The effects of leader succession and prior leader experience on postsuccession organizational performance. Human Resource Management, 55 (6).

In this article we strive to reconcile equivocal findings about the effects of top leader succession and prior leadership experience on postsuccession organizational performance. In doing so, we draw on insights from theories of human capital, learning, and asymmetric information to better understand the conditions under which leaders increase or decrease postsuccession performance. Employing a sample of 119 newly appointed leaders in the English Premier League (1996–2010), we find the following results. First, relating to the succession event, outside leaders that directly move between leadership positions are associated with higher postsuccession performance while the departure of a prior leader to a leadership position in another organization has a negative effect on postsuccession performance. Second, relating to prior leadership experience, leaders with domestic top leader experience are associated with lower postsuccession performance, while leaders with foreign top leader experience are associated with higher postsuccession performance.


POST-SUCCESSION ORGANIZATIONAL PERFORMANCE
Leader succession is a critical issue for organizations, which has important implications for organizational performance (Guthrie & Datta, 1998: Kesner & Sebora, 1994. While the performance and actions of leaders are often scrutinized, boards and owners may also be subject to criticism regarding the effectiveness of their succession decisions (Elsaid, Wang, & Davidson, 2011;Hamori & Koyuncu, 2014). Recent trends suggest that owners are increasingly looking to hire leaders with prior top leader experience, including directly hiring the leaders of other organizations (Hamori & Koyuncu, 2014;Karlsson & Neilson, 2009;Lucier, Kocourek, & Habbel, 2006). Owners may view the hiring of experienced leaders as a less risky decision than hiring inexperienced leaders, thereby avoiding the appointment of an unknown quantity. In this paper we explore the effect of two aspects of the effect of leader succession on post-succession organizational performance: (i) the actual succession event and (ii) the experience of the new leader. Karaevli (2007: 682), after reviewing the empirical evidence about the relationship between the leader succession event and post-succession performance, argued that: "scholars have failed to reach a consensus on whether succession events in general, and insider vs. outsider successions in particular, affect firm performance positively, negatively, or insignificantly." In addition, and in contrast to the body of scholarship focusing on the relationship between the leader succession event and post-succession performance, research into the effects of prior leadership experience on organizational performance is in its infancy (Hamori & Koyuncu, 2014). Interestingly, the two main studies find an absence of positive performance effects (or even negative performance effects) for the effect of prior leadership experience on post-succession performance (see: Elsaid et al., 2011;Hamori & Koyuncu, 2014). We suggest that these findings oppose conventional wisdom that having prior experience in the top job should enhance leaders' skills, which should lead to positive outcomes.
In this paper we contribute to theory by drawing on insights from human capital theory, learning and asymmetric information to better understand the relationship between the succession event and leadership experience on post-succession performance. First, we contribute to the succession literature, addressing Karaevli's (2007) finding of equivocal results between the leader succession event and post-succession performance, by decomposing outside appointments into active leaders hired directly from other organizations and those not currently in the top leader position. In doing so, we explore how differences in the origin of the newly hired outside leader may affect postsuccession performance and, in addition, examine the corollary in terms of the effects of how the nature of the departure of the predecessor leader affects post-succession performance. Second, we contribute to the emerging literature on prior top leader experience and post-succession performance by decomposing top leader experience into the experience of leading a domestic competitor and the experience of leading a foreign competitor. In doing so, we contribute to the debate as to the portability of leaders human capital across different contexts (Murphy & Zábojnik, 2004& 2007. To explore our ideas we focus on leaders of soccer organizations that operated in the English premier league (EPL) from 1996 to 2010. 1 We suggest that the EPL context is conducive for examining our ideas for two main reasons. First, the EPL is characterized by intense competition and high rates of successions (Brady, Bolchover, & Sturgess, 2008), with the average tenure of EPL leaders falling from 3.12 years (120 games) in 1993 to around 1.47 years (56 games) in 2009 (Bridgewater, 2010). Furthermore, it is common for leaders to be appointed from a range of different prior positions (including inside or outside of the organization, from abroad and/or directly from other organizations) and so we can examine the performance effects of mew leaders bringing in a diverse range of experiences to shape their new organizations. Second, the soccer industry intensifies and accelerates many of the problems faced by the leaders of business organizations (Brady et al., 2008). For example, like many other team sports, soccer teams are meticulously monitored by owners, fans and the media, and so their leaders are under enormous pressure to succeed and keep their jobs during performance difficulties (Rowe, Cannella, Rankin, & Gorman, 2005). Indeed, performance is assessed on a weekly, or a game-by-game, basis, in contrast to leaders of business organizations that are evaluated over a much longer time-scale (e.g., quarterly, mid-year, annually). Hence, the EPL provides an extremely useful context in which to examine issues of leader succession and organizational performance.

