Abstract
Prior research has identified two different sources of strategic imitation—through perceived organizational cluster similarity (cluster effects) and direct social connections (tied-to effects). In the research on tied-to effects, top executives’ social ties, such as outside directorships, have long been studied as a mechanism through which strategic imitation develops. However, are all ties the same? There has been little examination of whether some social ties have more influence than others. Using the attention-based view of the firm, we argue that certain social ties garner more attention by being salient to top executives. We empirically test this assertion by examining the effects of CEO outside directorships on R&D spending. Using panel data from large U.S. manufacturing firms, we find that CEOs imitate the R&D intensity of tied-to firms (i.e., a firm in which the CEO serves as an outside board member) in their own firm’s R&D decisions. Consistent with attention-based arguments, our results show evidence of selective imitation, as imitating relationships are stronger when the CEO has longer tenure as a director of a tied-to firm and the tied-to firm is performing well. In contrast to conventional institutional theory, our findings also show that CEOs imitate relatively smaller tied-to firms when they make R&D investment decisions. Not all social ties have equal influence on imitative strategic decision making; thus, they have different strategic implications.
Strategic imitation within and across organizations has long been an important topic in the management literature (e.g., DiMaggio & Powell, 1983; Garcia-Pont & Nohria, 2002; Haunschild & Miner, 1997; T. Levitt, 1966; Lieberman & Asaba, 2006; Lu, 2002). Even though imitation does not necessarily improve a firm’s performance (Barreto & Baden-Fuller, 2006), it can assist decision makers who are facing complex and uncertain decisions. Prior research, based mostly on institutional theory, has provided the following explanations as to why organizations imitate other organizations: (1) to learn from other firms (Argote, 1999; B. Levitt & March, 1988), (2) to respond to a competitor’s activities (Lieberman & Montgomery, 1988), (3) to acquire legitimacy for their actions (DiMaggio & Powell), and (4) to reduce risk associated with making decisions in an uncertain environment (Ordanini, Rubera, & DeFillippi, 2008).
While prior studies have been important in explaining the idea of institutional isomorphism (e.g., copying others to gain legitimacy) as a motivation behind strategic imitation (DiMaggio & Powell, 1983), our understanding of why and how executives decide to imitate other firms’ behaviors is still limited. In other words, existing research has not thoroughly investigated the underlying mechanisms of strategic imitation (Posen, Lee, & Yi, 2013: 149). Researchers have recently begun to perceive imitation as a strategic action available to firms (e.g., Csaszar & Siggelkow, 2010). Yet questions about how the strategic imitation process occurs remain unsolved.
In the past, researchers have generally hypothesized and examined two paths to strategic imitation. One approach is based on what may be called “cluster effects.” In cluster effect imitation, organizations tend to copy the strategies or decisions of other organizations that their managers believe are in the same, socially constructed, referent group as the imitating organization. Thus, organizations copy the strategies of organizations they believe are in the same industry, strategic group, geographic cluster, or group of firms that made similar decisions in the past. For example, studies of cluster effects have found strategic imitation of other organizations in plant locations (Henisz & Delios, 2001), curriculum changes in educational institutions (Kraatz, 1998), and the adoption of certain organizational structures (Burns & Wholey, 1993). In each of these cases, strategic imitation brought the strategies or structures of imitating firms into closer alignment with a cluster of firms that were social referents (e.g., the industry or group norm).
The cluster studies come from a sociological perspective where imitation is modeled on the basis of the similarity of organizations (e.g., firms to be imitated are in the same socially construed cluster as the imitating firm). A few studies have even examined “selective imitation” whereby organizations are more likely to imitate the decisions of certain other organizations to the extent that these organizations are more representative or prototypical of the referent cluster group. Representative or prototypical firms have been found to be those classified as the same reference group by regulators as the firm (Barreto & Baden-Fuller, 2006) or as being central in the network of organizations in the cluster (Kraatz, 1998; Lee & Pennings, 2002). As such, firms that are seemingly more representative or prototypical of the referent group are imitated more frequently. However, the cluster studies of imitation generally do not examine direct social ties (e.g., board interlocks) between an imitated firm and the firms it is imitating.
The other main stream of research examines strategic imitation through what may be called “tied-to effects.” This research stream assumes direct social ties as a conduit of strategic imitation, largely focusing on directorship ties between firms. These studies have examined, for example, how board ties across firms lead to the adoption of similar strategic decision-making processes (Westphal, Seidel, & Stewart, 2001), making acquisitions (Haunschild, 1993), and adopting certain governance practices (Davis, 1991; Westphal & Zajac, 1997).
Despite a seemingly more direct social connection in the tied-to effects studies of strategic imitation (e.g., imitation spreads through direct human social connections rather than through perceived organizational cluster similarity), research on imitation through board interlocks or ties has not focused on selective imitation. Thus, in this social connection research, all board ties or interlocks have the same influence on strategic imitation no matter what the characteristics of the tied-to firms are. All ties are equal in terms of being a possible source of strategic imitation, despite the heterogeneous nature of the firms on which executives can serve as external directors. For instance, highly profitable firms are just as likely to be copied through board ties as firms in financial trouble. This lack of selective imitation of socially tied firms therefore presents an empirically unexplored gap in our understanding of how board ties affect strategic imitation.
In this study, we attempt to fill this gap in strategic imitation research by examining selective imitation through board social ties, which has not been examined. We argue that selective imitation through social ties occurs because some firms garner a greater share of a manager’s attention than other tied-to firms (e.g., Ocasio, 1997). The firms that receive more attention have attributes that make them more salient than other tied-to firms. We test our argument by examining how CEOs’ service on other firms’ boards may lead to selective imitation of those firms’ R&D investment decisions.
In particular, we propose that CEOs engage in selective imitation by more likely copying the R&D investment decisions of more familiar, high-performing, and visible firms for which they serve as outside directors. Specifically, we expect that CEOs are more likely to imitate a tied-to firm’s R&D investment pattern if (1) the focal firm’s CEO has a long tenure as director of the tied-to firm, (2) the tied-to firm is performing well, (3) the tied-to firm is more visible in the organizational field as a result of its size, and (4) the focal CEO’s firm faces a high level of environmental uncertainty.
Using panel data from a sample of U.S. Fortune 500 manufacturing firms, we find that CEOs imitate the prior R&D investment decisions (i.e., industry-adjusted R&D intensity) of tied-to firms when they make the same decision for their own firms. Furthermore, we find evidence of selectivity in strategic imitation through board ties, suggesting not all ties are equally important. The R&D spending of focal firms is more strongly associated with the spending of tied-to firms when a firm’s CEO has a longer tenure as director of a tied-to firm and the tied-to firm is profitable. In contrast to conventional institutional theory arguments, we find that CEOs imitate relatively smaller firms than focal firms when they make R&D investment decisions. From these findings, we draw conclusions about when CEOs will rely on strategic imitation processes for important but uncertain strategic decisions, such as R&D spending.
