Chair–CEO age dissimilarity and firm value – evidence from China Open Access

This study investigates whether age dissimilarity between the Chairman (Chair) and Chief Executive Officer (CEO) affects firm value. Further, we examine whether board monitoring intensity varies depending on the age dissimilarity between the Chair and CEO. Finally, we investigate whether the board monitoring intensity moderates the association between Chair and CEO age dissimilarity and firm value. The Upper Echelon Theory (UET) proposes that the personal characteristics of top managers, including their age, can significantly impact organisational outcomes. Empirical research based on UET evidence that top management team (TMT) attributes affect financial reporting quality (Habib and Hossain, 2013), strategic decision-making (Amason and Sapienza, 1997) and firm performance (Wang et al., 2016). The age of the Chair and CEO indicates their wealth of experience and cognitive abilities, which may directly or indirectly impact individual work attitudes, behaviours, and values (Li et al., 2008). Therefore, it is increasingly important to understand the impact of age dissimilarity on firms (Rabl and Triana, 2014). Given the critical role of the Chair and CEO individuals in shaping the firm’s strategy, their age dissimilarity may be a crucial element to consider.

The impact of the Chair-CEO age dissimilarity on firm value is a topic that has received limited attention in existing academic research. A harmonious relationship between the Chair and CEO increases the alignment of interests, promoting the maximisation of firm value and personal benefits (Li and Roberts, 2017). Goergen et al. (2015) assert that the Chair-CEO age dissimilarity can lead to cognitive conflict, resulting in more effective board oversight and higher firm value in Germany. At the same time, Zhou et al. (2019) reiterate that the Chair-CEO age dissimilarity can reduce excessive risk-taking between firms. Moreover, the Chair-CEO age dissimilarity could lead to conflict in the decision-making process, such as task conflict and relationship conflict (Amason, 1996). Task conflict positively impacts team performance, while relationship conflict negatively impacts team performance (Jehn, 1997; Simons and Peterson, 2000). Both types of conflict are destructive to firm performance (De Dreu and Weingart, 2003). Task conflict has a far less positive impact on enterprises than relationship conflict’s negative impact (Korsgaard et al., 2008), even if conflict management methods are used (Choi and Cho, 2011).

However, previous studies on the Chair-CEO age dissimilarity (Goergen et al., 2015; Zhou et al., 2019) are conducted based on western developed economies. The cultural background of Chinese listed firms significantly differs from that of firms in developed economies, and such differences in cultural backgrounds may lead to conflicts in cognitive styles among individuals (Chow, 2023). For individual culture, the Confucian values emphasise respect for elders and authority figures, which may lead to a more hierarchical power structure within Chinese listed firms, called “paternalistic leadership” (Farh et al., 2014). It combines strong discipline and authority with benevolence and moral integrity, which are couched in a “personalistic” atmosphere and are common in Chinese listed firms (Farh and Cheng, 2000). This may make the relationship between the Chair and CEO age dissimilarity and firm value in Chinese listed firms more complex than in Western firms. On the other hand, emerging market business cultures are relatively risk-averse because these markets used to be closed economies with less competition between firms. Therefore, firms rarely engage in risk-taking investments (Gilpin, 2000). Thus, the applicability of conclusions based on Western advanced economies has yet to be explored to a large extent in emerging markets. Additionally, the way conflicts are addressed may vary across different cultural backgrounds. Westerners tend to confront conflicts directly, while Asians tend to use avoidance to handle conflicts (Tse et al., 1994; Weldon and Jehn, 2018), especially in Chinese firms (Chen et al., 2005). Many cross-cultural management studies show these differences because many Asians are collectivists (Triandis, 1990), and their values may lead to an emphasis on harmony and conflict resolution to maintain relationships (Morris et al., 1998).

Further, we analyse how firm monitoring influences the relationship between Chair-CEO age dissimilarities and firm value. We argue that board meetings are one of the critical strategies to monitor top management. When there is high trust among the TMT, task conflict is unlikely to lead to relationship conflict between the TMT (Simons and Peterson, 2000). Based on the “similarity attraction” phenomenon, Chairs and CEOs with low age dissimilarity are more likely to share information frequently and have stronger trust, leading to more frequent board meetings. Due to cultural differences, Chinese managers are more inclined to avoid conflict than their Western counterparts (Fu et al., 2008; Morris et al., 1998). Therefore, the Chair and CEO will avoid conflict escalation by reducing social connections, which will decrease the frequency of board meetings and firm performance.

Amidst competing arguments, this study asserts that the Chair-CEO age dissimilarity can impact firm value. The Chair-CEO age dissimilarity may result in power struggles and communication difficulties that impede decision-making and strategy execution (Camelo-Ordaz et al., 2014; De Dreu and Beersma, 2005). This could stem from the divergence in perspectives and priorities that arise from their age-related experiences, which is heterogeneity. Heterogeneity can give rise to viewpoint conflicts and affect firm value. In addition, “similarity attraction” suggests that humans tend to interact and communicate with those who share similar attributes (McPherson et al., 2001); such interactions would yield more positive feedback and limit the emergence of divergent and conflicting opinions. In corporate governance, the Chair-CEO age dissimilarity can decrease effective communication and increase differences in opinions, thereby hindering decision-making efficiency and creating a lack of coordination and cooperation that delays firm operations (Lee et al., 2014).

Using a sample of 37,200 firm-year observations in the Chinese Stock Exchange for 2001–2021, this study finds a negative association between Chair-CEO age dissimilarity and firm value. Chair and CEO age dissimilarity could increase relationship conflict and reduce the frequency of board meetings and the efficiency of business operations (Lee et al., 2014). Regarding economic significance, a 1% increase in the Chair-CEO age dissimilarity reduces the firm’s value by 0.5%. Also, we find a negative association between the Chair-CEO age dissimilarity and board monitoring intensity. Results show that board monitoring intensity negatively moderates the effect of Chair and CEO age dissimilarity on firm value. Afterwards, based on a cross-sectional analysis, firms with higher free cash flow face more severe agency conflicts (Gugler and Yurtoglu, 2003), which exacerbates the negative impact of Chair and CEO age dissimilarity on firm value. In a low industry concentration business environment, firms have higher market competitiveness but lower profitability (Pahi and Yadav, 2022), which worsens the negative impact of Chair and CEO age dissimilarity on firm value. We find that Chair and CEO age dissimilarity has a more significant negative effect on firm value in firms with dispersed ownership. The negative impact of Chair and CEO age dissimilarity on firm value in SOEs is insignificant compared to non-SOEs. The results still hold after controlling the impact of management committee meetings. Regarding the endogeneity concern, our results regarding Heckman’s two-stage least squares and propensity score matching are robust.