THEORETICAL MODEL
Drawing on extant research we argue that there are three main factors that will influence the relationship between leader succession and organizational performance: the knowledge and skills of the leader; the ability for the leader and organization to learn from one another; and the potential for asymmetric information in the hiring decision (Shen & Cannella, 2002b;Zhang & Rajagopalan, 2003).
First, leaders bring with them different skills and abilities depending on whether they come from inside or outside of the organization, which in turn has varying effects on performance. We root this idea in human capital theory (Becker, 1993;Castanias & Helfat, 1991). Human capital is defined as an individual's knowledge, skills, experiences and abilities (Bailey & Helfat, 2003;Harris & Helfat, 1997), which can be accumulated through education, employment, habits and activities (Becker, 1993). As the boundaries of our study are constrained to intra-industry successions, we build on Castinias & Helfat's (1991) classifications of generic skills, firm-specific and industry-specific skills. Firm-specific human capital is useful only to the firms which provide it and is not transferable whereas industry-specific human capital can be transferred within an industry but has less transferability across industries. Generic, or general, skills are those that can be transferred across organizations and industries. Another form of leader's human capital is international experience. The transferability, and in turn the usefulness, of such an experience will depend on the extent to which a leader's organization is located in a globalized industry, the international inter-dependence of the organization and the similarity in terms of national-level institutions (see : Crossland & Hambrick, 2007;Roth, 1995). We argue that prior top leader experience is a salient human capital attribute, which can affect the speed in which leaders can adapt to their new organization and the risks associated with their appointment.
Second, for a new leader to be successful both the individual leader and the organization need to learn to work with one another (Boeker, 1997;Rowe et al., 2005;Virany, Tushman, & Romanelli, 1992;Zhang & Rajagopalan, 2003). Learning can occur at the individual level and organizational level (Levitt & March, 1988). At the individual level, leaders need to learn about their environments absorbing salient information in order to make appropriate strategic decisions . At the organizational level, organizations can learn from the leader who brings with him/her new knowledge (Boeker, 1997;Zhang & Rajagopalan, 2003).
Third, each succession type is associated with some level of risk (Shen & Cannella, 2002b;Wiersema, 2002;Zhang, 2008). Risk arises because of the problem of asymmetric information, in that the leadership candidates will have more knowledge of their own abilities and willingness to work in the interests of the organization than those tasked with making the hiring decision (Zajac, 1990). The greater the problem of asymmetric information in relation to the succession decision the greater the potential for firms to make sub-optimal hiring decisions, which will have a negative effect on post-succession performance. Hiring organizations may perceive leaders with prior leadership experience, and especially those currently acting in the leader role, as less risky because they will have been better able to observe them in the top job prior to hiring. Although on some occasions the cause-effect outcomes of leader's decisions may be difficult to interpret, hiring organizations should still be able to assess their actions and behaviors.