We make several contributions to the existing research on strategic imitation and innovation. First, we illustrate that CEOs engage in imitative behavior through outside directorships when they make important organizational decisions, which is consistent with the research stream focusing on tied-to effects. Second, our hypotheses and findings provide a deeper understanding of the drivers that affect the inclination to imitate through directorship ties. Not all outside directorships have equal influence on imitative strategic decision making. Top executives do not simply mimic tied-to firms but, rather, pay more attention to tied-to firms that provide salient information. This greater attention to salient firms makes their strategies more likely to be adopted. Third, our findings also suggest that organizations learn from the experience of others by observing and imitating the decisions of their tied-to firms. Thus, our study may complement existing organizational learning models by showing how CEO directorships promote social learning. Lastly, while our study’s focus is on strategic imitation through board ties, our use of R&D investment as a dependent variable has implications for the innovation management literature. While past literature has long focused on the CEO’s role in setting R&D spending levels (e.g., Barker & Mueller, 2002), we expand that focus to the CEO’s social relationships with other firms as a source of information in making investment decisions.
Theoretical Background
When firms are faced with a high level of uncertainty, decision makers who have bounded rationality economize on search costs (Cyert & March, 1963), and this economizing effort sometimes leads to the imitation of other organizations. Because top managers may have incomplete or even too much information coupled with bounded rationality for processing that information (e.g., Simon, 1965), strategic imitation may be a form of learning. Strategic imitation may reduce the uncertainty in a situation by providing a realistic course of action that other firms are implementing.
As discussed previously, past research has generally studied two main mechanisms of imitation. One set of studies examines how a firm’s managers may imitate the decisions of clusters of firms that are perceived to be socially desirable referents (e.g., Burns & Wholey, 1993; Henisz & Delios, 2001; Kraatz, 1998; Lee & Pennings, 2002). From an organizational sociology perspective, socially desirable referent firms are imitated because such firms are assumed to have better knowledge or greater legitimacy. As a result, the copying firm’s managers decide to imitate the decisions of referent firms in order to become more like the socially desirable cluster. Ultimately, copying the cluster reduces uncertainty and increases the legitimacy of the imitating organization. However, “cluster effect” studies tend not to examine the social or human processes through which imitation occurs (see Porac, Thomas, & Baden-Fuller, 1989, for a notable exception to this point).
Studies examining strategic imitation through social ties between firms (e.g., what we call “tied-to effects” studies) are more about the process of strategic imitation through connections between firms or executives. Using the board of directors as an information conduit, these studies suggest that imitation comes from information sharing through either (1) having shared directors between firms (e.g., Davis, 1991; Westphal et al., 2001), (2) having CEOs sit on other firms’ boards (e.g., Haunschild, 1993) or (3) having outside board directors who are top managers at other firms (e.g., Ji & Oh, 2014; Westphal & Zajac, 1997). In these studies, uncertainty in strategic decision making is reduced by information coming from other firms through board ties. These ties lead to a greater chance of strategic imitation as firms share information.
Yet researchers in tied-to effects studies have generally treated or modeled all board ties as equally valuable in supplying information. This would seem to be problematic for strategic imitation research because all tied-to firms may not have characteristics, strategies, or performance levels that make them desirable to imitate. As discussed in the next section, this study focuses on CEO ties to other firms through outside directorships. Also, we argue that not all tied-to firms are given equal attention by CEOs.
Empirical Context: The CEO’s Decision on R&D Investment
In this study, the empirical context for examining strategic imitation is the CEO’s decision on the level of the firm’s R&D spending. CEOs make strategic choices about R&D investment in the hope of reaping positive financial returns and gaining a competitive advantage in the future. While a firm’s R&D spending does not guarantee successful innovation (Mansfield, 1968), investing financial resources in R&D is one of the most fundamental strategic decisions of firms that compete in industries where innovation may be a source of competitive advantage. A firm without appropriate R&D investment not only limits its capability to develop new technology (Helfat, 1997) but also restricts its capability to absorb new knowledge (Cohen & Levinthal, 1990). However, firms that overspend on R&D may deprive their shareholders of higher returns for the sake of increased management security (e.g., Jensen, 1993). Indeed, R&D investments have unique characteristics, such as outcome uncertainty, long payback periods, and the need for constant evaluation and commitment. Therefore, as a costly but critical input for the innovation process, the level of R&D investment is an important strategic decision that cannot be taken lightly by a firm’s managers. The uncertainty of R&D investment, coupled with a CEO’s discretion over its level of investment, creates a situation in which the CEO may look for cues about how much to spend. These cues may come at least partially from other organizations where the CEO serves as an outside board member.
Strategic Imitation: How CEO Directorships Matter
CEOs, as the heads of their firms, have the power and motivation to set the direction of their firms. Given CEOs’ unique position to influence the firm’s strategic agenda, they should have discretion to control the level of R&D investment (Li, Maggitti, Smith, Tesluk, & Katila, 2013). Empirical findings support this argument. For example, prior studies have found that CEO characteristics, such as age, career experience, educational background, and technical orientation, affect R&D spending (Barker & Mueller, 2002; Daellenbach, McCarthy, & Schoenecker, 1999). Other studies (e.g., Fong, 2010; Wu & Tu, 2007) also report that CEOs adjust the level of a firm’s R&D spending in response to attributes of their compensation packages, such as stock options.
CEOs influence what information matters to an organization and how this information is interpreted, and they actively engage in strategy formulation using this information (Yadav, Prabhu, & Chandy, 2007). In the information-collection process, CEOs may use their outside director ties as a source of information (Conyon & Read, 2006; Haunschild, 1993). CEOs gain insights regarding operations, agendas, and key decisions of other firms through their board service, which provide important conduits of information about strategic options for their firms. Thus, outside directorships provide access to strategic information and can channel the strategic imitation by determining which firms get imitated (e.g., Carpenter & Westphal, 2001; Galaskiewicz & Wasserman, 1989). As such, a CEO’s external directorships, as an important source of information, may play a significant role in determining strategic decisions at his or her own firm.
Attention-Based View of the Firm
While outside directorship ties may be an important source of information for CEOs, not all ties will likely have equal influence on CEOs’ decisions. The information CEOs bring to the strategic decision-making process is affected by where or on what their attention is focused (Cho & Hambrick, 2006; Yadav et al., 2007). Attention is a “cognitive process that involves the noticing, interpretation, and focusing of time and effort on the acquisition of knowledge and information” (Li et al., 2013: 894). Ocasio (1997: 197) specifically notes that the most critical players in organizational attention are typically the CEO and top executives. CEOs selectively allocate their attention because they are confronted with more information than they can possibly process (Kiesler & Sproull, 1982; Simon, 1965). Therefore, CEOs focus on “what they perceive to be key subsets of the available data” (Garg, Walters, & Priem, 2003: 725) while selectively ignoring others. Hambrick also specifically notes that executives “act on those phenomena to which their attention is drawn” (1981: 299).
