2.1 CEO’s functional experience breadth and firm performance

In recent years, scholars have begun to pay attention to the breadth dimension of CEOs’ functional experience. Studies of CEOs’ duality and functional experience have both produced different perspectives [19–21]. The critical roles of the CEO and board chairman in a company are different. The CEO is responsible for managing and operating the company; the board chairman is responsible for leading the board. Duality occurs when the CEO also serves as the board chairman. Agency theory suggests that such duality creates a powerful CEO. They may use their power to hinder the board’s decisions and, thus, affect the firm’s performance [9, 12]. Management theory has argued that a CEO’s duality establishes strong leadership through unified command. Therefore, CEOs may make better decisions that benefit firm performance [9, 12]. Other scholars have studied the effect of CEOs’ functional experience on firm performance. Hamori et al. investigate the impact of prior CEO experience on firm performance [22]. They find that experience in the CEO position is negatively related to firm performance. Also, Georgakakis et al. provide evidence on the relationship between CEO characteristics and organizational performance [23]. Their research demonstrates that the performance implications of a new CEO origin should not be considered in isolation but in interaction with multilevel characteristics. Chahyadi et al. examine how the type of CEOs’ industry experience (whether a CEO has cross-industry or specific-industry experience) on firm performance [24]. They find that CEOs with cross-industry experience tend to relatively lower firm performance. On the other hand, CEOs with specific-industry experience lead firms to higher performance, especially among high-growth firms. Zhang et al. showed that executives who have returned from working overseas could improve firm performance, especially in non-state-owned enterprises, large firms, and firms in East China. Further analysis of this mechanism shows that overseas-returned executives influence firm performance mainly through risk-taking mechanisms [25]. The results of [26] also supported the contention that firms experiencing low performance were more likely to recruit a CEO with a background in operations.

In addition to single functional experience studies, some scholars have also conducted studies of multiple functional experiences. Frederiksen et al. found a significant positive relationship between the number of roles experienced by a senior person in the labor market and their chances of career success [27]. A series of studies of different functional experiences were generated based on this. On the one hand, studies based on the domain expertise framework [27] argued that a CEO’s combined functional experience could leverage specific resources. These experiences can enhance adaptability [28], improve the firm’s relationship with investors [28], and help the firm to escape from financial difficulties [29] and transfer across companies [30], thus, contributing to the improvement of the firm performance. On the other hand, CEOs with combined functional experience face ambiguity when adjusting the external environment and the current resources available to the company, which may lead to lower firm performance, as the combination of experience with firm-specific resources achieves limited success when driven by different achievements [31]. In addition, Li and Patel investigate the impact of CEO generalist experience and firm performance [32]. They propose a negative association between more generalist CEO experience and firm performance, but one that is alleviated by longer tenure. In particular, knowledge and industry are less relevant in new environments. Based on these factors, the combined functional experience of CEOs may be negatively related to firm performance.

2.2 CEO’s functional experience depth and firm performance

Studies of the depth of CEOs’ functional experience have focused on the length of tenure and the hierarchical structure of political connections. Tenure-based studies mainly consider the effective accumulation of relevant expertise gained by managers as tenure increases, including the ability to be more thoughtful in strategy formulation and implementation [33], the ability to achieve implicit learning through a better understanding of practices and existing knowledge [34], and the ability to learn experientially through a combination of resources [35]. All three modes may enable firms to achieve better performance returns. Resource-based theory research suggests that a CEO accumulates a wealth of knowledge and organizational memory in the field during a long tenure and that these contribute to the firm’s performance [36]. Bergh et al. showed that a CEO’s tenure is positively related to firm performance [36, 37], and this finding is empirically supported by other researchers [38, 39]. Studies based on the upper echelons theory suggest that longer tenure is detrimental to firm performance. Hambrick et al. showed that long-tenured CEOs often rely on internal channels and ego models to achieve success [13]. They become risk-averse and increasingly distant from fresh and accurate external market information [36, 40]. Therefore, this is detrimental to firm performance [41]. Cai et al. examined the relationship between a CEO’s tenure and firm performance. Their findings also indicated that a longer CEO tenure positively affects firm performance.