The contribution of this paper is multi-fold. Firstly, the paper enriches the emerging research on the influence of Chair and CEO age dissimilarity on firm value and board monitoring intensity in the Chinese market. In addition, this study contributes to the very scarce literature on the effects of cooperation and conflict between the Chair and CEO. Thus far, extant research has primarily relied on the UET analysis, whereas this study incorporates “similarity attraction” theory, heterogeneity, age effect, cohort effect, social identity theory, social psychology theories, and socioemotional selectivity theory. Furthermore, this study contributes to future SOE-related research by including a comparison analysis between State-Owned Enterprises (SOE) and non-SOE in the cross-sectional analysis. An important implication of this study is that firms should consider aligning age diversity within top management to mitigate conflicts and enhance communication, thereby improving decision-making and firm value. For regulators and professional organizations, the study implies that governance frameworks and guidelines should account for the potential impact of leadership age diversity on firm performance. Regulators may recommend or encourage age alignment within top management to foster better communication and decision-making. Additionally, professional bodies could guide best practices for managing leadership dynamics, particularly in firms with significant age dissimilarity, to ensure more effective governance and oversight. These recommendations could help enhance corporate governance standards and improve overall firm performance.

This introduction is followed by a literature review and the development of hypotheses in Section 2. Section 3 discusses the methodology and sample selection process. Section 4 discusses empirical results and robustness tests. Finally, section 5 summarises and concludes this study.

The evolving TMT workforce, with an increasing presence of young members, has raised attention to age-related dynamics. Demographic factors like age, tenure, experience, education, nationality, and gender significantly influence behaviour and decision-making (Hambrick and Mason, 1984; Keck, 1997; Patzelt et al., 2008; Heenipellage et al., 2022; Kamran et al., 2023; Nielsen and Nielsen, 2013; Tonoyan and Olson-Buchanan, 2023). Age is a crucial basis for group classification, impacting team dynamics. Age dissimilarity in TMTs can manifest as Age Effects and Cohort Effects, influencing decision-making and strategic direction (Taylor, 1975). Older chairpersons, with more management experience, make accurate judgements and perform supervisory duties effectively (Hambrick et al., 2015). Age dissimilarity may impact the power dynamic between the CEO and Chair, leading to interpersonal conflicts, especially if the Chair is older than the CEO (Rahim, 2002).

Effective internal coordination and cooperation are critical for team performance, impacting firm outcomes. In the strategic decision-making process, conflicts within senior management teams are inevitable due to complexity and uncertainty (Simons and Peterson, 2000). Internal conflict is characterised by cognitive versus affective conflicts and task versus relationship conflicts (Amason, 1996; Jehn, 1997; Parayitam and Dooley, 2009). Task conflict, fostering diverse opinions, is theoretically positive for team performance (Mooney et al., 2007). However, empirical evidence varies, with some studies showing negative or insignificant correlations (De Dreu and Weingart, 2003; De Wit et al., 2012). Relationship conflict, fuelled by dissimilarity, harms team performance (Parayitam and Dooley, 2009). Age dissimilarity, particularly between Chairs and CEOs, may hinder personal connection and information sharing, leading to biased decision-making and ineffective performance (Gruenfeld et al., 1996). Both conflict types negatively impact firm performance, with relationship conflict potentially overshadowing the positive effects of task conflict (De Dreu and Weingart, 2003). Conflict management interventions addressing the root causes of relationship conflict may indirectly manage task conflict (Choi and Cho, 2011).

Nevertheless, both types of conflict can negatively impact firm performance (De Dreu and Weingart, 2003). Korsgaard et al. (2008) note that task conflict has a weak or non-existent impact beyond relationship conflict when examining both task conflict and relationship conflict in firms simultaneously. Moreover, the negative effects of relationship conflict on business performance may overshadow the positive effects of task conflict (De Dreu and Weingart, 2003). Therefore, conflict management interventions that focus on addressing the root causes of relationship conflict may indirectly lead to the management of task conflict (Choi and Cho, 2011).

Resource dependence theory demonstrates that dissimilarity can affect firm performance (Pfeffer and Salancik, 2015). Tanikawa et al. (2017) utilised the socioemotional selectivity theory ^[1] to investigate the impact of TMT age dissimilarity on a firm’s financial performance. Goergen et al. (2015) extended this investigation to explore the relationship between the Chair-CEO age dissimilarity and firm value. However, research on TMT age dissimilarity and firm value have produced conflicting theoretical arguments and mixed findings, with studies reporting a positive relationship (Kilduff et al., 2000), negative relationship (Ozer, 2010), and non-significant relationship (Nielsen and Nielsen, 2013).

Ning et al. (2022) have demonstrated a positive correlation between diverse boards and comparability of financial reports, which suggests that the age dissimilarity of the board can promote the achievement of corporate objectives and effectively enhance the quality of financial reporting (Hsieh et al., 2018; Thomsen and Conyon, 2012). Zhou et al. (2019) find that Chair and CEO age dissimilarity can positively affect bank loan portfolio risk. Similarly, Goergen et al. (2015) believe that Chair and CEO age dissimilarity can cause conflict, leading to improved decision-making quality and enterprise value. Some studies classify conflict resulting from age dissimilarity as task conflict, suggesting that age dissimilarity is a resource of information, knowledge, and professional experience (Williams and O'Reilly, 1998). From an information and decision-making standpoint, the cognitive conflict within the board caused by age dissimilarity could improve a firm’s moral development (Labelle et al., 2010), earnings quality (Hoang et al., 2017), and decision-making quality (Adams and Ferreira, 2007). Besides, lower Chair-CEO age dissimilarity is associated with weak corporate governance and lower firm value in US firms (Lee et al., 2014) and large French firms (Nguyen, 2012). Agency theory (Jensen, 1986) suggests that the Chair may not effectively monitor the CEO, as the trust and loyalty inherent in friendship can undermine the objectivity and willingness if the board is too friendly toward the CEO (Khedmati et al., 2020). This concept is referred to as a “friendly board”, which argues that a “friendly board” increases agency conflicts and information asymmetry, ultimately reducing firm value (Aebi et al., 2012). However, as noted by Milliken and Martins (1996), team heterogeneity with some aspects being beneficial and others detrimental. Some studies find a negative association between age diversity, corporate social responsibility disclosure (Khan et al., 2019), and earning management (Komal et al., 2023). Scholars also confirm the heterogeneity of executives’ views on risk, caution, and wealth, which makes knowledge sharing more difficult (Stahl et al., 2010), and is more likely to cause internal conflicts within the group during the decision-making process, resulting in a lack of attraction or even rejection which hinders the effectiveness of board meetings and weakens a firm’s profitability (Abt and Knyphausen-Aufseß, 2017).