Inside versus outside succession
When hiring a new leader, those tasked with the succession decision must decide whether to hire from inside or outside of the organization, and if they go outside, to hire a current leader of another organization or hire from the pool of talent not currently employed as a leader. We begin by examining performance effects of going for an inside or outside candidate, which is influenced by the availability of internal candidates reflecting the succession planning practices of the organization (Cappelli, 2008;Friedman, 1986;Friedman & Olk, 1995), and the informational dynamics surrounding the succession decision (Zhang, 2008).
The potential suitability of an internal candidate is, in part, shaped by the extent to which the board feels that the candidate has the suitable human capital to lead the organization. Scholars have argued that internal candidates have the advantage of firm-specific human capital, which may enhance firm performance (Bailey & Helfat, 2003). The firm-specific human capital helps to ensure that the internal candidate is better aligned with the organization, understanding their systems and procedures, as well as having the tacit knowledge of the organization. In contrast, the outside candidate will not have firm-specific human capital, and so may not represent a good fit with the needs of the hiring organization, or with its culture (Karaevli, 2007).
In terms of the informational dynamics surrounding leader succession, scholars have argued that boards, when making a leader selection decision, will have more information about the skills and abilities (i.e., human capital) of inside candidates than outside candidates due to a higher frequency of interactions and internal networks (Karaevli, 2007;Shen & Cannella, 2002a;Zajac, 1990;Zhang, 2008). We suggest, however, that the informational dynamics surrounding the leader succession decision are more complex. First, perceived informational problems associated with outside candidates will encourage boards to have a higher propensity to hire inside candidates, thereby making more marginal decisions (Coff, 1997;Lauterbach, Vu, & Weisberg, 1999). Second, boards face the risk that the status of an inside candidate as a "known quantity" may enable the individual to "sail through a lax due-diligence process" (Charan, 2005: 75). Relatedly, scholars have argued that the trend in hiring outside candidates may reflect greater board diligence in matters of corporate governance (Huson, Parrino, & Starks, 2001). Furthermore, Lauterbach et al. (1999) found that inside successors tend to perform worse than outside successors, suggesting that the internal selection decision may commonly be suboptimal.
Based on the above discussion we suggest that the potential positive performance effects of firm-specific human capital for inside candidates will be offset by problems associated with informational asymmetries in their appointment. Hence, when controlling for general leader human capital it is perhaps unsurprising that scholars have found, what Karaevli (2007) described as, equivocality of results in relation to post-succession performance. To address the equivocality of results, we de-compose the group of outside candidates into leaders hired directly from a leadership position at a competitor (i.e., poached) and leaders who were hired from a pool of available talent not currently in a leadership position and examine their impact on post-succession performance.
According to Friedman and Singh (1989), the positive effect of succession is most pronounced when the new leader's human capital is matched to the organization's task contingencies and when the leader is given the discretion to shape the organization. We argue that leaders that are poached away from their current employer will provide the hiring organization with more information about their ability, as compared to those who are not currently in a leadership position, thereby reducing the risk of adverse selection in selecting the new leader. Furthermore, the leader may have been poached because the they have the specific human capital that the hiring organization requires or strives for (Boeker, 1997). Indeed, where a leader is hired directly from another organization, the succession event is likely to be more costly due to search and compensation costs, which suggests that the leader is more likely to be the organization's number one choice, or close to it. Based on the above, we argue that leaders who are poached away from will be granted more autonomy to shape their new organization as compared to those who are not poached away. Relatedly, as poached leaders are outsiders to their new organizations, their early strategic are likely to be adaptive because they bring with them new ideas and are less likely to be influenced by the statusquo, which will likely have a positive effect on organizational performance (Zhang & Rajagopalan, 2010). Finally, being poached away is an indication that the leader was successful in their prior appointment, which may boost employee morale (Friedman & Saul, 1991) and motivate employees to take on board the leader's ideas, and assist the leader by getting him/her up to speed regarding organizational idiosyncrasies. On the basis that poached leaders will have the human capital and autonomy to influence their new organizations, we suggest that their post-succession performance will be greater than non-poached leaders. Hence: H1: Leaders who move directly from the position of leader in one organization to leader of another organization are associated with higher post-succession organizational performance than outside leaders who are not in a leadership position.
We now examine the reverse of the poaching decision to examine the effects of a departure of a predecessor leader when poached by a competitor. We suggest that having a leader poached away is an unplanned event (Friedman & Saul, 1991), and one that the organization will not have prepared for, which will influence the relationship between new leader succession and postsuccession performance.
We suggest that having a leader poached away will create an inertial and disruptive force because the organization is geared towards the predecessor's paradigm and decisions. The fact that the prior leader left voluntarily indicates that the organization was happy with his/her leadership, and therefore did not want to change their current strategy and/or the organization's current configuration. In contrast, a dismissal is an event that signals that the organization requires a change in direction (Friedman & Saul, 1991;Karaevli, 2007). Accordingly, the successor will likely struggle to imprint his/her ideas when taking charge and face, for example, entrenched cultures, institutionalized practices and investments that are specific to the prior leader (Fondas & Wiersema, 1997;Hannan & Freeman, 1977;Quigley & Hambrick, 2012). Relatedly, the departure of a poached predecessor leader will lead to a loss of firm-specific human capital, which will drain knowledge about how best to operate under the organization's current configuration. For instance, researchers have argued that high organizational performance outcomes are not merely due to the inherent properties of the resources at a leader's disposal, but rather are significantly influenced by the way in which a leader utilizes them (Brady et al., 2008;Sirmon, Hitt, & Ireland, 2007).
Hence, we argue that new leaders will face a dilemma in their efforts to shape the organization.
On the one hand, embracing the status-quo may prove to be problematic because the leader may not have the expertise and abilities to orchestrate the predecessor's resource allocations and need time to learn about the organization. On the other hand, leaders that attempt to detach the organization from the predecessor's decisions are likely to disrupt established structures and processes. As a consequence, having a leader poached away will result in the organization losing firm-specific knowledge about how to utilize its resources, which will have a negative impact on post-succession performance. Based on the above, we suggest that: H2: Leaders whose predecessor was poached will be associated with lower postsuccession organizational performance than those who follow a leader who was dismissed.
We now look past the origin of the leader to examine the nature of their human capital, focusing on the post-succession performance effects of domestic versus foreign top leader experience.

