CEO Characteristics, Family Ownership and Corporate Social Responsibility Reporting: The Case of Saudi Arabia

Abstract

Only a few studies have investigated the association between the characteristics of the chief executive officer (CEO) (i.e., tenure and local or expatriate) and corporate social responsibility (CSR) reporting. Our study adds to the fledgling literature by providing new evidence from Saudi Arabia. Given the dominance of family control among Saudi Arabian listed firms, additionally, this study examined the moderating effect of family ownership on the CEO-CSR relationship. Using CSR scores from Bloomberg database from 2010 to 2019 and ordinary least squares (OLS) regression, the findings reveal that the association between CEO tenure and CSR reporting is positively significant; however, the association between CEO nationality and CSR is not significant. In addition, the findings indicate that family ownership is an important contingency factor that explains the association between CEO tenure and CEO nationality, and CSR reporting. Our study contributes to an emerging line of CSR research that investigates the effects of foreign CEOs on CSR transparency, and supports prior evidence on the benefits to investors of having long-serving CEO and the costs of family entrenchment.

1. Introduction

In recent years, corporate social responsibility (CSR) has emerged as a critical and enduring topic that has attracted rapidly growing scholarly attention. This explosion reflects increasing awareness among firms of their roles and responsibilities to the community as well as the environment [1,2,3,4,5]. CSR participation by firms has been extensively scrutinized by the media, socially responsible investors, and numerous CSR rating agencies, such as Morgan Stanley Capital International, Sustainalytics, FTSE Russell, Bloomberg, and Thomson Reuters, and any CSR misconducts could have a significant impact on a firm’s reputation and its sustainability [6]. However, CSR reporting continues to vary greatly across firms, which has prompted several investigations into the determinants of a firm’s CSR commitment. Previous studies have shown that the CEO of a firm can significantly influence CSR reporting; thus, the incentives and characteristics of the CEO can explain the differences in the reporting of CSR activities across firms [6,7,8,9,10,11,12,13].

Nonetheless, very few studies are available in emerging economies that have explored the relationship between CEO characteristics and CSR reporting [7,11,14]. The aim of this study was to fill the gaps in the CSR literature by investigating how CEO tenure and CEO nationality, the two important dimensions of CEO characteristics, affect CSR reporting. We focused on these two profiles of the CEO for the following reasons. CEO tenure, the time a person spends in the CEO position, is multifaceted and complex, and its implication on CSR is still not well understood [15]. The presence of foreigners in the upper echelons of management is an intriguing topic of growing interest as many firms around the world have embarked on internationalization agenda. At the same time, there are negative biases, protest, and reservations from various stakeholders on the merit of appointing foreign CEOs, i.e., those CEOs born outside the country of the firms they lead [16], and the empirical evidence on the association between CEO nationality and CSR to date is very limited.

This study is underpinned by the Upper Echelons Theory (UET), which posits that the beliefs, values, and behavior of the top management greatly influence the business model and strategic choices of the firm. The UET has been used as the dominant research framework in a review study on the link between CEO incentives and characteristics and CSR outcomes [12]. This theory contends that certain observable personal characteristics can impede or initiate decision-making, for example, the decision to undertake CSR. CEOs are critical players in the firm, and they can guide firms and make decisions that affect the firm’s CSR activities. Ref. [17] argued that a firm’s sustainability and successful performance mainly depend on the performance and quality of the firm’s top managers. Clearly, CEO characteristics have been listed among the numerous factors that influence CSR [12].

Thus, this paper sheds light into the question of how the length of the CEO’s tenure in the firm affects its CSR reporting. The evidence to date is limited, with mixed findings. According to [18], there are both benefits (richer experience-based and knowledge sets) and costs (fixed paradigm) of CEO tenure, and the CEO tenure-CSR outcomes nexus is unclear, which prompted us to re-examine this issue. Another burning question is, if the CEO of the firm is of local or foreign origin, does this have any bearing on CSR reporting? This important research question has not been adequately addressed with the exception of [16], who argued that foreign CEOs step up their CSR commitment to counter the liability of foreignness bias faced by expatriates helming local firms. Additionally, this study examines if family ownership moderates the association be-tween CEO tenure and CEO nationality, and CSR involvement, which remains largely unexplored before. Our understanding of how CEO profiles interact with family ownership to influence CSR outcomes is deficient.

We use Saudi Arabia as the empirical setup for the following reasons. First, Saudi Arabia accounts for a quarter of the gross domestic product (GDP) of the Arab world and is a member of the G20 group, making it a fascinating and significant market to research. According to Forbes Middle East, Saudi Arabia is the largest Arab Economy in 2020 and 2021 based on GDP (see Figure 1). Further, Saudi Arabia is one of the world’s major oil exporters. During 2020 it accounted for 11% of global oil production (see

Table 1. The 10 largest oil producers and their share of total world oil production in 2020.

Country	Million Barrels per Day	Share of World Total
United States	18.60	20%
Saudi Arabia	10.82	11%
Russia	10.50	11%
Canada	5.26	6%
China	4.93	5%
Iraq	4.16	4%
Brazil	3.79	4%
United Arab Emirate	3.79	4%
Iran	3.01	3%
Kuwait	2.75	3%
Total top 10	67.60	72%
World total	94.20

Source: U.S. Energy Information Administration.

