Table 4 presents the descriptive statistics for the sample understudy during the period 2000 – 2019. The table reflects the number of observations (N), the mean, standard deviation, minimum, and maximum for the dependent variables (ATO and OCF), firm-specific variables (SIZE and LVG), economic variables (GDP, INF, UNEMP), institutional variables (CORR), and governance variables (%IND and DUA) under study. Statistics show that the mean of the two-year average effective ATO of the sample is 0.718 with a standard deviation of 0.665, a minimum of 0, and a maximum of 10.11, while the mean of the three-year average effective post-succession ATO is 0.719 with a standard deviation of 0.657, a minimum of 0 and a maximum of 9.489. Additionally, the mean of the two-year average effective OCF of the sample is 0.074 with a standard deviation of 0.115, a minimum of -2.031, and a maximum of 4.396, while the mean of the three-year average effective post-succession OCF is 0.076 with a standard deviation of 0.104, a minimum of -1.331 and a maximum of 2.928. For firm-specific variables, the average firm size is 13.847 with a standard deviation of 2.117, a minimum of 8.039, and a maximum of 18.891. Moreover, the average firm’s LVG is 0.547 with a standard deviation of 0.238, a minimum of 0.058, and a maximum of 1.278.
For the economic control variables, the average GDP is 27.531 with a standard deviation of 1.354, a minimum of 23.026, and a maximum of 30.27. The average inflation rate is 4.607 with a standard deviation of 5.932, a minimum of -4.863, and a maximum of 85.747. Further, the average unemployment rate is 8.584 with a standard deviation of 8.27, a minimum of 0.091, and a maximum of 33.473. The average rate for control of corruption, the institutional control variable in this paper, is 49.986 with a standard deviation of 14.8, a minimum of 9.596, and a maximum of 91.878. Finally, for the governance control variables, the average percentage of independent board members is 0.458 with a standard deviation of 0.188, a minimum of 0, and a maximum of 0.923. While the CEO duality has an average of 0.439 with a standard deviation of 0.496, a minimum of 0, and a maximum of 1.
To test the proposed hypotheses, the post-succession performance of opposite directions of change in CEO attributes is compared using three different models: the impact of the succession itself, the increase in the CEO successor attribute, and the decrease in the CEO successor attribute compared to the predecessor CEO. To get more insights into the impact of the changes in CEO successor attributes, the study tests the impact on two different periods: the two- and three-year average post-succession performance. Finally, to eliminate the potential bias of any existing multicollinearity, the maximum VIF score for each model is included.
Table 5 shows that the firm’s post-succession ATO is negatively related to CEO succession, hence (H1) is not rejected. For example, models 1 and 4 show -1.59 percent (significant at 5% significance level) and -1.27 percent (significant at 10% significance level) decrease in the firm’s post-succession ATO respectively. The results additionally show that hiring an insider CEO successor significantly reduces the negative impact of CEO succession on the firm’s post-succession ATO as reflected in the coefficient of model 2 denoting a -1.44 percent (significant at 10% significance level) decrease in the firm’s post-succession ATO. On the other hand, hiring an outsider CEO successor reveals mixed although insignificant results on the firm’s post-succession ATO.
Models 7 and 10 show that the firm’s post-succession OCF is negatively related to CEO succession, and hence (H2) is not rejected. The coefficients of Models 7 and 10 show -0.448 percent (significant at 10% significance level) and -0.665 percent (significant at 1% significance level) decrease in the firm’s post-succession OCF respectively. Models 8 and 11 show that hiring an insider CEO successor significantly reduces the negative impact of CEO succession on the firm’s post-succession OCF. The coefficient of Model 11 denotes a -0.439 percent (significant at 10% significance level) decrease in the firm’s post-succession OCF. On the other hand, hiring an outsider CEO successor significantly amplifies the negative impact of CEO succession on the firm’s post-succession OCF. For example, Models 9 and 12 show -1.16 percent (significant at 5% significance level) and -1.11 percent (significant at 1% significance level) decrease in the firm’s post-succession OCF respectively. The results presented in Table 5 conclude that outsider CEO successors have an insignificant impact on the firm’s post-succession asset utilization. Although, insider CEO successors have a significantly positive impact on the firm’s post-succession performance, (H3) is rejected. This conclusion is in support of scholars claiming that solely the origin of the CEO successor has an insignificant impact on the firm’s future performance (Ishak et al., 2013). Finally, the results are in support of Mohammad et al. (2020) concluding that CEO origin has an insignificant impact on the firm’s ROA, ROE, and ROS.
