firm must be able to assess accurately the future attractiveness or competitiveness of a segment. However, compared to the attractiveness and competitiveness of an industry, those of a market segment, including its future profitability, growth rate, and the degree of competitive threat, are much more difficult to estimate, not only because most available data and information are aggregated at the industry level, but also because other firms’ strategic actions at the segment level are much less observable than those at the industry level. This study therefore examines how firms make decisions about segment exit in a problematic decision-making context.

2.2. Aggregated actions of rivals as an uncertainty-reducing signal in exit decisions

Both the literatures of strategic management and organization theory suggest that when there is high uncertainty, a firm’s strategic decision is often influenced by actions of relevant other firms – i.e. rivals. Therefore, this paper views the rivals of a firm to constitute a type of organizational field that collectively influence the focal firm’s choices under uncertainty (DiMaggio and Powell, 1983). However, as earlier stated, these both of these pieces of management literature tend to focus on different aspects of interfirm influences from rivals.

Competitive interaction literature within the strategic management field treats rivals as a source of competitive threat (MacMillan et al., 1985; Porter, 1980; Smith et al., 1991). For instance, Cool and Dierickx (1993: 50) explicitly state that ‘high rivalry means firms adversely affect each other’ Porter (1980: 17) also highlighted the threat aspect of rivals by arguing that if moves and counter-moves between rivals escalate, all the agents involved may suffer. That is, most of these existing studies about interfirm interaction in strategy literature have focused on how a firm reacts to competitive actions by others by treating rivals primarily as threats (Grimm et al., 2006).

By contrast, most studies about interfirm interaction in the organization theory literature focus on why and how a firm imitates rivals’ actions. From an institutional theory perspective, DiMaggio and Powell (1983) argue that faced with uncertainty, firms often imitate actions that are popular among other firms belonging to the same organizational field because these actions are perceived to be legitimate, a process referred to as ‘mimetic isomorphism.’ A similar argument on interfirm imitation is also found in vicarious learning literature, which suggests that faced with uncertainty, firms often monitor, learn from, and imitate popular choices of others (Baum et al., 2000; Cohen and Levinthal, 1990). That is, unlike strategic management literature which mostly highlights the competitive threat effects of rival firms, organization theory literature views the organizational field of rivals as a source where the focal firm can find signal about the legitimacy of certain choices which it can imitate.

By integrating these two conflicting arguments from the perspective of signaling, we suggest that under circumstances of high uncertainty, aggregated actions of rivals are likely to affect the focal firm’s choice in both ways as a key source of signal. Signaling theory literature suggests that when an actor is faced with uncertainty or information asymmetry, that actor often tries to find signal that may provide clues for decision making, such as the choices made by others in similar situations (Heil and Robertson, 1991; Moore, 1992; Spence, 1973). As already discussed, exit from an existing market segment is a clear example of a situation with high uncertainty. Thus when there is insufficient information about the future attractiveness or competitiveness of a market segment, a firm may attempt to reduce uncertainty by finding signal from the aggregated actions of relevant or comparable firms, i.e. its rivals belong to the same organizational field (Guillen, 2003; Haunschild and Miner, 1997; Henisz and Delios, 2001).

2.3. The firm as an ‘intendedly rational’ actor

This paper’s conception of the firm departs not only from the model of a rational economic actor which is assumed in conventional economics and strategic management literatures, but also considerably from that of a non-rational social actor which is assumed in existing literature on imitative behaviors such as institutional theory. For instance, in institutional theory literature, it is posited that actors adopt highly popular actions which are unquestioned, legitimated, and taken for granted among the members of an organizational field by conforming to social pressures from others (DiMaggio and Powell, 1991). That is, although this paper manifestly disagrees with the excessively rationalist conception of the firm that is present in mainstream economics and strategy literatures, we also do not completely concur with the non-rational and excessively social conception of the firm often employed in institutional theory.

Although existing organizational theory tends to treat interfirm imitation among rivals under uncertainty as an example of non-calculative, non-rational, and social action (Denrell, 2003; DiMaggio and Powell, 1991), we view it as a kind of intelligent strategic action geared to utilizing existing knowledge at the segment or industry level (Nelson and Winter, 1982). From this perspective, Winter and Szulanski (2001) argue that the replication of past choices made by other firms contains a sound hypothesis about the causality of successful performances. DiMaggio and Powell (1983) also treat institutional isomorphism as a type of ‘collective rationality.’ It is similarly argued that rivals competing in the same industry tend to engage in a sort of collective decision making by monitoring and imitating one another (Guillen, 2002).

In this respect, we position our understanding of the firm in the tradition of the original bounded rationality thesis (March and Simon, 1958; Simon, 1955). The concept of bounded rationality is clearly different from that of irrationality or non-rationality. The initial definition of bounded rationality states that man is intendedly rational, but only limitedly so (March and Simon, 1958; Simon, 1955). That is, this influential view posits that although the consequences of organizational decisions are only partially or limitedly rational due to various factors such as ambiguous and unstable preferences, search costs, and imperfect computational capabilities, the organizational actor still does its best to make rational decisions at least at the intention level.

We argue that although a firm as a ‘boundedly rational’ actor may not be able to predict the future attractiveness or competitiveness of a segment perfectly on its own, it would still try to be as intendedly rational as possible, deriving further rational judgments from its interpretation of signals from the aggregated actions of rivals belonging to the same organizational field. In other words, we argue that a firm’s strategic responses to the aggregated actions of rivals as an uncertainty-reducing signal are more complicated, informed, and rational than is usually suggested in mimetic isomorphism and imitative behavior literatures (Denrell, 2003; DiMaggio and Powell, 1983).

2.4. Research model and hypotheses

We examine how signals from the organizational field of rivals will actually affect the focal firm’s segment exit decisions. This paper examines the following four types of behaviors that may reflect intended rationality in segment exit decisions: (1) estimating the competitiveness of a segment in a more refined way by taking into consideration the varying magnitude of competitive threat exerted by each rival in the aggregation of multiple rivals’ competitiveness; (2) making inferences about the attractiveness of a segment based on the performances of rivals operating in it; (3) giving a greater weight to recent actions by rivals than to older ones; and (4) adjusting sensitivity to influences from aggregated actions of rivals depending on the focal firm’s idiosyncratic conditions.

2.4.1. Aggregated competitive intensity and the competitiveness of the segment

The number of rivals in a market segment is probably one of the most frequently examined variables in the literatures of strategic management, industrial organization (IO) economics, and organizational ecology. Most existing studies in the literatures of strategic management and IO economics regard the number of rivals operating in a market segment or industry as a straightforward indicator of the strength of competition (Chen et al., 1992; Porter, 1980). That is, the larger the number of rivals operating in a segment, the more competitive the segment is. Since competition against other firms is regarded as the most important factor in exit decisions in existing strategy literature (D’Aveni, 1994; Porter, 1980), the general conception is that there is a positive relationship between the number of rivals operating in a segment and the likelihood of the focal firm’s exit from the segment.

In contrast, density dependence literature in the organizational ecology field usually presumes a curvilinear relationship between density and founding/mortality rates (Carroll and Hannan, 1989). However, recent organizational ecology studies which measured the scope of competition and legitimation in a much more refined way found a positive linear relationship between the number of rivals operating in the same segment, i.e. segment density rather than population density, and the level of competition in the segment (Dobrev and Kim, 2006). That is, regardless of disciplinary backgrounds, most existing studies predict a positive relationship between the number of rivals operating in a segment and the likelihood of the focal firm’s exit from the segment.

Although we concur that the number of rivals operating in a segment will positively affect subsequent segment exits, we extend this point and argue that even after controlling for the positive effect of segment density, the firm as an intendedly rational actor is likely to attempt to find a more refined and informed signal about the competitiveness of a segment from it’s the aggregated actions of rivals. We predict that the firm may pay attention to the heterogeneity among its rivals in terms of their impacts on the strength of competition when assessing the competitiveness of the segment. Most of the existing literature has implicitly reduced rival firms to a simple count number by treating all firms as if they are homogeneous in terms of their impacts on other firms. However, unless all firms in a segment have equal competitive impact on the others, the conventional measure of segment density that is based only on the number of rivals operating in a segment may not accurately indicate the strength of competition in the segment.