). This has resulted in pollution, which can potentially harm the economy as well as the health of the people [1,19]. Thus, this provides a valid reason for further investigation of CSR activities and reporting in Saudi Arabia. Second, the practices and reporting of CSR in Saudi Arabia are still in the infancy phase [1]; therefore, it is critical to examine the effect of CEO characteristics on corporate socially responsible behavior given that CEO influence can have a major impact on CSR activities [12].

Third, in 2016, Saudi Arabia embarked on a plan (Saudi Vision 2030) to diversify its economy beyond oil. In an effort to attract more foreign investment and intensify the privatization of state-run companies, the government has greenlighted the appointment of foreigners at the head of state-run companies [20,21]. Therefore, the role of foreign CEO towards firm CSR disclosure policy in Saudi Arabia is an empirical issue. Lastly, most Saudi Arabian firms are family-owned or controlled by families [22,23]. According to [24], three Saudi families control more than 41% of executive board positions in Saudi listed firms, and these families dominate the boards of 68 listed firms out of 168 Saudi listed firms. Moreover, 17 other families dominate the boards of other Saudi listed firms. Therefore, it would be interesting to look into how family ownership influences the association between CEO characteristics and CSR reporting.

Using a 10-year panel sample with 34 unique Saudi firms and 206 firm-year observations, a significantly positive relationship was found between the tenure of CEOs and CSR reporting, implying that CEOs with longer tenures are inclined to be more focused on undertaking and reporting CSR, while a negative and insignificant association was found between CEO nationality and CSR, suggesting there is no difference in CSR disclosure choice between local and foreign CEOs. As for the role of family ownership in moderating this relationship, we found the interactions between CEO tenure and CEO nationality with family ownership are negative and significant, indicating that increased family ownership in firms weaken the positive association between CEO tenure and CSR disclosures, consistent with the entrenchment hypothesis. These findings can help policymakers, investors, and other stakeholders better understand the role of family ownership in influencing the CEO characteristics-CSR reporting relationship.

This study advances the CSR literature on several fronts. This is one of the pioneering studies to examine the relationship between CEO characteristics and its impact on CSR in Saudi Arabia. We provide further evidence on the positive effects of CEO tenure on the practices of CSR. Secondly, it advances the discourse on the effect of foreign CEOs on CSR reporting, an area that has received scant attention by CSR and international business scholars hitherto. Thirdly, it analyzes the moderating role played by family ownership in the association between CEO tenure and CEO nationality and CSR involvement, which, to the best of our knowledge, has not previously been addressed. Finally, to our knowledge, this is the first research in Saudi Arabia to use Bloomberg’s environmental, social, and governance (ESG) score to investigate the relationship between CEO characteristics and CSR reporting.

The rest of this research paper is organized as follows: Section 2 reviews the theoretical framework; Section 3 provides a review of related studies and formulates the hypotheses. In Section 4, the research method is provided, while the findings are discussed in Section 5. Section 6 checks the robustness of the findings, while Section 7 provides a conclusion, as well as the implications, study limitations, and suggested avenues for researchers wishing to undertake similar studies in future.

2. Theoretical Framework

The literature on CEOs has shown that the UET is the most widely used underpinning theory, which posits that managers perceive the reality of the strategic situation of their firms based on experience, values, and personalities, resulting in specific strategic decisions, which ultimately influence firm outcomes [25]. According to [10], the characteristics of a firm’s CEO are a partial predictor of organizational outcomes. As explained by [26], the main tenet of the UET is that the demographics of the upper echelons, which are the “reflections of the values and cognitive bases of powerful actors”, have implications for organizational outcomes. UET contends that corporate strategic choices and decision outcomes can be predicted by the decision maker’s cognitive biases and personal values. Ref. [27] argued that the CEOs are powerful individuals and responsible for strategic decision making and outcomes in organizations. They play a significant role in the decisions related to firm’s disclosures [28]. It is therefore valuable to scrutinize the CEO’s cognitive frames, whether their prior knowledge, experience, and values inform strategic decision-making, and ultimately, corporate strategy, particularly in the context of CSR participation. According to the stakeholder theory, one of the primary objectives of firms is to balance the competing demands of various stakeholders and not just the shareholders, and the CEO plays a critical role in assessing the importance of balancing and satisfying stakeholder requirements through judicious CSR participation toward achieving the firm’s objectives, namely to create long term value and therein contributing to the greater good, not only in pure business terms but also to society at large [29].