Based on the “similarity-attraction” phenomena, McPherson et al. (2001) state the negative association between TMT members’ demographic diversity and firm performance. Individuals like to associate with similar people since it assures greater input while reducing the possibility of discrepancies and various perspectives. As a result, the Chair and CEO’s resemblance can help them to receive more favourable feedback and reduce disputes during communication (Lee et al., 2014). Directors of similar ages may have similar experiences, attitudes, opinions, and beliefs, which improves the operational efficiency and effectiveness of the organisation (age effect) (Wagner et al., 1984). In contrast, dissimilarity ages could heighten cognitive conflict and lead to relational conflict. According to social identity theory, group identification (cohort effect) is key to influencing commitment and motivation (Tajfel and Turner, 1979), while group differences could lead to social fragmentation (Guillaume et al., 2012). Diversity can impede collaboration among senior management (Schmid and Mitterreiter, 2020). Homogeneity may increase communication, foster successful working relationships, and promote team cohesiveness, which is favourably connected with team performance. TMTs’ internal consistency improves in avoiding internal setbacks and enables them to make good strategic decisions rapidly (Hambrick and Mason, 1984). Thus, TMTs with high age similarity among members will have higher collaboration efficiency and frequency, resulting in timely business decisions and increased firm value. Further, in executing corporate strategy, TMTs’ commitment to decision consistency is crucial for corporate performance (Pitcher and Smith, 2001).

Social categorisation (Knapp et al., 2011) and the similarity-attraction paradigm (Byrne, 1997) suggest that the Chair-CEO age dissimilarity negatively affects firm value. The categorisation of subgroups based on similarities and differences in work groups results in forming in-groups and out-groups, and people are more willing to cooperate with the former. Cooperation frequency and efficiency are stronger among TMT members with high age similarity than those with low similarity based on age effects, cohort effects, and UET. Furthermore, the greater the Chair-CEO age dissimilarity, the slower the decision-making process, leading to untimely business decisions and reduced firm value (Van Knippenberg and Schippers, 2007). Therefore, the Chair-CEO age dissimilarity may result in reduced communication, lower frequency of information sharing, and more communication errors, leading to poor business decision-making and a decrease in firm value (Williams and O'Reilly, 1998). Based on these studies, we developed the following hypothesis:

Board monitoring intensity is crucial to corporate governance, as it ensures that the board of directors can effectively oversee top management’s actions and protect shareholders’ interests. Age dissimilarity is one factor that has been shown to impact board functioning and monitoring intensity, particularly in the context of Chair and CEO relationships. Age dissimilarity can bring various perspectives and experiences to the boardroom, improving decision-making and reducing groupthink. On the other hand, age dissimilarity can also lead to communication difficulties and conflicts due to generational differences and varying experience levels. Based on the “similarity-attraction” phenomena, interactions and frequency of interpersonal relationships are higher among people with similarities (McPherson et al., 2001). Age dissimilarity between the Chair and CEO could negatively affect board monitoring intensity, leading to a power struggle and undermining the board’s effectiveness, specifically when the Chair and CEO have a significant age dissimilarity and career horizon. Also, the Chair may be less (or more) willing or able to challenge the CEO and hold the other accountable for their actions. This can lead to a lack of oversight and a lower monitoring intensity. Therefore, the following hypothesis is proposed based on the above discussion:

Board monitoring intensity significantly impacts firm operations and performance (Jaffar and Abdul-Shukor, 2016). Empirical Research indicates that increased monitoring is associated with improved performance, helping identify and rectify issues promptly and deterring misconduct (Vafeas, 1999). However, excessive monitoring may hinder innovation and risk-taking, which are crucial for long-term growth (Adams and Mehran, 2005). A highly effective board can navigate challenges of age diversity, leveraging its benefits like diverse opinions and in-depth discussions. Effective boards ensure age diversity contributes to decision-making and strategy creation, enhancing firm performance and value. In contrast, less effective boards overseeing senior management may amplify age diversity drawbacks, leading to conflicts and poor decision-making, negatively affecting firm value. Enhancing board monitoring efficiency could impact the relationship between age diversity and firm value. Increased board meeting frequency may improve communication and collaboration, particularly for chairs and CEOs with significant age dissimilarity, leading to better decision-making and improved firm performance (Ntim and Osei, 2011). Yet, it could also escalate conflicts, especially if age dissimilarity is large, potentially harming firm performance and value (Adams and Ferreira, 2007). Therefore, we develop the following hypothesis:

We employ two widely used firm value measures, namely Tobin’s Q (TOBINQ) and Market-to-Book ratio (MKTBK) (Hilliard and Zhang, 2015; Sucuahi and Cambarihan, 2016). TOBINQ is defined as the market capitalisation plus the difference between the book value of total assets and the book value of equity, all divided by the book value of total assets. If Tobin’s Q is greater than 1, it suggests that the firm is perceived to have higher growth potential and firm value because the asset’s market value is more than its replacement cost and vice versa. In comparison, the market-to-book ratio is another widely used measure of firm value. A higher MKTBK suggests greater perceived firm value and growth potential, as the market value of the firm’s equity is more than the book value suggests. This study measures MKTBK, the share price divided by the net book value per share.

Empirical research does not show any consensus on the measure of age dissimilarity. Therefore, we consider several measures for Chair and CEO age dissimilarity consistent with the existing literature (Goergen et al., 2015; Zhou et al., 2019; Zhu et al., 2021). This study uses two variables to measure the dissimilarity between the age of the Chair and the CEO. GAPS is the age dissimilarity between the Chair and CEO, calculated as the Chair’s age minus the CEO’s age. The purpose is to test whether the Chair is older or younger than the CEO. GAPU is the absolute value of GAPS. This variable captures the age difference between the Chair and CEO, regardless of who is older. In addition, GAPSQU is the squared GAPS used to test whether the regression model has a non-linear relationship (Younsi and Bechtini, 2020).

We use two proxies to measure board monitoring intensity. Studies have shown that the number of board meetings indicates board monitoring intensity (Brick and Chidambaran, 2010). Therefore, we follow prior studies (Goergen et al., 2015) and use board meetings (BMEET) to measure board monitoring intensity, defined as the sum of board meetings and the number of conferences held by the board during the fiscal year (Adams and Mehran, 2005). Following Brick and Chidambaran (2010), we use another proxy (LNBMEET), the natural logarithm of the number of board meetings per year, to measure board monitoring intensity.

Following Brick and Chidambaran (2010), we control for several Chair characteristics, CEO characteristics, supervisory board characteristics, and firm characteristics that may affect FV and BMI.