Given this, our study pays attention to competitive intensity literature, which suggests that larger firms have a stronger competitive effect on rivals than smaller ones owing to various advantages, such as scale and scope, resources and capability, and market control power (Barnett, 1997; Baum, 1995). Therefore, a firm as an intendedly rational actor may focus on the composition of segment density in terms of each firm’s competitive intensity, i.e. ‘the magnitude of the effect that an organization has on its rivals’ life chances’ (Barnett, 1997: 130), rather than to the simple number of its rivals. According to this argument, ceteris paribus, larger firms have a greater competitive intensity than smaller ones. Therefore, a more accurate indicator for the strength of competition in a segment would be segment density, i.e. the number of firms operating in the segment, weighted with each firm’s size, which this paper refers to as ‘aggregated competitive intensity’ at the segment level. Hence, we predict that segment exit decisions by the firm as an intendedly rational actor will be positively affected by aggregated competitive intensity, which is a more accurate indicator of competition at the segment level.

H1: The aggregated competitive intensity of a market segment will be positively associated with the likelihood of the focal firm’s subsequent exit from the segment.

2.4.2. Rivals’ performances as a signal of segment attractiveness

The previous hypothesis suggests that firms may find competitiveness signal based on aggregated competitive intensity. Then, where will a firm acquire signal for the attractiveness of a segment? We propose that segment attractiveness signal, on the other hand, will come from the performances of rivals operating in the segment.

In institutional theory literature of imitative behavior, it is suggested that not all firms in an organizational field will have the same degree of impact on the perceived legitimacy of an action (Haunschild and Miner, 1997). In the trait-based imitation model of institutional theory, the legitimacy of an action is affected more strongly by specifically which firms adopted it, rather than by simply how many firms did so. That is, imitation is often selective (Haunschild and Miner, 1997). Haveman (1993) argues that firms mainly imitate high-performing rivals, rather than indiscriminately imitating all rivals. The status-based imitation argument similarly suggests that organizations tend to imitate higher-status others (Fombrun and Shanley, 1990). Therefore, if many high-performing firms stay in a segment, the focal firm will imitate it and is also more likely to stay than exit.

That is, the performances of rivals operating in a segment may serve as a signal of the segment’s attractiveness. However, this study departs slightly from existing selective and trait-based imitation studies in that we factor in performance and size as different elements. In existing studies, performance and size have often been used interchangeably when defining ‘high-performing’ or ‘high-status’ others. For instance, Haunschild and Miner (1997) maintained that ‘legitimacy is inferred from traits like large size and success.’ Fombrun and Shanley (1990) similarly suggested that high-status organizations are usually large and successful. In their simulation study of imitative behavior, Mezias and Lant (1994) explicitly set up a decision rule in which the focal agent imitates large organizations. The main logic underlying these existing studies is that the high visibility of large firms induces imitation by others. For instance, Kraatz (1998) points out that firms tend to imitate especially those actions taken by ‘large or prominent’ firms, which are by nature highly visible.

However, we argue that, strictly speaking, large firms and high-performing firms exert different kinds of influences. As discussed when forming Hypothesis 1, since larger firms have a greater competitive intensity, they exert a stronger competitive threat than smaller ones do. Therefore, if there are many large rivals operating in a segment, firms are likely to exit the segment by interpreting this fact as a signal of the segment’s competitiveness.

Yet, we argue that the effects of visibility are not always consistent with those of legitimacy. For instance, although an obviously irrational action taken by a large firm would be still visible, it would not be likely to be imitated. That is, if large firms yield poor performances, their practices and actions are not likely to induce imitation by others, although they will still be visible. Following this, it can be inferred that if there were many poor-performing large firms in a market segment, their presence would not send a signal of the segment’s attractiveness. We argue that the firm as an intendedly rational actor is likely to look for a more informed and refined indicator of the segment’s attractiveness when making an exit decision, rather than simply imitating the choices of visible others.

Specifically, we argue that the focal firm is likely to take into consideration the actual performances of rivals in assessing the attractiveness of a segment. If many firms operating in a segment yield high performances, the segment is likely to be perceived as munificent and attractive by the focal firm (Haunschild and Miner, 1997). According to prominent debates on the comparison between firm effects and industry effects (McGahan and Porter, 1997; Rumelt, 1991), the performance of each firm is often determined by the characteristics of larger units or groups, such as segment or industry as a whole, as well as by the characteristics of individual firms. Therefore, unlike the size distribution of other firms, which indicates competitive intensity, the performance distribution of others operating in a segment is likely to signal the attractiveness of the segment.

Applying this argument to the strategic dynamics of interfirm influence in segment exit decisions, we may predict that if there are many high-performing but not necessarily large rivals in a market segment, the focal firm is likely to decide not to exit the segment, based on the attractiveness signal that comes from the performance distribution of its rivals. This line of argument is especially consistent with our depiction of a firm as an intendedly rational actor. Although imitative behaviors can be viewed as an indicator of bounded rationality, the fact that firms ‘follow the leader’ (Haveman, 1993) suggests that firms are at least intendedly rational enough to monitor, analyze, and take into consideration the performance distribution of rivals. Therefore, we predict that if a segment is crowded with high-performing rivals, it will be perceived to be attractive, which will in turn discourage the focal firm’ exit from the segment.

H2: The performance of rivals operating in a market segment will be negatively associated with the likelihood of the focal firm’s subsequent exit from the segment.

2.4.3. Exit bandwagon as a signal of ongoing segment deterioration

Most existing literatures from strategic management, IO economics, organizational ecology, and institutional theory pay special attention to the number of firms which have a common characteristic at a given time point, either as an indicator of competition or legitimacy (Carroll and Hannan, 1989; Haunschild and Miner, 1997; Porter, 1980). When applying the same reasoning to the assessment of segment attractiveness in exit decisions, the total number of firms operating in a certain market segment, i.e. the density, at a particular point of time will be regarded as an effective measure of either the segment’s competitiveness or its attractiveness. As a matter of fact, this is exactly how most existing studies operationalize the degree of competition or legitimacy (Hannan and Freeman, 1989; Tolbert and Zucker, 1983).

However, it is important to distinguish between the density of a certain action and the density of the outcome of that action at each point of time. For instance, it should be noted that the number of firms operating in a segment at a certain time point may not exactly indicate the current attractiveness of the segment, as this information measures the accumulated outcome of numerous past entries into and exits from the segment, many of which may have occurred long before. Therefore, the firm as an intendedly rational actor which has to make a choice between staying and exiting is likely to search for a more accurate signal to assess the attractiveness of the segment than the simple number of firms operating in the segment at the time.

It should also be noted that although the outcome of each entry (or exit) is counted equally in cumulative density, a recent entry/exit and a past entry/exit may exert different influences on the focal firm’s decision. Two cumulative densities with exactly the same value may have entirely different implications, when considering the time distribution of events in each of the two cumulative densities. That is, the focal decision maker may receive one set of signals from a cumulative density in which events have been spread widely over a long period of time and another set from a cumulative density in which events have been concentrated in a recent period.

Therefore, we may infer that despite the limits of rationality in decision making, the firm will still try to be as intendedly rational as possible by taking into consideration the contemporariness of relevant events. A recent event is likely to be taken more seriously by the focal firm than an event that occurred long before, not only because the effect of a recent event is more vividly remembered, but also because its context is perceived to be ongoing and to remain valid (Barnett et al., 2003; Belderbos et al., 2011). The literature of behavioral decision making also suggests that an actor tends to give more weight to recent events than to past events (Tversky and Kahneman, 1986). Immediacy plays an especially important role in the perception of infrequent events (Henderson and Cool, 2003), which include segment exit. The lesson learned from a segment exit event that took place in the distant past is likely to be subject to serious depreciation by the time of the focal firm’s decision making (Henderson and Cool, 2003; Tversky and Kahneman, 1986).

As an attempt to include this effect in our research model, we pay particular attention to the unique dynamics of the boom or wave of segment exits that have occurred recently, which this paper refers to as the exit bandwagon. In strategic management literature, it is suggested that bandwagon pressures affect various strategic decisions, such as diversification (Fligstein, 1991; Haveman, 1993), the adoption of technological or organizational innovations (Chiles et al., 2004; Suárez and Utterback, 1995), globalization and foreign market entry (Henisz and Delios, 2001), and product market expansion (Henderson and Cool, 2003).