However, outsider CEO successors have significantly worse post-succession liquidity than insider CEO successors, amplifying the negative impact of CEO succession on the firm’s post-succession OCF. While insider CEO successors significantly reduce the negative impact of CEO succession on the firm’s post-succession OCF. Accordingly, there is a significant relationship between the change in CEO successor origin (hiring an outsider CEO successor) and the firm’s post-succession liquidity and hence (H4) is not rejected. The results are consistent with the agency theory and information asymmetry stating that outsider CEOs’ previous behavior and performance is unknown to the hiring firm, which may cause adverse selection problems and increase agency costs. Unlike insider CEO successors who are promoted based on their observed performance in the firm, and accordingly hiring an outsider CEO may lead to a negative firm performance (Zhang and Qu, 2016; Garcia-Blandon et al., 2019; Saidu, 2019; Wang, 2021).
The results are also in support of studies stating that hiring outsider CEO successors amplifies succession disruptions and leads to lower post-succession performance. While insider CEO successor’s familiarity with the firm’s culture and employees grants them an edge while reducing disruptions, and hence enhances the firm’s post-succession performance especially if proper relay succession planning is considered (Kesner and Sebora, 1994; Shen and Cannella, 2002; Zhang and Rajagopalan, 2004; Zhang and Qu, 2016; Georgakakis and Ruigrok, 2017; Garcia-Blandon et al., 2019; Saidu, 2019; Wang, 2021; Choi et al., 2022). Moreover, Serra et al. (2016) conclude that outsider CEOs do not enhance the firm’s profitability, productivity, and operational efficiency. Finally, Qi (2011) concludes that to enhance post-succession firm performance, an insider successor with an appropriate early set succession plan is the solution.
Table 6 shows that while controlling for CEO gender, both the firm’s post-succession ATO and OCF are negatively related to CEO succession, hence (H1 and H2) are not rejected. Hiring a CEO successor of a different gender reduces the negative impact of CEO succession on the firm’s post-succession ATO although the results are insignificant. On the other hand, hiring a CEO successor of the same gender significantly amplifies the negative impact of CEO succession on the firm’s post-succession ATO. For example, the coefficient of Model 3 shows a -1.64 percent (significant at 5% significance level) decrease in the firm’s post-succession ATO. The results conclude that the change in the CEO successor’s gender has an insignificant impact on the firm’s post-succession performance. Although, CEO successors of the same gender have a significantly negative impact on the firm’s post-succession performance, (H4) is rejected.
In addition, the coefficients of Models 8 and 11 show that hiring a CEO successor of a different gender significantly amplifies the negative impact of CEO succession on the firm’s post-succession OCF. For example, the coefficient of Model 11 is -1.48 percent (significant at 5% significance level) decrease in the firm’s post-succession OCF. On the other hand, the coefficient of Model 12 is -0.562 percent (significant at 5% significance level) decrease in the firm’s post-succession OCF, which demonstrates that hiring a CEO successor of the same gender significantly reduces the negative impact of CEO succession on the firm’s post-succession OCF.
The results conclude that the change in CEO successor's gender has a significantly negative impact on the firm’s post-succession liquidity amplifying the negative impact of CEO succession on the firm’s post-succession liquidity and thus (H4) is not rejected. These results are in support of the pre-mentioned literature stating that male and female CEOs have different leadership, managerial, and communication styles. Thus, hiring a CEO of a different gender may increase CEO succession disruptions and negatively affect the firm’s post-succession performance (Finkelstein et al., 2009; Zhang and Qu, 2016; Zhang et al., 2019).