Based on the agency theory, separating control and ownership in a firm creates agency problems [30]. One suggestion for resolving the agency issue (Type I) is to align managerial interests with the rest of the shareholders through managerial or family ownership [31,32]. In contrast, Refs. [33,34] argued that managerial or family ownership may not be the best solution to the agency issue because another agency problem (Type II) is created between controlling insider shareholders and minority outsider shareholders. Ref. [35] posited a higher level of insider ownership provides management with deeper entrenchment, and hence, more opportunity for self-dealing. Ref. [36] showed that ownership structure has strong effects on a firm’s disclosures policies either directly or indirectly. Entrenched family owners may have an incentive to extract wealth from minority shareholders by exerting pressure on firm management to influence CSR reporting; family-owned firms may also practice lower levels of corporate sustainability than non-family-owned firms. Consequently, we employed these theories to investigate the impact of CEO tenure and CEO nationality on a firm’s CSR disclosure policy with the moderating role of family ownership.

3. A Review of Literature and the Development of Hypothesis

3.1. Corporate Governance and Corporate Social Responsibility in Saudi Arabia

The implementation of corporate governance in Saudi listed firms is supported by a number of rules and regulations. In 1985, the disclosure and transparency standard were issued by the Ministry of Commerce and Industry. Since disclosure and transparency are some of the most essential components of corporate governance best practices, the standard of corporate governance would be enhanced with the issuance of these standards [37,38]. Over the last two decades, Saudi Arabia has seen significant developments in corporate governance, starting with the publication of the internal control standards in 2000, followed by the publication of a corporate governance code in 2006, with the latter becoming mandatory for all Saudi listed firms beginning in 2010 [39]. Saudi’s code of corporate governance in its current status, issued in 2017, was intended to align Saudi Arabia with international best practices such as the Organization for Economic Co-operation and Development (OECD) principles [1]. The Code placed more emphasis on the social responsibility initiatives by Saudi firms. For example, Article 87: Social Responsibility on page 46 stated that “The Ordinary General Assembly, based on the Board recommendation, shall establish a policy that guarantees a balance between its objectives and those of the community for purposes of developing the social and economic conditions of the community” [40]. In addition, Article 88: Social Initiatives on page 46 mentioned that “The Board shall establish programmes and determine the necessary methods for proposing social initiatives by the Company, which include: establishing indicators that link the company’s performance with its social initiatives and comparing it with other companies that engage in similar activities; disclosing the objectives of the company’s social responsibility to its employees and raising their awareness and knowledge of social responsibility; disclosing plans for achieving social responsibility in the periodical reports on the activities of the company and establishing awareness programmes to the community to familiarize them with the company’s social responsibility” [40].

3.2. CEO Tenure and Corporate Social Responsibility Reporting

The UET suggests that the demographic traits of executives might affect how they make strategic decisions [10]. Scholars have contended that tenure is a major characteristic of CEOs which has an effect on a firm’s strategic decisions as well as performance [10,41,42]. Furthermore, the length of the tenure of a CEO can indicate his or her understanding of the firm and the ownership [43]. Therefore, long-serving CEOs may have a positive relationship with CSR reporting for several reasons. First, long-term CEOs may have greater power to make the best decisions, which, in turn, can positively affect the performance of the firm because they can build on previous decisions and successful strategies [44,45]. Second, over the course of their tenure at the firm, CEOs gain increasing amounts of firm-specific skills, knowledge, and power, as well as greater insights into the firm’s culture, structure, and resources, including those of external parties [46]. Therefore, they will feel more connected to the firm as members, accepting its vision and objectives, and be better positioned to integrate stakeholders, such as government agencies, customers, partners, and other global activist groups to gain access to external opportunities [47,48]. Third, since CSR creates long-term benefits to firms [6], CEOs with longer tenure can develop more bridging social capital [47]. These ties can help long-tenured CEOs to become more involved in CSR and, ultimately, be inspired to ensure their firm’s survival and success over the long-term. On the other hand, long CEO tenure engenders lack of flexibility, i.e., fixed paradigm syndrome, which is detrimental in managing multiple stakeholders’ interests, and can lead to inferior CSR engagement [18].

Prior studies have found conflicting effects of CEO tenure on CSR reporting. Ref. [43] posited that longer-tenured CEOs have higher scores when corporate social performance is measured. Ref. [7] found that the tenure of CEOs has a significantly positive effect on CSR performance. Ref. [49] postulated that the tenure of CEOs is positively related to the environmental performance of firms. Ref. [50] demonstrated that sustainable development is significantly and positively impacted by the tenure of CEOs. A negative relationship has been found by prior studies between CEO tenure and CSR. For instance, [14,42] verified that increasing CEO tenure negatively impacts social and environmental performance of firms. Likewise, [6] discovered that CEO tenure has a negative influence on CSR performance. In a similar vein, [51] documented that CEOs with longer tenures in the firm present weaker ESG performance. Nonetheless, we expect that CEO tenure will significantly impact CSR because CEOs are in a better position to integrate stakeholders, such as government agencies, clients, partners, and other global activist groups so as to contribute to the firm’s survival and success over the long-term. Therefore, the first hypothesis is as follows:

Hypothesis 1 (H1).

The association between CEO tenure and corporate social responsibility reporting is positive.

3.3. CEO Nationality and Corporate Social Responsibility Reporting

According to [52], the tenure of CEOs who hail from different countries and nationalities can result in the business culture being different as well. According to the UET, the cognitive and demographic characteristics of the top management team can impact their decision-making [10], and thus we anticipate that the nationality of CEOs will have two main impacts on CSR reporting as follows. First, the UET contends that the psychological preferences of CEOs will significantly impact their strategic decisions and results of a firm [53].