Chair and CEO characteristics include the Chair and CEO’s different gender (DIFGEN), different educational backgrounds (DIFEDU), cultural differences (DIFNAT), and length of time working together (JONTENU). Following Andres et al. (2014), we measure their corresponding power through the replacement (CHANGE1/CHANGE2) and the tenure (TENURE1/TENURE2) of the Chair and CEO (Kaplan et al., 2012). All of these are used to measure the possibility of tacit understanding or conflict between the Chair and CEO. Supervisory board characteristics include Board size (BODSIZE), which refers to the number of directors on the board. Board age (BODAGE) refers to the average age of board members. CV Board Age (CVBA) controls for age diversity on the board committee and supervisory board. These variables have also impacted firm value (Brick and Chidambaran, 2010; Yermack, 1996).

Firm characteristics include firm size (FSIZE), leverage (LEVER), beta coefficient (BETA), stock volatility (STOCKVOL) ^[2], sales growth (SALEGR), R&D expenses (R&D), the ratio of capital expenditure (CapEx/TA), firm listing age (FIRMAGE), and the family firm (FFIRM) (Audretsch et al., 2013; Bouslah et al., 2018; Burney et al., 2021; Custódio and Metzger, 2014; Dittmann et al., 2010; Le et al., 2006; Peng and Jiang, 2010; Popov and Ongena, 2011). Following Goergen et al. (2015), this study also includes a set of management committee characteristics that may affect the dependent variable as an additional control for models. As an essential part of the TMT, the board of supervisors plays a supervisory role on the board of directors and the CEO. Management committee characteristics include the management board committee size (MBSIZE), the average tenure of management board committee members (MBTENU), the age of the management board committee members (CVMBA), and the supervisory board size (SBSIZE) (Vafeas, 1999). We refer to  Appendix for further details on the variable definitions and the sources.

In H1, we examine the association of Chair and CEO age dissimilarity and firm value. The following Ordinary Least Square (OLS) regression tests the H1.

To test H1, the primary variable of interest is $β1GAPi,t$⁠. GAP_i,t is the proxy for the age dissimilarity between the Chair and CEO. As shown in Section 3.2, we consider two different measures of age dissimilarity: GAPS and GAPU. Firm value is the dependent variable in H1. This study uses two primary measures of firm value (FV), such as Tobin’s Q (TOBINQ) and Market-to-Book ratio (MKTBK). In equation (1), a positive (negative) coefficient of $β1GAPi,t$ will indicate that age dissimilarity increases (decreases) firm value.

In H2, we posit that the age dissimilarity between the Chair and CEO will likely affect the firm’s monitoring needs. Therefore, to test H2, we develop the following OLS regression model is developed:

In H2, the primary variable of interest is $γ1GAPi,t$⁠. A positive association on $γ1GAPi,t$ will indicate that age dissimilarity increases board monitoring intensity and vice versa. As shown in Section 3.3, we use BMEET and LNBMEET to measure the intensity of the board monitoring. We conduct both OLS and Poisson regression equations to test H2. Consistent with Goergen et al. (2015), the Poisson regression equation is an appropriate regression model for a variable measured on count data, such as BMEET (Hayat and Higgins, 2014).

In H3, we hypothesise that board monitoring intensity moderates the association between Chair and CEO age dissimilarity and firm value. We develop the following OLS regression model to examine the H3:

The primary variable of interest is $α3GAPi,t*BMIi,t$⁠. When α₁ is a positive coefficient, a positive association on α₃ would suggest that an increase in the board monitoring intensity could promote the positive impact of the Chair and CEO’s age dissimilarity on firm value (positive moderator). In contrast, a negative association on α₃ could indicate that an increase in the board monitoring intensity would negatively moderate the positive effect of the Chair and CEO’s age dissimilarity on firm value (negative moderator). Conversely, when α₁ is a positive coefficient, a positive association of α₃ could suggest that an increase in the intensity of board monitoring would reduce the negative impact of the Chair and CEO age dissimilarity on firm value (negative moderator; weakening the main effect), while a negative association on α₃ would suggest that an increase in the intensity of board monitoring would enhance the negative impact of the Chair and CEO age dissimilarity on firm value (negative moderator). In addition, this positive association on α₂BMI_i,t would suggest that firm value could increase with higher board monitoring intensity (Wang et al., 2018).

This study extracts the relevant financial and firm governance data of 5,049 Chinese listed firms from 1998 to 2021 from the China Stock Market & Accounting Research Database (CSMAR) (Du and Boateng, 2015), we drop the missing firm-year observations. Table 1 illustrates the sample selection process. After calculating the data for some firm-specific financial variables, we exclude 3,134 observations from 1998 to 2000, which are not part of the study period. In addition, we exclude missing values and outliers in the remaining sample data. Following Fang et al. (2021), we excluded financial institutes (SIC codes between J66 to J69), because of their unique regulatory environments. The final sample of this study is 37,200 firm-year observations from 3,870 unique firms. All the continuous variables are winsorised, top and bottom, at a 1% level to rule out any potential biases from outliers. Table 2 reports a distribution of samples based on industry classification. We follow the SIC code of CSMAR to classify the firms consistent with the prior study (Xie et al., 2021). The sample is well distributed. Sample observations of this study include a total of 17 industries. As an economic pillar industry in the Chinese market, industrial products account for the majority of the sample in this study.

Table 3 reports the descriptive statistics of all the variables used in this study. The mean (median) of TOBINQ, MKTBK, NAVPS, and MKTVAL are 2.45(1.87), 3.12(3.11), 4.40(3.75) and 22.65(22.50). Sample firms hold an average of 11.84 board meetings (BMEET) annually. Most CEOs are younger than Chairs. The GAPS (GAPU) mean is 3.59 (5.94) years, ranging from 0 to 28. A total of 6% of the sample firm-year has a 20-year age gap between Chair and CEO (GAP20), and the mean of GAP15 (GAP10) is 12% (24%) of the sample firm. The mean age of the Chair and CEO is 52.15 (48.56) years. Regarding control variables, 84% of the Chairs and CEOs have different educational backgrounds. A total of 8% of the sample firm-year observations have gender diversity. Finally, 43% of the sample firms are SOE firms, and 20% have a majority shareholder with over 50% ownership.

Table 4 reports the results of the difference in mean for all the variables for the lower age dissimilarity (GAPU ≤ median) and higher age dissimilarity (GAPU > median) categories. Firm value is lower, and the board meets less frequently when the Chair and CEO’s age dissimilarity is higher. The findings are consistent across all the measures of firm value. Also, differences in Chair and CEO education (DIFEDU), nationality (DIFNAT), and gender diversity (DIFGEN) are greater within the group of firms with higher Chair and CEO age dissimilarity (GAPU > median). The reported variables are statistically significant at the 10% level. This result tentatively supports our hypothesis.