Since CEOs are constrained by bounded rationality (Cyert & March, 1963; Simon, 1965), their attention cannot be dedicated to all tied-to firms equally. In other words, their cognitive limits make them pay more attention to some of their directorship ties than others as sources of information. In this sense, managerial attention provides a filter through which CEOs identify which tied-to firms are better models than others to imitate.
Cognitive psychologists (e.g., Bandura, 1977; Nisbett & Ross, 1980) and researchers studying managerial cognition (e.g., Kiesler & Sproull, 1982) have long argued that decision makers allocate more attention to stimuli or cues that are salient. Salient stimuli are those that stand out from other potential stimuli because they present themselves more vividly (Nisbett, 1987; Nisbett & Ross), stand out from what decision makers normally perceive (Kahneman, 1992), or affect the high-priority goals of the perceiving individual (Fiske & Taylor, 1991: 248).
In this paper, we argue that tied-to firms (1) where the CEO has a long tenure as director, (2) that are performing better than others, and (3) that are large relative to the focal firm are more likely to be imitated by a CEO. Information coming from such firms will be more salient because that information stands out and is more connected to the CEO’s goal of making the focal firm successful. As such, CEOs pay more attention to tied-to firms with the above characteristics. In addition, a tied-to firm’s past decisions become more important (4) when the focal firm faces a high level of environmental uncertainty. Environmental uncertainty will cause CEOs to scan their environments more frequently and thoroughly (Daft, Sormunen, & Parks, 1988), thus producing greater focus on tied-to firms.
Hypotheses Development
Strategic Imitation in R&D Investment Decisions
As discussed in the paper’s introduction, strategic imitation has received attention from multiple social science disciplines. Strategic imitation usually occurs when there is a high level of complexity, uncertainty, and ambiguity (Lieberman & Asaba, 2006). Under such conditions, strategic imitation occurs because prior decisions by other organizations increase the legitimacy of similar decisions (Haunschild & Miner, 1997). While it is very difficult for one firm to imitate another firm’s competitive advantage (e.g., Barney, 1991), CEOs may be able to imitate an input to competitive advantage (i.e., R&D spending) as a second-best alternative.
Researchers have examined the imitation of strategic decision processes or organizational decisions facilitated by board ties across firms (e.g., Davis, 1991; Haunschild, 1993; Ji & Oh, 2014; Westphal et al., 2001; Westphal & Zajac, 1997). We earlier referred to these studies as being theoretically based on tied-to effects. Some studies have specifically examined the imitation of strategic decisions facilitated by a CEO’s service on other firms’ boards of directors (e.g., Haunschild). Indeed, knowledge obtained from top executives’ outside directorships can be used at their own firms (Conyon & Read, 2006). In particular, given the high failure rate of R&D investment (Mansfield, 1968) due to complexity, uncertainty, and causal ambiguity, the R&D investment decision may be one where top executives look to other firms for cues about what investment levels are appropriate.
Therefore, when the focal firm is tied to other firms through its CEO having an outside directorship, we expect that the focal firm’s R&D spending will be positively associated with the prior spending levels of tied-to firms. In other words, we propose the idea that “what you see is what you do” (Greve, 1998). CEOs are more likely to rely on the decisions of tied-to firms rather than other noninterlocked organizations, which suggests the following:
Hypothesis 1: There will be a positive association between the R&D intensity of focal firms and the R&D intensity of tied-to firms at which the focal firms’ CEOs serve as outside directors.
Unequal Ties: The Role of Managerial Attention
While we predict that CEOs will engage in imitative behavior in R&D investment decisions through their outside directorships, CEOs may selectively choose which organizations are the appropriate models to imitate. Without selective imitation, all board ties have equal influence on the firm, which would seem to be a suboptimal information-processing strategy for CEOs. Do CEOs focus more on information from struggling firms or profitable firms? Do CEOs pay more attention to firms they are just beginning to understand or to firms for which they know more about their strategy and operations? On the basis of the attention-based view of the firm (Cho & Hambrick, 2006; Ocasio, 1997), it is reasonable to expect that CEOs do not imitate other firms’ decisions simply because tied-to firms may be more or less legitimate in an organizational field. Rather, certain tied-to firms’ R&D expenditures may capture the attention of a CEO because those firms stand out (e.g., Nisbett, 1987; Nisbett & Ross, 1980).
Therefore, top managers engage in selective imitation, with prior decisions at tied-to firms that are salient in CEOs’ minds being more likely to affect decisions at focal firms. Information from salient tied-to firms is more likely to attract the CEO’s limited attentional resources. Specifically, we propose three different moderating factors that increase the salience of cues from tied-to firms: (1) greater attention to familiar firms (i.e., CEOs will pay more attention to tied-to firms when they have served on that firm’s board for a long time), (2) greater attention to high-performing firms (i.e., CEOs will pay more attention to tied-to firms that are financially successful), and (3) greater attention to larger firms (i.e., CEOs will pay more attention to tied-to firms that are large relative to the focal firms). Information from tied-to firms with these characteristics will be more salient to CEOs and will receive more attentional resources. Greater attentional focus will increase the CEOs’ propensity to imitate these firms.
We also propose a fourth moderator based on the focal firms’ own industry environments. CEOs are more likely to imitate tied-to firms when their own firms face a high level of environmental uncertainty. Under these conditions, CEOs will be more vigilant in scanning their environments for information to help them make decisions (Daft et al., 1988). This increased attention to the external environment will increase the likelihood that information from tied-to firms is noticed and used in strategic imitation.
Greater attention to familiar firms
Research based on the cognitive decision theory has shown that people recognize familiar information or cues in a complex decision situation more frequently and give familiar information more weight in decision making (Nisbett & Ross, 1980; Tversky & Kahneman, 1973). As such, CEOs pay more attention to information from familiar sources.
A CEO’s tenure as director of a tied-to firm is an indication of his or her familiarity with the tied-to firm. Therefore, extended tenure increases the salience of tied-to firms in the CEO’s mind for the following reasons. First, CEOs with longer tenure as director have more knowledge about the strategic decision processes of tied-to firms (e.g., Carpenter & Westphal, 2001). In terms of R&D investment decisions, CEOs who have served as directors of tied-to firms for a long time have a greater understanding of the complex decision-making process and rationale behind the tied-to firm’s R&D spending decisions. In contrast, their shorter-tenured counterparts may know the dollar amount of R&D spending but not the decision-making process and rationale behind it. In addition, CEOs with longer tenure as directors are likely to have a stronger commitment to the tied-to firms (Buchanan, 1974). This increased commitment will make the tied-to firms’ choices more salient and will further increase the CEOs’ attention to the tied-to firms for which they have served longer as outside directors. Hence, we hypothesize:
Hypothesis 2: The length of CEOs’ outside directorship tenures at tied-to firms will positively moderate the relationship (i.e., increase the strength) between the R&D intensity of focal firms and the R&D intensity of tied-to firms at which the focal firms’ CEOs serve as outside directors.
