Horizon problem and firm innovation: The influence of CEO career horizon, exploitation and exploration on breakthrough innovations

Research Policy xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Research Policy journal homepage: www.elsevier.com/locate/respol

Horizon problem and firm innovation: The influence of CEO career horizon, exploitation and exploration on breakthrough innovations Sam Yul Choa, Sang Kyun Kimb, a b

⁎

Department of Strategy and Entrepreneurship, College of Business, Oregon State University, Corvallis, OR 97331, United States Department of Organization and Human Resources, School of Business, Sungkyunkwan University, Seoul, 03063, Korea

A R T I C L E I N F O

A B S T R A C T

JEL classification: M12 O31 O32 D23 D81

Building on labor market evaluations and legacy conservation motivation perspectives, we propose a mechanism to explain the relationship between CEO career horizons and breakthrough innovations. Using 10-year panel data from 681 U.S. firms, we find that firms that have a CEO with a short career horizon tend to produce fewer breakthrough innovations. We also find that the relationship between CEO career horizon and breakthrough innovation is partially mediated by R & D spending, and also moderated by organizational learning behavior (exploration vs. exploitation). This study highlights how a CEO’s motivation to protect success in the short term affects the firm’s innovativeness.

Keywords: CEO career horizon Breakthrough innovation Risk aversion Legacy conservation Labor market evaluation Organizational learning behavior

1. Introduction Strategy scholars view a company's chief executive officer (CEO) as the most powerful actor in a firm; as such, they have examined how CEOs affect a firm’s strategic decisions and subsequent performance (e.g., Chatterjee and Hambrick, 2011; Hayward and Hambrick, 1997; Hayward et al., 2004). Not surprisingly, studies have also explored how such decisions and performance are influenced by certain CEO-related factors, including cognition, personalities, and levels of compensation (Chatterjee and Hambrick, 2011; Hayward and Hambrick, 1997; Seo et al., 2014). In order to examine CEO influence on firm performance, the literature has begun to shed light on the impact of a CEO's “career horizon,” a term that encompasses the time left as a CEO approaches retirement (Kang, 2015; Matta and Beamish, 2008), on a firm’s strategic decisions. Researchers have found that CEOs with a short career horizon tend to choose risk-averse strategies, resulting in a CEO ‘horizon problem’ (Hambrick and Mason, 1984). This line of research has shown that a CEO’s career horizon significantly affects such risky strategies as acquisitions (Matta and Beamish, 2008), research and development (R & D) (Barker and Mueller, 2002; Zona, 2016), capital expenditures (Cheng, 2004; Dechow and Sloan, 1991), and corporate social

⁎

responsibility (CSR) (Kang, 2015). Drawing on the literature, this study explores the impact of career horizon on breakthrough innovations. A breakthrough innovation has been defined as a subset of innovation that can change the competitive landscape and create new market opportunities (Abernathy and Utterback, 1978; Gatignon et al., 2002). While breakthrough innovations are an important factor for creating value (e.g., Ahuja et al., 2005; Hall et al., 2005; Phene et al., 2006) and are influenced by how CEOs manage innovation streams (e.g., Galasso and Simcoe, 2011; Hirshleifer et al., 2012; Smith and Tushman, 2005; Tang et al., 2012; Wu et al., 2005), the literature on CEO career horizons has paid scant attention to the influence on breakthrough innovations and the boundary conditions of this relationship. This study fills this gap by examining important research questions: How does a CEO’s career horizon affect a firm’s ability to generate breakthrough innovations? How do a firm’s organizational learning activities (exploitation and exploration) contribute to the relationship between CEO career horizon and breakthrough innovations? Building on labor market evaluations and legacy conservation motivation perspectives that explain risk aversion by CEOs facing a short career horizon, our study seeks to unpack the mechanisms linking a CEO's career horizon to a firm's breakthrough innovations. Specifically,

Corresponding author. E-mail addresses: [email protected] (S.Y. Cho), [email protected] (S.K. Kim).

http://dx.doi.org/10.1016/j.respol.2017.08.007 Received 15 May 2016; Received in revised form 24 August 2017; Accepted 26 August 2017 0048-7333/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Cho, S.Y., Research Policy (2017), http://dx.doi.org/10.1016/j.respol.2017.08.007

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

conserve his legacy: greater equity ownership and unexercised in-themoney options would amplify concerns over success, motivating CEOs to avoid taking risky actions (Matta and Beamish, 2008). The degree to which a CEO pays attention to the managerial labor market may also affect leaders with a short career horizon. Pressure from the labor market would likely induce CEOs to avoid making risky investments, since they would want to protect their job prospects (Kang, 2015). Committing to breakthrough innovation is a risky strategy that may jeopardize a firm’s short-term performance, due to its uncertain outcomes (e.g., Anderson and Tushman, 1990; Teece, 1986). Although breakthrough innovations can create substantial value (Hall et al., 2005), contribute to a CEO's reputation and legacy (Heider, 2013; Ross, 1977), and be considered a credible signal of a CEO’s ability (Westley and Mintzberg, 1989), delivering such innovations requires a firm to combine internal and external resources and capabilities, making it difficult for the firm to generate profits quickly (Ahuja and Lampert, 2001; Hill and Rothaermel, 2003; Phene et al., 2006). Thus, the financial return on a commitment to breakthrough innovation is not immediate, with further innovations requiring substantial time and resources. Pursuing such innovations would be considered a risky move for CEOs facing retirement, since a firm commitment to breakthrough innovations may harm profits and put executive prospects and shortterm success at risk. For these reasons, protecting their own success may dampen the willingness of CEOs with a short career horizon to allocate resources toward risky innovation strategies. From our research context, theoretical arguments would suggest that a CEO with a short career horizon (i.e., older CEO) is less likely to deliver breakthrough innovations. Thus, we propose the following hypothesis:

we propose that a short career horizon induces a CEO to become risk averse and thus forego investing in risky breakthrough innovations as they could harm the firm's short term performance, endangering job prospects and CEO legacies in the short term. We also suggest that the impact of a short career horizon on breakthrough innovations is partially mediated by a reduction in R & D spending. Furthermore, different performance implications associated with a firm’s exploitation and exploration activities affect a CEO's willingness to commit to such breakthrough innovations. That is, when a firm leverages internal knowledge within a familiar technological domain (i.e., a focus on exploitation), this mitigates the behavioral tendencies of a CEO with a short career horizon, such as not pursuing breakthrough innovations, while such behavioral tendencies are exacerbated when a firm applies external technologies with an unfamiliar trajectory (i.e., a focus on exploration). This study contributes to the literature on CEO decision making and firm innovation. First, as a contribution to literature on CEO decision making (Chatterjee and Hambrick, 2011; Hayward et al., 2004; Seo et al., 2014), our study explains the mechanisms of how and why a CEO’s short career horizon influences firm innovations. This is consistent with prior studies that have examined the impact of myopic CEOs with a short career horizon on risky strategies (Barker and Mueller, 2002; Cheng, 2004; Dechow and Sloan, 1991; Matta and Beamish, 2008; Zona, 2016). Second, our study contributes to innovation literature by emphasizing the important role of CEOs in breakthrough innovation, which is consistent with recent studies that have explored the impact of CEOs on firm innovation (Galasso and Simcoe, 2011; Hirshleifer et al., 2012; Tang et al., 2012; Wu et al., 2005). Finally, this research provides practical implication to people in the field who desire to develop firm innovation and manage innovation process. Our study highlights the finding that a strategic alignment between a CEO’s willingness to take risks and a firm’s engagement in exploitation and exploration enables the firm to achieve greater firm innovations. Hence, our major contribution to the literature is how a CEO’s motivational factor plays an important role in decision making for innovation strategy. The rest of this paper consists of five sections. The next section develops a conceptual model that examines a mechanism that links a CEO’s career horizon to firm innovation based on our review on the theoretical and empirical studies. Section three presents our methodology, followed by an interpretation of our findings. In section five we discuss implications for theory and practice, elucidating potential directions for future research. This paper ends with concluding remarks about the study of CEO influence on firm innovation.