Table 5 presents the Pearson correlations between the dependent and independent variables included in the baseline regression. The test variables GAPU and TOBINQ, MKTBK, NAVPS, and MKTVAL are all significantly and negatively correlated (correlation coefficients −0.04, −0.03, −0.04, −0.03, p < 0.01, respectively). In addition, the test variables GAPU and BMEET, and LNBMEET are also significantly and negatively correlated (correlation coefficients of −0.04, −0.02, p < 0.01). This provides univariate support for the negative correlation between the Chair and CEO age dissimilarity and firm value and monitoring intensity, respectively. Similarly, GAPS is also significantly negatively correlated with all dependent variables. Firms with a longer co-work time between Chair and CEO (JONTENU), firms with high R&D, and firms with high capital expenditures (CapEx/TA) have better performance and weaker board monitoring intensity (Full table in  Appendix). Firms with high sales growth rates (SALEGR) correlate positively with firm value and monitoring intensity.

This section uses the empirical model presented in section 3.1 to test the hypotheses, using sample data for OLS linear regression and Poisson regression analysis.

Table 6 reports the results of the OLS regression for H1. In the H1, we examine the association between Chair and CEO age dissimilarity and firm value. Column (1) only includes the control variables. In columns (2) and (3), we examine the impact of age dissimilarity (GAPS, GAPU) on the firm value (TOBINQ). Columns (5) and (6) report the impact of age dissimilarity (GAPS, GAPU) on the alternative measure of firm value (MKTBK). There is a sensitivity test in columns (4) and (7), both GAPU and GAPSQU are taken in the regression (Haans et al., 2016). We examine whether the association between Chair and CEO age dissimilarity and firm value has a non-linear relationship in columns (4) and (7). Finally, columns (8) and (9) illustrate whether the results are still available in a complex working environment.

Consistent with the expectation, the association between GAPS and GAPU on TOBINQ shows a negative coefficient (−0.003, −0.005; t-value = −3.854, −4.202; p < 0.01), which indicates that firm value reduces due to the Chair and CEO age dissimilarity. The economic significance of the findings suggests that a 1% increase in the Chair and CEO age dissimilarity causes a corresponding 0.5% reduction in the firm’s value. Also, we examine the impact of Chair and CEO age dissimilarity on the MKTBK and find that the association of both GAPS and GAPU are negative (coefficient = −0.0000337 & −0.0000616, t-value = −3.517 & −4.341) and the findings are statistically significant at a 1% level. There is a non-linear effect of Chair-CEO age dissimilarity on firm value. Column (4) shows the coefficient for GAPU is significantly negative, while the coefficient for GAPSQU is not statistically significant (−0.007, t-value = −2.480, p < 0.05; 0.0001001, t-value = 0.805, p > 0.10). Column (7) also shows similar results. The original inference remains unchanged when models include the four proxies (CVMBA, MBSIZE, MBTENU, SBSIZE) that reflect management board committee characteristics as additional control variables in columns (8) and (9). In other words, the significant negative associated relationship between Chair and CEO age dissimilarity and firm value are not impacted by the involvement of a firm’s management board committee.

Regarding the control variables, we find that a firm with a Chair and CEO from different nationalities (DIFNAT) and a firm with higher average age of board members (BODAGE) positively affects firm value. Firms with higher capital investment (CapEx/TA), higher R&D investment (R&D), and higher sales growth (SALEGR) increase firm value. Also, we find that family firms (FFIRM) have higher firm value (Bebchuk et al., 2009; Custódio and Metzger, 2014; Dittmann et al., 2010). Overall, we find that a negative coefficient between Chair and CEO age dissimilarity and firm value suggests an age dissimilarity between the Chair and CEO has a negative consequence on firm value. Thus supports H1.

Table 7 reports the results of H2. Column (1) shows the control variables on the board monitoring intensity. In columns (2), (3), (4), and (5), we examine the impact of age dissimilarity (GAPS, GAPU) on the board monitoring intensity (BMEET and LNBMEET). Finally, columns (6) and (7) examine the influence of Chair and CEO age dissimilarity on board monitoring intensity, including several additional attributes of management board committee members (CVMBA, MBSIZE, MBTENU, SBSIZE).

In contrast with prior research (Goergen et al., 2015), the results indicate that the association between GAPS and GAPU on BMEET presents a significantly negative association (coefficient = −0.001, −0.001; t-value = −4.301, −4.412; p < 0.01). Similarly, we examine the impact of Chair and CEO age dissimilarity on the LNBMEET and find the association of both GAPS and GAPU a negative coefficient (−0.001, −0.001; t-value = −5.406 & −4.837) and the findings are statistically significant at a 1% level. Specifically, columns (1) and (2) show that for every 1% increase in the age dissimilarity between the Chair and CEO, the strength of internal monitoring decreases at 0.1% accordingly. After adding additional characteristics of management board committee members in columns (6) and (7), Chair and CEO age dissimilarity still has a significant negative impact on board monitoring intensity (BMEET and LNBMEET). The significant negative relationship associated with the Chair and CEO’s age dissimilarity and the board monitoring intensity is not affected regardless of whether the firm’s management board committee members are involved. Thus, H2 is supported.

Table 8 shows the results for H3. We test whether the effect of the association between Chair and CEO age dissimilarity and firm value is related to board monitoring intensity. Therefore, we use TOBINQ and MKTBK as proxies for FV; we infirm BMEET_median or LNBMEET_median as proxies for BMI; we use GAPS as proxies for GAP. BMEET_median (LNBMEET_median) refers to a dummy variable of value 1 assigned if the value of BMEET (LNBMEET) is higher than the median and 0 otherwise. This variable’s 1(0) values mean that BMI is relatively stronger (weaker). The primary interest variable is the interaction term between GAP and BMI (i.e. GAPS*BMEET_median, GAPS*LNBMEET_median). Columns (1) and (2) of Table 8 show the results of a moderation test using TOBINQ to measure firm value. Consistent with the results in 4.4.1, the values of coefficient α1 in columns (1) and (2) are both −0.004, which is significant at the 1% level (t-value = −3.881 & −4.228). Also, the association of both BMEET_median and LNBMEET_median has a positive coefficient (0.072, 0.079; t-value = 3.877 & 4.604), and the findings are statistically significant at a 1% level. Importantly, in these two columns, the coefficients for the variables (i.e. GAPS*BMEET_median and GAPS*LNBMEET_median) are positive in columns (1) and (2), respectively and are statistically significant at the 10% and 5% levels. Similarly, we performed substitution tests on columns (1), and (2), respectively, and replaced the variable TOBINQ, which measures the value of the firm, with MKTBK (columns (3), (4)), keeping the moderator and independent variables unchanged. Similar to columns (1), and (2), GAPS*BMEET_median and GAPS*LNBMEET_median still maintain a significant positive correlation with MKTBK. As a result, the moderating effect of board monitoring intensity as a moderating variable again proves to be significant. Therefore, Chair and CEO age dissimilarity has a greater effect on firm value in firms with lower internal monitoring intensity, and firm performance is reduced due to the generation problem and cognitive washout between Chair and CEO. Thus, the results support H3.