Domestic versus foreign experience
Human capital may be viewed as consisting of a hierarchy of skills and knowledge with varying degrees of transferability across contexts (Castanias & Helfat, 1991). The extent to which human capital is portable across organizations and contexts is subject to debate (Hamori & Koyuncu, 2014). Researchers have suggested that CEO jobs have a considerable proportion of general human capital (Murphy & Zábojnik, 2004& 2007, and that even the specific human capital associated with CEO positions has become easier to develop due to the computerized provision of firmspecific information (Murphy & Zábojnik, 2007). Consequently, Murphy and Zábojnik (2004) argue that there has been a shift in the importance of general relative to firm-specific human capital, which has led to a reduction in the number of inside successions and more outside successions.
Existing evidence about the effect of prior job-specific experience on an individual's job performance, however, has been met with equivocal results. In terms of a positive effect, research has found that the prior acquisitions experience of outside directors was positively associated with the performance of their focal firm's acquisitions and that prior entrepreneurial experience of actors was associated with successful new ventures (McDonald, Westphal, & Graebner, 2008;Stuart & Abetti, 1990). In contrast, evidence suggests that there is a negative relationship between prior jobspecific experience, in relation to being CEO, and post-succession performance (Elsaid et al., 2011;Hamori & Koyuncu, 2014), which is explained in terms of prior experience slowing down learning in a new context because some knowledge and skills need to be "unlearned" before learning in the new context can take place (Morrison & Brantner, 1992). Finally, Bragaw and Misangyi (2013) found no significant relationship between prior CEO experience and performance.
We suggest that the presence of equivocal results stems from a lack of precision as to the nature of prior job-specific experience. In considering the performance effects of general human capital, which is not specific to a particular firm, we delineate between two forms of leadership experience, that of being a leader at a domestic competitor firm (i.e., a firm of similar standing domestically) and that of being a leader at a foreign competitor (again of similar standing). In examining the potential for performance differences to accrue to different forms of leader-specific human capital we draw on two related theoretical explanations.
First is the extent to which a new leader may bring with them new skills and knowledge into the organization, which is complementary to the organization's own, thereby promoting performance improvements. Drawing on ideas from the resource-based view of the firm we argue that given the same set of resources, different leaders may see different possibilities for their utilization (Kor, Mahoney, & Michael, 2007;Penrose, 1959). Where leaders' prior experience is very closely related to the new firm then this will limit the potential for learning outside of areas where the firm already holds prior knowledge (Cohen & Levinthal, 1990;Teece, Pisano, & Shuen, 1997). For example, in an industry there may be common approaches to leadership, so hiring a leader with leadership experience with another domestic competitor will do little to introduce new knowledge and perspectives into the firm. Rather, hiring someone with experience of leading a domestic competitor may merely reinforce existing practices and reduce innovation in organizational practices. In contrast, hiring a leader with experience of leading a foreign competitor may lead to a greater diversity of leadership knowledge in the organization, which may stimulate innovation in organizational practices.
Second, and related to the issue of knowledge diversity, is the ability of the new leader to learn about their new position. Hamori and Koyuncu (2014), drawing on the position of Morrison and Brantner's (1992) work on actor's ability to learn their new job, suggest that where the job-specific experience is very similar to the new role the leader will perform, then issues relating to "unlearning" will be particularly pronounced. A leader's perception of having job-specific experience may mean that they underestimate differences between the organizations that they have led, and now will lead (Hamori & Koyuncu, 2014). We suggest that two domestic competitors, although they may have many common attributes, may have very different values, norms and standards, and/or salient operational and cultural differences (Dokko, Wilk, & Rothbard, 2009;Finkelstein & Haleblian, 2002;Hamori & Koyuncu, 2014).
We suggest that where actors have leadership experience with a domestic competitor they may be less able to learn about their new role, as they will merely seek to transfer existing practices as they underestimate the potential differences between organizations. In contrast, new leaders with experience of leading a foreign competitor will be more likely to be aware of the differences between organizations, having a greater variety of experience, and so be better able to understand the need to learn about the new job. We argue that leaders with foreign experience will be able to do so because they have been exposed to different values, cultural norms, institutional environments and hence new ways of learning and responding to stimuli (Ricks, Toyne, & Martinez, 1990). Studies that have examined the prior international experience of CEOs that were obtained when they were at lower level positions (e.g., international assignments at executive levels) have found a positive association with organizational performance (Carpenter, Sanders, & Gregersen, 2001;Daily, Certo, & Dalton, 2000).
Based on the arguments above we suggest that where leaders are recruited with experience of leading a domestic competitor then innovation in organizational practices may be stifled, and the new leaders ability to learn about their new job diminished, which will negatively affect postsuccession organizational performance. In contrast, where leaders are recruited with experience of leading a foreign competitor then innovation in organizational practices may be increased, and their ability to learn about their new job enhanced, which will positively affect post-succession organizational performance. Hence: H3: Leaders with experience of leading domestic competitors are associated with lower post-succession organizational performance than leaders who do not have experience of leading domestic competitors.

H4:
Leaders with experience of leading international competitors are associated with higher post-succession organizational performance than leaders who do not have experience of leading international competitors.

DATA AND METHOD
We test our model on leaders of soccer organizations that operate in the English Premier League (EPL). Professional team sports are an interesting research site for examining strategic management and leadership phenomena (Berman, Down, & Hill, 2002;Bloom, 1999;Day, Gordon, & Fink, 2012;Katz, 2001), which has long been used as an empirical site for scholars of succession (Giambatista, 2004;Grusky, 1963;Ndofor, Priem, Rathburn, & Dhir, 2009;Pfeffer & Davis-Blake, 1986;Rowe et al., 2005). In addition to the suitability of the EPL context for testing our model, as outlined in the introduction, our context provides data that is reliable, objective and readily available, and researchers can identify and measure organizational performance and the human capital of leaders and team members (Audas, Dobson, & Goddard, 2002;Hughes, Hughes, Mellahi, & Guermat, 2010;Pazzaglia, Flynn, & Sonpar, 2012;Werner & Mero, 1999).

Sample
Our sample consists of longitudinal data from 1995/96 to 2009/10 of 119 leaders at 31 soccer organizations that operated in the EPL on a game-by-game basis. In total we have 4452 individual game observations. 2 Consistent with prior studies (Graffin, Boivie, & Carpenter, 2013;Shen & Cannella, 2002b;Zhang, 2008), we use a post-succession (or early leader tenure) window to capture the performance effects of succession. We define our post-succession period as the first two years (i.e., seasons) of tenure. While we could focus specifically on the first season, some leaders, in our case 66 percent, are appointed within-season and so not all leaders would have been in place for a full first season of tenure. 3 We gathered performance data, on a game-by-game basis, from various football almanacs including Rothmans and Sky Sports Football Yearbooks. Financial and takeover data were obtained from Companies House and Financial Analysis Made Easy (FAME). Data on leader tenure, succession and origin were obtained from Lexis-Nexis. Team member acquisitions data (of which there were over 1000 in total) were obtained from www.soccerbase.com and Lexis-Nexis.