Accordingly, their nationality, a psychology partiality that arises from the identity of the place, plays an essential role in influencing firm strategies. Individuals’ pro-environmental intentions and behavior are related to their local attachment [54,55]. Moreover, individuals will form emotional ties with their local natural environment through their daily behavior, and these emotional ties can govern individuals’ attitude and behavior toward the environment, such as day-to-day resource and environmental conservation; thus, they will exhibit friendly behavior toward their environment and community [55,56]. Ref. [55] found that the CEOs with a hometown identity are more likely to be worried about environmental issues and to be actively involved in green innovation. Secondly, individuals are more inclined to support local causes because of their nationality [54]. Extending this hometown CEO analogy to local versus expatriate CEOs, as a result of emotional ties to their nationality, nonforeign CEOs tend to make decisions based on economic factors and social interests [57]. Thus, nationality may elicit the pro-social incentive of individuals and direct them toward the objective of helping others out of concern for the welfare of their local groups. Local CEOs can assume greater responsibility for their communities’ health and well-being, and thus they may feel morally obliged to help mitigate or resolve environmental-related problems [55].

On the other hand, as argued by [16], foreign CEOs face disadvantage and are confronted with liability of CEO foreignness as they are perceived by the locals as outgroup members. As leaders who represent their firms before local stakeholders, the nonprototypical outgroup leaders are perceived as “not one of us’’ and therefore are distrusted, disliked, and not supported and thus local stakeholders are less inclined to collaborate with them compared with local CEOs. One way for the foreign CEOs to collaborate successfully with local stakeholders is to display their benevolence by allocating more resources for CSR activities in order to overcome the intergroup bias and mitigate the negative discrimination.

Although there are divergent views on the likely effect of CEO nationality and CSR outcomes, we therefore hypothesize as follows:

Hypothesis 2 (H2).

The association between Saudi CEOs and corporate social responsibility reporting is positive.

3.4. The Moderating Roles of Family Ownership

According to the agency theory (Type I agency cost), Refs. [30,58] assumed that family ownership has a substantial impact on mitigating agency problems. Family members, according to the alignment hypothesis, have a significant financial incentive to align their interests with other stakeholders’ interests as they know more about the firm’s activities, thus allowing them to implement effective supervision over the managers. Ref. [59] showed that equity-holding families have longer investment horizons than other investors. On the other hand, family ownership may create Type II agency problems. Refs. [60,61] suggested that conflicts of interest between majority and minority owners may provide incentives for the majority to seize the assets of minority shareholders. According to the entrenchment hypothesis, family members may benefit personally at the expense of other shareholders, as dominant shareholders have less incentives to be interested in CSR disclosure because they have direct access to the firm’s information [62,63].

Despite a large body of literature on family businesses, empirical research on family-controlled firms and CSR is still in its infancy [3]. Ref. [64] argued that dominating families can use their voting power to direct the firm’s CSR resources to other programs that benefit them, resulting in family firms exhibit lower CSR performance, in support of the expropriation hypothesis. In addition, Ref. [65] claimed that because of the Type II agency problem, socio-emotional wealth maximization in family businesses overwhelms shareholder wealth maximization, resulting in non-value-enhancing CSR engagement. Empirical studies by [63,64,66,67,68,69,70,71] found a negative association between family ownership and CSR, which is consistent with the entrenchment notion that family firms are less committed to CSR than non-family firms. In Saudi Arabia, previous studies found that family firms have lower financial reporting quality and are more involved in accruals and real-earnings management activities [23,72]. Ref. [1] found a negative association between family ownership and CSR disclosures in Saudi Arabia. Therefore, the effect of Type II agency problems on family businesses in Saudi Arabia may be pervasive. Hence, the following hypotheses are put forth:

Hypothesis 3 (H3a).

The positive association between CEO tenure and corporate social responsibility reporting is attenuated by family ownership.

Hypothesis 3 (H3b).

The positive association between Saudi CEO and corporate social responsibility reporting is attenuated by family ownership.

4. Research Methodology and Data

4.1. Data and Sample

This study’s sample comprises all Saudi stock market firms with ESG reporting in the Bloomberg database from 2010 to 2019. Our sample was constructed by integrating several sets of data. Data on CSR were obtained from the Bloomberg database (i.e., the extent of ESG disclosure of a firm annually). We manually extracted data on CEO tenure, CEO nationality, family ownership, institutional investors, and board characteristics from Saudi firms’ annual reports and reports of corporate governance which were downloaded from the www.saudiexchange.sa website (accessed on 31 May 2021). As for the other financial control variables, data were downloaded from DataStream. Following these procedures, the final sample comprised 206 firm-year observations from 34 firms, encompassing the 2010 to 2019 period.

Table 2. Sample selection.