Greater attention to high-performing firms
In a strategic imitation process, CEOs will pay more attention to information from more profitable tied-to firms. Bandura (1977) suggests that individuals observe the outcomes of others’ actions and imitate the actions that generate positive outcomes and avoid those that produce negative outcomes. Cognitive decision theory also suggests that individuals pay more attention to vivid informational cues (Nisbett, 1987; Nisbett & Ross, 1980) that stand apart from what they normally encounter (Kahneman, 1992) and may affect high-priority goals (Fiske & Taylor, 1991: 248). Because profitability is a high priority for CEOs, the actions and decisions of highly profitable firms to which the CEO has access should stand out as information available to the CEO for decision making at his or her own firm. Thus, more profitable firms are likely to serve as desirable models for other firms (Burns & Wholey, 1993; Csaszar & Siggelkow, 2010). This argument is similar to the notion of “outcome-based imitation” (Haunschild & Miner, 1997; Williamson & Cable, 2003), such that decision makers will mimic the practices they believe have produced positive outcomes for other firms.
Empirical findings outside of the director ties’ studies also support that executives pay more attention to firms with high levels of profitability in strategic imitation processes. For example, Haveman (1993) finds that managers imitate profitable organizations in their industry when they make diversification decisions. Similarly, Argote, Beckman, and Epple (1990), in examining shipyard productivity, show that firms starting production later imitate the practices observed to be successful in earlier start-ups. Thus, while CEOs may have the same opportunity to perceive the decisions of all firms on whose boards they sit, their attention will more likely be focused on the successful firms’ R&D investment levels. Therefore, we hypothesize:
Hypothesis 3: Tied-to firms’ profitability will positively moderate the relationship (i.e., increase the strength) between the R&D intensity of focal firms and the R&D intensity of tied-to firms at which the focal firms’ CEOs serve as outside directors.
Greater attention to larger firms
Executives tend to pay greater attention to organizations with high visibility as a result of the scale of their activities (e.g., Burns & Wholey, 1993). Previous literature has suggested that firm size plays an important role in imitation (e.g., Greve, 2005; Haveman, 1993). We propose that large tied-to firms in particular may attract considerable attentional resources of CEOs. Because large firms receive more coverage by journalists and other arbiters of attention, the amount of general stimuli related to these firms will be greater and thus make them more prominent in the wave of information processed by CEOs. This should make information that is personally available to the CEOs through their ties to large firms easier to recall (Tversky & Kahneman, 1973) and therefore the focus of greater attention. Also, if large size, similar to profitability, is a desirable goal for firms, the actions and decisions of larger firms to which the CEO has access should stand out (Fiske & Taylor, 1991) as information available for making decisions at his or her own firm.
The empirical findings from cluster effect strategic imitation studies support this assertion. For instance, in a study of capacity expansion decisions by chemical producers, Gilbert and Lieberman (1987) find a pattern of small firms following larger firms’ capacity decisions. Haveman (1993) also reports that large savings institutions tended to copy other large savings institutions in entering certain markets on the U.S. West Coast. In the R&D intensity context, it is expected that CEOs will pay greater attention to the decisions of larger tied-to firms because large firms have grown in size as a result of past success and have more information created about them by social arbiters (i.e., the press, industry analysts, etc.). Thus, larger firms will draw more attentional focus of CEOs than smaller firms. Therefore, we hypothesize:
Hypothesis 4: Tied-to firms’ relative size will positively moderate the relationship (i.e., increase the strength) between the R&D intensity of focal firms and the R&D intensity of tied-to firms at which the focal firms’ CEOs serve as outside directors.
Greater attention to tied-to firms under environmental uncertainty
In addition to the tied-to firm’s attributes, as described above, the industry environment influences the structure of managerial attention (Nadkarni & Barr, 2008). From the attention-based perspective (Ocasio, 1997), environmental uncertainty should also increase the likelihood of strategic imitation of tied-to firms’ decisions. However, the mechanism for this moderator is not the characteristics of the tied-to firms (e.g., familiarity, profitability, or size) but, rather, how vigilantly the focal firms’ CEOs scan the environment. It has long been argued that top executives will increase their examination or scanning of the external environment when their firm’s situation seems more uncertain (Garg et al., 2003; Thompson, 1967). Daft and colleagues (1988) empirically verify this argument by examining top executives’ environmental scanning behaviors. They find that CEOs in highly uncertain industries scan their environments more frequently and thoroughly. Thus, this increased search activity focused on the external sources of CEOs in uncertain environments will increase the chances of imitating tied-to firms.
Institutional theorists also argue that uncertainty in the environment increases the need for imitative decision making; thus, CEOs tend to see imitation as a natural and safe response (Lieberman & Asaba, 2006). As such, increased uncertainty enhances the importance of strategic imitation such that CEOs may use social comparisons as a basis for strategic decision making in highly uncertain environments. For example, Haunschild (1994) finds that managers’ tendency to mimic the paying of acquisition premiums is stronger when acquiring firms face a high level of environmental uncertainty. Geletkanycz and Hambrick (1997) also note that imitating other firms’ decisions is beneficial when the decision is associated with a high level of uncertainty.
Lieberman and Asaba specifically note that “in highly uncertain environments, where quick action is necessary, imitating others becomes an attractive decision rule” (2006: 373). Thus, we expect that it is difficult for CEOs to decide the optimal level of R&D spending solely on the basis of technological criteria if their firms operate under a high level of uncertainty. Since these CEOs more actively search for information, they are more likely to pay attention to the decisions of the firms for which they serve as outside directors. Because of this increased attention to tied-to firms under environmental uncertainty, we hypothesize:
Hypothesis 5: The level of dynamism in a focal firm’s industry will positively moderate the relationship (i.e., increase the strength) between the R&D intensity of focal firms and the R&D intensity of tied-to firms at which the focal firms’ CEOs serve as outside directors.