Dependent Variable
Performance was measured as the points gained by a leader for each game (i.e., competitive event).
In the EPL, teams are awarded three points for a win, one point for a draw and zero points for a loss. Using points as our dependent variable is consistent with prior studies that use a game-bygame approach in a soccer context (Audas et al., 2002;Hughes et al., 2010). The frequency of observations for each of the performance categories for 3, 1, 0 were 1,645, 1,243 and 1,891 respectively.

Model Variables
Inside succession relates to leaders who were promoted from inside of the organization. We measure this as a binary variable coded one if a leader was hired from inside of the organization and coded zero otherwise. We identified 32 leaders that were inside appointments.
Leader to Leader entry relates to organizations where a leader was directly hired from another organization where they were a leader. We measure this as a binary variable coded one if a leader was hired (or resigned in order to be hired) from another organization and coded zero otherwise.
In our sample 47 leaders were poached leader entrants from other organizations. 4 Among these poached entrants, there were cases in which leaders came from outside of the sample (i.e., outside of the EPL). Consistent with Boeker (1997), we treat these entrants just like those that move to other organizations that are within the sample.
Leader to Leader exit relates to organizations where a predecessor leader has left to join another organization. We measure this as a binary variable coded one if the predecessor leader left to go to another organization and coded zero otherwise. We identified 21 leaders that were appointed post a poached leader exit.
Domestic experience. We measure this as a binary variable coded one if the leader has been a top leader in the hiring organization's domestic industry (i.e., the EPL) and coded zero otherwise.
In our sample, we identified 56 leaders that had prior domestic leadership experience.
Foreign experience. We measure this as a binary variable coded one if the leader has been a top leader at a foreign competitor and coded zero otherwise. We define foreign competitor in terms of organizations that competed in the top four European leagues outside of the EPL (France, Germany, Italy, and Spain) and/or had competed in the UEFA Champions league (the premier pan-European competition which is restricted to the leading competitors in each domestic league) over the period.
Employing this definition we are able to exclude foreign experience of leading organizations that are much smaller and less complex than EPL organizations. In our sample, 25 leaders had foreign leadership experience.

Control Variables
Lagged Performance was measured by lagging our performance (points) variable, by three games.
Lag performance t-1, Lag performance t-2, and Lag performance t-3. 5 Consistent with prior research, these values can be inferred as default expectations or momentum affects (Giambatista, 2004;Hughes et al., 2010).
Tenure was measured as the count of the number of games a leader was in position from appointment.
Takeovers were measured as a binary variable, which we coded one if there was a takeover event and coded zero otherwise. We created a lead of 50 games (i.e., around one and one-half seasons) from the date of the takeover to fully capture the disruptions which may affect performance.
Within-season succession relates to leaders appointed during the playing season. The timing of when a leader takes charge may affect post-succession performance, as within-season succession being more problematic for leaders as they do not have the off-season to shape their new organizations (Giambatista, 2004;Rowe et al., 2005). We measured this as a binary variable coded one if the leader was hired within the soccer season and coded zero otherwise.
Promoted leaders relates to those leaders who have joined the EPL through their organization being promoted 6 . We control for promoted leaders as promoted organizations tend to have lower levels of resources as compared to established EPL organizations, which may affect performance.
We measure this as a binary variable, which we coded 1 if the leader of a soccer organization was promoted and coded 0 otherwise.
Strategic changes were measured as the total number of team member entrants, which we employed as a rolling average over the first two seasons of a leader's tenure. As new leaders enter their organizations, they may seek to align areas close to their world views (Gabarro, 1987;Hambrick & Fukutomi, 1991).
Team quality was measured as a soccer organization's total wage cost, updated on a seasonal (yearly) basis. Wage costs are a well-accepted measure of team quality in the literature (Brady et al., 2008;Szymanski & Smith, 1997).
Total experience is measured in terms of the total number of seasons a leader has held a leader position prior to appointment, and is independent of where the experience was gained. We employ this as a measure of general leader human capital.

Model Estimation
We use an ordered probit model (using STATA) as our dependent variable takes three outcomes, which follow a natural ordering -win (3 points), draw (1 point) and loss (0 points). Initially, a latent variable * can be given as follows: * = ′ + where, * is an unobservable variable that signifies the game result, is a vector of explanatory variables for observation , is a vector of unknown parameters that we aim to estimate and is a random error term. As we assume that the term is normally distributed, our unobservable variable * produces the three observed variables: where 1 and 2 are the unknown threshold parameters that define the boundaries of the different levels of performance that our model seeks to estimate. Given that 2 > 1, the appropriate probabilities can be generated as follows: where F is the normal cumulative distribution function. Ordered probit estimates for the threshold parameters and coefficients are generated by the maximum likelihood method. Employing an ordered probit approach is consistent with studies that use a game-by-game approach in such a context (Audas et al., 2002;Hughes et al., 2010). As our data are organized by team, in some cases we have two observations for each match. We allow for correlation of the error term in such cases by clustering standard errors by game. Table 1 shows the means, standard deviations and correlation coefficients for the variables that were employed in this study. Of the total number of leaders (119), 26.9 percent were inside appointments, 39.5 percent were hired directly from another organization and 33.6 percent were hired from a non-top leader outside position. In addition, 17.6 percent of leaders were appointed after the predecessor leader was poached away. In terms of prior leadership experience, 47.1 percent of leaders had domestic top leader experience and 21.0 percent had foreign leadership experience at an organization of a similar stature. Table 1 about here ------------------------------------- Table 2 presents the results of our hypotheses tests, which were analyzed using an ordered probit approach with performance (points) as the dependent variable. Seven models were estimated in Hypothesis 1 predicts that leaders that move directly from one organization to another organization (i.e., were poached) will have a positive association with post-succession organizational performance as compared to those hired that were not directly hired from a leadership position. The results show the coefficient is positive and significant (b= 0.09, p <0.05 in model 3 and b= 0.10, p <0.05 in model 7). The marginal effects (table 3) are positive and significant at 1 point and are even more so at 3 points; therefore it is more likely to be in the win category. Specifically, all else equal, the probability of winning for poached leaders, as compared to non-poached leaders, increases by around 4 percent in any given game. As such, we find evidence to support H1.