In addition to tenure, scholars have made a series of explorations regarding the depth of functional experience through the hierarchical structure of CEOs’ political connections. Almost all executives are politically connected [41]. Resource dependence theory suggests that CEOs with political affiliations help firms obtain resources to cope better with challenging contingencies. Political connections may, thus, benefit firm performance. Wu et al. suggested that CEOs’ political connections positively impact firm performance. This impact is stronger in less-developed regions [42]. In contrast, Ling et al. used a sample of real estate firms. They showed that political connections are negatively related to firms’ financial performance. Cultivating political connections may be risky and not always reward performance [43].

In summary, existing research demonstrates the centrality of CEOs’ functional experience in studying firm performance. However, there are some shortcomings in the existing studies. Breadth-based studies do not consider the diversity and repetition of functions. An approximate count of the number of positions is not sufficient to reveal the breadth of CEOs’ experience. Depth-based studies do not consider the length of tenure at each stage, and these studies do not contain a hierarchy of individual positions. Therefore, this paper combined breadth and depth to portray CEOs’ cumulative perceptions and investigate whether CEOs’ functional experience can affect firm performance.

3.1 CEO’s functional experience breadth and firm performance

Hambrick and Mason first proposed the upper echelons theory in 1984. This theory links the demographic characteristics of executives to firm outcomes and argues that executive characteristics are ultimately reflected in corporate strategy choices and firm performance [43]. Executives play a dominant role in firm management. Managers may have different options when faced with corporate decisions because of the potential influence of their characteristics.

The upper echelons theory suggests that in complex corporate environments, individual executives’ decisions are driven by "limited rationality" in the face of complex and uncertain situations [1]. Recent studies have found that CEO’s experience hinders subsequent firm performance [22, 44]. The rationale is that prior experience is too heavily laden with the specific environments in which it was gained and therefore is not as beneficial to the new firms as the CEOs believe it will be. In short, available evidence suggests that past experience is a detriment to success in a subsequent job, mainly because experienced CEOs have to give up firm-specific skills that are not useful to a new firm [22]. It is exacerbated by the tendency of CEOs to become less adaptable with increasing experience [40, 45, 46]. Therefore, as the CEO’s functional experience increases, they face more and more instability in the environment and decision-making, which is not conducive to improving the firm performance [47]. The upper echelons theory believes that the individual’s experience characteristics also affect risk preference [1]. When a listed company hires a CEO, their experience is usually considered. In reality, more and more listed companies hire experienced CEOs because they can provide more professional policy advice for the company’s development. On the one hand, experienced CEOs can apply their expertise to practical decisions, helping to make sound decisions. On the other hand, reliable and prudent practice requirements will also prompt them to form corresponding characters and habits with a lower risk appetite [48]. Suppose the incentive mechanism is not complete and the supervision system is not perfect, CEOs may make the pursuit of maximizing their interests their goal, borrowing managerial authority and information advantage for their development to gain more benefits, regardless of improving company performance.

However, improved firm performance is associated with market share expansion and technological innovation. High firm performance may be accompanied by more significant risk. With sufficient functional experience, the CEO has ample time and opportunity to bond with the company’s staff and mechanisms to mitigate the company’s nurturing costs. After the CEO and the other members have connected, exceptional productivity is created, in addition to a working understanding, showing a "special cooperation" relationship. Furthermore, the wealth of experience enables them to replicate success stories. It is used to resist customer churn, technological conservatism, and declining performance, ensuring a steady improvement in company performance [1].

Based on the above analysis, firm performance weakens with increased CEO richness. However, with further expansion of functional experience richness, human capital and risk resistance are strengthened, enhancing firm performance. As a result, the following hypothesis was proposed.