Table 7 shows that while controlling for CEO age, both the firm’s post-succession ATO and OCF are negatively related to CEO succession, and hence (H1 and H2) are not rejected. The results show that hiring an older CEO successor reduces the negative impact of CEO succession on the firm’s post-succession ATO, although the results are insignificant. While hiring a younger CEO successor significantly amplifies the negative impact of CEO succession on the firm’s post-succession ATO. For example, the coefficient of Model 3 shows a -2.26 percent (significant at 5% significance level) decrease in the firm’s post-succession ATO. Additionally, the findings presented in Models 7 and 10 show that hiring an older CEO successor reduces the negative impact of CEO succession on the firm’s post-succession OCF although the results are insignificant. On the other hand, hiring a younger CEO successor significantly amplifies the negative impact of CEO succession on the firm’s post-succession OCF. For example, the coefficients of Models 9 and 12 show -0.625 percent (significant at 5% significance level) and -0.842 percent (significant at 1% significance level) decrease in the firm’s post-succession OCF respectively.
The results conclude that the direction of change in the CEO successor’s age is significantly positively related to the firm’s post-succession asset utilization and liquidity. Accordingly, hiring a younger CEO successor results in amplifying the negative impact of CEO succession on the firm’s post-succession asset utilization. Thus, (H3 and H4) are not rejected, signifying that hiring older CEO successors have better post-succession performance than hiring younger CEO successors. The results are in support of the UET and the human capital theory, stating that the CEO’s age affects his human capital quality and capabilities, and thus the firm’s performance. The results also support studies stating that older CEOs have more knowledge, greater experience, and better managerial skills and are thus considered better human capital with better-expected firm performance (Hambrick and Mason, 1984; Henderson et al., 2006; Herrmann and Datta, 2006; Zhang, 2010; Knutsson, 2015; Von den Driesch et al., 2015; Serra et al., 2016; Wang et al., 2016; Burney et al., 2021; Minh Ha et al., 2021; Zhao, 2021). Likewise, these results are in support of Peni (2014), Kokeno and Muturi (2016), and Wang et al. (2016).
Table 8 shows that while controlling for the CEO successor’s number of qualifications and postgraduate studies as presented in Panels A and B respectively, both the firm’s post-succession ATO and OCF are negatively related to CEO succession, thus (H1 and H2) are not rejected. As shown in Panel A, hiring a CEO successor with a lower education level significantly amplifies the negative impact of CEO succession on the firm’s post-succession ATO and OCF. For example, the coefficients of Models 3, 9, and 12 are -2.38 percent (significant at 10% significance level), -2.07 percent (significant at 1% significance level), and -1.87 percent (significant at 1% significance level) decrease in the firm’s post-succession performance respectively.
Furthermore, panel B shows that post-graduate degrees on the other hand have an insignificant impact on ATO. Accordingly, the direction of change in the CEO successor’s education reveals mixed results and hence (H3) is rejected. This shows that the number of CEO certificates influences the CEO’s cognitive ability and competence, which affects the firm’s asset utilization efficiency, however having a post-graduate degree does not have a significant impact on the CEO’s management efficiency. The results are in support of Ofe (2012), Serra et al. (2016), Ghardallou et al. (2020), and Mukherjee and Sen (2022) concluding that holding any postgraduate certificate has an insignificant impact on a firm’s performance. Finally, the results demonstrate that hiring a CEO successor without postgraduate degrees following a degree-holder CEO predecessor significantly amplifies the negative impact of CEO succession on the firm’s post-succession OCF. For example, the coefficients of Models 9 and 12 are -2.10 percent (significant at 1% significance level) and -1.15 percent (significant at 5% significance level) decrease in the firm’s post-succession OCF respectively. The results presented in Table 8 show that the direction of change in both CEO successor’s education measures is significantly positively related to post-succession liquidity. Thus, (H4) is not rejected.
The results are in support of the UET and the human capital theory claiming that education is an important managerial qualification, and hence highly educated executives are better human capital and of higher quality to the firm and thus can enhance the firm’s future performance. The results also support the RDT claiming that highly educated CEOs have better skills and cognitive ability to gather and integrate internal and external resources. Such capabilities help highly educated CEOs make better decisions and enhance the firm’s performance (Hambrick and Mason, 1984; Rajagopalan and Datta, 1996; Barker and Mueller, 2002; Serra et al., 2016; Wang et al., 2016; Garcia-Blandon et al., 2019; Saidu, 2019; Ghardallou et al., 2020; Zappalà, 2020; Minh Ha et al., 2021; He et al., 2022). Likewise, the results support Bertrand and Schoar (2003) claiming that CEOs with postgraduate degrees enhance the firm’s performance. Finally, the results are in support of Datta and Rajagopalan (1998) and Kokeno and Muturi (2016).