RESULTS
Hypothesis 2 suggests that predecessor leaders that move directly to another organization (i.e., were poached away) will have a negative association with post-succession organizational performance (i.e., at the departed organization) as compared to predecessor leaders that were dismissed. The results show that the coefficient is negative and significant (b= -0.09, p <0.10 in model 4 and b= -0.12, p <0.05 in model 7). The marginal effects are negative at 1 point and negative and significant at 3 points, hence it is more likely to be in the loss category. Specifically, all else equal, the probability of losing when a predecessor was poached away, as compared to when the predecessor was dismissed, increases by around 5 percent in any given game. As such, we find evidence to support H2.
Hypothesis 3 states that hired leaders with domestic top leader experience will have a negative association with post-succession organizational performance as compared to leaders with no such experience. The results show the coefficient is negative and significant (b= -0.13, p <0.01 in model 5 and b= -0.10, p <0.05 in model 7). The marginal effects are negative at significant at 1 point and are even more so at 3 points, hence it is more likely to be in the loss category. Specifically, all else equal, the probability of losing for leaders with prior domestic leadership experience, as compared to those with no such experience, increases by around 4 percent in any given game. As such, we find evidence to support H3.
Hypothesis 4 predicts that leaders that have foreign top leader experience will be positively associated with post-succession organizational performance as compared to leaders that have no such experience. The results show the coefficient is positive and significant (b= 0.16, p <0.01 in model 6 and b= 0.12, p <0.05 in model 7). The marginal effects are positive and significant at 1 point and are even more so at 3 points; thus it is more likely to be in the win category. Specifically, all else equal, the probability of winning for leaders with prior foreign leadership experience, as compared to those with no such experience, increases by around 4 percent in any given game.
Hence, we find evidence to support H4.

DISCUSSION AND CONCLUSION
In this study we have examined two aspects of leader succession on post-succession organizational performance: (i) the actual succession event (at the organization level); and (ii) organizations that hired leaders with prior top leader experience (at the individual level). Grounding our ideas in theories of human capital, learning and asymmetric information, and testing our ideas in the context of English soccer organizations, our empirical findings suggest that the majority of our hypotheses have salient implications on post-succession organizational performance. Our work contributes to the work on leader succession and the nascent literature on the effects of leaders with prior top leader experience (see: Elsaid et al., 2011;Hamori & Koyuncu, 2014) as detailed below.