1. H1: As a CEO’s functional experience breadth increases, it has a u-shaped relationship with firm performance.

3.2 CEO’s functional experience depth and firm performance

As a firm’s business grows, the scope and complexity of its CEO’s unconventional strategic decisions become more extensive. However, most CEOs lack the necessary information-processing skills to deal with unexpected events [49]. As the depth of experience increases, the CEO accumulates sufficient resources from higher levels and longer tenure beyond the current company. It is easier for them to make strategic decisions to improve firm performance. This view is supported by scholars who argue that CEOs’ knowledge of the growth of their tenure is conducive to enhancing organizational competitiveness and firm performance. At the same time, the market law of survival of the fittest means that CEOs with poor performance are eliminated early. Those with greater in-depth experience are relatively more capable talents. CEOs with deep experience mark personal ability and better firm performance [50]. It has also been shown that CEO promotion positively affects firm performance [51]. Greater career opportunities improve CEOs’ managerial motivation. The career concern theory suggests that CEOs have two sources of reward for their efforts. One is the monetary reward of the current job, and the other is the career reward of the current job [51]. These two components together motivate CEO effort, and in the context of this paper, the expectation was that a CEO’s depth of tenure would result in higher returns for firm performance.

However, as the depth of experience increases, the CEO’s aversion to risk increases because they represent a significant investment in the company’s future, both on a psychological level and in terms of actual commitment [40]. Risk aversion leads CEOs to resist changes in the status quo, which may be detrimental to firm performance. On the other hand, there is an increasing reliance on internal resources. As these resources have often learned how to cater to the CEO’s information preferences, it isn’t easy to accept new information [13]. At the same time, due to limited rationality constraints [1], CEOs are less likely to consistently address the information and knowledge inputs required for firm performance [13]. In particular, as the depth of a CEO’s tenure increases, they are prone to directional thinking, progressively more information asymmetry, and reduced information processing power. As demonstrated in Kale et al.’s study, CEOs are at the top of the firm for long periods and lack promotion-based incentives [52]. Second, as their tenure increases, overconfidence from existing achievements increases. They may blindly maintain their own decisions, reducing the company and industry fit. Furthermore, if they remain in the same hierarchy for a long time, they are limited by the information sources window. The content of the information they are exposed to and receive becomes increasingly monotonous and is not conducive to their firm’s performance improvement.

Based on the above analysis, firm performance is enhanced with the depth of a CEO’s tenure. However, as this depth deepens, information asymmetry and solidification of thinking bring increasing numbers of problems, and the focus of attention changes, which may hurt firm performance. As a result, the following research hypothesis was proposed.

1. H2: As a CEO’s functional experience depth increases, it has an inverted u-shaped relationship with firm performance.

4.1 Data Source

The sample data of this paper includes all listed companies in China between 2006 and 2020. This paper mainly focused on CEOs of listed companies due to the following considerations. First, the information disclosure requirements of listed companies in China ensure the availability and reliability of data. Second, CEOs of listed companies have a certain degree of mobility, reflecting the mobility of China’s managerial market to a certain extent. The study can provide a preliminary understanding of the current situation of CEOs, which can then be extended to the study of other types of executives. Third, this paper is based on the upper echelons theory. The CEOs’ organizational environment and industry experience of China’s listed companies are relatively wealthy.

The Measures for the Administration of Information Disclosure of Listed Companies was considered and adopted by the China Securities Regulatory Commission in 2006. The data disclosure is complete, so the research data in this paper starts in 2006. This paper automatically extracted data related to functional experience from CEOs’ CVs based on the deep learning model for CEOs’ functional experience richness. Among the data, the resume information was obtained from the CSMAR database. However, there was a problem in that some resume information was missing or hidden. Building a targeted training corpus and maximizing the model’s performance was necessary to obtain more comprehensive and accurate information about CEOs’ tenure. Therefore, this paper also used other information channels to supplement executive CVs. All the CVs obtained from different channels were cleaned using data fusion, and finally, a relatively complete resume text was obtained.