Year	Freq.	Percent	Cum.
2010	7	3.40	3.40
2011	17	8.25	11.65
2012	23	11.17	22.82
2013	23	11.17	33.98
2014	21	10.19	44.17
2015	25	12.14	56.31
2016	24	11.65	67.96
2017	26	12.62	80.58
2018	27	13.11	93.69
2019	13	6.31	100.00
Total	206	100.00

shows the annual distribution of our sample, which indicates a rising trend in sample size, and that Bloomberg has covered an increasing number of firms in recent years.

4.2. Model Specification

The following models were used in this study to examine our hypotheses. In Model 1, ordinary least squares (OLS) regression model was used to estimate the influence of CEO tenure and CEO nationality on CSR reporting.

$$\begin{array}{c}{\mathrm{CSR}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{CEO}\_{\mathrm{TEN}}_{\mathrm{it}}+{\mathsf{\beta }}_2\mathrm{CEO}\_{\mathrm{NAT}}_{\mathrm{it}}+{\mathsf{\beta }}_3\mathrm{TO}\_{\mathrm{IO}}_{\mathrm{it}}+{\mathsf{\beta }}_4{\mathrm{BMET}}_{\mathrm{it}}+{\mathsf{\beta }}_5{\mathrm{BSIZ}}_{\mathrm{it}}+\\ {\mathsf{\beta }}_6{\mathrm{BIND}}_{\mathrm{it}}+{\mathsf{\beta }}_7{\mathrm{MVTB}}_{\mathrm{it}}+{\mathsf{\beta }}_8{\mathrm{LEVERG}}_{\mathrm{it}}+{\mathsf{\beta }}_9{\mathrm{ROA}}_{\mathrm{it}}+{\mathsf{\beta }}_{10}{\mathrm{FSIZE}}_{\mathrm{it}}+\mathrm{Year}\\ \mathrm{dummies}+{\mathsf{\epsilon }}_{\mathrm{it}}\end{array}$$

(1)

In model 2, an interaction term of CEO tenure and CEO nationality with family ownership (FMOWN) was included to examine if family ownership plays a moderating role in the association between CEO tenure and CEO nationality, and CSR.

$$\begin{array}{l}{\mathrm{CSR}}_{\mathrm{it}}={\mathsf{\beta }}_0+{\mathsf{\beta }}_1\mathrm{CEO}\_{\mathrm{TEN}}_{\mathrm{it}}+{\mathsf{\beta }}_2\mathrm{CEO}\_{\mathrm{NAT}}_{\mathrm{it}}+{\mathsf{\beta }}_3{\mathrm{FMOWN}}_{\mathrm{it}}+{\mathsf{\beta }}_4\mathrm{FMOWN}\ast \\ \hspace{1em}\hspace{1em}\hspace{1em}\mathrm{CEOTEN}\mathrm{it}+{\mathsf{\beta }}_5{\mathrm{FMOWN}\ast \mathrm{CEONAT}}_{\mathrm{it}}+{\mathsf{\beta }}_6\mathrm{TO}\_{\mathrm{IO}}_{\mathrm{it}}+{\mathsf{\beta }}_7{\mathrm{BMET}}_{\mathrm{it}}\\ \hspace{1em}\hspace{1em}\hspace{1em}+{\mathsf{\beta }}_8{\mathrm{BSIZ}}_{\mathrm{it}}+{\mathsf{\beta }}_9{\mathrm{BIND}}_{\mathrm{it}}+{\mathsf{\beta }}_{10}{\mathrm{MVTB}}_{\mathrm{it}}++{\mathsf{\beta }}_{11}{\mathrm{LEVERG}}_{\mathrm{it}}+{\mathsf{\beta }}_{12}{\mathrm{ROA}}_{\mathrm{it}}\\ \hspace{1em}\hspace{1em}\hspace{1em}+{\mathsf{\beta }}_{13}{\mathrm{FSIZE}}_{\mathrm{it}}+\mathrm{Year}\mathrm{dummies}+{\mathsf{\epsilon }}_{\mathrm{it}}\end{array}$$

(2)

A regression diagnostics test was performed to investigate estimation issues, such as normality, heteroscedasticity, and multicollinearity. To solve the issue of outliers, all variables of extreme values of 1% and 5% at the top and bottom were winsorized. The heteroscedasticity test was performed according to Breusch-Pagan/Cook-Weisberg and the results indicated that it was present. The Durbin-Watson statistical test was also used to identify potential autocorrelation of the research model and this problem was confirmed by the results. For these two econometric problems, we used the OLS with robust standard errors, as it is an appropriate estimator offering more reliable estimates and non-biased standard errors.

The measurements of all the studied variables are presented in

Table 3. Variable definitions.