The concept of a bandwagon captures the idea of a recent dynamic process in which a large number of firms have suddenly engaged in a certain activity that has become recently popular within a short period (Abrahamson and Rosenkopf, 1993; Barnett et al., 2003; Xia et al., 2008). In other words, the concept of a bandwagon, also referred also as a ‘boom’ or ‘wave,’ highlights the dynamic aspect of pressure from currently ongoing collective actions taken by other firms (Barnett et al., 2003). Therefore, the focal firm as an intendedly rational actor may take the signal stemming from recent actions taken by other firms, such as an exit bandwagon, particularly seriously.

In terms of the effect on segment exit decisions, an exit bandwagon in a market segment may send a special kind of signal about the ongoing deterioration of the segment. In such cases, the focal firm will be pressured to escape from the rapidly deteriorating segment by jumping on the bandwagon of segment exit as quickly as possible. Therefore, we predict that even after controlling for the effects of cumulative density, the number of rivals who have recently exited from a market segment will positively affect the likelihood of subsequent exit from the segment by the focal firm.

H3: The magnitude of a recent exit bandwagon from a market segment will be positively associated with the likelihood of the focal firm’s subsequent exit from the segment.

2.4.4. Interaction between an exit bandwagon and the focal firm’s performance

An exit bandwagon in a segment may affect each of the firms operating in the segment differently, since the firm as an intendedly rational actor is likely to take into consideration not only the heterogeneity among other firms in terms of competitive intensity, but also its own idiosyncratic situation and characteristics. A number of existing studies in strategic management literature suggest that high-performing firms are less affected by bandwagon pressures than low performers. For instance, Gilbert and Lieberman (1987) showed that, compared to larger firms, smaller firms were more likely to jump on the bandwagon of size expansion due to their disadvantages in gathering and interpreting information. Buzzell et al. (1975) argued that firms with large market shares are less influenced by rivals’ moves, since they possess substantial power to influence the conditions of competition.

In exit decisions, high-performing firms are also less likely than low-performing firms to jump blindly on the bandwagon of segment exit without carefully looking into their own unique situations. Furthermore, since poor performances often push firms to seek clues that may be helpful in overcoming performance problems and to consider the execution of new strategic options, including exits from existing segments or industries (Chang, 1996; Karakaya, 2000; March and Simon, 1958), lower performers are likely to be affected more strongly by an exit bandwagon than higher performers. On the other hand, if a firm is confident about its future performance in a certain segment, it will be less influenced by an exit bandwagon when making an exit decision.

That is, we argue that although never perfectly rational, exit decisions by the focal firm are at least intendedly rational enough to take into consideration their own idiosyncratic conditions. Therefore, we can predict a negative moderating effect of the focal firm’s performance on the relationship between the presence of an exit bandwagon and the likelihood of the firm’s subsequent segment exit.

H4: The positive effect of an exit bandwagon in a market segment on the likelihood of the focal firm’s subsequent exit from the segment will be greater for low-performing firms than for high-performing firms.

Figure 1 shows our research model that summarizes our discussion of the relationships between aggregated actions of rivals belonging to the same organizational field and high-tech ventures’ segment exits. Our research model is composed of three main-effect hypotheses and one interaction hypothesis.

Figure 1.

Research model.

3. Empirical setting

To test our hypotheses empirically, we chose high-tech venture firms to examine the dynamics of segment exits, since they tend to change their niche compositions more frequently than established firms (Katila and Shane, 2005). Delacroix and Solt (1988) submit that ventures often perform better than large established firms in radical organizational change, since they are relatively free from structural inertia. A venture firm’s prototypical reaction to the increase of competition in an existing market segment is to escape flexibly from the worsening situation (Anand and Singh, 1997), rather than to stick rigidly to the existing segment (Lévesque and Shepherd, 2004).

Our empirical study specifically examines the conditions that may affect segment exits by high-tech ventures in the Korean systems integration (SI) industry. The SI industry of Korea was born in the mid-1990s. Owing significantly to the government’s aggressive policies to nurture high-tech ventures as a new economic engine of Korea and also to the worldwide IT bubble in the late 1990s, the Korean SI industry grew rapidly (Shin et al., 2014). Moreover, because even entrepreneurs with little capital could easily start their own ventures as long as they possessed needed technologies during the period, a large number of entrepreneurs with technology backgrounds founded their own IT ventures and fiercely competed against one another. As a consequence, ecological dynamics such as founding and disbanding were extremely volatile in the Korean SI industry (Shin et al., 2014). Furthermore, most Chaebols founded in-house SI companies to satisfy the IT needs of their member companies. As a consequence, both segment entries and exits have been highly active in the Korean SI industry.

4. Methods

4.1. Data and measurement

We coded the data from The Annual Directory of Korean Systems Integration Companies published by the Korean Association of Systems Integration Companies. Although the first issue of the directory was published in 1994, it did not provide information on market segments until 1998. Thus, the dataset actually used in our analysis goes back only to 1999.

As this study focuses on exit decisions by high-tech ventures, we separated SI ventures from large established SI firms for the dependent variable side. Thus, the number of our sample only covers exit decisions by SI ventures. To implement this procedure, out of the 415 SI firms mentioned in SI industry directories 1999 to 2006, we sorted out non-venture SI firms, such as Chaebol companies or MNC subsidiaries for the dependent variable side. However, in the calculation of the independent and control variables, we included those non-venture SI firms, because exit decisions by ventures are likely to be influenced not only by other ventures, but also by large established firms. After excluding observations with missing values and applying a one-year lagging for the independent and control variables, the number of the SI ventures in the dependent variable side of our data matrix from 2000 to 2006 was 199. That is, the empirical study of this paper analyzes exit decisions by these 199 SI ventures, which may be influenced not only by actions of other SI ventures, but also by those of large SI firms.

Since this study estimates the proportional hazard of a firm’s exit from a market segment at each time point, our level of analysis is a firm-segment-year. That is, our N is calculated as “Number of Firms (199 in total, but varies with years due to new founding or disbanding) X Number of Segments (39 in total, but varies with years) X Number of Years (7 years).” In this equation, although the number of years (7) is fixed, the number of segments (39) varies with years since we include only those segments where at least one SI venture operates in each given year. That is, although there are 39 market segments in the official categorization scheme provided by the Korean Association of Systems Integration Companies, we included only those segments in which at least one of the 199 venture firms operated in the focal year in the risk set of our statistical analysis. If all SI ventures exit from a certain segment in a certain year, that particular segment is not counted as N anymore after the time point. The number of firms also varies by year due to the founding of new firms and the disbanding of existing firms each year, which is a typical characteristic of unbalanced panel data. The result of this calculation, i.e. the total number of firm-segment-years, was 3097, which was N for our statistical estimation.

The variables included in our research model were measured in the following ways.

4.1.1. Dependent and independent variables

4.1.1.1. Segment exit

We measured our dependent variable by examining whether the focal firm exited the focal market segment during the focal time spell. According to the standard classification scheme of SI market segments provided in The Annual Directory of Korean Systems Integration Companies, the Korean SI industry has 39 different market segments. We coded the ‘exit’ variable from information that each SI firm declares as its business areas every year. That is, this declaration by each firm means that whether or not it actually has a customer in a certain segment each year, it still wants and is capable of doing business in that particular segment. On the other hand, if a firm suddenly drops a particular segment from its declared business areas in a certain year, we regarded such event as the firm exit from the segment in that year.

4.1.1.2. Aggregated competitive intensity at the segment level

To measure this independent variable, we first calculated the size-weighted density (by sales) of each segment in which the focal venture operated following the methodological practice in the mass dependence literature of organizational ecology (Barnett, 1997; Barnett and Amburgey, 1990). However, instead of putting ‘aggregated competitive intensity’ and ‘segment density’ together in the model, we constructed a composite measure by combining these two variables which are highly correlated with each other. For this purpose, we divided the former by the latter. However, the value of this new variable was still too large compared to other variables, so we divided this value by 10,000,000. Because this value was also skewed, we used the log-transformed value of this variable as the measure of aggregated competitive intensity in our statistical estimation.