This paper used two variables, return on total assets (ROA) and return on equity (ROE), to measure firm performance. Data for other corporate governance and control variables, such as financial indicators, were obtained from the CSMAR and Wind databases. The initial sample was processed as follows: (1) sample companies from the financial industry were excluded; (2) sample companies with missing data were excluded; (3) to reduce the influence of individual extreme values, all continuous variables were subjected to tailoring at the 1% level. Finally, this paper obtained 15 years of data from 1,862 companies with 27,930 samples. Table 1 shows the variable definitions, and the descriptive statistics of the variables after the tailoring process are shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Variable definition.

https://doi.org/10.1371/journal.pone.0281866.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Descriptive statistics.

https://doi.org/10.1371/journal.pone.0281866.t002

4.2 Variable design

Firm performance: This paper followed the common practice in the literature to construct firm performance indicators. This refers to the performance and efficiency of a firm’s operations, which reflects their effectiveness and is generally reflected by a financial indicator or a set of financial indicators. Typical measures of firm performance in the literature are based on accounting-level calculations, commonly represented by measures such as ROA and ROE [53, 54]. Therefore, in this paper, ROA and ROE represent firm performance.

A CEO’s functional experience: This paper analyzed the functional experience of CEOs in each period. It can be seen from Fig 1 that most CEOs have undergone the evolution of different hierarchies of positions before reaching the highest position. With the support of junior levels such as workers, station managers, and workshop technicians, they enter the core circle of senior-level positions such as directors, managers, and boards of directors through the experience of intermediate-level positions like higher-level project managers, section chiefs, and directors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Evolution of CEO’s position promotion chart.

https://doi.org/10.1371/journal.pone.0281866.g001

Breadth: func_rich was constructed by the functional experience richness of the CEO. This paper used the information entropy formula to calculate the functional experience richness, as shown in Eq (1). Where p_i is the frequency of the ith category of tenure position in the career, and the value range of i is [7, 50]. Positions in the resume were divided into nine categories: production, R&D, design, human resources, management, marketing, finance, legal, and others. For example, if a CEO held 13 positions in their career, of which three were in the finance category, the frequency of positions in the finance category is 3/13. The higher the final calculated score, the more experience the CEO has in each category of positions and the broader the CEO’s breadth.

(1)

Depth: A CEO’s functional experience depth was measured by the function hierarchy (level_score) and the average tenure of each position (avg_emp). Where p_i is the ith category of tenure frequency in the career, r_i is the number of times the ith tenure hierarchy appears, and а_i is the weight of ith level score. The levels were divided into senior, middle, and junior. Combined with the structural ratio of senior, intermediate, and junior positions [55], this paper gave a weighted score of 6:3:1 to senior, intermediate, and junior levels.

On August 31, 2006, the Ministry of Human Resources and Social Security of the People’s Republic of China issued the Implementation Opinions of the Tentative Measures for the Management of Post Establishment in Public Institutions. According to Article 13 of the setting of professional and technical post levels states: the structural proportion of senior, intermediate, and junior posts shall be controlled according to the regional economic and social development level and industrial characteristics, as well as the function, specification, subordination and professional and technical level of the public institution. The overall national control target is 1:3:6 for the proportion of senior, intermediate, and junior posts. Combined with the structural ratio of senior, intermediate, and junior positions, this paper gave a weighted score of 6:3:1 to senior, intermediate, and junior levels.

(2)

4.3 Model design

According to the characteristics of panel data, we test the validity of the fixed-effects model and the random-effects model on the validity of the regression model. Usually, the Hausman test is used to judge the applicability of the two models. The p-value of the Hausman test result is 0.0000, so the null hypothesis is strongly rejected, and a fixed-effects model should be used instead of a random-effects model.

This paper adopts the two-way fixed effects model. Yeardummy and Industrydummy represent the year dummy and industry dummy variables, respectively. Both of them are 0–1 dummy variables. μ_it is the model error term. i is the ith firm. t is the t year. The above control and dummy variables are added to the model to control the influence of other factors further. In this paper, two models were designed to test the hypotheses. The first model (1) is shown in Eq (3), and the second model (2) is shown in Eq (4). Model (1) tests the direct effect of a CEO’s breadth of experience and depth of experience on firm performance. Model (2) tests the relationship between a CEO’s breadth of functional experience and the depth of functional experience and firm performance. In addition to ROA, this paper also examined ROE as a measure of firm performance, replacing ROA with ROE and keeping the other variables the same as these two models.