Variables	Acronym	Descriptions
Dependent variable
Corporate Social Responsibility	CSR	Bloomberg proprietary score—based on the amount of the ESG disclosure of a firm.
Independent variables
CEO Tenure	CEO_TEN	The number of years as CEO in the firm.
CEO Nationality	CEO_NAT	A dummy variable for CEO nationality, with ‘1′ for Saudi Arabian nationality and ‘0′ otherwise.
Moderating variable
Family Ownership	FMOWN	Proportion of outstanding shares owned by family members on the board (fulfilling at least one of the requirements: (i) a firm in which a person or a group is related by family ties or blood (i.e., share same surname; and (ii) holds directly or indirectly at least 5% of the total number of outstanding common shares).
Family*CEO Tenure	FMOWN*CEOTEN	Interaction of family ownership and CEO tenure.
Family*CEO Nationality	FMOWN*CEONAT	Interaction of family ownership and CEO nationality.
Control variables
Total institutional investors	TO_IO	Percentage of shares owned by institutional investors divided by the outstanding number of shares.
Board meeting	BMET	Number of meetings held by the board of directors.
Board size	BSIZ	Number of directors on the board.
Board independence	BIND	Number of independent directors divided by board size.
Market to book ratio	MVTB	Market capitalization scaled by book equity of the firm.
Leverage	LEVERG	Ratio of total debt to total assets.
Return on assets	ROA	Net profit scaled by total assets.
Firm size	FSIZE	Natural logarithm of total assets.

.

4.3. Variable Measurement

4.3.1. Dependent Variable: Corporate Social Responsibility

In this study, the dependent variable was CSR reporting with a 0 to 100 scale. This estimates the amount of CSR data disclosed by Saudi firms between 2010 and 2019. Market transparency can be enhanced by CSR reporting and also can provide an avenue for evaluating the future of the firm and its investment performance. The reason for selecting the ESG Bloomberg database was because of its several advantages compared with other open-ended databases as follows: (i) the Bloomberg Sustainalytics team states that ESG data can “provide investors with a macro-level assessment of how firms are managing their ESG capital”, allowing investors to “integrate ESG factors into their fundamental analysis” [73]; (ii) according to [74], “the ESG database adopts the most comprehensive methodology to evaluate firms’ environmental, social, and governance activities and outcomes”. It makes a clear distinction between CSR-related governance mechanisms and outcomes; and (iii) it provides a 0–100 score for each ESG category [75]. Therefore, the ESG database is considered as having adequate data to investigate the CSR reporting and CSR activities.

4.3.2. Independent and Moderating Variables

CEO tenure and CEO nationality were the independent variables in this study. CEO tenure is measured by the total number of years an executive has been the CEO from the year of entry [76,77]. CEO nationality was estimated using a dummy variable with a value of ‘1’ if the CEO was a Saudi citizen, and ’0’ otherwise [7,50]. In the second model, family ownership (the moderating variable) was measured as the proportion of outstanding shares held by “family members on the board (meeting at least one of the requirements): (i) a firm in which a person or group is related by blood or family ties (i.e., share the same surname); and (ii) holds at least 5% of the total number of outstanding common shares, directly or indirectly” [23,72,78].

4.3.3. Control Variables

In our model, we controlled for several corporate governance variables. As external directors are delegates of various stakeholders and are eager to fulfill the expectations of stakeholders, firms that have frequent board meetings (BMET), a larger board size (BSIZ), and a greater number of directors who are independent (BIND) should have better CSR reporting [79,80]. The BMET is measured as the total number of meetings held by board of directors, BSIZ is measured as total number of directors on the board of the firm, and BIND is measured as the ratio of independent directors over the total board [37,81,82,83,84,85,86,87,88,89,90,91,92]. We also controlled for institutional investors ownership (TO_IO), since TO_IO significantly affects CSR disclosure [93]. TO_IO is measured as shares owned by institutional investors divided by the outstanding number of shares [94,95]. We further controlled for several financial variables at the firm level. We predict that firms with higher market-to-book (MVTB) ratios, higher profitability (ROA), and larger size (FSIZE) will be more inclined to invest in CSR as they have more resources to support CSR. It is predicted that the association between leverage (LEVERG) and CSR is negative because firms can fund CSR more quickly by utilizing underutilized financial resources [6].

5. Empirical Results

5.1. Descriptive Statistics

Table 4. Descriptive statistics.

Variables	Mean	Median	Std. Dev.	Min	Max
CSR	14.621	11.280	9.685	2.193	48.347
CEO_TEN	11.449	8.000	9.650	0.164	39.000
CEO_NAT	0.777	1.000	0.417	0.000	1.000
FMOWN	9.189	0.000	18.438	0.000	94.996
TO_IO	28.904	15.300	26.802	0.000	98.180
BMET	6.049	6.000	2.195	2.000	16.000
BSIZ	9.660	9.000	0.890	7.000	12.000
BIND	42.229	36.364	11.920	20.000	100.000
MVTB	2.061	1.625	1.367	0.660	10.430
LEVERG	23.576	22.774	18.323	0.000	74.430
ROA	4.861	2.435	5.957	−3.420	38.200
Size	120,600,000	58,598,352	154,300,000	1,848,221	1,481,000,000
FSIZE	17.828	17.886	1.422	14.430	21.116
		Freq.	Percent	Cum.
CEO_NAT	0	46	22.33	22.33
	1	160	77.67	100.00
Total		206	100.00

Note: All variables are defined in Table 2.