Table 9 demonstrates that while controlling for different CEO experience measures (CEO total number of boards to date and time in company) as shown in Panels A and B respectively, the firm’s post-succession ATO and OCF are negatively related to CEO succession, therefore (H1 and H2) are not rejected. The results additionally show that hiring a CEO successor with higher experience reduces the negative impact of CEO succession on the firm’s post-succession ATO, although the results are insignificant. Quite the opposite, the results show that hiring a CEO successor with less experience significantly amplifies the negative impact of CEO succession on the firm’s post-succession ATO. For example, the coefficients of Models 3 and 6 Panel A demonstrate -3.00 percent (significant at 1% significance level) and -3.14 percent (significant at 1% significance level) decrease in the firm’s post-succession ATO respectively. Additionally, the coefficient of Model 6 Panel B demonstrates a -1.72 percent (significant at 10% significance level) decrease in the firm’s post-succession ATO.
Moreover, hiring a CEO successor with larger board experience significantly reduces the negative impact of CEO succession on the firm’s post-succession OCF. While, hiring a CEO successor with more company tenure reduces the negative impact of CEO succession on the firm’s post-succession OCF, although the results are insignificant. However, hiring a CEO successor with less board experience or less company tenure significantly amplifies the negative impact of CEO succession on the firm’s post-succession OCF. The results show that hiring a CEO successor with less experience significantly amplifies the negative impact of CEO succession on both the firm’s ATO and OCF, hence (H3 and H4) are not rejected. The results are in support of the UET stating that experience affects the individual’s decision-making process and hence the firm’s performance; as it shapes the CEO’s wisdom, skills, knowledge, and cognitive ability required for efficient information interpretation and utilization. (Hambrick and Mason, 1984; Henderson et al., 2006; Herrmann and Datta, 2006; Serra et al., 2016; Wang et al., 2016; Farag and Mallin, 2018; Minh Ha et al., 2021).
In addition, the results support the RDT claiming that experienced CEOs facilitate the firm’s access to different required resources through their connections which subsequently enhances the firm’s performance. Moreover, experienced CEOs are considered more prestigious and powerful attracting better human capital to the firm and thus improving the firm’s performance as the human capital theory claims (Bigley and Wiersema, 2002; Serra et al., 2016; Wang et al., 2016; Minh Ha et al., 2021). The results also support the claim that CEOs with higher experience decrease agency problems and agency costs. This is explained by the implication that they can better understand the surrounding environment and the stakeholders’ needs which will subsequently enhance the firm’s performance (Bigley and Wiersema, 2002; Henderson et al., 2006; Herrmann and Datta, 2006; Serra et al., 2016; Wang et al., 2016; Farag and Mallin, 2018). Finally, the results are in support of Harymawan et al. (2019) and Wang et al. (2016).
Table 10 demonstrates that while controlling for the CEO social network, both the firm’s post-succession ATO and OCF are negatively related to CEO succession, and hence (H1 and H2) are not rejected. The findings show that hiring a CEO successor with a larger social network reduces the negative impact of CEO succession on the firm’s post-succession ATO, although the results are insignificant. On the other hand, the results show that hiring a CEO successor with a smaller social network significantly amplifies the negative impact of CEO succession on the firm’s post-succession ATO as demonstrated in Model denotes -1.89 percent (significant at 10% significance level) decrease in the firm’s post-succession ATO. Finally, the results demonstrate that hiring a CEO successor with a larger social network significantly reduces the negative impact of CEO succession on the firm’s post-succession OCF. For example, the coefficient of Model 11 shows a -0.638 percent (significant at 5% significance level) decrease in the firm’s post-succession OCF. On the other hand, the coefficients of Models 9 and 12 show -1.09 percent (significant at 1% significance level) and -0.770 percent (significant at 5% significance level) decrease in the firm’s post-succession OCF respectively.