(3)

(4)

5.1 CEO’s breadth, depth, and firm performance

The regression results of model (1) are shown in column (1) in Table 3. The regression results of model (2) in column (3) are in Table 3. Both are the results of the impact of a CEO’s breadth of tenure and depth of tenure on firm performance in the case of firm performance expressed by ROA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Regression results.

https://doi.org/10.1371/journal.pone.0281866.t003

At the breadth level, the regression results in column (1) show that the regression coefficient of func_rich is significantly negative at the 1% level. The regression results in column (3) show that the regression coefficient of func_rich is significantly negative at the 1% level, and the regression coefficient of its quadratic term func_rich2 is significantly positive at the 5% level. Therefore, there is a u-shaped change in the relationship between a CEO’s functional experience richness and firm performance as it increases, and hypothesis one of this paper is verified.

At the depth level, the regression results in column (1) show that the regression coefficients of level_score and avg_emp are both significantly positive at the 1% level. The regression results in column (3) show that level_score and avg_emp are significantly positive at the 1% level. The regression coefficients of their quadratic terms are significantly negative at the 5% level. There is an inverted u-shaped change in the relationship between a CEO’s depth and firm performance, increasing from weak to strong. Hypothesis two of this paper is verified.

Similarly, columns (2) and (4) both show the results of the effect of a CEO’s breadth of tenure and depth of tenure on firm performance when firm performance is expressed in terms of ROE. The regression results also support the two hypotheses of this paper.

The regression coefficient of the CEO’s gender is significantly positive, which indicates that male CEOs are better at promoting firm performance. The regression coefficient of CEOs’ education (degree) is significantly positive, indicating that CEOs accumulate sufficient knowledge with their increase in education, which benefits the company’s performance. The regression coefficient of CEOs’ sharend is significantly positive, indicating that the higher the level of a CEO’s shareholding, the better the company’s performance. Second, at the firm level, the regression coefficient of firm debt (dar) is significantly negative, indicating that the higher the firm’s gearing, the greater the financial risk it faces, and the resulting proxy for debt service pressure conflicts with the continued need for cash flow for firm performance. The regression coefficient of firm size is significantly positive, indicating that the larger the firm is, the more resources it has to carry out performance improvement activities. The regression coefficients of the number of employees (nstaff) and executive team members (exe_num) are significantly positive. The higher the number of team members, the more information is available. The CEO can obtain timely information, such as technical signals related to company performance, through bottom-up information gathering. The regression coefficient of found_time is significantly positive, indicating that the company accumulates greater resources as time increases, which is conducive to improving firm performance.

5.2 Robustness tests

5.2.1 Alternative approach to firm performance variables.

Firm performance studies often use ROA, ROE, and TobinQ to represent firm performance. Among them, ROA and ROE are described in this paper’s experiments, and TobinQ is the ratio of the market value of the firm to total assets. We use TobinQ measure of firm performance in the robustness test as our dependent variable. TobinQ ratio (or Q as it is commonly known), defined as the capital market value of the firm divided by the replacement value of its assets incorporates a market measure of firm value which is forward-looking, risk-adjusted, and less susceptible to changes in accounting practices [63, 64]. In order to better measure a company’s performance, this paper used PCA (Principal Component Analysis) to analyze the three indicators of ROA, ROE, and TobinQ. PCA is to try to recombine the original many related indicators (such as P indicators) into a new set of comprehensive indicators independent of each other to replace the original ones. The results of the principal component analysis are shown in Table 4. Among them, (Component) represents the name of the principal component extracted by the system. It can be found that a total of 5 principal components have been extracted. The Eigenvalue column represents the eigenvalue of the principal component extracted by the system. The size of the eigenvalue means the explanatory power of the principal component. The larger the eigenvalue, the stronger the explanatory power. The Proportion column represents the variance contribution rate of the principal component extracted by the system, and the variance contribution rate is also the explanatory power of the principal component.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Regression results.