presents the descriptive statistics for the main variables of this study. The mean (median) CSR value in Saudi Arabia was 14.621 (11.280), indicating relatively low CSR reporting (15 out of a possible 100). The standard deviation of CSR was 9.685, indicating that there are significant differences in the performance of CSR across the firms sampled in this study. The mean (median) CEO tenure was 11.449 (8), indicating that a CEO in a Saudi firm has an average tenure of 11 years. Table 3 also shows that the percentage of CEOs of Saudi nationality is 77.67%, while the percentage of other nationalities is 22.33%. For family ownership, the mean shareholding by family members was around 9.19%. The mean of board size (BSIZ) among Saudi firms was 9.660, while for board independence (BIND), it was 42.23%. This implies that there are about 42% independent directors on Saudi firm boards. The mean values of ROA and LEVERG were 4.86% and 23.58%, respectively.

5.2. Correlation Analysis

In

Table 5. Correlation matrix.

Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)	VIF
(1) CSR	1.000
(2) CEO_TEN	0.280 ***	1.000										1.27
(3) CEO_NAT	0.010	0.198 ***	1.000									1.30
(4) TO_IO	0.209 ***	−0.028	0.127 *	1.000								2.73
(5) BMET	−0.086	−0.168 **	0.138 **	0.381 ***	1.000							1.28
(6) BSIZ	0.038	0.007	−0.133 *	−0.372 ***	−0.034	1.000						1.44
(7) BIND	−0.066	0.016	0.012	−0.086	0.087	0.185 ***	1.000					1.24
(8) MVTB	−0.022	0.258 ***	0.060	−0.364 ***	−0.163 **	0.143 **	−0.088	1.000				2.07
(9) LEVERG	0.042	−0.077	0.175 **	−0.069	−0.026	−0.315 ***	−0.269 ***	−0.064	1.000			1.45
(10) ROA	−0.024	0.359 ***	0.158 **	−0.247 ***	−0.187 ***	0.089	−0.016	0.668 ***	−0.143 **	1.000		2.43
(11) FSIZE	0.193 ***	−0.078	−0.208 ***	0.652 ***	0.265 ***	−0.070	−0.122 *	−0.502 ***	−0.200 ***	−0.519 ***	1.000	3.27

Note(s): ***, **, * Significant at the 1%, 5% and 10% levels, respectively.

, the correlation matrix for the variables is presented. CEO_TEN had significant and positive correlations with CSR reporting, hence supporting H1. In addition, the correlation between CEO_NAT and CSR was positive but insignificant. The association between the control variables (institutional ownership and firm size) and CSR of Saudi listed firms was positive and significant. The highest significant correlation coefficient (0.668) among the independent variables was between MVTB and ROA. Table 5 also indicates that multicollinearity is not a problem as the variance inflation factor (VIF) value was significantly lower compared with the 5-cutoff threshold. According to the data, the highest VIF value was 3.27.

5.3. Regression Results

Table 6. Ordinary Least Square (OLS) Regression Results.

Independent Variables	Model (1)		Model (2)
	Coef.	t-Value	Coef.	t-Value
CEO_TEN	0.289	4.14 ***	0.243	3.37 ***
CEO_NAT	−1.921	−1.22	−2.982	−1.81 *
FMOWN			1.801	2.35 **
FMOWN*CEO_TEN			−1.776	−2.63 ***
FMOWN*CEO_NAT			−1.970	−1.66 *
TO_IO	0.134	4.09 ***	0.130	3.78 ***
BMET	−0.632	−2.48 **	−0.567	−2.33 **
BSIZ	2.145	2.03 **	1.710	1.48
BIND	−0.010	−0.23	−0.009	−0.18
MVTB	0.339	0.40	−0.166	−0.20
LEVERG	0.095	2.57 **	0.111	2.84 ***
ROA	−0.072	−0.40	0.011	0.05
FSIZE	0.068	0.09	0.050	0.06
Constant	−14.955	−1.12	−8.693	−0.63
Year Dummies	Yes		Yes
N. of Obs	206		206
R-squared	0.3568		0.3928
Prob > F	<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

displays the hypotheses test results. We used robust standard errors STATA to run OLS regressions to test the hypotheses. It also indicates that the OLS regression’s F-test was at the 1% significance level, suggesting that CEO_TEN and CEO_NAT variables in this analysis are important for explaining CSR variation. In Column 1, the summary of the regression results of Equation (1) and on the impact of CEO_TEN and CEO_NAT on corporate CSR is shown. The CEO_TEN coefficients were significantly positive (0.289 with a t-value of 4.14), meaning that longer-tenured CEOs have a positive relationship with a firm’s CSR reporting. These findings also suggest that CEOs who have been with the firm for a longer time are more interested in undertaking CSR practices. This is consistent with H1 of this study as well as with previous studies [7,49,50]. The findings are also in tandem with the stakeholder theory, i.e., longer tenured CEOs are more stakeholder-centric.