Leader Succession and Organizational Performance
Research has long found equivocal results pertaining to the effect of outside successions on postsuccession organizational performance (see: Karaevli, 2007). Given organizations' propensity to directly appoint leaders with prior leadership experience has increased over time (Karlsson & Neilson, 2009;Lucier et al., 2006), we decomposed outside successions into those that were hired directly from the leader position at another organization and those that were not. In addition, we examined the corollary in terms of the effect on post-succession organization performance when a predecessor leader has departed to take a leadership position at another organization.
First, we found support for the argument that newly appointed leaders that moved directly from a leader position at another organization (i.e., were poached) have better post-succession organizational performance as compared to those that were not in a leadership position. We suggest that hiring a leader of another organization will allow the organization to make a more informed selection decision and hence identify the human capital they require. In addition, the associated search and compensation costs will indicate that the candidate is either their first choice, or thereabouts. Consequently, in being perceived as having the required human capital to lead the organization, the new leader will be better able to gain discretion from owners to influence and reshape the organization (Friedman & Singh, 1989), which will enhance organizational performance. Furthermore, given our empirical focus is that of a single industry study, the finding may reflect the greater ease with which leaders can move between organizations, as compared to studies in which the sample of organizations are more heterogeneous in nature.
In addition, the above finding contributes to recent work on prior top leader experience that has examined leaders that move directly between top leader positions, and runs counter to Hamori and Koyuncu's (2014) finding of partial support for the reverse. We believe the difference in findings across the two studies is due to the fact that we focus on the decision to appoint an outside candidate and compare the performance effects against those not currently in a leadership position. In contrast, Hamori and Koyuncu (2014) compare direct CEO-CEO moves against individuals that held an executive position (e.g., a COO) between CEO jobs at their new organization, which in effect are inside candidates (i.e., they have firm-specific experience). Furthermore, our finding perhaps helps to highlight Friedman and Singh's (1989) assertion that post-succession performance is most likely to lead to performance increments when the selection process is rational in that the potential candidate's competencies are better matched to the organization's requirements and when they are given the discretion to shape the organization.
Second, we find that the departure of predecessor leaders to go to another organization (i.e., leaders who are poached away) has a negative relationship with post-succession organizational performance. To our knowledge, ours is the only study that has examined such a relationship. We suggest that the new leader will enter an organization that is highly anchored in the decisions of the predecessor and so will likely face inertial pressures (Fondas & Wiersema, 1997;Hannan & Freeman, 1977;Quigley & Hambrick, 2012). Furthermore, the loss of firm-specific human capital will be drain knowledge that likely led to high performance outcomes. These factors will negatively affect leader's ability to shape their new organization and performance accordingly. In highlighting the post-succession performance effects of how the prior leader stepped down from the leader role we contribute to the literature on the performance consequences of succession, which may shed new light on the previous equivocal findings (see: Giambatista, Rowe, & Riaz, 2005;Karaevli, 2007;Kesner & Sebora, 1994).
In addition, we contribute to the literature on succession as an adaptive mechanism. Prior research has shown factors that prevent a leader from adapting to their new organizations such as longer predecessor leader tenure (Shen & Cannella, 2002b) and predecessor leader remaining as board chair (Quigley & Hambrick, 2012). While there may be many organizational and environmental factors that prevent a leader from adapting (see: Hannan & Freeman, 1977), our findings show that an important reason could be the predecessor leader departing to go to another organization. Furthermore, our finding illuminates Lucier et al's (2006: 9) argument about how poaching a leader, which they refer to as "beggar thy neighbor" recruitment, is problematic for the raided organization because they also have to find a replacement as a consequence. As such, these organizations will likely face lost opportunities and disruptions while the new leaders are being selected and being brought up to speed.

Prior Top Leader Experience and Organizational Performance
We examined the relationship between prior top leader experience and post-succession organizational performance because although the number of newly appointed leaders with top leader experience has increased, studies have shown mixed performance effects (Bragaw & Misangyi, 2013;Elsaid et al., 2011;Hamori & Koyuncu, 2014). We suggest that the presence of equivocal results stems from a lack of precision as to the nature of prior job-specific experience, and to address this we decomposed leader experience into leader experience at a domestic competitor firm (i.e. a firm of similar standing domestically) and that of being a leader at a foreign competitor (again of similar standing).
First, we found that experience at a domestic competitor is negatively related to post-succession organizational performance. Our results suggest that having similar experience (i.e., in the same industry and, in our case the same industry domestically) may do little to introduce new insights into the organization and hence leaders will persist with similar practices. In addition, leaders' belief of the transferability of leadership experience and their oversight as to the importance of idiosyncratic differences organizations possess may also relate to low performance. This finding is consistent with the work on job-specific (i.e., top leader) experience and post-succession organizational performance that has found that prior leadership experience in the same industry will harm performance rather than improve it (Hamori & Koyuncu, 2014).
Second, we found that prior leadership experience at a foreign competitor is positively related to post-succession organizational performance. Our result suggests that having foreign experience will likely enrich the human capital of leaders, which will allow them to not only have leadership exposure but also be innovative and think independently of their organization's context. The finding is consistent with work on CEO's prior international experience (at non-CEO levels) that has found such experience will likely be a key organizational resource that can positively impact performance (Carpenter et al., 2001;Daily et al., 2000). Furthermore, because soccer is a globalized industry (Brady et al., 2008), this finding may provide evidence that foreign experience is perhaps most pronounced in such contexts. Accordingly, our finding is in line with current debates about the benefits of international experience on organization's bottom-line performance (see : Carpenter, Sanders, & Gregersen, 2000).
Overall, our findings about domestic and foreign prior leader experience hold important insights into the relationship between leader human capital and performance, and the debate on whether or not general managerial human capital is portable or context specific (Hamori & Koyuncu, 2014;Murphy & Zábojnik, 2004& 2007.