Hypothesis 1. CEOs with a short career horizon (i.e., older CEOs) have a negative effect on breakthrough innovations. 2.2. CEO career horizon, R & D spending, and breakthrough innovations The literature on how CEOs affect a firm’s strategic decisions suggests that broad discretion for choices and actions is afforded to CEOs (e.g., Hambrick and Finkelstein, 1987) and that CEO-related factors shape their attitude toward risk, which could affect their choices regarding risky strategies (e.g., Chatterjee and Hambrick, 2011; Hayward and Hambrick, 1997; Hayward et al., 2004). Drawing on this perspective, we propose a partial mediation model where CEO career horizon can affect R & D spending and in turn, how this affect the way a firm performs in terms of breakthrough innovations. R & D spending has been shown to be one of discretionary decisions monitored and controlled by CEOs (e.g., Barker and Mueller, 2002; Cheng, 2004; Dechow and Sloan, 1991; Zona, 2016), although the levels of R & D expenditure are relatively stable over time (Hambrick et al., 1983). Some management scholars have suggested that CEOs with a short career horizon actually reduce a firm’s spending on R & D. Those empirically examining the influence of CEO age on R & D spending have found that older CEOs are likely to reduce a firm’s R & D spending (Barker and Mueller, 2002), which suggests that older CEOs who are approaching retirement would reduce expenditures in order to pursue their own interests (Barker and Mueller, 2002; Zona, 2016). Similarly other scholars claim that CEOs approaching retirement age are less willing to maintain their firm’s commitment to R & D since the benefit of long-term expenditures are unlikely to be realized in the short term (Cheng, 2004; Dechow and Sloan, 1991). Thus, given the latitude that CEOs receive, CEOs with a short career horizon are likely to reduce discretionary R & D expenditure, due to a concern about short-term success. CEOs can influence other managerial decisions that affect a firm’s ability to generate innovation in different ways (Damanpour, 1991). CEOs influence innovation by paying attention to forthcoming events and communicating these to the rest of the firm (Yadav et al., 2007). CEOs also affect firm innovation through the establishment of an

2. Theory and hypotheses 2.1. CEO career horizon and breakthrough innovations CEO career horizons represent the time it takes for a CEO to reach retirement age (Kang, 2015; Krause and Semadeni, 2014; Matta and Beamish, 2008). A shorter career horizon represents a CEO who is nearing retirement. The literature suggests that a decision maker who is facing a short career horizon tend to become risk-averse, based on a desire to protect his or her success in the short term. A short career horizon likely leads a decision maker to avoid taking risks that amplify uncertainty over short-term firm performance. Since CEOs with a shorter career horizon have a limited amount of time to recoup investments and reverse performance shortfalls, they are likely to forego making risky long-term investments which could hurt firm performance in the near term. Prior studies on career horizons suggest that a CEO’s concern over legacy conservation and labor market evaluations may influence a CEO’s decision making and firm risk behavior (Kang, 2015; Matta and Beamish, 2008). Thus, the behavioral tendencies of a CEO with a short career horizon may depend on the degree to which he is willing to 2

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

stemming from exploitation tend to be more stable and proximal. Given the performance outcomes and unique nature of each organizational learning behavior, we posit that the degree to which a firm engages in exploration or exploitation shapes a CEO's commitment to firm innovation by affecting their concern about legacy conservation and labor market evaluations.

organizational structure that utilizes knowledge resources across business units (Ettlie et al., 1984). Since there is high uncertainty associated with delivering innovation (Anderson and Tushman, 1990; Teece, 1986), CEOs’ risk tolerance affects a firm’s innovation process it engages in (Hage and Dewar, 1973). Thus, CEO career horizon can affect the way a firm performs regarding breakthrough innovations, through adjusting R & D spending and other strategic decisions. In order to justify the partial mediation hypothesis, we explain how R & D spending can affect firm innovations. Prior studies have shown that a firm’s R & D spending is one of the most important strategic inputs that directly contributes to firm innovation (Ahuja et al., 2008). A firm that devotes more resources to R & D is likely to deliver more patent outputs (e.g., Ahuja and Lampert, 2001; Griliches, 1990) and develop better capabilities of acquiring new ideas, learning new technologies, and combining internal and external resources to generate more innovative output (e.g., Cohen and Levinthal, 1990). These studies show a positive relationship between R & D expenditures and innovations, suggesting that a reduction in R & D spending results in a lower level of innovation. Based on our proposition that CEOs with a shorter career horizon may reduce their commitment to delivering breakthrough innovations (hypothesis 1), that a reduction in R & D spending may be negatively related to firm innovations, and that other strategic decisions that CEOs make may be associated with firm innovations, we suggest that R & D spending may partially mediate a relationship between a shorter CEO career horizon (i.e., older CEOs) and a firm’s breakthrough innovations. Therefore, we advance the following hypothesis:

2.3.1. The moderating role of exploitation Exploitation involves leveraging existing knowledge within a familiar domain of technological trajectories. It allows a firm to increase efficiency by relying on available technologies, reducing errors and rates of failure (Levinthal and March, 1981). Reusing knowledge in a proximate technological trajectory may make organizational learning more reliable, generating more predictable short-term performance outcomes. Although exploitation enables a firm to enhance its efficiency within a specific domain of technology, repeating knowledge in a narrow domain can cause core rigidity and inertia, which could reduce a firm's ability to adapt to technological change (Benner and Tushman, 2003; Leonard, 1992) and lead to competency traps (Levinthal and March, 1993). An overemphasis on exploitation can thus hurt a firm’s ability to generate breakthrough innovations. These arguments suggest that exploitation can generate predictable short-term performance but can be detrimental to a firm’s accessing breakthrough innovations. CEOs with a shorter career horizon may wish to take advantage of the benefits of exploitation in order to preserve their short-term success. Stable and predictable returns based on variance-reducing exploitation can withstand uncertainty or fluctuation in firm performance in the near term, mitigating CEO concerns about preserving short-term success. CEOs with a shorter career horizon may consider exploitation as an opportunity to preserve their legacy of success without compromising short-term financial performance. Because older CEOs have deeper knowledge and experience in specific technological trajectories (Hambrick and Mason, 1984) and they would be willing to utilize their expertise to preserve their success, the firm may identify valuable knowledge about a specific domain which in turn contributes to firm innovations. Thus, when a firm engages in exploitation activities, a CEO with a short career horizon may take advantage of technological exploitation and mitigate a detrimental impact on breakthrough innovations. Together, these theoretical arguments suggest that CEOs with a short career horizon are less likely to commit firm resources to generating breakthrough innovations (hypothesis 1). However, a firm’s greater exploitation may mitigate CEO concerns about firm performance in the short term and enable CEOs to utilize their knowledge of a technology domain to preserve their short-term success. We suggest that the negative influence on breakthrough innovations of CEOs approaching retirement may be mitigated given a greater degree of exploitation. Thus, the following hypothesis is proposed:

Hypothesis 2. The negative relationship between CEOs with a short career horizon (i.e., older CEOs) and breakthrough innovations is partially mediated by R & D spending. 2.3. CEO career horizon, exploitation and exploration, and breakthrough innovations The literature on CEO career horizons provides theoretical and empirical guidance that suggest that legacy conservation concerns and labor market evaluations may contribute to the behavioral tendencies of CEOs who are facing a short career horizon (Kang, 2015; Matta and Beamish, 2008). Drawing on these studies, we consider whether the extent to which a firm engages in exploitation and exploration shapes the influence of CEOs facing a short career horizon on firm innovation (Cyert and March, 1963; March 1991), since how a firm leverages its internal and external technological knowledge can cause a CEO concern about securing future prospects and preserving short-term success. March (1991) proposed two types of organizational learning behaviors: exploitation and exploration. ‘Exploitation includes such things refinement, choice, production efficiency, selection, implementation, execution [while] exploration includes things captured by terms such as search, variation, risk taking, experimentation, play, flexibility, discovery, innovation concerted variation, planned experimentation and play’ (March, 1991, p.71). Scholars have conceptualized the relationship between exploitation and exploration as two ends of a continuum versus an orthogonal choice (Gupta et al., 2006). By following past research (Baum et al., 2000; Beckman et al., 2004; Katila and Ahuja, 2002; Uotila et al., 2009), this study views exploration and exploitation as distinct organizational learning activities (i.e., orthogonal choice). Specifically, we follow the conceptualization and operationalization envisioned by Katila and Ahuja (2002), who use search scope (citing new patents) and search depth (citing existing patents) to approximate exploration and exploitation. Proponents of the view that exploration and exploitation are distinct activities have argued that they rely on different organizational structures, processes, resources, and capabilities (Deeds et al., 2000; Dosi et al., 2000; He and Wong, 2004; Jansen et al., 2006), resulting in different performance outcomes. Those associated with exploration are more uncertain and distant, while those

Hypothesis 3. A firm’s exploitation moderates the negative relationship between CEOs with a short career horizon (i.e., older CEOs) and breakthrough innovations, such that the negative relationship between CEOs with a short career horizon and breakthrough innovations is weaker for firms where the degree of exploitation is high and stronger for firms where the degree of exploitation is low. 2.3.2. The moderating role of exploration In comparison, exploration seeks knowledge that goes beyond existing technology domains and across technologies. It allows a firm to enrich its knowledge pool by discovering new technologies (March, 1991), enabling it to produce innovations by combining new technologies with existing ones (Nelson and Winter, 1982) which results in breakthrough innovations. While exploration may promote breakthrough innovations and positively affect CEO reputations and 3

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

after winsorizing. This allowed us to build 10-year panel data. A final sample consisted of 4691 observations from 681 U.S. firms during 1992–2001. Average observation per firm was 6.9 years.

shareholder values (Hall et al., 2005), it can be costly to develop technologies in an unfamiliar domain (Levinthal and March, 1993) and increase the complexity and decrease the reliability of knowledge integration (Katila and Ahuja, 2002). Experimenting with knowledge in technologically distant trajectories also brings about more variation in short-term performance and makes it difficult for firms to generate quick profits, endangering success in the near term (He and Wong, 2004; McGrath, 2001). Given a retiring CEO's limited amount of time to make profitable returns on exploratory activities and recover from performance shortfalls, exploration may pose a significant challenge. When exploration inflates uncertainties and fluctuation regarding financial performance, concerns about preserving success will more likely outweigh seeking the long-term benefits associated with breakthrough innovations that might accrue after a CEO's retirement. Such CEOs tend to avoid risk in order to preserve success in the short term rather than enhancing their reputation and firm value in a long run. Nor are they able to absorb technological knowledge within an unfamiliar trajectory (Hambrick and Mason, 1984). Their knowledge and experience with a specific trajectory may hinder a firm’s ability to identify which knowledge is valuable, leading to fewer breakthrough innovations being generated. Therefore, exploration is less likely to motivate those CEOs with a shorter career horizon to commit firm resources toward breakthrough innovations. Overall, it may be expected that CEOs with a short career horizon, who desires to protect a firm's short-term success, are less likely to take risks when a firm engages in greater exploration, even though it could enhance the generation of breakthrough innovations that would benefit the firm in a long run. From our research context, such arguments suggest that CEOs who are approaching retirement age are less likely to deliver breakthrough innovations (as proposed in hypothesis 1), and that the detrimental effects on breakthrough innovations may increase, given a greater degree of exploration. Therefore, the following hypothesis is proposed:

3.2. Dependent variable Breakthrough innovations represent the degree to which a firm develops novel technology that has a high impact on the field. Relying on the previous findings that the number of forward citations is highly associated with the technological importance of a patent (e.g., Trajtenberg, 1990), breakthrough innovations were measured as the number of patents compiled five years after the focal year, belonging to the top five percent in number of forward citations (Ahuja and Lampert, 2001; Phene et al., 2006; Srivastava and Gnyawali, 2011). To operationalize this variable, the number of each patent’s forward citations was divided by the mean value of forward citations based on all patents in the same technological subcategory for each year. Such a fixed effect approach using average-adjusted citations for a group of patents addresses the potential issue of truncation (Hall et al., 2001) which occurs when an earlier patent is cited more often than a recent one. Patents with a value higher than the 95th percentile were considered breakthrough innovations. 3.3. Independent variables CEO career horizon refers to the time leading up to a CEO's retirement. Prior studies on career horizons suggest that CEO age represents a consideration of the time horizon for decision-making. CEO age is used as a proxy measure for career horizons (Matta and Beamish, 2008; McClelland et al., 2012). Consistent with these studies, we measured CEO age in years; it was expected that the older a CEO was, the shorter the career horizon. In addition to age, alternative measurements of career horizon were operationalized as dummy variables, coded 1 if a CEO for firm i in year t was greater than or equal to 60, 61, 62, 63, 64 and 65 years of age, and zero otherwise (Brickley et al., 1999).

Hypothesis 4. A firm’s exploration moderates the negative relationship between CEOs with a short career horizon (i.e., older CEOs) and breakthrough innovations, such that the negative relationship between CEOs with a short career horizon and breakthrough innovations is stronger for firms where the degree of exploration is high and weaker for firms where the degree of exploration is low.

3.4. Mediating variables The mediating variable was R & D spending, and was operationalized as the log value of total R & D spending subtracted from the industry’s average R & D spending. To calculate average R & D spending, industries were defined using North American Industry Classification System (NAICS) codes at the two-digit level. The alternative measurement, R & D intensity, was operationalized as the ratio of R & D expenditure over total revenue and then adjusted by the industry’s average R & D intensity (Schilling and Phelps, 2007).