4.1.1.3. Segment attractiveness

We measured the attractiveness of a market segment by the average ROS of all firms operating in the segment at time point t-1. We had to use ROS instead of the more frequently used ROA due to the problem of limited data availability.

4.1.1.4. Exit bandwagon

Our third independent variable, exit bandwagon, was measured by the count of recent exit events by other firms in each market segment. We calculated this variable by counting the total number of firms which had exited a segment during a one year period between time point t-1 and time point t-2, again following the methodological practice found in organizational ecology (Barnett et al., 2003).

4.1.2. Control variables

4.1.2.1. Population density

To control for the potential effects of population-level dynamics, we included the density of the entire Korean SI companies in our models. We measured population density by finding the total number of SI firms, including member companies of Chaebol groups and subsidiaries of MNCs as well as SI ventures in Korea each year.

4.1.2.2. Societal-level discourse

We controlled for potential effects of societal-level discourse about the SI industry. We measured this variable with the annual frequency of news articles about the SI industry in all the daily newspapers published in Korea. Since this particular variable indicates the magnitude of nation-wide interest in the SI industry, it may affect the focal SI venture’s perception of the industry’s attractiveness, which in turn may influence its segment exit decision.

4.1.2.3. Firm age

Firm age was measured by the number of years that had elapsed since the focal venture’s founding until time t.

4.1.2.4. Firm size

We also controlled for the potential effects of firm size for similar reasons. While larger firms may have advantages in competition and survival compared to smaller firms, they may at the same time have disadvantages in change due to structural inertia (Hannan and Freeman, 1989). We measured firm size in the following two ways. First, we measured firm size by their annual sales, following existing studies in strategic management literature (Brush et al., 2000; Gomez-Mejia, 1992). However, since the distribution of this value was skewed, we used log-transformed values for our estimation. Second, we also measured firm size by the number of full-time employees. Since systems integration is a type of professional service industry, the most important strategic resource for SI firms is their human resources (Barney, 1991). Human resources may serve as a key source of internal capabilities required for the exploration of new niches. We measured this variable with the number of employees working for the focal firm on a full-time employment basis. Since the distribution of this value was skewed, we log transformed it in the estimation of our model.

4.1.2.5. Firm performance

We measured the performance of a firm through ROS at time point t-1. As already explained, we had to rely on ROS as a performance measure due to the problem of limited data availability.

4.1.2.6. Strategic alliances

We controlled for the potential effects of strategic alliances, since alliances may provide a buffer to competitive threats and segment-specific risks (Dyer and Singh, 1998). We measured this variable as the total number of alliances that the focal venture has at the end of each year.

4.1.2.7. Entry bandwagon

We controlled for the potential effect of new entries to each segment. Exits and entries often occur simultaneously in certain segments. As our research model implies, exit and entry decisions are affected not only by segment-level factors, but also by firm-level factors. Since a certain segment that is perceived to be deteriorating by a firm can be evaluated to be attractive by another firm, exits by certain firms from and entries by other firms into the same segment can happen simultaneously. Thus, we included an ‘entry bandwagon’ variable in our model to control for the potential effect of new entries into the focal segment. This variable was measured by the count of recent entry events by other firms in each market segment. We calculated this variable by counting the total number of firms which had entered a segment during a one year period between time point t-1 and time point t-2.

4.1.2.8. Niche width

This control variable was measured by the number of segments in which the focal venture operated in each year (Carroll, 1985).

4.2. Statistical analyses and results

Since the current research was designed as a longitudinal study that could estimate the proportional hazard rate of each firm’s exit from each segment at each point of time, an event-history model was employed to analyze the segment-exit rates of Korean SI ventures. The segment-exit rate is here the instantaneous rate of a firm’s exit from a segment in which it had operated during the previous period. The unit of analysis for segment-exit rate analyses is firm-segment-year, and a firm’s tenure in a segment serves as the clock to measure duration. The segment-exit rate is predicted by regression models that include various firm-level, segment-level and industry-level covariates. Among the various possible duration-dependence models in event history analysis, the current study used Cox proportional hazard model, which does not have an assumption about the distribution of durations (Blossfeld and Rohwer, 1995). All standard treatments used in event history analysis were also applied.

While Table 1 summarizes descriptive statistics and correlations, Table 2 summarizes the results of event history analysis.

Table 1.

Descriptive statistics and Pearson correlation coefficients.

Variables	Mean	SD	Min	Max	1	2	3	4	5	6	7	8	9	10	11	12
l. Population density	178.0634	20.8082	120	202	1
2. Societal-level discourse	2513.2130	719.6968	881	3454	0.4203 ***	1
3. Firm age	9.9530	5.6800	1	25	0.0072	−0.181 ***	1
4. Firm sales (log)	8.5569	1.4900	2.5649	12.4935	−0.1001 ***	−0.0943 ***	0.4326 ***	1
5. No. of employees (log)	3.2821	1.0644	0.9031	6.1203	0.1539 ***	0.0576 ***	0.2791 ***	0.4549 ***	1
6. Firm performance	0.0895	0.1575	−2.7473	1	0.0987 ***	0.217 ***	−0.2194 ***	−0.265 ***	−0.0689 ***	1
7. Strategic alliances	0.9574	1.4351	0	7	0.0582 ***	0.0561 ***	−0.077 ***	0.1555 ***	0.0813 ***	0.0157	1
8. Entry bandwagon	8.4855	10.5694	0	44	0.1015 ***	0.7231 ***	−0.1077 ***	−0.0901 ***	0.0496 ***	0.1626 ^***	0.0055	1
9. Niche width	9.2231	6.1374	1	28	0.1721 ***	0.0189	0.195 ***	0.2963 ***	0.1479 ***	−0.0289	−0.0536 ^***	0.0268 ^*	1
10. Agrregated competitive intensity (log)	3.3411	1.0120	−16.1181	5.6788	0.1651 ***	0.0453 ***	0.111 ***	0.0351 **	−0.0445 ***	−0.0681 ^***	−0.0227	0.0715 ^***	0.2246 ^***	1
11. Segment attractiveness	−0.0617	0.1033	−0.9206	0.1977	−0.2107 ***	−0.2514 ***	−0.0229	0.0317 ^*	0.0474 ^***	−0.0387 ^**	−0.0068	−0.247 ^***	−0.0954 ^***	−0.1167 ^***	1
12. Exit bandwagon	1.8639	1.9926	0	9	0.0267 *	0.48 ***	−0.1197 ***	−0.0885 ***	0.0142	0.0965 ^***	0.0166	0.4153 ^***	−0.0462 ^***	−0.0265 ^*	−0.0726 ^***	1

N=3097.

p<0.1, **p<0.05, ***p<0.01.

Table 2.

Results of Cox proportional hazard model.