Method
Data and Sample
This study’s sample was drawn from Fortune 500 manufacturing firms (two-digit Standard Industrial Classification, SIC, Codes 20 through 39) for the years 2004 to 2007. We limited our sample firms to manufacturing industries because R&D expenditures are often not reported in other industries (e.g., retail and service industries). To be included in the sample, a firm’s CEO had to have served as an outside director on another publicly traded firm that also reported R&D expenditures. Therefore, our sample was not randomly selected from Fortune 500 manufacturing firms. In order to correct this nonrandom sampling, we used the Heckman selection model in our analysis, as we discuss later (see Heckman, 1979). Along with missing data, these conditions provided a sample of 199 firm-year observations from 66 focal firms over the period from 2004 to 2007, which created unbalanced panel data. We used a number of archival sources (i.e., Who’s Who in Finance and Business, Standard & Poor’s COMPUSTAT, and Corporate Library), firm annual reports (i.e., Securities and Exchange Commission, SEC, Form DEF 10K), and proxy statements (i.e., SEC Form DEF 14A) to collect the data.
Measurement of Variables
Dependent variable
Creating R&D intensity measures, as opposed to flat total dollar R&D expenditures, is the norm in empirical studies because it corrects for firm size effects (e.g., Barker & Mueller, 2002). We used an industry-adjusted R&D intensity measure by calculating a Z score for the focal firm as follows: Z = (X – μ) / σ, where X is the firm R&D spending as a percentage of sales, µ is the industry average R&D spending as a percentage of sales, and σ is the industry standard deviation in R&D spending as a percentage of sales. This variable was labeled industry-adjusted R&D intensity of the focal firms. Industry average R&D intensity and its standard deviation were calculated by using each firm’s industry (two-digit SIC code) for all firms from COMPUSTAT with more than 1,000 employees. Thus, our measure for R&D intensity, which is adjusted in terms of industry norms, has either a positive value if the firm has greater R&D intensity than the industry average or a negative value if the firm has less than the industry average. Also, by using Z scores, we can compare across a broad cross-section of manufacturing industries because a common scale is created, regardless of the base level of R&D spending in a particular industry.
Independent variable
CEO outside directorship data were mainly collected from firms’ annual reports (i.e., SEC Form DEF 10K), proxy statements (i.e., SEC Form DEF 14A), and other data sources (e.g., Who’s Who in Finance and Business). As discussed earlier, when CEOs (1) did not have any outside directorships or (2) served as a director of a nonmanufacturing firm that did not report R&D spending, we did not include those firms in our analysis, since sources of imitation from tied-to firms do not exist. The independent variable tied-to firm’s R&D intensity was calculated similarly to how the dependent variable was calculated: We subtracted the tied-to firm’s industry average R&D intensity from the tied-to firm’s R&D intensity before dividing by the tied-to firm’s industry standard deviation in R&D intensity (e.g., industry again defined at the two-digit SIC code level). For focal firms with CEOs with directorships at multiple firms, the average value of all tied-to firms was utilized.
Moderating variables
We predict that there will be a higher likelihood of imitation when the CEO has a longer tenure as director of a tied-to firm (Hypothesis 2), when the tied-to firm is more profitable (Hypothesis 3), when the tied-to firm is larger than the focal firm (Hypothesis 4), and when the focal firm’s industry is more dynamic (Hypothesis 5). CEO directorship tenure was measured by the difference between the focal year (i.e., between 2004 and 2007) and the nominating year as director of the tied-to firm. Tied-to firm’s ROA (return on assets) was measured by the ROA of the tied-to firm, stated as a percentage. We measured the relative size difference (between the focal and tied-to firms) in terms of number of employees. This is calculated by subtracting the number of employees of the focal firm from that of the tied-to firm after a log transformation of each number, due to skewed distributions. Thus, when the tied-to firm is larger (smaller) than the focal firm, the value is positive (negative). Our hypothesis is based on the assumption that a larger firm’s decision, compared to that of the focal firm, will attract more of the CEO’s attention, so we used the “relative” size difference (as opposed to the absolute value of a tied-to firm’s size 1 ). We calculated industry dynamism, following Keats and Hitt (1988), by a two-step procedure. First, the natural logarithm of sales for each industry at the two-digit SIC level for 5 years is regressed against time (as an independent variable), and then antilogarithms of the regression slope coefficient and standard errors from these models were calculated. The standard errors from the regression models indicate the variability of industry growth and were used as industry dynamism.
Control variables
We controlled for other firm-level and CEO individual-level factors that could influence R&D expenditures: firm age, firm size, proportion of outside directors, institutional ownership, focal firm’s ROA, organizational slack, debt ratio, CEO age, CEO tenure, and industry relatedness between focal and tied-to firms. First, we controlled for firm age, since older firms are more likely to be inert (Hannan & Freeman, 1977) and thus likely to spend less on R&D than younger firms. The previous literature finds that firm size is positively (Baysinger, Kosnik, & Turk, 1991) or negatively (Hansen & Hill, 1991) associated with R&D investment. Thus, we included firm size in the analysis by taking the logarithm of the number of the firm’s employees, due to a skewed distribution. Board composition also has significant influence on a firm’s R&D investment (Dalziel, Gentry, & Bowerman, 2011), so we controlled for the proportion of outside directors by dividing the number of outside directors on a board by the total number of board members. Institutional ownership is measured by the percentage of a firm’s shares held by institutions. It is often argued that institutional investors have a relatively short time horizon, which forces firms to underinvest in R&D (e.g., Graves, 1988), although this assertion has been challenged by the opposite findings (e.g., Hansen & Hill). We included the focal firm’s ROA, organizational slack, and debt ratio since past performance and financial resources may have a positive (e.g., Hundley, Jacobson, & Park, 1996) or negative (e.g., Hitt, Hoskisson, Ireland, & Harrison, 1991) association with R&D spending. For this study, the focal firm’s ROA was calculated by net income divided by total assets stated as a percentage. Organizational slack was calculated by working capital divided by total sales (Bourgeois, 1981). Lastly, we measured debt ratio as a percentage of the focal firm’s long-term debt relative to its total assets. All of these firm-level variables were collected from COMPUSTAT and the Corporate Library database.
Prior studies have found that R&D expenditures are influenced by the characteristics of the CEO (e.g., Barker & Mueller, 2002; Fong, 2010; Wu & Tu, 2007). Therefore, we controlled for CEO age and CEO tenure. Researchers examining executive age find that older managers tend to be more conservative and risk averse (Barker & Mueller) and thus are likely to spend less on R&D. Also, longer-tenured CEOs may have less interest in pursuing an innovative strategy (Miller, 1991). Furthermore, since similar industry conditions may drive similar R&D spending patterns in both tied-to and focal firms, we created the dummy variable industry relatedness. This variable was coded as 1 if two firms had the same two-digit SIC code or if they belonged to buyer or supplier industries and as 0 otherwise. 2 Finally, we included the inverse Mills ratio, calculated by a Heckman selection model to correct the nonrandomness of our sample. The Heckman selection model details are included in the appendix. In order to provide stronger causal inferences and reduce the possibility of reverse causality, all predictor variables (i.e., hypothesized and control variables) were lagged by 1 year.