We use Net Asset Value Per Share (NAVPS) and Market Value (MKTVAL) as alternative measures for TOBINQ and MKTBK to check the robustness of the findings (Andrikopoulos et al., 2022; Huang and Lu, 2022). The higher the NAVPS of a firm, the more assets the shareholders own (Li and Xia, 2013). MKTVAL can directly reflect the level of firm value (Black and Kim, 2012). We re-estimated equation (1). Table 9 shows the measurement results using alternative measures of firm value. Columns (1) and (2) examine the associations of NAVPS, MKTVAL, and GAPS, respectively. Columns (3) and (4) examine the associations of NAVPS, MKTVAL, and GAPU, respectively. Column (5) shows a non-linear analysis between Chair and CEO age dissimilarity and firm value involving MKTBK, GAPU, and GAPSQU. Overall, all findings support H1.

Following prior research, we performed additional tests using the inter-generational age gap. In line with the sociology literature, Strauss and Howe (1997), Goergen et al. (2015), and Zhou et al. (2019) used Gap20 as the primary research variable to study firm performance in European business settings. Among listed firms in the Chinese market, SOEs have strict restrictions on the length of employment and promotion of the Chair and CEO. It is rare for the Chair and CEO’s age dissimilarity to exceed 20 years old. Zhu et al. (2021) used Gap10 as the primary research variable to study the Chinese market. We combined these two classification methods and broadened them. We shortened the inter-generational age dissimilarity to 5 years as the dividing line and used the four different age dissimilarity dummy variables of Chair and CEO Gap20/15/10/5 (GAP20/GAP15/GAP10/GAP5) as the alternative measures of Chair and CEO age dissimilarity. These four dummy variables are used to more comprehensively and accurately test the influence trends of different age dissimilarity between the Chair and CEO on firm value and board monitoring intensity. We re-estimated equation (1) and equation (2).

Table 10 Panel A shows the relationship between the Chair and CEO inter-generational age dissimilarity (GAP20/15/10/5) with the primary firm value proxies TOBINQ (columns (1) to (4)) and MKTBK (columns (5) to (8)). The GAP20, GAP15, GAP10, and GAP5 coefficients consistently maintained significant negative associations in all the above models (p-values level at the 1% level). As with the previous findings, all findings support H1.

After that, Panel B shows the association between the Chair and CEO inter-generational age gap (GAP20/15/10/5) with the primary board monitoring intensity proxies BMEET (columns (1) to (4)) and LNBMEET (columns (5) to (8)). As expected, whether it is the Poisson regression of columns (1) to (4) or the OLS regression of columns (5) to (8), the results are consistent with those in Table 7. The impact of the Chair and CEO age dissimilarity on the BMEET and LNBMEET is that the association of GAP20/15/10/5 is negative, and the findings are statistically significant at a 1% or 5% level. As with the previous findings, all findings support H2. Of note, the results of all regressions after adding management board committee characteristics (CVMBA, MBSIZE, MBTENU, SBSIZE) as control variables are similar to the existing regressions’ results, and the original negative significance results remain negative significant (untabulated). This suggests that the research model is broadly reasonable.

To further validate the conclusions drawn in 4.4.1, this study conducts a cross-sectional analysis of the firm’s operating status and financial characteristics (Goergen et al., 2015). We attempt to distinguish between the firm’s operating status and the financial characteristics through a comparative analysis of four sub-samples (including agency conflicts, market competitiveness, ownership concentration, and state ownership attribute). To obtain more concise and accurate results, Table 11 lists the regression coefficients of the two main independent variables, GAPS and GAPU, involved in this study. The main indicator of firm value used in the study is TOBINQ.

Whether agency conflicts could influence the association between age dissimilarity and firm value is measured by whether the firms have below-average-level free cash flow (captured by the variable FCF/TA). Firms with high free cash flow and stronger relationships between dividends and governance will also face high levels of external monitoring (threat of takeovers). This will strengthen the firm’s internal governance intensity. High-intensity internal governance (board of directors, major institutional shareholders) will also increase agency conflicts among managers within the enterprise. According to agency costs and the free cash flow (FCF) hypothesis, managers can reduce free cash flow by shifting dividends to their personal interests or funding less profitable projects (Jensen, 1986). According to the rent extraction hypothesis (Gugler and Yurtoglu, 2003), the firm manager may use some means (i.e. dividend payment) to reduce the FCF controlled by major shareholders and share FCF with all shareholders in equal proportion. This behaviour could reduce the deprivation of rights of large shareholders to small shareholders (Nur and Karnen, 2014).

In Table 11, columns (1) to (4), following Wang (2010), groups free cash flows according to values greater than the median (FCF > Median; 12,140 observations) and less than or equal to the median (FCF ≤ Median; 25,060 observations). Firms with higher free cash flows have more significant agency conflict. The impact of the Chair-CEO age dissimilarity on firm value is significantly stronger in the FCF > Median group than in the FCF ≤ Median group, and all two Chair-CEO age dissimilarity indicators corresponding to the FCF > Median group are negative and significantly affected firm value. Conversely, the Chair-CEO age dissimilarity in the FCF ≤ Median group does not significantly impact firm value. Therefore, the regression findings suggest that the negative effect of the Chair-CEO age dissimilarity on firm value is only observed for those firms with higher significant agency conflict.

Whether a firm’s market competitiveness could affect the association between the Chair and CEO’s age dissimilarity and firm value is measured by whether the firm belongs to an industry monopoly in the financial year (measured by the Herfindahl-Hirschman index, which is higher than the industry average in the same year). Allen and Gale (2000) argue that competition between firms may be the most effective form of firm governance compared to board monitoring mechanisms (board, institutional, and bank oversight) and external mechanisms (market controls). Pahi and Yadav (2022) show a positive association between profitability and industry concentration (HHI). Therefore, based on Harjoto et al. (2015), this study divides HHI into two groups: HHI > Mean (monopolistic or non-competitive enterprises) and HHI ≤ Mean (low concentration or competitive enterprises) in the product market competition.

Table 11, columns (5) and (6), show that there is no significant negative association between Chair-CEO age dissimilarity and firm value in monopolistic or non-competitive firms (HHI > Mean; 19,548 observations). However, columns (7) and (8) demonstrate that the negative impact of the Chair-CEO age dissimilarity on firm value is highly significant in low-concentration or competitive firms (HHI ≤ Mean; 17,652 observations). This suggests that the significant negative impact of the Chair-CEO age dissimilarity on firm value is limited to low-concentration or competitive firms.