3. Methods 3.1. Data and sample Data were retrieved from the ExecuComp database and National Bureau of Economic Research (NBER) patent citations file. The ExecuComp database contains compensation data for top executives, and the NBER patent file compiles data on patents, application years, and citations from 1975 to 2006 (Hall et al., 2001). We initially identified CEOs for all firms listed in the ExecuComp database who served from 1992–2001, and matched firm information to patent filings during 1987–2006 as well as financial information obtained from COMPUSTAT. Following prior studies that examined CEO influence on firm outcomes (Galasso and Simcoe, 2011; Hirshleifer et al., 2012; Kang, 2015; Krause and Semadeni, 2014; Matta and Beamish, 2008; Smith and Tushman, 2005; Tang et al., 2012; Wu et al., 2005), our study introduces a lag structure into the regression model to account for time lag between CEO decisions and firm outcomes. CEO career horizons, R & D spending, and other variables were measured for focal year (t), while exploration and exploitation variables were based on patent backward citations for the past five years (t–5 ∼ t-1) and breakthrough innovations were measured using patent forward citations up to three years after the focal year (t + 1 ∼ t + 3). As in previous studies, outliers were replaced using 99th or 1 st percentiles, and mean-centered

3.5. Moderating variables Two moderators of interest in this study were exploitation and exploration. Following previous studies (e.g., Benner and Tushman, 2003; Rosenkopf and Nerkar, 2001; Tzabbar and Kehoe, 2014), we used patent data to assess a firm’s learning behavior. Exploitation was operationalized as the ratio of the number of self-citations and repeat citations over total number of backward citations made by the focal firm during the past five years (Benner and Tushman, 2003). This ratio reflects the extent to which a firm’s patenting efforts built on knowledge it has developed and used in the past. Exploration was measured as the ratio of the number of patents applied in the new technological class to the total number of applied patents during the past five years (Tzabbar and Kehoe, 2014). A new technological class was defined as a technological domain in which a firm had not applied for patents over the past five years. This allowed us to assess the extent to which a firm was exploring technology in an unfamiliar domain. 4

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

negative effect on breakthrough innovations by the firm. In Model 4, the coefficient of CEO age was negative and statistically significant (β = −0.006, p < 0.05). We performed additional tests with the alternative measurements of CEO career horizon. The results of these analyses with dummy variables also showed a negative impact for a short career horizon on breakthrough innovation. These results support Hypothesis 1, suggesting that a firm whose CEO has a short career horizon tend to produce fewer breakthrough innovations. Hypothesis 2 proposed that R & D spending partially mediates the relationship between CEOs with a short career horizon and breakthrough innovations. According to Baron and Kenny (1986), several conditions must be met in order to establish the partial mediation suggested by Hypothesis 2. First, CEO age (the independent variable) must be negatively associated with R & D spending (the mediators) as established in Model 2 (β = −0.002, p < 0.05). Second, CEO age must be negatively associated with breakthrough innovations (dependent variable), as established in Model 4 (β = −0.006, p < 0.05; Hypothesis 1). Third, R & D spending must positively affect breakthrough innovations, as confirmed in Model 5 (β = 0.076, p < 0.001). As seen in Model 6, CEO age is negatively associated with breakthrough innovations in the presence of R & D spending (β = −0.005, p < 0.05; β = 0.068, p < 0.001). Finally, the effects of CEO age on breakthrough innovations must be reduced when R & D spending levels are included in the regression equation. This condition was also confirmed when CEO age drops from Model 4 (β = −0.006, p < 0.05) to Model 6 (β = −0.005, p < 0.05). Based on Models 2, 4, 5, and 6, we find support for the argument that R & D spending partially mediates the negative relationship between CEO age and breakthrough innovations. To insure a mediation effect of R & D spending on breakthrough innovations, we conducted additional analyses. First, the results of the Sobel test (Sobel, 1982) show an indirect effect of CEO age on breakthrough innovations (Sobel z-statistic = −4.31, p < 0.001). We also employed STATA’s ‘cmp’ and ‘nlcom’ commands to test for an indirect effect of CEO age on breakthrough innovations. These results also provide evidence of an indirect effect (z-statistic = −1.89, p < 0.059). All of the analyses thus support the proposition that CEOs with a short career horizon have a negative, indirect effect on breakthrough innovations through R & D spending. We performed the same analytical process to test for indirect influence of R & D intensity, the alternative mediating variable, on breakthrough innovations. Consistent with the results of models using R & D spending, a Sobel test (Sobel z-statistic = −11.20, p < 0.001) and STATA’s ‘cmp’ and ‘nlcom’ command test (z-statistic = −2.04, p < 0.041) confirmed an indirect effect of CEO age on breakthrough innovations, supporting Hypothesis 2. Before testing the moderation effect of organizational learning behavior on the relationship of CEO age and breakthrough innovations, we checked the assumption proposed by the orthogonal view of exploration and exploitation in order to justify conceptualization and operationalization of these variables. Following an approach similar to prior studies (He and Wong, 2004; Katila and Ahuja, 2002), we created an interaction term between exploration and exploitation and tested its effect on breakthrough innovations. The coefficient of the interaction was positive and significant (β = 0.452, p < 0.05, Model 10), suggesting that exploration and exploitation represent two distinct organizational learning activities. The finding also confirms the argument that a firm can achieve ambidexterity by undertaking exploitation and exploration simultaneously (Griliches, 1990; Katila and Ahuja, 2002). Upon confirmation of the orthogonal view, Model 7 examines the moderating effect of exploitation on the relationship between CEO age and breakthrough innovations. The coefficient of interaction term between CEO age and exploitation in Model 7 is positive and significant (β = 0.016, p < 0.01), supporting Hypothesis 3. To better understand the moderation effect, we plotted this relationship in Fig. 1, using ± 2 standard deviations from the mean for CEO age and exploitation at the mean values of all other variables. Fig. 1 shows a weaker slope for the