Statistical Analysis
We used the panel data regression techniques to examine CEOs’ imitative behavior in making R&D investment decisions. Since some of the explanatory variables may be correlated with the unobserved random effects, we adopted the Hausman–Taylor panel data regression for endogenous covariates (Hausman & Taylor, 1981), an estimator that uses an instrumental variable method. For the purpose of this modeling, we assumed that industry relatedness, a focal firm’s financial performance, and our hypothesis testing variables might be endogenous. A necessary condition for Hausman–Taylor estimation is that the number of exogenous time-varying covariates is equal to or greater than the number of endogenous time-invariant covariates (Hausman & Taylor: 1385). Our data set satisfies this condition because most of our control variables are exogenous. The Hausman–Taylor estimation method shows improvement over the fixed-effects or random-effects models in that it controls for the endogeneity problems and produces estimates for the effects of time-invariant variables (Baltagi, Bresson, & Pirotte, 2003). In addition, the Hausman–Taylor estimation method is efficient in the use of instruments since all instruments are derived from within the model.
In order to assess the multicollinearity among variables, the full models based on pooled data were tested for multicollinearity by calculating variance inflation factors (VIFs). The mean value of VIFs is 2.27, and the VIF of each variable is below the conventional criteria (VIF < 10); thus, we do not have multicollinearity concerns (e.g., Belsley, Kuh, & Welsch, 1980). We used mean-centered variables for all continuous variables in order to reduce the potential effects of multicollinearity. Lastly, as we stated previously, our sample consists of only firms with CEOs who served as directors of other public manufacturing firms that reported R&D expenditure data during the sampling timeframe. To correct this bias from nonrandom sampling, we used the Heckman selection model (Heckman, 1979), a two-stage procedure that corrects for sample selection bias (see the appendix).
Results
Table 1 shows the descriptive statistics, including means, standard deviations, and correlations, of the variables. Table 2 reports the results of our analyses. We conducted the Hausman–Taylor estimation in a stepwise manner. Model 1 included only control variables, and Model 2 tested the main effect of strategic imitation in R&D investment decisions. Models 3, 4, 5, and 6 added each interaction term in order to test the moderation hypotheses. Model 7 is the fully specified model with all interaction terms.
Descriptive Statistics: Means, Standard Deviations, and Correlations
Note: Two-tailed coefficient tests. Correlations with absolute values greater than .14 are significant at p < .05, and absolute values greater than .18 are significant at p < .01. ROA = return on assets.
Effects of CEO Outside Directorship on Industry-Adjusted R&D Intensity of Focal Firms: Hausman–Taylor Estimation
Note: Two-tailed coefficient tests. ROA = return on assets.
p < .10.
p < .05.
p < .01.
p < .001.
In terms of specific hypotheses, Model 2 examines the main effect of a tied-to firm’s R&D intensity on a focal firm’s R&D investment decision. We find that a tied-to firm’s R&D intensity is positively associated with the R&D intensity of a focal firm (p < .001); thus, Hypothesis 1 is supported.
We also predict selective imitation such that CEOs pay more attention to, and thus are more likely to imitate, a tied-to firm’s decision when they have longer tenure as director (Hypothesis 2), when the tied-to firm is more profitable (Hypothesis 3), when the tied-to firm is large relative to the focal firm (Hypothesis 4), and when the focal firm is under conditions of environmental dynamism (Hypothesis 5). The tests of these hypotheses are shown in Models 3, 4, 5, 6, and 7 in Table 2. The coefficient for the interaction term between the tied-to firm’s R&D intensity and CEO directorship tenure is positive, and the relationship is statistically significant in Model 3 and Model 7 (p < .001), indicating support for Hypothesis 2.
In Hypothesis 3, we predict the moderating effect of a tied-to firm’s profitability on the CEO’s imitative decision making. Both Model 4 and Model 7 (p < .001) provide supporting evidence for our hypothesis, indicating that the positive relationship between industry-adjusted R&D intensity and a tied-to firm’s R&D intensity is stronger when a tied-to firm is more profitable.
In Hypothesis 4, we predict that CEOs are likely to imitate firms that are larger than the focal firms, which is consistent with the logic of institutional theory. In contrast to our hypothesis, the coefficient of the interaction terms of a tied-to firm’s R&D intensity and relative size difference is negative and significant in both Model 5 (p < .05) and Model 7 (p < .001). This finding suggests that CEOs may imitate smaller firms (as opposed to larger counterparts) when they make R&D investment decisions.
Finally, Hypothesis 5 predicts that CEOs are more likely to imitate their tied-to firm’s decisions when they are in highly uncertain environments. However, the coefficient of the interaction terms of a tied-to firm’s R&D intensity and industry dynamism is not statistically significant. Therefore, we do not find support for Hypothesis 5.
In order to understand the moderating effects, we plotted the relationship between the focal firm’s R&D intensity and the tied-to firm’s R&D intensity on the basis of the fully specified estimation. We graphed the interaction effect at different levels of CEO directorship tenure (see Figure 1), tied-to firm’s ROA (see Figure 2), and relative size difference (see Figure 3). In these figures, high (and low) levels of a moderator or independent variable are plus (and minus) 1 SD from the mean. The vertical axis represents the industry-adjusted R&D intensity of the focal firm, which is specified as a Z score.
Moderating Effect of CEO Directorship Tenure
Moderating Effect of Tied-to Firm’s Profitability
Moderating Effect of Relative Size Difference
Figure 1 indicates that strategic imitation effects are stronger when CEOs have longer tenure as director. However, when CEOs have shorter tenure, such imitative behavior does not exist (i.e., the line has a negative slope for short tenure of CEO directorship). Similarly, in Figure 2, CEOs show a greater tendency to imitate a tied-to firm’s R&D investment decisions when the tied-to firms are more profitable. In contrast, CEOs do not seem to imitate the R&D spending decisions of less profitable tied-to firms. Rather, they show an increased tendency to move in the opposite decision: If unprofitable tied-to firms have higher (lower) R&D intensity, CEOs tend to have lower (higher) R&D intensity. Figure 3 shows a positive relationship between the industry-adjusted R&D intensity of a focal firm and a tied-to firm’s R&D intensity when the tied-to firms are smaller than focal firms. On the other hand, the slope for larger tied-to firms suggests that their R&D spending decisions may have less influence on the focal firms’ decisions. This finding, which is counter to our hypothesis, is discussed in greater detail in the Discussion section.