	Model 1		Model 2		Model 3	Model 4	Model 5		Model 6		Model 7		Model 8		Model 9
Control variables
Population density	0.0124	***	0.0116	***	0.0123 ^***	0.0149 ***	0.0116	***	0.0142	***	0.0150	***	0.0143	***	0.0141	***
	(0.003)		(0.003)		(0.003)	(0.0031)	(0.003)		(0.0031)		(0.0032)		(0.0031)		(0.0031)
Societal-level discourse	−0.0002	*	−0.0002	**	−0.0003 ^**	−0.0004 ^***	−0.0003	**	−0.0004	***	−0.0004	***	−0.0005	***	−0.0005	***
	(0.0001)		(0.0001)		(0.0001)	(0.0001)	(0.0001)		(0.0001)		(0.0001)		(0.0001)		(0.0001)
Firm age	0.0018		0.0011		0.0012	0.0023	0.0004		0.0016		0.0016		0.0007		0.0005
	(0.007)		(0.007)		(0.007)	(0.007)	(0.007)		(0.007)		(0.007)		(0.007)		(0.007)
Firm sales (log)	−0.1345		−0.1402		−0.1373 ^***	−0.1255 ^***	−0.1428		−0.1311		−0.1283		−0.1339		−0.1314
	(0.0269)		(0.027)		(0.0269)	(0.027)	(0.027)		(0.0271)		(0.027)		(0.0271)		(0.0274)
No. of employees (log)	−0.0061		0.0034		0.0029	−0.0172	0.0130		−0.0079		−0.0078		0.0023		−0.0005
	(0.0407)		(0.0407)		(0.0409)	(0.0409)	(0.0409)		(0.0409)		(0.041)		(0.0411)		(0.0411)
Firm performance	−0.1275		−0.1173		−0.1174	−0.1059	−0.1066		−0.0958		−0.0925		−0.0816		0.2668
	(0.1687)		(0.1711)		(0.1694)	(0.1727)	(0.1719)		(0.1755)		(0.1738)		(0.1767)		(0.2747)
Strategic alliances	−0.0105		−0.0080		−0.0105	−0.0109	−0.0081		−0.0082		−0.0110		−0.0084		−0.0090
	(0.0251)		(0.0251)		(0.0251)	(0.0252)	(0.0251)		(0.0251)		(0.0252)		(0.0251)		(0.0252)
Entry bandwagon	0.0127	*	0.0129		0.0136 ^**	0.0136 ^**	0.0138		0.0141		0.0149	**	0.0154		0.0149	**
	(0.0066)		(0.0066)		(0.0066)	(0.0068)	(0.0066)		(0.0068)		(0.0068)		(0.0069)		(0.0068)
Niche width	0.0117	*	0.0097		0.0119 ^*	0.0120 ^*	0.0098		0.0099		0.0122	*	0.0100		0.0108	*
	(0.0063)		(0.0063)		(0.0063)	(0.0063)	(0.0063)		(0.0064)		(0.0063)		(0.0063)		(0.0064)
Independent variables
Aggregated competitive intensity (log)			0.0945				0.0963		0.1029				0.1055		0.1061	***
Aggregated competitive intensity (log)			(0.0331)				(0.0332)		(0.0339)				(0.0341)		(0.0341)
Segment attractiveness					−0.8707 ^**		−0.8949				−0.9716	***	−1.0045		−0.9852
					(0.3449)		(0.3474)				(0.3468)		(0.3497)		(0.3495)
Exit bandwagon						0.0780 ^***			0.0826		0.0816		0.0864		0.0983
						(0.0203)			(0.0205)		(0.0202)		(0.0205)		(0.0212)
Exit bandwagon * firm performance															−0.1468
															(0.0734)
Log-likelihood	−6012.134	***	−6007.213	***	−6009.192 ^***	−6004.984 ***	−6004.155	***	−5999.395	***	−6001.387	***	−5995.621	***	−5993.67	***
Likelihood ratio test			9.8414		5.8834	14.3006	6.1172		15.6372		15.6102		17.0682		3.901

N=3097.

Values in parentheses are standard errors.

p<0.1, **p<0.05, ***p<0.01.

All likelihood-ratio data are significant at the α= 0.05 level.

Table 1 shows the descriptiwve statistics and correlations of the variables included in the statistical estimation of our model. Among the correlations, the two measures of firm size, sales and number of employees, have a positive correlation with each other, which is a typical characteristic of a labor-intensive field like the SI industry. Firm size measured with sales also has a positive correlation with niche width measured as the number of segments where the focal venture operates, which is also logically straightforward in that the operation of a larger number of businesses is likely to result in bigger sales. Another positive correlation is observed between exit bandwagon and entry bandwagon. We may explain this particular correlation as follows. Potential entrants may have interpreted the exit of incumbents from a segment as a creation of new opportunities stemming from reduced competition. For instance, in his work on chains of opportunities, White (1970) insightfully suggested that the exit of an incumbent from a position creates a new opportunity for potential entrant.

Model 1 of Table 2 is the baseline model that includes only control variables. Among the industry-level control variables, the result of Model 1 suggests that population density positively affects segment exit. The positive effect of population density remains significant in all of the nested combinations and the full model. This result suggests that segment exits are affected not only by segment-level competition, but also by industry-level competition. Also according to the result, societal-level discourse about the IT industry turned out to affect segment exit negatively. This result suggests that societal-level discourse about the SI industry as an indicator of nation-wide interest in this particular industry is likely to reduce exits in general regardless of segments by increasing the perceived attractiveness of the entire industry. That is, this result implies that a firm’s strategic action like segment exit is influenced by social factors like discourse as well as by economic and managerial factors.

Among the firm-level control variables, firm age was not significant. This result is inconsistent with the liability of newness argument in the organizational ecology literature (Hannan and Freeman, 1989). We may attribute this particular result to the uniqueness of our sample composed mostly of young high-tech ventures. Since most of the sample firms are young ventures with the average age of 9.9 years, the variable of age itself may not have enough variations that may make this variable statistically significant.

We included two control variables that capture effects of firm size: sales and number of employees. Firm size measured by sales turned out to affect segment exit negatively. This result can be interpreted that abundant resources of relatively larger firms played a role as a buffer against environmental threats. However, the number of employees was not significant. We may attribute this result to the idiosyncratic nature of HR composition in high-tech venture firms. Unlike large manufacturing firms where the number of employees stands for their capacity of scale economy, high-tech ventures often depend heavily on a very small number of talented engineers who are usually founders. Therefore, the simple size of employees may not have a significant meaning in high-tech ventures’ strategic capability.

The result that niche width positively affects segment exit may indicate the limit of firm size growth suggested by Williamson (1985). Another firm-level control variable which did not turn out to be significant was firm performance. This particular result may not be consistent with the prediction of the problematic search argument of organizational learning (Cyert and March, 1963). We may attribute this result to the limit of our data. Because information on segment-level performance was unavailable, this control variable was measured with the annual performance of the firm which usually operates over multiple segments with varying performances.

We also included entry bandwagon as a segment-level control variable to control for potential effects of new entries. However, this particular control variable was not significant. We may interpret this particular result that segment exit and segment entry are separate processes relatively independent from each other, rather than are mirror images of each other.

Models 2, 3, and 4 show the results of analysis in which we entered each independent variable separately. Models 5, 6, and 7 demonstrate the robustness of our results by testing various combinations of variables in nested models. Model 8 includes the control variables as well as all the independent variables. Finally, Model 9 is the full model that includes the interaction term, as well as all the independent and control variables.

Models 2, 5, 6, 8, and 9 show the positive effect of aggregated competitive intensity on segment exit. As predicted in Hypothesis 1, the aggregated competitive intensity of a market segment was shown to have a positive relationship with the likelihood of the focal firm’s exit from the segment. The results of Models 3, 5, 7, 8, and 9 indicate that segment attractiveness has a negative effect on segment exit as was predicted in Hypothesis 2. Models 4, 6, 7, 8, and 9 show the positive effect of the presence of an exit bandwagon on segment exit. Finally, Model 9 shows that the interaction between an exit bandwagon and firm performance is significant in the negative direction, as was predicted in Hypothesis 4. Thus, as we can clearly see from the results of event history analysis and effect size analysis, all of our four hypotheses were strongly supported.

Related to the measurement of variables, a coding scheme that merits further examination is our dependent variable, because an event coded as ‘exit’ may have been caused by the lack of demand in a particular segment, rather than by the focal firm’s voluntary decision to exit from it. However, if other firms still operate in a segment after the focal firm exits it, we may assume that there is still considerable demand in that particular segment. Then, the event we coded as ‘exit’ can be regarded as a consequence of the focal firm’s exit decision. For this purpose, we examined the yearly trend of ‘exits,’ ‘entries,’ and ‘stays’ in three of the most representative SI segments, which is summarized in the Appendix. As clearly shown in the table, even after many firms exit a segment, a considerable number of other firms still remain in the segment.

Another statistical issue that deserves further discussion is the effect of sample size. The statistics literature suggests that since large sample size tends to exaggerate the statistical power of a model (Howell, 2010), simple p-value statistics are often insufficient for large sample data. Thus, we examined whether the significances of our results are substantive from a more rigorous statistical criterion. According to the statistics literature, the robustness of effect size across different models can mitigate this potential limit of large sample data (Cohen, 1992; Kraemer and Theimann, 1987). Therefore, we conducted effect size analyses using a likelihood ratio test, which is suitable for effect size analysis between different models in Cox proportional hazard model used for our estimation. Unlike OLS regressions where an adjusted R square is usually used, the effect-size analysis of Cox proportional hazard model requires a likelihood ratio test. The results of the test showed that all likelihood-ratio data are significant at α=0.05. We included likelihood ratio values in Table 2.