https://doi.org/10.1371/journal.pone.0281866.t004

Ultimately, the characteristic root of the first principal component is 2.4820, which is greater than 1. The first principal component explains 82.73% of the total variables. The expression of the composite indicator of firm performance (PEF) obtained by expressing the first principal component as a linear combination of the individual variables is: (5)

The empirical test was conducted with TobinQ and PEF as the dependent variable. As shown in columns (5) and (6) in Table 3, the regression results support hypotheses one and two.

5.2.2 Control CEO endogeneity.

A potential problem with our findings is that listed companies have other executives besides the CEO. According to our statistics, the top management team changes almost every year. If the top management team is highly fluid, it is difficult to measure whether the CEO or other executives bring about an improvement in firm performance.

Therefore, this paper analysis the stability of the CEO and top management team each year. Fig 2 shows the CEO and top management team changes over the statistical period as a group of company representatives. The blue dotted line indicates whether the CEO has changed in the current year, 1 if the CEO changes, and 0 if there is no change. The solid red line represents the rate of change in the top management team.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The change rate of the CEO and top management team.

https://doi.org/10.1371/journal.pone.0281866.g002

As can be seen from the figure, most listed companies have changed their CEOs during the statistical period. The changes in the top management team are different. Some companies have undergone substantial changes in the top management team during the statistical period, while the changes were relatively small in others. We further make a descriptive statistic of the change rate of the top management team. The rate of change of the top management team is distributed between [0, 1]. The average rate was 13%. Therefore, this paper believes that if the current year’s change rate of the top management team is less than 13%, it is relatively stable. According to this standard, this paper filtered the data of CEO changes and a relatively stable top management team for model testing. The part of the sample is shown in Fig 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. The change rate of the CEO and Top management team in the filtered samples.

https://doi.org/10.1371/journal.pone.0281866.g003

The filtered samples are smaller but cleanest. In these samples, it is guaranteed that the CEO changes, but the top management team remains relatively stable. This method can better describe the relationship between CEOs’ functional experience and company performance. The model empirical analysis results for this sample are in column 7 in Table 3. The model results still support Hypotheses 1 and 2.

Second, Instrumental Variable Analysis. Although the fixed effect model can solve the problem of missing variables to a certain extent. There may still be two-way causal problems. Without verification, we cannot rule out the possibility of a two-way causal relationship between firm performance and the CEO’s functional experience. For example, due to the improvement of the firm’s performance, it is possible to achieve a leading position in the whole industry or even more extensive cross-industry cooperation. As a result, CEOs have more opportunities for cross-border cooperation, thus enriching their functional experience. Of course, the above inference is only a possible guess. As for the possible two-way causal problem, this paper uses the instrumental variable method to test it further. An appropriate instrumental variable needs to meet two basic requirements: it is related to an endogenous variable (func_rich), and it is independent of other variables that cannot be observed but will affect the dependent variable (firm performance, ROA). This paper uses the average CEOs’ functional richness (func_mean) in the same province and industry in the same year as an instrumental variable of the func_rich.

Table 5 reports the regression results using the instrumental variable. Column (1) of Table 5 reports the regression results of the first stage of instrumental variables. It can be seen from column (1) of Table 5 that the func_mean regression coefficient is significantly positive at 1%, meeting the correlation conditions. At the same time, Cragg Donald Wald F statistic result shows 1736.93, and Kleibergen Paap Wald rk F statistic result shows 1006.22, both of which reject the hypothesis of weak instrumental variables. Column (2) of Table 5 reports the regression results of the second stage of instrumental variables. It can be seen from column (2) of Table 5 that the func_rich regression coefficient is significantly positive at the level of 1%, which indicates that the research conclusion of this paper is still valid after using instrumental variables to solve potential endogenous problems.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Regression results of instrumental variable.

https://doi.org/10.1371/journal.pone.0281866.t005