Supplemental Analyses
We also conducted a number of supplemental analyses to assess the robustness of our findings. First, to control for time and industry effects on R&D investment decisions, given a multiyear, multi-industry sample structure, we ran a set of regression analyses using year and industry dummy variables. This analysis did not change the pattern of findings reported in Table 2 and is available from the authors. Second, we conducted both random- and fixed-effects regression models on the basis of the assumption of a lack of endogeneity in our sample. Our findings for Hypotheses 1, 2, 3, and 4 are all confirmed across these alternate regression methods but with slightly different levels of significance (but all still significant at p < .05 or smaller). Lastly, we also created an alternative inverse Mills ratio, using the probability that CEOs have any outside directorships of other firms (including firms without R&D investment reporting), as a dependent variable in the first stage. The results from the second stage regression models are not different from the reported tables.
Discussion
In this study, we examined the association between firms’ R&D spending decisions and their CEOs’ board service at other firms (e.g., tied-to firms). We found that CEOs imitate the R&D spending of tied-to firms in their own firm’s R&D decisions. Imitation may occur when innovations are not well understood. Thus, CEOs are likely to consider other firms’ behaviors when they make risky strategic decisions, such as R&D investment. In this sense, uncertainty and risk reduction may come from following other firms’ prior decisions. If a tied-to firm’s executives feel confident enough in the market conditions to spend more heavily on R&D, that situation might provide a vital cue to a CEO sitting on the tied-to firm’s board that his or her own firm can spend more in the future. Conversely, when a firm reduces its spending, it might send a signal to a well-placed observer that future returns from R&D spending may be less. While we are not arguing that CEOs blindly imitate other firms’ decisions, our data suggest that they do pay attention to the actions of other firms where they have information about R&D spending decisions.
More importantly, we find that CEOs are likely to pay more attention to signals that are salient—the imitating relationships are stronger when the CEO has had a longer tenure as director and the tied-to firm is financially successful. We also find that CEOs imitate smaller tied-to firms, in contrast to what is suggested by conventional institutional theory. A CEO’s service on another firm’s board with extended tenure increases familiarity, and such familiarity, in turn, makes the tied-to firm’s past decisions more salient in the CEO’s mind. CEOs pay more attention to familiar sources of information and give more weight to that information in complex decision-making processes, such as R&D investment decisions. In addition, when tied-to firms are financially successful from their past decisions, a CEO may think that imitating their decisions can be justified in the strategic context. As a result of the increased salience in the CEO’s mind as vivid informational cues, the decisions of financially successful firms increase the likelihood that their R&D spending patterns are likely to be imitated by CEOs. Hence, decisions made by familiar and financially successful tied-to firms are perceived as better role models to imitate in R&D spending decisions.
These findings highlight the main contribution of our study to strategic imitation, decision-making research, and social learning perspective. As discussed previously, to our knowledge, no studies have examined selective imitation through CEOs’ external board ties. While imitation through board ties has been examined by others (e.g., Davis, 1991; Haunschild, 1993; Westphal et al., 2001; Westphal & Zajac, 1997), these studies have generally modeled all board ties as having equal potential to influence strategic imitation. As such, we add to this research stream by finding that the type of linked firm matters for strategic imitation. Some tied-to firms may draw more attention than others from top executives. While some cluster effects strategic imitation studies do examine selective imitation (e.g., Barreto & Baden-Fuller, 2006; Haveman, 1993; Williamson & Cable, 2003), these studies (1) do not examine board connections between firms and (2) do not really model the social process by which imitation may occur. Therefore, this study may be the first to examine selective imitation through the social process of board ties. If not all ties have the same effect on strategic decision making, it suggests the need for further research delving into different types of board ties and trying to replicate or extend our results both inside and outside the boundaries of strategic imitation issues. Thus, scholars examining tied-to effects may want to explore further the idea of selective imitation.
On a different topic, our study shows one finding that is counter to the cluster effects studies of strategic imitation (e.g., Burns & Wholey, 1993; Gilbert & Lieberman, 1987; Haveman, 1993). Our sample CEOs imitated smaller tied-to firms when they made R&D investment decisions. Why would CEOs imitate smaller tied-to firms’ decisions? While the nature of our data allows us to speculate only, this finding could stem from the perception that small firms may be more innovative than large firms. Larger firms are generally associated with economies of scale, market power, and organizational slack. In contrast, small firms have advantages, such as flexibility, agility, speed, and risk-seeking behavior (Chen & Hambrick, 1995; Fiegenbaum & Karnani, 1991). These attributes are normally associated with innovation, which is the intended objective of R&D investment. Thus, CEOs will pay greater attention to the decisions of smaller (rather than larger) tied-to firms, assuming the decisions of smaller and possibly more innovative firms provide better information for CEOs to incorporate into their own firms’ R&D spending decisions. Therefore, this finding implies that the strategic context (i.e., the kind of strategic decision being made by a firm) can determine which tied-to firms are desirable models to imitate.
An interesting way to contrast the theoretical notions of legitimacy versus attention is to consider the findings for strategic imitation of smaller firms. Institutional theory’s focus on mimicking larger firms has come from the assertion that organizations seek to become more like the norm for a group of referent organizations. Accordingly, large firms are imitated to the extent that they are seen as representative of a socially desirable group. However, decision makers pay more attention to the decisions of tied-to firms that produce outcomes that are normatively appropriate for the focal firm. For example, our data show that profitable firms are imitated. Being profitable is normatively appropriate for publicly traded firms; thus, it makes sense to pay more attention to profitable firms. However, since large firms tend to be more inertial (Hannan & Freeman, 1984), spend more resources on process versus product R&D (Cohen & Klepper, 1996), and may have difficulty introducing large-scale product innovations (Ettlie & Rubenstein, 1987), it may be more normatively appropriate for CEOs to devote more attentional resources to the smaller firms on whose boards they serve. While the nature of our data does not allow a more fine-grained examination of this assertion, the idea of what makes a focal firm a candidate to be copied may go well beyond the traditional ideas of legitimacy.
Our findings also suggest that organizations learn from the experience of others by observing and imitating the decisions of their tied-to firms. Thus, this study may complement existing organizational learning models (e.g., Argote, 1999; Greve, 2005; B. Levitt & March, 1988) by showing how CEO directorships promote social learning. Innovation-seeking behavior through R&D investment is influenced by firms’ current performance, existing knowledge, and routines (e.g., Helfat, 1997), but the prior decisions of the tied-to firms also shape the firms’ search and learning behavior. CEOs’ social ties influence their access to different sorts of information, thus affecting organizational-level strategic decision making. Specifically, executives’ directorship ties can promote discriminant social learning depending on the attributes of tied-to firms. CEOs not only observe the prior decisions of their tied-to firms but also assess tied-to firms’ outcomes vicariously and try to benefit from the lessons they have learned. This learning perspective suggests that CEOs are selective in the strategic imitation process (i.e., selective imitation); thus, learning occurs only when it makes sense to learn from others and doing so can be justified in terms of strategic contexts.