3.6. Control variables In order to exclude alternative explanations, our analysis model contained several control variables at both the CEO level and firm level. At the CEO level, we controlled for CEO influence in a firm by measuring ownership (Makri et al., 2006) and tenure (Wu et al., 2005). CEO ownership was calculated as the number of shares owned by a CEO divided by the number of shares outstanding, while CEO tenure was calculated using the number of years a CEO had served in that role for the firm. We controlled for the influence of CEO compensation by measuring CEO salary, CEO bonuses, and CEO in-the-money options (Matta and Beamish, 2008). Cash compensation for CEOs was measured as the natural logarithm of CEO salary and bonuses. Stock options holdings were calculated by taking the natural logarithm of the Black-Scholes value of vested and unvested in-the-money options held by a CEO. CEO chairman duality was controlled for by establishing that a CEO who holds a chairman position may influence the level of R & D expenditures and overall innovations (Kor, 2006). CEO chairman duality was operationalized as a dummy variable, where 1 denotes a CEO holding the chairman position, and zero otherwise. We also sought to control for other control variables at the firm level. Profitability, measured as return on assets, was chosen because firm performance influences a firm’s R & D efforts and overall breakthrough innovations (e.g., Cyert and March, 1963). Given the previous finding of its positive effect on innovation (Chaney and Devinney, 1992; Damanpour, 1989), firm size, measured as the natural log of total assets, was included. A diversified firm tends to have a mindset both for exploration and exploitation, along with high R & D investment (Chen, 1996) and delivers more innovation (e.g., Kim et al., 2013). We thus controlled for the level of firm diversification by using the entropy measure (Palepu, 1985). Finally, year dummies were included to control for any potential patenting trends over time. 3.7. Estimation method Breakthrough innovation, our dependent variable, had a mean of 10.35 and a standard deviation of 21.85, indicating violation of the equal-dispersion assumption for the Poisson model. To address the issue of over-dispersion, we used negative binomial regression for the panel data in order to test our hypotheses (Cameron and Trivedi, 2009). The Hausman test (Hausman et al., 1984) specifies a fixed effect model over a random effect model (p < 0.001). Therefore, estimates from the negative binomial regression with a fixed effect model for the panel data were reported. Since the fixed effect model has an omitted variable bias at the firm level (Baltagi, 2008) and industry level, we tested negative binomial regression with a random effect model in order to control for a time invariant variable (i.e., industry effects), and found that the estimates were the same as those for the fixed effect model. 4. Results Table 1 presents descriptive statistics and correlation between any pair of the variables in our model. The average age of CEO in the sample was 55.98 years old with a standard deviation of 7.49. The variance inflation factor analysis (VIF) shows that VIF values ranged from 1.00 to 2.47 with a mean of 1.69, indicating no problem with multicollinearity (Neter et al., 1996). Table 2 presents the results of negative binomial regression with a fixed effect model. Models 1 and 2 in Table 2 used R & D spending as the dependent variable, while breakthrough innovation was tested in Models 3 through 10. Consistent with previous studies, control variables for both the firm level and CEO level accounted for a variation of breakthrough innovation in Model 3. Models 4 through 10 show the effect of CEO career horizon and its interaction effect with exploitation and exploration on breakthrough innovation. Hypothesis 1 proposed that CEOs with a short career horizon have a 5

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

Table 1 Descriptive statistics and correlations.a

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.

a

Breakthrough innovations Firm size Firm diversification R & D spending Profitability CEO ownership CEO Salary CEO Bonus CEO in-the-money option CEO tenure CEO chairman duality CEO age Exploration Exploitation Mean S.D. Min Max

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

1.00 0.43 0.10 0.55 0.10 −0.06 0.21 0.15 0.21 −0.07 0.07 −0.03 −0.19 0.21 10.35 21.85 0.00 102.00

1.00 0.21 0.49 0.19 −0.17 0.46 0.34 0.21 −0.15 0.27 0.18 −0.12 0.07 7.40 1.71 0.00 13.37

1.00 0.12 −0.01 −0.06 0.13 0.04 0.08 −0.07 0.06 0.11 −0.04 0.08 0.79 0.78 0.00 3.19

1.00 0.02 −0.09 0.27 0.17 0.26 −0.15 0.13 −0.03 −0.20 0.21 0.91 1.89 −4.10 6.93

1.00 0.06 0.10 0.33 0.19 0.07 0.05 0.11 0.02 −0.05 0.04 0.12 −0.58 0.25

1.00 −0.14 −0.11 −0.22 0.26 0.06 0.07 0.02 −0.06 1.97 5.17 0.00 33.98

1.00 0.32 0.14 −0.06 0.18 0.15 −0.08 0.10 6.26 0.83 0.00 8.29

1.00 0.34 −0.06 0.12 0.07 −0.04 0.02 5.12 2.52 0.00 10.68

1.00 −0.11 0.06 −0.10 −0.05 0.09 11.35 6.52 0.00 28.69

1.00 0.09 0.24 0.05 −0.07 0.69 0.46 0.00 1.00

1.00 0.26 −0.03 0.01 9.95 18.31 0.00 99.00

1.00 −0.01 0.03 55.98 7.49 29.00 87.00

1.00 −0.31 0.20 0.29 0.00 1.00

1.00 0.33 0.27 0.00 0.92

obs = 4691; n = 681; p < 0.05 for bold.

relationship between CEO age and breakthrough innovations when a firm engages in more exploitation. This indicates that the detrimental influence of short CEO career horizon on a firm’s breakthrough

innovation is reduced as the degree of exploitation increases, and supports the argument that greater exploitation mitigates CEO concerns about labor market evaluations and legacy conservation, reducing the

Table 2 Results of negative binomial regression with fixed effect model for the relationship between CEO age, R & D spending, exploitation, exploration and breakthrough innovations.a Dependent variable

Intercept Firm size Firm diversification Profitability CEO ownership CEO salary CEO bonus CEO in-the-money option CEO tenure CEO chairman duality

R & D spending Model 1 −0.146*** (0.014) 0.274*** (0.009) 0.036*** (0.009) −0.59*** (0.046) 0.000 (0.001 0.030*** (0.007) −0.001 (0.002) 0.003*** (0.001) 0.000 (0.000) 0.008 (0.011

CEO age

Breakthrough Innovation Model 2 −0.154*** (0.014) 0.283*** (0.010) 0.036*** (0.009) −0.577*** (0.047) 0.000 (0.001) 0.031*** (0.007) −0.001 (0.002) 0.003*** (0.001) 0.000 (0.000) 0.017 (0.012) −0.002* (0.001)

Model 3 1.665*** (0.071) 0.066*** (0.022) 0.029 (0.024) 0.187† (0.134) 0.011* (0.005) −0.071*** (0.020) 0.002 (0.006) 0.007** (0.002) 0.001† (0.001) −0.073* (0.034)

Model 4 1.634*** (0.073) 0.073*** (0.022) 0.036† (0.025) 0.206† (0.136) 0.014*** (0.006) −0.068*** (0.020) 0.002 (0.006) 0.007*** (0.002) 0.001 (0.001) −0.041 (0.036) −0.006* (0.003)

Model 5 1.558*** (0.077) 0.024 (0.024) 0.023 (0.024) 0.269* (0.135) 0.010* (0.005) −0.078*** (0.019) 0.002 (0.005) 0.005* (0.002) 0.001† (0.001) −0.073* (0.034)

0.076*** (0.021)

R & D spending

Model 6 1.545*** (0.077) 0.033† (0.025) 0.029 (0.025) 0.279* (0.137) 0.013* (0.006) −0.075*** (0.019) 0.001 (0.006) 0.006** (0.002) 0.001 (0.001) −0.045† (0.035) −0.005* (0.003) 0.068*** (0.021)

Exploitation CEO age*Exploitation

Model 7 1.655*** (0.082) 0.042† (0.025) 0.033† (0.024) 0.259* (0.137) 0.012* (0.005*** −0.078*** (0.019) 0.001 (0.005) 0.006** (0.002) 0.001 (0.001) −0.033 (0.035) −0.009** (0.003) 0.067*** (0.021) −0.281*** (0.066) 0.016** (0.007)

Exploration

Model 8 1.543*** (0.078) 0.039† (0.025) 0.032† (0.025) 0.240* (0.137) 0.013** (0.006) −0.080*** (0.019) 0.001 (0.006) 0.006** (0.002) 0.001 (0.001) −0.037 (0.035) −0.004† (0.003) 0.069*** (0.021)

0.196*** (0.058 −0.019** (0.007)