The results conclude that the direction of change in the CEO successor’s social network is significantly positively related to both the firm’s ATO and OCF and thus (H3 and H4) are not rejected. Moreover, the results are consistent with Burt (2000) stating that social capital is an “advantage” to its owners, and thus highly connected CEOs have differential successes in their business and activities. Likewise, the human capital theory claims that socially connected CEOs with good quality network connections are considered prestigious, helping their firms attract better human capital quality (Chuluun et al., 2014; Romano et al., 2020b; Griffin et al., 2021). The results also support the claim that CEOs with larger networks are more pressured to meet the market’s expectations, enhance the firm’s performance, and prove themselves worthy (Griffin et al., 2021).
Furthermore, the results are in support of both the social network theory and the RDT claiming that the CEO’s social network facilitates the CEO’s accessibility to different information and resources through his connections. Thus, it is considered a critical resource and an asset as it improves the CEO’s decision-making process and increases his power in the firm, which improves the firm’s performance. In other words, the CEO’s social network allows him to access cheaper resources and hence effectively manage the firm’s expenses and enhance the firm’s performance (Harjoto and Wang, 2020; Kirchmaier and Stathopoulos, 2008; Von den Driesch et al., 2015; Romano et al., 2020b). The results also support the literature claiming that hiring highly connected CEOs reduces information asymmetry and adverse selection costs and consequently decreases agency costs by decreasing the stakeholders’ concerns about the firm and its future performance (Huang and Shang, 2019; Wang, 2021; Zhao, 2021). Finally, the results are in support of Kirchmaier and Stathopoulos (2008), Horton et al. (2012), and Sun et al. (2020).
Table 11 shows that while controlling for CEO busyness, both the firm’s post-succession ATO and OCF are negatively related to CEO succession, accordingly (H1 and H2) are not rejected. Additionally, the coefficients of Models 2, 5, 8, and 11 show -4.56 percent (significant at 1% significance level), -4.1 percent (significant at 1% significance level), -1.37 percent (significant at 1% significance level), and -1.2 percent (significant at 1% significance level) decrease in the firm’s post-succession ATO and OCF respectively. The results signify that hiring a busier CEO successor significantly amplifies the negative impact of CEO succession on the firm’s post-succession asset utilization and liquidity. However, hiring a less busy CEO reveals mixed significant results on the firm’s post-succession ATO, and thus (H3) is rejected.
While hiring a less busy CEO successor significantly reduces the negative impact of CEO succession on the firm’s post-succession liquidity, as shown by the coefficient of Model 12, signifying a -0.848% percent (significant at 5% significance level) decrease in the firm’s post-succession OCF, thus (H4) is not rejected. The results are in support of the agency, busyness, and over-boarding theories claiming that the large number of board meetings and commitments imposed on busy CEOs will eventually force them to compromise some of their duties to fulfill others. Accordingly, CEO busyness hinders the firm’s performance (Sarkar and Sarkar, 2009; Andres et al., 2013; Kutubi et al., 2018; Harymawan et al., 2019; Ratri et al., 2021). Finally, the results are in support of Mukherjee and Sen (2022) and Harymawan et al. (2019).
Using FE regression models, the findings show that while controlling for different CEO attribute measures across different models, CEO succession has a significant negative impact on all firms’ post-succession performance. Moving to the firm-specific control variables, for both ATO and OCF, the pre-succession firm performance is significantly positively related to the firm’s post-succession performance supporting Kim (2005), Horton et al. (2012), Hamori and Koyuncu (2015), Zhang and Qu (2016), Georgakakis and Ruigrok (2017), Le and Kroll (2017), Zhang et al. (2019), and Sun et al. (2020) results. Merely for some of the three-year post-succession OCF models, the pre-succession firm performance is positively, although insignificantly, related to the post-succession OCF. In addition, SIZE is significantly negatively related to both performance measures, contradicting the aforementioned literature.
Moreover, concerning the asset utilization measure (ATO), in all models, LVG is significantly positively related to the firm’s post-succession ATO. The results are consistent with Ishak et al. (2013) and Zhang et al. (2019) results. On the contrary, for the liquidity measure (OCF), in all models, LVG is significantly negatively related to the post-succession OCF in support of Peni, (2014), Saidu (2019), and Ghardallou et al. (2020). Finally, multi-collinearity is not a problem; all variables’ scores are less than 2.5, which is within all differently stated acceptable parameters (Menard, 2002; Vittinghoff et al., 2006; Gareth et al., 2013; Johnston et al., 2018).