In addition, while our study’s focus is on strategic imitation through board ties, our use of R&D investment as a dependent variable has implications for the innovation management literature. While prior studies have long focused on the manager’s role in setting the level of R&D spending (e.g., Barker & Mueller, 2002), we expand that focus to the CEO’s social ties with other firms as a source of information in making strategic decisions. Our study indicates that the strategic imitation perspective adds explanatory power to existing explanations of the determinants of R&D investment. To date, relatively little consideration has been given to exploring the effects of the social context surrounding firms and top managers on R&D expenditures. Indeed, executives’ social ties are an important factor in determining firms’ innovation-seeking behaviors.
In sum, consistent with the attention-based view of the firm, our findings suggest that CEOs are selective when they imitate the actions of tied-to firms. Not all CEO ties have equal importance in terms of imitative decision making; thus, they have different strategic implications and values. Given the constraints of managers’ bounded rationality (Cyert & March, 1963), blind imitation of other firms could be only a suboptimal strategy (e.g., Ordanini et al., 2008). The optimal R&D investment decision for firms is not easy. Thus, CEOs may try to “rationalize” imitation by selectively paying attention to the tied-to firm’s decisions (e.g., mimicry of more familiar, more profitable, or smaller firms) that can be justified within the strategic context of R&D spending decisions. In this rationalizing process, managerial attention functions as a filter through which CEOs identify role models to imitate.
Practical Implications
As discussed earlier, empirical studies of the effects of CEOs’ characteristics and compensation on innovation or R&D spending have been a staple in the innovation management literature leading to a focus on the backgrounds and experiences of executives (e.g., Barker & Mueller, 2002; Daellenbach et al., 1999; Fong, 2010; Wu & Tu, 2007). Yet most of this research suggests that career experiences, education, and compensation are strong drivers of a CEO’s R&D spending or innovation decisions. We present evidence that organizational decisions are affected by the CEOs’ outside directorships, and some tied-to firms garner greater attention by being salient to CEOs.
Our results confirm that hiring and developing CEOs is important in order to have the best-matched CEO with a firm’s innovation strategies. Ideally, CEOs should be selected with consideration of their external social ties, as well as their backgrounds, which may affect their cognitive models. Specifically, as our results indicate, the role of CEOs’ external ties will not be the same in every situation; rather, the role of such ties will vary according to the strategic context. Therefore, CEOs’ external ties, their managerial attention, and firms’ strategic contexts should receive more attention.
Limitations
This study is not without limitations. First, our modeling was limited by using archival data. While archival data are generally the norm in studies of strategic imitation, can decision-making processes like imitation be successfully examined with archival data? 3 We believe they may be modeled with archival data if researchers are careful to develop theory-driven models and control for some of the potential problems, such as possible reverse causality and endogeneity issues. Similar to other streams of research in strategic management and organization theory, it is often difficult for researchers to directly see strategic decision processes at firms (e.g., Godfrey & Hill, 1995). However, from a realist perspective, carefully designed, theory-driven studies that test hypotheses with visible outcomes, and whose outcomes are consistent with what the theory predicts on how invisible elements should act, add value to scientific endeavors. Such studies form an empirical base for supporting or rejecting a model or theory with hard-to-see elements. We believe our study adds to knowledge about directorship ties in that regard. It is a theory-consistent extension of existing perspectives on strategic imitation through social ties.
In addition, we tried to rule out alternative explanations for the findings with study design elements and statistical modeling. First, all independent variables were lagged by 1 year, so they occurred in time prior to the dependent variable. Second, we performed a Heckman selection procedure to reduce the likelihood of biased sampling driving the results. Third, we controlled for the possible endogeneity issues through Hausman–Taylor estimation of the regression coefficients. Fourth, by industry-norming R&D spending at both the focal and tied-to firms, we avoided the artifactual finding that firms with high (low) R&D spending have CEOs who serve on boards in similar industries with high (low) R&D spending per firm. Because we industry-normed the R&D spending variable, we are instead capturing the possible copying of the within-industry strategic posture (in particular, R&D spending) of tied-to firms, rather than just industry similarity. Finally, the moderated results do not make sense theoretically when one considers their findings in the context of reverse causality. Thus, while we can never rule out reverse causality due to the nature of our archival data and our inability to view the imitative process up close, an abundance of precautions makes this a highly unlikely explanation of our findings.
One could argue that strategic decision making should be examined at a different type of analysis, possibly through qualitative methods (e.g., Eisenhardt, 1989; Mintzberg, Raisinghani, & Theoret, 1976). While our study lacks the rich theoretical explanations that may emerge from qualitative data analyses, it does have empirical testing of theory-based hypotheses that can inform strategic imitation and decision-making researchers. Thus, going forward, researchers can test hypotheses about attention, board ties, and strategic imitation with methods that may replicate, extend, or falsify the theoretical mechanisms in this study.
Another limitation of this study is the lack of diverse interfirm connections. We measured the effects of CEO outside directorships only. Therefore, this study does not capture the effects of directors of focal firms who might be the top managers of other firms. In addition, we do not consider types of social ties other than directorships. Director ties are not likely to be the only information conduit through which strategic imitation occurs. For example, hiring managers from other firms, joining trade associations, and having professional association ties may also function as avenues of imitation. Finally, our study, in terms of its sample, is based on observations from large Fortune 500 manufacturing firms. Therefore, increasing the sample size (e.g., including smaller firms, nonmanufacturing firms, and non-U.S. firms) will help increase the generalizability of the findings.
Conclusion
CEO outside directorships have long been regarded as a mechanism through which strategic imitation occurs. In this study, we have shown that CEO outside directorships are one channel for firms to imitate other firms’ strategic decisions. On the basis of the attention-based view of the firm, we show that certain social ties may attract more attention in imitation and find evidence of selective imitation, as the strategic imitation in R&D investment decision is stronger when the CEO has longer tenure as director of a tied-to firm and the tied-to firm is performing well. In contrast to conventional institutional theory arguments, we also find that CEOs imitate smaller tied-to firms. Our study illustrates how the influence of a CEO’s outside directorship on strategic imitation is moderated by the situations of the tied-to firm, thus leading to selective imitation. From these results, we make several contributions to the existing research on strategic imitation, social learning, and innovation. Not all outside directorship ties of CEOs have equal influence on imitative strategic decision making; thus, they have different strategic implications and values.
Footnotes
Appendix
Acknowledgements
This article was accepted under the editorship of Patrick M. Wright. An earlier version of this paper was presented at the 2012 Academy of Management Conference in Boston. We would like to thank the action editor, Dr. Sucheta Nadkarni, as well as two anonymous reviewers for their insightful comments. Additionally, we thank Dave Wangrow, Karl Kammerer, and Zheng Cheng for their comments. Any errors or omissions are the authors’ responsibility alone.