Following Goergen et al. (2015), whether ownership concentration could impact the association between the Chair and CEO age dissimilarity and firm value is measured by whether a shareholder holds more than 50% of the common shares outstanding in the firm. According to institutional theory, the ownership structure of a firm often affects the market behaviour of the firm (Buckley et al., 2018; Hoskisson et al., 2000). The level of decentralisation of firm control can affect firm value and board monitoring intensity (Goergen et al., 2015). A dispersed shareholding structure can help address issues related to controlling shareholders in a firm, but a dispersed ownership structure may invite hostile takeovers. On the other hand, compared to firms with concentrated ownership, firms with dispersed ownership are more prone to internal power struggles (He et al., 2023). Therefore, Lambert et al. (2007) show a negative correlation between a firm’s equity structure concentration and its performance. On the contrary, from a managerial perspective, the higher the concentration of firm equity, the better the firm’s performance (Botosan and Stanford, 2005). Therefore, an increase in the equity of major shareholders will positively impact the firm’s performance.

In Table 11, Columns (9) to (12), the sample is divided into firms without major shareholders (Blockholder 50% = 0; 29,662 observations) and firms with major shareholders (Blockholder 50% = 1; 7,538 observations). When there are major shareholders in the firm, the Chair-CEO age dissimilarity does not have a significant effect on the firm value (columns (9) and (10)). On the contrary, when there is no major shareholder in the firm, the Chair-CEO age dissimilarity will have a significant impact on the firm value (columns (11) and (12)). This shows that compared with firms with concentrated ownership, the Chair-CEO age dissimilarity will more significantly negatively affect the firm value in the firms with dispersed ownership.

Whether the state ownership attribute of the firm affects the association between the Chair and CEO age dissimilarity and firm value is measured by whether the firm is an SOE. Based on previous studies, this study classifies the ownership type of enterprises into two categories: SOE (state-run or state-owned holding enterprise) and non-SOE (the other enterprises) (Chen et al., 2011).

The literature examining whether state ownership positively or negatively impacts firm value is mixed (Sun et al., 2002; Wei et al., 2005). Existing literature suggests that state ownership harms firm value because government intervention slows down the firm’s decision-making process in an increasingly competitive environment. Conversely, it has been argued that SOEs are more likely to receive preferential treatment from the government, thereby increasing their value (Blanchard and Shleifer, 2001). SOEs benefit from government support through VAT exemptions and preferential financing (Xiao and Sun, 2005). After 1979, China gradually adopted market-based economic reforms, including corporatising former SOEs. As the majority shareholder of former SOEs, the Chinese government maintained two key controls over the firms: asset disposal and mergers and acquisitions, and appointed TMTs (e.g. CEOs) of these SOEs. Non-SOEs’ financing decisions are mainly based on the firm’s financial situation rather than on political considerations. Chen et al. (2011) claim that non-SOEs’ access to financial decisions depends more on the quality of financial statement information. SOEs are managed by the State-owned Assets Supervision and Administration Commission (SASAC), which acts as the board of directors, evaluates management performance, and approves investment projects. Meanwhile, performance evaluation, salary increases, and career development in SOEs often depend on political connections and the level of the SOE (Du and Boateng, 2015; Haans et al., 2016). Based on this, some of the problems arising from the generational gap between the Chair and CEO would be avoided more or less; for example, there would be relatively few disagreements between the two on the direction of the firm’s financial decisions.

In Table 11, columns (13) to (16), the sample size of the SOEs makes up 43% (16,944), and that of non-SOEs makes up 57% (21,909). When firms are non-SOE, there is a negative associated relationship between all independent variables (GAPS/GAPU) and firm value (TOBINQ), and the coefficients are all negative and significant at the 1% level. However, when the firm is SOE, the empirical findings are the opposite (all coefficients are insignificant with a positive associate relationship). In SOE, the Chair-CEO age dissimilarity does not significantly affect firm value.

Thus far, this study has not determined the channel through which the Chair-CEO age dissimilarity reduces the firm value. According to the conclusion of section 4.5, this study evidences that board monitoring intensity can be tried as an intermediary channel. LNBMEET was chosen because previous research proved that excessive monitoring could reduce firm value (Vafeas, 1999). The reason given by Adams and Ferreira (2007) is that high-frequency board meetings increase the CEO’s work pressure. However, Eluyela et al. (2018) confirmed that the frequency of board meetings is positively associated with firm performance, and at least four board meetings per year can ensure the stable development of firm performance. To test the mediation effect (H3) of board monitoring intensity on the association between Chair-CEO age dissimilarity and firm value, we followed the causal step mediation test method of Baron and Kenny (1986) and developed the following equations:

This study uses the following three-step step-by-step test of mediation channels (equations (4.1) to (4.3)). First, it was confirmed that there is a significant associated relationship between the independent variables (GAPS, GAPU) and the dependent variables (TOBINQ, MKTBK) to confirm the possibility of a mediating effect (equation (4.1)). The second step shows how changes in the independent variables (GAPS, GAPU) explain changes in the mediator (LNBMEET) (equation (4.2)). Finally, it is verified that the mediator variable significantly affects the dependent variable in equation (4.3).

Table 12 reports the results of the mediation test. Table 12 Panel A reports the intermediary test results using TOBINQ to measure firm value. Column (1) shows that the coefficient for GAPS is negative and significant (total mediation effect β1 is −0.003, t = −3.687, p < 0.01). Column (2) records a significant negative associated relationship between the Chair-CEO age dissimilarity (GAPS) and board monitoring intensity (LNBMEET), indicating that the Chair-CEO age dissimilarity leads to a decrease in board monitoring intensity (γ1 is −0.001, t = −5.406, p < 0.01). Column (3) shows that there is a significant positive associated relationship between LNBMEET and TOBINQ (χ² is 0.104, t = 4.257, p < 0.01), which means that an increase in board monitoring intensity will promote better firm performance. The value of γ1*χ² is −0.000104, not equal to 0, and has the same sign as χ1 (−0.003). Therefore, the mediation effect is established, and M_i,t is a partial mediation effect, accounting for 3.456% (γ1*χ²/β1). Correspondingly, when GAPU replaces the proxy variable of GAP_i,t, the results are consistent with GAPS. M_i,t is still a partial mediation effect, accounting for 2.08%. We also performed a similar analysis using MKTBK as a proxy for the FV, and our findings remain consistent (untabulated for brevity).