5. Conclusion

This paper attempted to integrate the rational actor model of strategic management and the social actor model of institutional theory by examining how a firm’s exit from an existing market segment is affected by the aggregated actions of rivals belonging to the same organizational field from an intended rationality perspective. We argued that the aggregated actions of rivals significantly influence the exit decisions of the focal firm by providing signals about the competitiveness and attractiveness of the segment as suggested by institutional theory. However, we argued that while exit decisions by firms are not perfectly rational due to various constraints such as high uncertainty and limits of rationality, they are still far more complicated, informed, and intendedly rational than is usually presumed in institutional theory (March and Simon, 1958). Our findings empirically confirmed this argument. We have found that the competitiveness of a segment signaled by the aggregated competitive intensity of rivals has a positive relationship with exit decisions, the attractiveness of the segment evaluated by the performances of rivals operating in the segment has a negative relationship with exit decisions, and a recent exit bandwagon has a positive relationship with exit decisions. We also found that the performance of the focal firm reduces the positive effect of the exit bandwagon on the likelihood of subsequent segment exit.

The findings of this paper may have following contributions and implications to the future development of management studies.

First, this paper demonstrated how the rational actor model of strategy and economics and the social actor model of institutional theory can be integrated with each other in a mutually fruitful way. To accomplish this daunting task, we returned to the original definition of bounded rationality thesis which distinguished between intention and consequence in the level of rationality (March and Simon, 1958). The original formulation of this influential concept which serves institutional theory as the key behavioral assumption (DiMaggio and Powell, 1991) clearly states that while organizational actions are rational at the intention stage, their actual consequences are only limitedly rational (March and Simon, 1958). In this regard, this paper paid attention to and empirically showed specifically how intended rationality is acted out in strategic decision making related to segment exit. Thus, the findings of this paper suggest that a balanced stance that can integrate the two competing views of organizational rationality is always called for in management studies. As clearly exemplified in the hegemonic rational choice argument (Becker, 1976; Bernheim, 1984), mainstream scholars in most social sciences used to view organizational decision making as a rational process geared to the maximization of self-interests based on economic calculation (Batson, 1998; Miller, 1999). Yet more recent approaches, as suggested by the literature of institutional theory (DiMaggio and Powell, 1983; Meyer and Rowan, 1977), highlight the non-rational and social nature of organizational action, such as the ways in which firms conform to social pressure from others, imitate others to deal with uncertainty, or take certain forms and practices for granted regardless of their technical efficiency. The findings of this paper clearly show that the two competing conceptions of organizational rationality are not necessarily contradictory with each other, but can be fruitfully integrated with each other. Thus, we call for management scholars’ conscious endeavor to strike a balance between competing images of the organizational actor.

Second, the findings of this paper have enriched our understanding of signaling dynamics by highlighting the social nature of signaling and also by vividly illustrating dual signaling effects of rivals’ actions. Unlike the original signaling literature in economics which focused mainly on how an economic actor may overcome information asymmetry in markets from the focal actor’s perspective (Spence, 1973), this paper has shifted the focus of attention to signaling effects of other firms’ actions by examining the aggregated actions of rivals as a source of uncertainty-reducing signal. That is, our focus in this paper is not an individual firm’s signaling strategy, but the social sources of signaling from other firms. This paper further highlighted the social aspect of signaling by arguing that the aggregated actions of multiple rivals at the field level, rather than actions of an individual rival firm, collectively affect the focal firm’s exit decision. Moreover, this paper underlined the dual signaling effects of rivals’ actions by showing that the aggregated actions of rivals belonging to the same organizational field may serve the focal firm as an important source of signal not only about the risk of each market segment, but also about its attractiveness. That is, neither like strategic management that has treated rivals mainly as threat nor like institutional theory that have viewed rivals mostly as an indicator of legitimacy, this paper viewed rivals as a source of multifaceted signals. The findings of this paper show that either as a signal of competitive threat or as a signal of attractiveness, the aggregated actions of rivals serve as an important source of signals in exit decisions made under uncertainty. In fact, rival firms are likely to play both roles simultaneously, since an actor plays multiple roles in the network of complex interfirm relations as White et al. (1976) suggested. By further extending this line of inquiry, future studies may systematically compare the multifaceted roles of rivals acting as sources of actual competitive threats and sources of signals.

Third, the findings of this paper demonstrate that exit is a highly intriguing and promising future research area of management studies. Regardless of theoretical backgrounds, existing management studies have paid far greater attention to entry than to exit. In most of the existing literatures, it has been taken for granted that the dynamics of exit will be a mirror image of entry. This paper contributed to understanding of this underexplored issue by pointing out and empirically showing that although exit decisions are subject to high uncertainty, firms still strive to be as rational as possible at least at the intention level. Moreover, contradicting the widespread perception that the dynamics of exit are a mirror image of those of entry (Brauer, 2006), these two types of strategic actions may in fact be governed by different decision rules. Hence, systematically comparing the underlying dynamics of exits with those of entries can be a highly intriguing academic inquiry. Thus, we call for management scholars’ attention to the issue of exit decision, which has been largely ignored in the field of management studies (Brauer, 2006).

Fourth, the findings of this study may have shed new light on studies on bandwagon phenomena by underlining each firm’s differing response to the same bandwagon pressure. Although there have been a considerable number of studies on bandwagon behavior in the field of management studies (Abrahamson and Rosenkopf, 1993; Barnett et al., 2003; Xia et al., 2008), those existing studies have paid scant attention to the possibility that the strength of the bandwagon pressure may vary depending on each firm’s idiosyncratic characteristics or situations. By contrast, the result of Hypothesis 4 in this paper suggests that each firm’s performance significantly moderates the effect of bandwagon pressure in its exit decision. In future studies, similar moderating effects from a variety of other factors can be widely explored and tested, such as network position, alliance relation, resource endowment, competency composition, core technology, business portfolio, and niche composition.

Fifth, this paper may contribute to the understanding of new economic dynamics in emerging-market countries. The conventional image of emerging-market economies in the global academic community tends to have been preoccupied with the crucial role of a few large firms such as Korean Chaebols. By contrast, this paper has shifted attention to the strategic dynamics of high-tech ventures that may significantly substitute or complement Chaebol groups as a new growth engine of Korea in the 21^st century (Shin et al., 2014). However, the flipside of this new focus on high-tech ventures by this paper is the possible limit of generalizability. Therefore, to see whether the findings of this paper can be generalized beyond the organizational field of high-tech ventures, future studies will need to systematically compare strategic dynamics of high-tech ventures and those of large established firms.

On the other hand, this paper may suffer from the following shortcomings that need to be solved by future studies.

First, one should take caution in generalizing the findings of this study beyond the current research context. Our results might have been affected by the idiosyncratic conditions of the Korean SI industry during the early 2000s. For instance, the environment of global IT industries around this time was exceptionally volatile owing to the IT bubble (Ljungqvist and Wilhelm, 2003). Moreover, the Korean government at this point was investing a huge amount of funding to the IT industries: the central economic policy of the government during the period was to nurture high-tech ventures that could compete with the large Chaebols that had dominated the entire economy until that point. These idiosyncratic environmental conditions may have affected the results of our empirical analysis. To examine this possibility, a systematic cross-cultural study that addresses different national contexts at different time periods is called for.

Second, the empirical part of this paper suffers from the limited availability of data. This data problem can be attributed at least partially to the life-cycle of the Korean SI industry at that time. A significant proportion of the SI companies that comprised the Korean SI industry at that time were young and small entrepreneurial ventures that did not have full-fledged management systems. For instance, we could not use ROA as our performance measure because many ventures were reluctant to disclose the necessary information. Thus, above all, future studies will have to make efforts to build a comprehensive dataset that contains rich and complete information about this highly interesting and dynamic industry.

Another limitation of this paper is the unavailability of segment-level data. Since segment exit, which is our dependent variable, can be affected by the decrease of demand or the worsening of economic conditions as well as by the independent and control variables that we included in the research model, it is important to control for the potential effects of segment-level economic factors, such as market size, segment munificence, or segment attractiveness. Therefore, future studies on this theme need to make a special effort to collect complete data on each segment.

Footnotes

Appendix

Table A1.

Frequencies of exits, entries, and stays in 3 SI segments.