CEO age*Exploration

Model 9 1.637*** (0.082) 0.044* (0.026) 0.035† (0.024) 0.236* (0.137) 0.013* (0.005) −0.080*** (0.019) 0.001 (0.005) 0.006** (0.002) 0.001 (0.001) −0.030 (0.035) −0.007* (0.004) 0.068*** (0.021) −0.237*** (0.067) 0.010† (0.007) 0.146** (0.060) −0.014* (0.007)

Exploitation*Exploration Year effects Firm fixed effects F-value Log likelihood Wald chi2 †

Yes Yes 120.35***

Yes Yes 109.54***

Model 10 1.661*** (0.083) 0.043* (0.026) 0.036† (0.025) 0.242* (0.137) 0.013** (0.005) −0.079*** (0.019) 0.001 (0.005) 0.007** (0.002) 0.001 (0.001) −0.033 (0.035) −0.004* (0.003) 0.067*** (0.021) −0.261*** (0.068)

0.102† (0.067)

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

Yes Yes

0.452* (0.229) Yes Yes

−8053.74 260.05***

−7930.24 270.01***

−8046.98 280.33***

−7925.15 286.10***

−7912.75 313.45***

−7914.52 307.60***

−7907.24 324.29***

−7910.02 317.65***

p < 0.10; * p < 0.01; ** p < 0.01; *** p < 0.001; Standard errors are in parentheses. a obs = 4691; n = 681.

6

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

results shown in Table 2. We also ran a negative binomial regression with a random effect model in order to account for time-invariant industry effects, and obtained similar results (β = −0.006, p < 0.01, Model 4; β = 0.022, p < 0.001, Model 7; β = −0.019, p < 0.01, Model 8 in the random effect model). Such similarities derived from different model specifications that underscore the robustness of our results. Finally, in order to check the robustness of our results, we created a retention on board variable that distinguishes between those connected to management and those CEOs who simply retire. We tested our model with the retention on board as an additional control variable, and found that results did not change substantially. Fig. 1. The moderating influence of exploitation on the relationship between CEO age and breakthrough innovations.

5. Discussion In this study, we explored how a CEO's career horizon affects firm innovations. We also examined the mediating mechanism that links a CEO career horizon to breakthrough innovations, and explored how the impact of the career horizon is contingent on the extent to which a firm engages in exploitation and/or exploration. We investigated the complex interrelationships of CEO career horizon and a firm’s organizational learning activities (i.e., exploitation and exploration), as these may influence its ability to generate innovations. Consistent with our theoretical argument, we found that a firm whose CEO has a short career horizon tends to produce fewer breakthrough innovations, and that this relationship is partially mediated by the level of R & D spending. Regarding organizational learning behavior, we found that the degree of exploitation or exploration to which a firm commits affects a CEO’s willingness to maintain a commitment to breakthrough innovations. Specifically, greater exploitation mitigates the detrimental impact of a short career horizon on breakthrough innovations, while greater exploration exacerbates this negative influence on breakthrough innovations. It is interesting to note that although CEOs’ short career horizon tends to lower breakthrough innovations, such detriments of horizon problem on firm innovation could be reduced when a firm leverages existing knowledge within a familiar technological domain. Overall, we find support for the theoretical argument that CEOs' desire to preserve their legacy of success can play an important role in how a shorter career horizon affects firm innovations.

detrimental influence of ‘horizon problem’ on a firm’s ability to produce breakthrough innovations. In terms of exploration, the coefficient of the interaction term between CEO age and exploration in Model 8 was negative and significant (β = −0.019, p < 0.01), supporting Hypothesis 4. We plotted this relationship in Fig. 2, using ± 2 standard deviations from the mean for CEO age and exploration at the mean values of all other variables. A stronger slope in Fig. 2 indicates that the negative effect of a short CEO career horizon on a firm’s breakthrough innovations is exacerbated, given a greater degree of exploration. This result can be interpreted that a higher degree of exploration increases CEO concerns over labor market evaluations and performance shortfalls in the short term, convincing them to forego allocating resources in breakthrough innovations and in turn lowering innovation outcomes. These results shed light on the contrasting moderation effects that a firm’s organizational learning has on a CEO’s willingness to adopt risky strategies. Specifically, each learning activity shapes the CEO’s willingness or reluctance to choose risky strategies: a firm’s engagement in exploitation weakens a CEO’s tendency to avoid risk, while the firm’s engagement in exploration strengthens the propensity to avoid risk. Overall, any alignment and misalignment between CEO and organizational learning behavior regarding a tendency toward risk could strongly influence the firm’s ability to generate breakthrough innovations. Additional analyses were conducted to confirm these results. First, we operationalized the measurement of breakthrough innovations using different percentiles (e.g., 97th, 90th, 85th, 80th), different time windows (three years, four years, and five years following focal year), and different lag times (two-year lag as opposed to one-year lag between dependent and other variables). We also created another alternative measure for breakthrough innovation as a dummy variable (coded 1 if a firm had breakthrough innovations, and zero otherwise) and ran a logistic regression. The models using different measurements of dependent variables showed no significant difference regarding the

5.1. Contribution to literatures The current research contributes to the literature on CEO decision making and firm innovation in a number of ways. First, our study contributes to the literature on CEO decision making (Chatterjee and Hambrick, 2011; Hayward et al., 2004; Seo et al., 2014) by elucidating a mechanism that can explain the influence of a CEO’s career horizon on a firm’s ability to generate breakthrough innovations. Consistent with the literature on CEO career horizon, we highlight how a CEO’s career horizon can shape their unwillingness to undertake risky strategies (Cheng, 2004; Dechow and Sloan, 1991; Kang, 2015; Matta and Beamish, 2008). Additionally, supporting studies such as Kang (2015) and Matta and Beamish (2008), we show that a firm’s organizational learning activities (exploitation and exploration) can accentuate or mitigate whether CEOs avoid taking risks. Thus, when a CEO’s future prospects and legacy of success are at risk, short-sighted CEOs tend to seek to avoid firm risk in order to protect their own interests. Our study also contributes to the literature of innovation by underscoring the important role of CEOs in generating breakthrough innovations. We extend prior work that examines the antecedents of breakthrough innovations (e.g., Hill and Rothaermel, 2003; Miller et al., 2007; Smith and Tushman, 2005; Zhou and Wu, 2010), articulating how CEO motivations can influence a firm’s ability to generate breakthrough innovation. Consistent with recent studies that explore the role of CEOs on firm innovation (Galasso and Simcoe, 2011; Heider, 2013; Tang et al., 2012; Wu et al., 2005), our findings suggest that

Fig. 2. The moderating influence of exploration on the relationship between CEO age and breakthrough innovations.

7

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

motivational factors can shape a CEO’s willingness to allocate firm resources to risky projects, which in turn can affect firms achieving breakthrough innovations.

Acknowledgments This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2016S1A5A8019728).