Managerial Implications
Our findings hold important early tenure performance implications for organizations when considering succession decisions. We suggest that the distinction between an inside and outside appointment may not be particularly relevant, rather it is the experience of the leader, and his/her congruence with the hiring organization, that has the greatest effect on performance.
First, we begin by considering the negative post-succession implications of a leader being hired away by a competitor. We argue that the hiring away of a leader constitutes a disruptive event for the organization, and one they may not be able to prepare for. In order to minimize the probability of such an occurrence we suggest that owners need to look carefully into the different means by which they can ensure that their leader does not have their head turned. Different retention strategies may include the provision of extended contracts, enhanced pay, and also performance related pay that is contingent on the leader remaining in post. Also, as owners cannot prevent a leader from leaving they need to be vigilant in terms of being clear about what they want from any leader, whether in post of the successor, and the identification of potential talent (both internal and external). We argue that not only should owners develop their own talent as potential heirs to the incumbent after mandatory retirement as well as having an understanding of the external talent market; they also need to bear in mind that the incumbent could depart at any time. Consequently, owners need to plan for such an occurrence, and think through who would be put into place in the event of a sudden departure (Dutra & Griesedieck, 2010). Although turbulence may be inevitable, by being clear about what the requirements are for the leader, and identifying potential candidates, owners may be able to make more informed hiring decisions that are less influenced by the problems of asymmetric information and adverse selection.
Second, the findings about the relationship between prior top leader experience and postsuccession performance hold important insights for owners when hiring new leaders. Our results suggest that appointing leaders with leadership experience in the same domestic industry has a negative effect on post-succession performance; conversely leaders with prior foreign industry leadership experience have a positive effect on post-succession performance. We suggest, therefore, that owners need to think through what skills and knowledge they want the new leader to bring to their organization, and how they diversify the existing knowledge base of the organization. For example, having the same domestic industry experience may limit the individual's ability to bring new knowledge to an organization. Once owners are clear about their requirements, they also need to manage the post-succession integration process carefully. Whether or not the hire has domestic and/or foreign experience, a big challenge for any new leader is the transition to their new company's ways of operating (Davis, 2010). Accordingly, owners need to take steps to ensure that the new leader is acclimatizing to, for example, the new organization's processes, culture and history (Davis, 2010).

Generalizability and Limitations
Our paper follows the well-established tradition of employing sports industries, and sporting organizations, to study issues of leadership and human resource management (e.g., Berman et al., 2002;Pazzaglia et al., 2012). Sporting organizations, and soccer in particular, can be generalizable to other types of business organizations as they share a range of common features as detailed below (Bolchover & Brady, 2002;Brady et al., 2008).
First, soccer and business organizations are performance oriented. Although the performance of business organizations is commonly based around financial metrics, whereas performance of soccer organizations is based on winning, research suggests that on-field performance is significantly correlated with financial performance (Lewis, 2004;Giambatista et al., 2005).
Second, soccer organizations follow a common set of rules and relegations, which are monitored and enforced by an impartial authority . The rules and regulations govern the nature of competitive interactions between organizations and players, including aspects such as how games are played and the recruitment and retention of human resources (Ndofor et al., 2009).
Indeed, these rules can administer competition similar to those experienced by business organizations that prohibit certain actions through regulation, competition (anti-trust) policy and employment legislation (Ndofor et al., 2009). Third, while leaders of soccer organizations are not identical to their CEO counterparts in business organizations, they share many common characteristics: being accountable to owners for performance and strategic decisions; needing to satisfy salient stakeholder groups (e.g., owners, fans, players, agents, media, community); and have a relatively high degree of control depending on board structure (e.g., duality or split-leadership function) (Hughes et al., 2010;Ndofor et al., 2009;Rowe et al., 2005). Fourth, soccer organizations in the EPL are commercial organizations in that they are privately held, experience corporate activities such as takeover bids, and need to raise revenue from a range of different sources including game-day receipts, television rights, sponsorship and the trading of players. Furthermore, given that players are the key resource of soccer organizations (the majority of any cost base), and that soccer organizations are increasingly taking serious the issue of corporate social responsibility through, for example, community based initiatives such as social and cultural inclusion, then they are very much mirroring business organizations in being oriented towards the triple bottom line.
Based on the above we argue that sporting organizations, and particular soccer organizations, are comparable to business organizations ensuring that our results should generalize beyond the soccer context. Furthermore, it is arguable that business leaders are now arriving where soccer (and other sporting) organization's leaders have always been: open to immense scrutiny, at the mercy of disparate and demanding stakeholders and high levels of transparency (Bolchover & Brady, 2002). In addition, ideas developed in the management of sporting organizations are stimulating thought about their applicability in business and commerce (see : Bryan & Rafferty, 2006;Wolfe, Wright, & Smart, 2006a& 2006b. As with all studies our work does have limitations, which we hope can provide directions for future research. First, like much research in the succession domain, we use archival data from published sources. Future research may use other research methodologies such as direct observations and/or field studies to better ascertain leader origins and the mechanics pertaining to leader initiated departures. Second, leaders that were appointed from outside of the organization in a non-leadership position consisted of dismissed prior leaders. Although the pool of readily available experienced leaders may be greater in the EPL than other contexts, the reduction in leader tenure rates will mean that other industries may mirror the EPL over time. We encourage scholars to examine different industries, and hence other modes of outside successions, as compared to poached leaders. Third, we were unable to compare the effects of leaders being poached away with those that leave due to mandatory retirements because leader exits in the EPL largely consist of dismissals, being poached away and, in some cases, health related departures. Future research using different samples and contexts may be able to shed light on the matter. Finally, our study is conducted in only one industry. Consistent with prior single-industry research in succession and migration (e.g., Boeker, 1997;Wu, Levitas, & Priem, 2005), we argue that such a context allows us to better control for industry effects and have clear boundary conditions for our ideas. Given the data advantages as outlined above, and that the EPL is rife with top leader movements across organizations, we suggest that the benefits dominate any drawbacks present in our study. However, we encourage scholars to explore the issues we raise in this paper across different industry and organizational contexts to assess formally the extent to which our results are generalizable.    -4785.417 -4785.309 -4782.756 -4783.850 -4780.422 -4780.159 -4773.190 Robust standard errors in parentheses. Significant levels: † p<0.10, * p<0.05, ** p<0.01, *** p<0.001