		2001	2002	2003	2004	2005
Finance information system	Stay	16	16	27	35	41
	Entry	2	17	10	8	0
	Exit	0	2	6	2	2
Groupware/electronic document management system (EDMS)	Stay	10	9	10	46	45
	Entry	3	6	38	1	0
	Exit	3	4	5	2	2
Multi-media/web homepage/web solutions	Stay	2	7	38	74	75
	Entry	7	40	36	3	0
	Exit	0	2	9	0	2

Final transcript accepted 26 February 2015 by Peter Liesch (AE Strategy and International Business).

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Abrahamson

Rosenkopf

(1993) Institutional and competitive bandwagons: Using mathematical modeling as a tool to explore innovation diffusion. Academy of Management Review 18: 487–517.

Anand

Singh

(1997) Asset redeployment, acquisitions and corporate strategy in declining industries. Strategic Management Journal 18: 99–118.

Barnett

(1997) The dynamics of competitive intensity. Administrative Science Quarterly 42: 128–160.

Barnett

Amburgey

(1990) Do larger organizations generate stronger competition? In: Singh

(ed.) Organizational Evolution: New Directions. Newbury Park, CA: SAGE, pp.78–102.

Barnett

Swanson

Sorenson

(2003) Asymmetric selection among organizations. Industrial and Corporate Change 12: 673–695.

Barney

(1991) Firm resources and sustained competitive advantage. Journal of Management 17: 99–120.

Batson

(1998) Altruism and prosocial behavior. In: Gilbert

Fiske

Lindzey

(eds) The Handbook of Social Psychology. New York: McGraw-Hill, pp.282–316.

Baum

(1995) The changing basis of competition in organizational populations: The Manhattan hotel industry, 1898-1980. Social Forces 74: 177–204.

Baum

Usher

(2000) Making the next move: How experiential and vicarious learning shape the location of chains’ acquisition. Administrative Science Quarterly 45: 766–801.

10.

Becker

(1976) The Economic Approach to Human Behavior. Chicago, IL: The University of Chicago Press.

11.

Belderbos

Olffen

Zou

(2011) Generic and specific social learning mechanisms in foreign entry location choice. Strategic Management Journal 32: 1309–1330.

12.

Bergh

(1997) Predicting divestiture of unrelated acquisitions: An integrative model of ex ante conditions. Strategic Management Journal 18: 715–731.

13.

Bergh

Lawless

(1998) Portfolio restructuring and limits to hierarchical governance: The effects of environmental uncertainty and diversification strategy. Organization Science 9: 87–102.

14.

Bernheim

(1984) Rationalizable strategic behavior. Econometrica 52: 1007–1028.

15.

Blossfeld

Rohwer

(1995) Techniques of Event History Modeling. Mahwah, NJ: Lawrence Erlbaum Ass.

16.

Bowman

Singh

(1993) Corporate restructuring: Reconfiguring the firm. Strategic Management Journal 14: 5–14.

17.

Brauer

(2006) What have we acquired and what should we acquire in divestiture research? A review and research agenda. Journal of Management 32: 751–785.

18.

Brush

Bromiley

Hendrickx

(2000) The free cash flow hypothesis for sales growth and firm performance. Strategic Management Journal 21: 455–472.

19.

Bruyaka

Durand

(2011) Sell-off or shut-down? Alliance portfolio diversity and two types of high tech firms’ exit. Strategic Organization 10: 7–30.

20.

Buzzell

Gale

Sultan

(1975) Market share-key to profitability. Harvard Business Review 33: 97–106.

21.

Cabantous

Gond

Johnson-Cramer

(2010) Decision theory as practice: Crafting rationality in organizations. Organization Studies 31: 1531–1566.

22.

Carroll

(1985) Concentration and specialization: Dynamics of niche width in populations of organizations. American Journal of Sociology 90: 1262–1283.

23.

Carroll

Hannan

(1989) Density dependence in the evolution of populations of newspaper organizations. American Sociological Review 54: 524–541.

24.

Chang

(1996) An evolutionary perspective on diversification and corporate restructuring: Entry, exit, and economic performance during 1981-1989. Strategic Management Journal 17: 587–611.

25.

Chang

Singh

(1999) The impact of modes of entry and resource fit on modes of exit by multi-business firms. Strategic Management Journal 20: 1019–1035.

26.

Chatterjee

Harrison

Bergh

(2003) Failed takeover attempts, corporate governance and refocusing. Strategic Management Journal 24: 87–96.

27.

Chen

M-J

Smith

Grimm

(1992) Action characteristics as predictors of competitive responses. Management Science 38: 439–455.

28.

Chiles

Meyer

Hench

(2004) Organizational emergence: The origin and transformation of Branson, Missouri’s musical theaters. Organization Science 15: 499–519.

29.

Cohen

(1992) Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ: Erlbaum.

30.

Cohen

Levinthal

(1990) Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly 35: 128–152.

31.

Cool

Dierickx

(1993) Rivalry, strategic groups and firm profitability. Strategic Management Journal 14: 47–59.

32.

Cyert

March

(1963) A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice Hall.

33.

D’Aveni

(1994) Hypercompetition: Managing the Dynamics of Strategic Management. New York: Free Press.

34.

Delacroix

Solt

(1988) Niche formation and founding in the California wine industry 1941-84. In: Carroll

(ed.) Ecological Models of Organizations. Cambridge, MA: Ballinger, pp.53–75.

35.

Denrell

(2003) Vicarious learning, undersampling of failure, and the myths of management. Organization Science 14: 227–243.

36.

DiMaggio

(1988) Interest and agency in institutional theory. In: Zucker

(ed.) Institutional Patterns and Organizations: Culture and Environment. Cambridge, MA: Ballinger, pp 3–21.

37.

DiMaggio

Powell

(1983) The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review 48: 147–160.

38.

DiMaggio

Powell

(1991) Introduction. In: Powell

DiMaggio

(eds) The New Institutionalism in Organizational Analysis. Chicago, IL: The University of Chicago Press, pp.1–38.

39.

Dobrev

Kim

(2006) Positioning among organizations in a population: Moves between market segments and the evolution of industry structure. Administrative Science Quarterly 51: 230–261.

40.

Dyer

Singh

(1998) The relational view: Cooperative strategy and sources of interorganizational competitive advantage. Academy of Management Journal 23: 660–679.

41.

Fligstein

(1991) The structural transformation of American industry: The causes of diversification in the largest firms, 19191-1979. In: Powell

DiMaggio

(eds) The New Institutionalism in Organizational Analysis. Chicago, IL: The University of Chicago Press, pp.311–336.

42.

Fombrun

Shanley

(1990) What’s in a name? Reputation building and corporate strategy. Academy of Management Journal 33: 233–258.

43.

Gilbert

Lieberman

(1987) Investment and coordination in oligopolistic industries. RAND Journal of Economics 18: 17–34.

44.

Gimeno

Woo

(1996) Hypercompetition in a multimarket environment: The role of strategic similarity and multimarket contact in competitive de-escalation. Organization Science 7: 322–341.

45.

Gomez-Mejia

(1992) Structure and process of diversification, compensation strategy and firm performance. Strategic Management Journal 13: 381–397.

46.

Greve

(2003) Organizational Learning from Performance Feedback: A Behavioral Perspective on Innovation and Change. Cambridge: Cambridge University Press.

47.

Grimm

Lee

Smith

(2006) Strategy as Action: Competitive Dynamics and Competitive Advantage. Oxford: Oxford University Press.

48.

Guillen

(2002) Structural inertia, imitation, and foreign expansion: South Korean firms and business groups in China, 1987–95. Academy of Management Journal 45: 509–525.

49.

Guillen

(2003) Experience, imitation, and the sequence of foreign entry: Wholly owned and joint-venture manufacturing by South Korean firms and business groups in China, 1987–1995. Journal of International Business Studies 34: 185–198.

50.

Guiltinan

Gundlach

(1996) Aggressive and predatory pricing: A framework for analysis. Journal of Marketing 60: 87–102.

51.

Hamilton

Chow

(1993) Why managers divest – Evidence from New Zealand’s largest companies. Strategic Management Journal 14: 479–484.

52.

Hannan

Freeman

(1989) Organizational Ecology. Cambridge, MA: Harvard University Press.

53.

Haunschild

Miner

(1997) Modes of interorganizational imitation: The effects of outcome salience and uncertainty. Administrative Science Quarterly 42: 472–500.