5.2. Practical implications References This research provides insights to those seeking to develop breakthrough innovations and manage organizational learning and innovation process for their firms. The alignment between a CEO’s willingness to take risks and a firm’s emphasis on organizational learning in both familiar and unfamiliar technological trajectories has an influence on firm innovations. Specifically, high levels of exploration reduce innovations when a CEO approaches retirement, whereas high levels of exploitation mitigate the harmful effect of career horizon on breakthrough innovations. These findings point to the fact that CEOs play an important role in producing breakthrough innovations, and that a strategic alignment between the willingness to take risks and the firm’s engagement in risky strategies can enable the firm to achieve greater innovations

Abernathy, W.J., Utterback, J.M., 1978. Patterns of industrial innovation. Technol. Rev. 80, 40–47. Ahuja, G., Lampert, C.M., 2001. Entrepreneurship in the large corporation: a longitudinal study of how established firms create breakthrough inventions. Strateg. Manage. J. 22, 521–543. Ahuja, G., Coff, R.W., Lee, P.M., 2005. Managerial foresight and attempted rent appropriation: insider trading on knowledge of imminent breakthroughs. Strateg. Manage. J. 26, 791–808. Ahuja, G., Lampert, C.M., Tandon, V., 2008. Moving beyond schumpeter: management research on the determinants of technological innovation. Acad. Manage. Ann. 2, 1–98. Anderson, P., Tushman, M.L., 1990. Technological discontinuities and dominant designs: a cyclical model of technological change. Adm. Sci. Q. 604–633. Baltagi, B., 2008. Econometric Analysis of Panel Data. John Wiley & Sons. Barker III, V.L., Mueller, G.C., 2002. CEO characteristics and firm R & D spending. Manage. Sci. 48, 782–801. Baron, R.M., Kenny, D.A., 1986. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J. Pers. Soc. Psychol. 51, 1173–1182. Baum, J.A., Li, S.X., Usher, J.M., 2000. Making the next move: how experiential and vicarious learning shape the locations of chains' acquisitions. Adm. Sci. Q. 45, 766–801. Beckman, C.M., Haunschild, P.R., Phillips, D.J., 2004. Friends or strangers? Firm-specific uncertainty, market uncertainty, and network partner selection. Organiz. Sci. 15, 259–275. Benner, M.J., Tushman, M.L., 2003. Exploitation, exploration, and process management: the productivity dilemma revisited. Acad. Manage. Rev. 28, 238–256. Brickley, J.A., Linck, J.S., Coles, J.L., 1999. What happens to CEOs after they retire? New evidence on career concerns, horizon problems, and CEO incentives. J. Financial Econ. 52, 341–377. Cameron, A.C., Trivedi, P.K., 2009. Microeconomics Using Stata. Stata Press Books, Lakeway Drive, TX. Chaney, P.K., Devinney, T.M., 1992. New product innovations and stock price performance. J. Bus. Finance Account. 19, 677–695. Chatterjee, A., Hambrick, D.C., 2011. Executive personality, capability cues, and risk taking how narcissistic CEOs react to their successes and stumbles. Adm. Sci. Q. 56, 202–237. Chen, R., 1996. Technological expansion: the interaction between diversification strategy and organizational capability. J. Manage. Stud. 33, 649–666. Cheng, S., 2004. R & D expenditures and CEO compensation. Account. Rev. 79, 305–328. Cohen, W.M., Levinthal, D.A., 1990. Absorptive capacity: a new perspective on learning and innovation. Adm. Sci. Q. 35, 128–152. Cyert, R.M., March, J.G., 1963. A Behavioral Theory of the Firm. (Englewood Cliffs, NJ 2). Damanpour, F., 1989. The relationship between organizational size and innovation. In: Annual Meeting of the Academy of Management. Washington DC. Damanpour, F., 1991. Organizational innovation: a meta-analysis of effects of determinants and moderators. Acad. Manage. J. 34, 555–590. Dechow, P.M., Sloan, R.G., 1991. Executive incentives and the horizon problem: an empirical investigation. J. Account. Econ. 14, 51–89. Deeds, D.L., DeCarolis, D., Coombs, J., 2000. Dynamic capabilities and new product development in high technology ventures: an empirical analysis of new biotechnology firms. J. Bus. Venturing 15, 211–229. Dosi, G., Nelson, R., Winter, S., 2000. The Nature and Dynamics of Organizational Capabilities. Oxford University Press. Ettlie, J.E., Bridges, W.P., O'keefe, R.D., 1984. Organization strategy and structural differences for radical versus incremental innovation. Manage. Sci. 30, 682–695. Galasso, A., Simcoe, T.S., 2011. CEO overconfidence and innovation. Manage. Sci. 57, 1469–1484. Gatignon, H., Tushman, M.L., Smith, W., Anderson, P., 2002. A structural approach to assessing innovation: construct development of innovation locus, type, and characteristics. Manage. Sci. 48, 1103–1122. Griliches, Z., 1990. Patent Statistics as Economic Indicators: A Survey. National Bureau of Economic Research. Hage, J., Dewar, R., 1973. Elite values versus organizational structure in predicting innovation. Adm. Sci. Q. 279–290. Hall, B.H., Jaffe, A.B., Trajtenberg, M., 2001. The NBER Patent Citation Data File: Lessons, Insights and Methodological Tools. National Bureau of Economic Research. Hall, B.H., Jaffe, A., Trajtenberg, M., 2005. Market value and patent citations. Rand J. Econ. 36, 16–38. Hambrick, D.C., Finkelstein, S., 1987. Managerial discretion: A bridge between polar views of organizational outcomes. In: Cummings, L.L., Staw, B.M. (Eds.), Research in Organizational Behavior. JAI Press Greenwich, CT, pp. 369–406. Hambrick, D.C., Mason, P.A., 1984. Upper echelons: The organization as a reflection of its top managers. Acad. Manage. Rev. 9, 193–206.

5.3. Limitations and future research This study is not without limitations. Our analysis utilized secondary data retrieved from the ExecuComp database and NBER patent citation files. Since the ExecuComp database also contains compensation data for executives of established organizations, it might be fruitful to examine the influence of a CEO’s career horizon on firm outcomes within the context of smaller organizations in order to generalize the findings. NBER patent citation files contain only patent information. It would therefore be interesting to explore the effect of short career horizons using a different measure of firm innovation. New product introductions could be a meaningful dependent variable used to generalize our findings. In addition, our study measured CEO career horizon as CEO age based on the premise that an older CEO has a shorter career horizon, which leads to risk aversion. A limitation of the measure is that it does not directly assess an older CEO’s attitude toward risky actions. For instance, if an older CEO has psychological ownership of the firm, he or she may be willing to maintain a commitment to risky innovation strategies that will lead to breakthrough innovations for the long-term benefits of the organization. Thus, it might be beneficial to examine the influence of psychological ownership on the behavior of a CEO who is nearing retirement. There are other avenues of interest for examining the influence of CEO career horizon and decision-making within various contexts. If CEOs approaching retirement have a strong desire to preserve a record of success, they may divest poorly performing projects in order to improve the short-term performance of the firm. Additionally, it would be interesting to examine how cognitive characteristics influence CEOs’ decision-making as they approach retirement. Overconfidence may overcome a CEO's risk aversion and lead them to take risks even given a shorter career horizon. We look forward to examining these issues in the future.