To further check whether the indirect effect is statistically significant, we choose bootstrap as an alternative measure to strengthen the results of Barron and Kenny’s three-stage mediation test (Preacher and Hayes, 2004). Table 12 Panel C reports the test results of the bootstrap approach. The mediation effect test uses two dependent variables (TOBINQ, MKTBK), two independent variables (GAPS, GAPU) and one mediator variable (LNBMEET) for the mediation effect test. The bootstrap process for all groups is resampled with 5,000 repetitions. The results show that the confidence intervals of all direct effects and indirect effects do not contain 0. Therefore, tests for all groups revealed a partial mediation effect between the mediator variable and the dependent and independent variables. All mediation effect test results support H3. Therefore, the reduction of firm value by the Chair-CEO age dissimilarity is regulated through the board monitoring intensity as a part of the intermediary channel.

The sample selection may have unobservable factors that could affect the regression models used to examine the hypothesis. Since the appointment of the Chair and CEO in the firm is not randomly determined, some unobservable factors that affect the Chair-CEO age dissimilarity may cause a concern of endogeneity in the current findings. To effectively control the endogenous concerns caused by sample selection bias, we perform the Heckman (1979) test as follows:

First, following Zhu et al. (2021), we designed CRG (Cultural Revolution Generation, dummy variable set to 1 if the Chair or the CEO is at least 16 years old during the Cultural Revolution (1966–1976), and 0 otherwise.) as instrumental variables for Heckman’s two-stage test. Older generations who have fully experienced this turbulent historical period (people with a sense of personal independence since 1966) will have a certain cognitive conflict with the new generation of business managers (different values and thinking patterns), which will be reflected in firm governance. This variable is related to age but not firm value and board monitoring intensity. It effectively reflects the specific characteristics of the historical background experienced by the Chair and CEO and can prevent multicollinearity between explanatory variables and inverse Mills ratio. In Table 3, the mean values of ChairNCN and CEONCN are both 0.98. Therefore, the nationality of the Chair and CEO in 98% of observations are both Chinese. This instrumental variable is available for this study sample.

Secondly, we introduced GAPU_median or GAPS_median as the dependent variable in the first stage of the model. The Inverse Mills Ratio (IMR) is calculated in the first-stage probability model and brought into the second-stage regression model as an additional independent variable to judge whether there is a sample selection problem in the original conclusion. Table 13 Panel A, column (1) shows the regression result of the first stage. We did a Logit regression with GAPU_median, CRG, and other control variables. The association coefficient between GAPU_median and CRG is positive (coefficient 1.876, t-value 57.725) and significant at the 1% level. Columns (2) to (5) show the regression results of the second-stage. The coefficient of IMR on TOBINQ (column (2)) is positive and not significant, while the coefficients of IMR on MKTBK (column (3)), BMEET (column (4)), and LNBMEET (column (5)) are significantly positive and different from 0. Therefore, there may be a sample selection problem in this study. Sample selection bias could cause the equation to overestimate the effect of Chair and CEO age dissimilarity on firm value. However, the results of columns (2) and (3) indicate that the association of GAPU on firm value presents a negative association, all significant at the 1% level. Meanwhile, the results of columns (4) and (5) report that the association between GAPU and board monitoring intensity shows a negative coefficient, and the findings are statistically significant at a 1% and 5% level. Therefore, after correcting the sample selection bias, the regression results are consistent with the previous conclusions. In addition, in Table 13 Panel B, we replaced GAPU_median and GAPU with GAPS_median and GAPS, respectively. The results obtained are similar to those of Panel A. Finally, we also find that all results remain consistent in column (1) if we use probit regression model. In summary, after correcting for sampling bias, all research results still support the hypothesis.

False specification of the functional form (FFM) used in OLS will likely lead to a biased estimation of research conclusions. Propensity Score Matching Analysis (PSM) can alleviate the existing FFM problem by reducing the dependence on the functional form setting. This study uses the PSM model to test self-selection bias (Rosenbaum and Rubin, 1985). Firstly, the processing variables used need to be grouping variables. Therefore, this section introduces a new dummy variable, GAPD, to replace GAPS and GAPU. GAPD indicates whether there is an age dissimilarity between the Chair and CEO (Zhu et al., 2021). Each observation in the two groups (GAPD = 1, treatment group; GAPD = 0, control group) is matched using the nearest neighbour matching rule with TOBINQ or LNBMEET as the outcome variable. The PSM results are reported in Table 14.

Table 14 demonstrates the matched sample’s proper balance between treatment and control groups. First, the bias of most variables is less than 10% after matching. Second, only some of the covariates were statistically significant. Between GAPD and all dependent variables, are a significant negative association. The results prove that when there is an age gap between the firm’s Chair and CEO, the firm’s value and board monitoring intensity will decrease accordingly. Therefore, the propensity score matching analysis confirms our previous inference.

This study examines the association between the Chair-CEO age dissimilarity and firm value in Chinese listed firms and the moderating effect of board monitoring intensity. Overall, the result is consistent with previous research that shows a wider age dissimilarity is likely to cause cognitive conflict. Age dissimilarity leads to an age and cohort effect between the Chair and CEO, increasing the conflict between different age groups. The conflict effect occurs between the Chair and CEO, which exacerbates the principal-agent problem. This, in turn, negatively impacts the firm’s internal monitoring (board meetings) and strategic decision-making design and implementation. Since Chinese firms are still building a modern enterprise system and have not yet formed a sound coordination mechanism and formal communication system, the potential impact of TMT heterogeneity on the firm value within Chinese listed firms is inconclusive. In addition, because the Chinese market is in a transitional economy, and the social structure and constitutional environment are changing rapidly, managers with huge age dissimilarities are prone to conflicts in values and concepts. Therefore, the coherence of the Chair and CEO being held by members of similar ages is higher than that of members of different ages, the cooperation between members is better, and effective strategies can be formulated more quickly. All of this can help the team better handle emergencies and improve the firm’s operations and decision-making efficiency. Under the impact of economic transformation and globalisation, the external environment of Chinese firms has become increasingly complex, and the competition has become more intense. Therefore, the ability of the Chair and CEO to make rapid decisions together helps improve firm performance.

In addition, this study combines age effects and cohort effects, effectively avoiding the research bias that cohort effects only exist in cross-sectional data and mitigating the threat of research data to internal validity. In addition, the longitudinal research on age effects is group-oriented, ensuring that group differences will not threaten its internal validity. The findings of this research should be interpreted with caution as the association between Chair-CEO age dissimilarity and firm value could have been impacted by another macroeconomic scenario, such as the global financial crisis. Future research could extend this line of research by examining the impact of such a macro-economic crisis.

We sincerely thank the Editor, Phillip de Jager, and the two anonymous reviewers for their insightful comments and recommendations on an earlier version of the manuscript. We also thank the Massey University School of Accountancy Research Seminar participants and the 2024 Quantitative Accounting Research Symposium for their valuable suggestions. This paper is part of the first author’s PhD thesis at Massey University, New Zealand.