54.

Haveman

(1993) Following the leader: Mimetic isomorphism and entry into new markets. Administrative Science Quarterly 38: 564–592.

55.

Heil

Robertson

(1991) Toward a theory of competitive market signaling: A research agenda. Strategic Management Journal 12: 403–418.

56.

Henderson

Cool

(2003) Corporate governance, investment bandwagons and overcapacity: An analysis of the worldwide petrochemical industry, 1975-95. Strategic Management Journal 24: 349–373.

57.

Henisz

Delios

(2001) Uncertainty, imitation, and plant location: Japanese multinational corporations, 1990-1996. Administrative Science Quarterly 46: 443–478.

58.

Hoskisson

Hitt

(1994) Downscoping: How to Tame the Diversified Firm. New York: Oxford University Press.

59.

Hough

White

(2003) Environmental dynamism and strategic decision-making rationality: An examination at the decision level. Strategic Management Journal 24: 481–489.

60.

Howell

(2010) Statistical Methods for Psychology. Belmont, CA: Cengage Wadsworth.

61.

Ilmakunnas

Topi

(1999) Microeconomic and macroeconomic influences on entry and exit of firms. Review of Industrial Organization 15: 283–301.

62.

Karakaya

(2000) Market exit and barriers to exit: Theory and practice. Psychology & Marketing 17: 651–668.

63.

Katila

Shane

(2005) When does lack of resources make new firms innovative? Academy of Management Journal 48: 814–829.

64.

Kraatz

(1998) Learning by association? Interorganizational networks and adaptation to environmental change. Academy of Management Journal 41: 621–643.

65.

Kraemer

Theimann

(1987) How Many Subjects? Statistical Power Analysis in Research. Newbury Park, CA: SAGE.

66.

Kuratko

Ireland

Hornsby

(2001) Improving firm performance through entrepreneurial action. Academy of Management Executive 1: 60–71.

67.

Lévesque

Shepherd

(2004) Entrepreneurs’ choice of entry strategy in emerging and developed markets. Journal of Business Venturing 19: 29–54.

68.

Ljungqvist

Wilhelm

(2003) IPO pricing in the dot-com bubble. Journal of Finance 58: 723–752.

69.

Loomes

Sugden

(1982) Regret theory: An alternative theory of rational choice under uncertainty. Economic Journal 92: 805–824.

70.

Lovejoy

(1971) Divestment for Profit. New York: Financial Executives Research Foundation.

71.

McGahan

Porter

(1997) How much does industry matter, really? Strategic Management Journal 18: 15–30.

72.

McGrath

Nerkar

(2004) Real options reasoning and a new look at the R&D investment strategies of pharmaceutical firms. Strategic Management Journal 25: 1–21.

73.

MacMillan

Siegel

Narisimha

(1985) Criteria used by venture capitalists to evaluate new venture proposals. Journal of Business Venturing 1: 119–129.

74.

March

Olsen

(1976) Ambiguity and Choice in Organizations. Bergen: Universitetsforlaget.

75.

March

Simon

(1958) Organizations. New York: Wiley.

76.

Markides

(1992) Consequences of corporate refocusing: Ex ante evidence. Academy of Management Journal 35: 398–412.

77.

Meyer

Rowan

(1977) Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology 83: 340–363.

78.

Mezias

Lant

(1994) Mimetic learning and the evolution of organizational populations. In: Baum

Singh

(eds) Evolutionary Dynamics of Organizations. New York: Oxford University Press, pp.179–198.

79.

Miller

(1999) The norm of self-interest. American Psychologist 54: 1053–1060.

80.

Miller

Arikan

(2004) Technology search investments: Evolutionary, option reasoning, and option pricing approaches. Strategic Management Journal 25: 473–485.

81.

Moliterno

Wiersema

(2007) Firm performance, rent appropriation, and the strategic resource divestment capability. Strategic Management Journal 28: 1065–1087.

82.

Montgomery

Thomas

(1988) Divestment: Motives and gains. Strategic Management Journal 9: 93–97.

83.

Moore

(1992) Signals and choices in a competitive interaction: The role of moves and messages. Management Science 38: 483–500.

84.

Naman

Slevin

(1993) Entrepreneurship and the concept of fit: A model and empirical tests. Strategic Management Journal 14: 137–153.

85.

Nelson

Winter

(1982) An Evolutionary Theory of Economic Change. Cambridge, MA: Harvard University Press.

86.

O’Brien

Folta

(2009) Sunk costs, uncertainty and market exit: A real options perspective. Industrial and Corporate Change 18: 807–833.

87.

Porter

(1980) Competitive Strategy: Techniques for Analyzing Industries and Competitors. New York: Free Press.

88.

Powell

DiMaggio

(1991) The New Institutionalism in Organization Analysis. Chicago, IL: The University of Chicago Press.

89.

Rumelt

(1991) How much does industry matter? Strategic Management Journal 12: 167–185.

90.

Shin

Kim

Jung

(2014) Is the traditional Asian development model still valid for entrepreneurial ventures? Asian Business & Management 13: 117–141.

91.

Simon

(1955) A behavioral model of rational choice. Quarterly Journal of Economics 69: 99–118.

92.

Smith

Grimm

Gannon

Chen

M-J

(1991) Organizational information processing, competitive responses and performance in the U.S. domestic airline industry. Academy of Management Journal 34: 60–85.

93.

Spence

(1973) Job market signaling. Quarterly Journal of Economics 87: 355–374.

94.

Suárez

Utterback

(1995) Dominant designs and the survival of firms. Strategic Management Journal 16: 415–430.

95.

Tirole

(1988) The Theory of Industrial Organization. Cambridge, MA: The MIT Press.

96.

Tolbert

Zucker

(1983) Institutional sources of change in the formal structure of organizations: The diffusion of civil service reform, 1880–1935. Administrative Science Quarterly 28: 22–39.

97.

Tong

Reuer

(2007) Real options in strategic management. Advances in Strategic Management 24: 3–28.

98.

Tversky

Kahneman

(1986) Rational choice and framing of decisions. Journal of Business 59: 251–284.

99.

Van Kranenburg

Palm

Pfann

(2002) Exit and survival in a concentrating industry: The case of daily newspapers in the Netherlands. Review of Industrial Organization 21: 283–303.

100.

White

(1970) Chains of Opportunity: System Models of Mobility in Organizations. Cambridge, MA: Harvard University Press.

101.

White

Boorman

Breiger

(1976) Social structure from multiple networks. I. blockmodels of roles and positions. American Journal of Sociology 81: 730–880.

102.

Williamson

(1985) The Economic Institutions of Capitalism. New York: Free Press.

103.

Winter

Szulanski

(2001) Replication as strategy. Organization Science 12: 730–743.

104.

Xia

Tan

(2008) Mimetic entry and bandwagon effect: The rise and decline of international equity joint venture in China. Strategic Management Journal 29: 195–217.

Intended rationality in exit decisions: The organizational field of rivals as a source of signal in segment exit decisions by Korean SI ventures,2000–2006

Abstract

Keywords

1. Introduction

2. Theory

2.1. Segment exit and uncertainty

2.2. Aggregated actions of rivals as an uncertainty-reducing signal in exit decisions

2.3. The firm as an ‘intendedly rational’ actor

2.4. Research model and hypotheses

2.4.1. Aggregated competitive intensity and the competitiveness of the segment

2.4.2. Rivals’ performances as a signal of segment attractiveness

2.4.3. Exit bandwagon as a signal of ongoing segment deterioration

2.4.4. Interaction between an exit bandwagon and the focal firm’s performance

3. Empirical setting

4. Methods

4.1. Data and measurement

4.1.1. Dependent and independent variables

4.1.1.1. Segment exit

4.1.1.2. Aggregated competitive intensity at the segment level

4.1.1.3. Segment attractiveness

4.1.1.4. Exit bandwagon

4.1.2. Control variables

4.1.2.1. Population density

4.1.2.2. Societal-level discourse

4.1.2.3. Firm age

4.1.2.4. Firm size

4.1.2.5. Firm performance

4.1.2.6. Strategic alliances

4.1.2.7. Entry bandwagon

4.1.2.8. Niche width

4.2. Statistical analyses and results

5. Conclusion

Footnotes

Appendix

Funding

References