6. Concluding remark This study has provided a conceptual framework to explain why and how CEOs’ behavioral tendencies have an influence on firm innovations and strategies for leveraging internal and external knowledge. We hope it will inspire academics to explore the role of top executives in a firm’s innovation process, and explain why, how, and under what conditions top executives strengthen or lessen a firm’s ability to generate innovations. 8

Research Policy xxx (xxxx) xxx–xxx

S.Y. Cho, S.K. Kim

Miller, D.J., Fern, M.J., Cardinal, L.B., 2007. The use of knowledge for technological innovation within diversified firms. Acad. Manage. J. 50, 307–325. Nelson, R.R., Winter, S.G., 1982. An Evolutionary Theory of Economic Change Harvard. University Press, Cambridge, Mass. Neter, J., Kutner, M.H., Nachtsheim, C.J., Wasserman, W., 1996. Applied Linear Statistical Models. Irwin, Chicago. Palepu, K., 1985. Diversification strategy, profit performance and the entropy measure. Strateg. Manage. J. 6, 239–255. Phene, A., Fladmoe-Lindquist, K., Marsh, L., 2006. Breakthrough innovations in the US biotechnology industry: the effects of technological space and geographic origin. Strateg. Manage. J. 27, 369–388. Rosenkopf, L., Nerkar, A., 2001. Beyond local search: boundary-spanning, exploration, and impact in the optical disk industry. Strateg. Manage. J. 22, 287–306. Ross, L., 1977. The intuitive psychologist and his shortcomings: distortions in the attribution process. Adv. Exp. Soc. Psychol. 10, 173–220. Schilling, M.A., Phelps, C.C., 2007. Interfirm collaboration networks: the impact of largescale network structure on firm innovation. Manage. Sci. 53, 1113–1126. Seo, J., Gamache, D.L., Devers, C.E., Carpenter, M.A., 2014. The role of CEO relative standing in acquisition behavior and CEO pay. Strateg. Manage. J. 36, 1877–1894. Smith, W.K., Tushman, M.L., 2005. Managing strategic contradictions: a top management model for managing innovation streams. Organiz. Sci. 16, 522–536. Sobel, M.E., 1982. Asymptotic confidence intervals for indirect effects in structural equation models. Sociol. Methodol. 13, 290–312. Srivastava, M.K., Gnyawali, D.R., 2011. When do relational resources matter? Leveraging portfolio technological resources for breakthrough innovation. Acad. Manage. J. 54, 797–810. Tang, Y., Li, J., Yang, H., 2012. What I see, what I do: how executive hubris affects firm innovation. J. Manage. 41, 1698–1723. Teece, D.J., 1986. Profiting from technological innovation: implications for integration, collaboration, licensing and public policy. Res. Policy 15, 285–305. Trajtenberg, M., 1990. A penny for your quotes: patent citations and the value of innovations. Rand J. Econ. 21, 172–187. Tzabbar, D., Kehoe, R.R., 2014. Can opportunity emerge from disarray? An examination of exploration and exploitation following star scientist turnover. J. Manage. 40, 449–482. Uotila, J., Maula, M., Keil, T., Zahra, S.A., 2009. Exploration, exploitation, and financial performance: analysis of S & P 500 corporations. Strateg. Manage. J. 30, 221–231. Westley, F., Mintzberg, H., 1989. Visionary leadership and strategic management. Strateg. Manage. J. 10, 17–32. Wu, S., Levitas, E., Priem, R.L., 2005. CEO tenure and company invention under differing levels of technological dynamism. Acad. Manage. J. 48, 859–873. Yadav, M.S., Prabhu, J.C., Chandy, R.K., 2007. Managing the future: CEO attention and innovation outcomes. J. Market. 71, 84–101. Zhou, K.Z., Wu, F., 2010. Technological capability, strategic flexibility, and product innovation. Strateg. Manage. J. 31, 547–561. Zona, F., 2016. Agency models in different stages of CEO tenure: the effects of stock options and board independence on R & D investment. Res. Policy 45, 560–575.

Hambrick, D.C., MacMillan, I.C., Barbosa, R.R., 1983. Business unit strategy and changes in the product R & D budget. Manage. Sci. 29, 757–769. Hausman, J.A., Hall, B.H., Griliches, Z., 1984. Econometric Models for Count Data with an Application to the Patents-R & D Relationship. National Bureau of Economic Research Cambridge, Mass., USA. Hayward, M.L., Hambrick, D.C., 1997. Explaining the premiums paid for large acquisitions: evidence of CEO hubris. Adm. Sci. Q. 42, 103–127. Hayward, M.L., Rindova, V.P., Pollock, T.G., 2004. Believing one's own press: the causes and consequences of CEO celebrity. Strateg. Manage. J. 25, 637–653. He, Z.-L., Wong, P.-K., 2004. Exploration vs. exploitation: an empirical test of the ambidexterity hypothesis. Organiz. Sci. 15, 481–494. Heider, F., 2013. The Psychology of Interpersonal Relations. Psychology Press. Hill, C.W., Rothaermel, F.T., 2003. The performance of incumbent firms in the face of radical technological innovation. Acad. Manage. Rev. 28, 257–274. Hirshleifer, D., Low, A., Teoh, S.H., 2012. Are overconfident CEOs better innovators? J. Finance 67, 1457–1498. Jansen, J.J., Van Den Bosch, F.A., Volberda, H.W., 2006. Exploratory innovation, exploitative innovation, and performance: effects of organizational antecedents and environmental moderators. Manage. Sci. 52, 1661–1674. Kang, J., 2015. Labor market evaluation versus legacy conservation: what factors determine retiring CEOs' decisions about long-term investment? Strateg. Manage. J. 37, 389–405. Katila, R., Ahuja, G., 2002. Something old, something new: a longitudinal study of search behavior and new product introduction. Acad. Manage. J. 45, 1183–1194. Kim, S.K., Arthurs, J.D., Sahaym, A., Cullen, J.B., 2013. Search behavior of the diversified firm: the impact of fit on innovation. Strateg. Manage. J. 34, 999–1009. Kor, Y.Y., 2006. Direct and interaction effects of top management team and board compositions on R & D investment strategy. Strateg. Manage. J. 27, 1081–1099. Krause, R., Semadeni, M., 2014. Last dance or second chance? Firm performance, CEO career horizon, and the separation of board leadership roles. Strateg. Manage. J. 35, 808–825. Leonard, D., 1992. Core capabilities and core rigidities: a paradox in managing new product development. Strateg. Manage. J. 13, 111–125. Levinthal, D., March, J.G., 1981. A model of adaptive organizational search. J. Econ. Behav. Organiz. 2, 307–333. Levinthal, D.A., March, J.G., 1993. The myopia of learning. Strateg. Manage. J. 14, 95–112. Makri, M., Lane, P.J., Gomez-Mejia, L.R., 2006. CEO incentives, innovation, and performance in technology-intensive firms: a reconciliation of outcome and behavior-based incentive schemes. Strateg. Manage. J. 27, 1057. March, J.G., 1991. Exploration and exploitation in organizational learning. Organiz. Sci. 2, 71–87. Matta, E., Beamish, P.W., 2008. The accentuated CEO career horizon problem: evidence from international acquisitions. Strateg. Manage. J. 29, 683–700. McClelland, P.L., Barker, V.L., Oh, W.-Y., 2012. CEO career horizon and tenure: future performance implications under different contingencies. J. Bus. Res. 65, 1387–1393. McGrath, R.G., 2001. Exploratory learning, innovative capacity, and managerial oversight. Acad. Manage. J. 44, 118–131.

9