Furthermore, the results show that the coefficient of CEO_NAT was negative but not significant (−1.921 with t-value of −1.22). This finding suggests that a local Saudi CEO has little interest in CSR, whereas a foreign national may be more concerned. Our finding is in line with [7], who found that CEO_NAT and CSR performance are not correlated. In model (2) (Table 6), the interaction between CEO_TEN and FMOWN is negatively significant (−1.776 with t-value of −2.63), suggesting that family-owned firms play a negatively moderating role in the association between CEO_TEN and CSR, consistent with the entrenchment hypothesis. Therefore, H3a is supported. In addition, the interaction between CEO_NAT and FMOWN was negative and significant (−1.970 with a t-value of −1.66), indicating that increased family ownership in firms strengthens the negative association between CEO nationality and CSR, which supports H3b and is also consistent with the entrenchment hypothesis. This result implies that in family firm, longer tenured CEO and local CEO negatively affect CSR engagement.

For the control variables, the results in Table 6 show that TO_IO had a positive and significant relationship with CSR reporting at 1% level in both models, suggesting that more ownership by institutional investors are associated with more CSR reporting by Saudi firms, consistent with prior Saudi studies [1,96]. BMET had a negative and significant association with CSR reporting, where BSIZ was positively and significantly connected to CSR. LEVERG had a positive and significant link to CSR reporting. With regards to the other control variables BIND, MVTB, ROA, and FSIZE, the results showed no significant association between these variables and CSR reporting.

6. Further Investigations

6.1. Alternative Measurement of CEO Tenure

To ensure the robustness of our findings, further tests were performed using alternative measures of CEO_TEN (i.e., CEO tenure as the natural logarithm of the number of years a CEO has been in office) (LNCEO_TEN) [6]. The results of replacing the new variable of interest (LNCEO_TEN) with early CEO_TEN measurement in the main empirical model are shown in

Table 7. OLS Regression Results of Alternative measurement of CEO Tenure.

Independent Variables	Model (1)		Model (2)
	Coef.	t-Value	Coef.	t-Value
LNCEO_TEN	2.013	3.42 ***	1.699	2.67 ***
CEO_NAT	−1.577	−1.00	−1.840	−1.18
FMOWN			1.876	2.20 **
FMOWN*CEO_TEN			−1.067	−1.79 *
FMOWN*CEO_NAT			−2.080	−1.73 *
TO_IO	0.141	4.07 ***	0.131	3.54 ***
BMET	−0.581	−2.21 **	−0.543	−2.11 **
BSIZ	2.318	2.16 **	2.069	1.77 *
BIND	−0.005	−0.11	0.008	0.17
MVTB	0.451	0.54	−0.213	−0.24
LEVERG	0.103	2.75 ***	0.106	2.79 ***
ROA	−0.007	−0.04	0.098	0.45
FSIZE	0.179	0.22	0.298	0.36
Constant	−20.601	−1.56	−18.872	−1.41
Year Dummies	Yes		Yes
N. of Obs	206		206
R-squared	0.3369		0.3641
Prob > F	<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

. The coefficient of LNCEO_TEN was positively significant (2.013 with a t-value of 3.42), and the interaction between LNCEO_TEN and family ownership had a negative relationship with CSR. As for the interaction between CEO_NAT and family ownership, it had a negative relationship with CSR. Overall, the findings concur with our main findings.

6.2. Endogeneity Checks

Endogeneity, which is the result of concurrent findings explaining variables and omitted variables, is a major problem in accounting research. The previous financial performance of firms can simultaneously influence CSR performance and CEO tenure [6], which may bias the empirical results using OLS regression. In this research, two approaches were used to address this problem. The first approach was to employ the instrumental variables (IV) approach [6,97]. The previous year’s average CEO tenure (LNCEO_AVRRS) was used as an IV in the first-stage regression to derive the estimated value of LNCEO_AVRRS, which may have an impact on CEO_TEN, but is not likely to be associated with a firm’s CSR. In the second stage, the estimated value of LNCEO_AVRRS was utilized to replace the “CEO_TEN” variable.

Table 8. Regression Results: Instrumental variable in the two-stage least squares (IV-2SLS).

Independent Variables	Model (1)		Model (2)
	Coef.	t-Value	Coef.	t-Value
CEO_TEN_AVRG	2.619	3.02 ***	2.803	3.02 ***
CEO_NAT	−0.373	−0.24	−1.892	−1.14
FMOWN			2.123	2.72 ***
FMOWN*CEO_TEN			−2.146	−2.83 ***
FMOWN*CEO_NAT			−2.712	−2.25 **
TO_IO	0.126	3.62 ***	0.117	3.31 ***
BMET	−0.885	−3.21 ***	−0.732	−2.85 ***
BSIZ	2.029	1.90 **	1.546	1.34
BIND	0.015	0.35	0.010	0.23
MVTB	0.562	0.65	−0.090	−0.11
LEVERG	0.097	2.57 **	0.113	2.84 ***
ROA	0.124	0.72	0.178	0.89
FSIZE	0.731	0.93	0.620	0.81
Constant	−23.452	−1.79 *	−14.964	−1.09
Year Dummies	Yes		Yes
N. of Obs	206		206
R-squared	0.3239		0.3842
Prob > F	<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

shows the results of the IV-2SLS regression based on the estimated LNCEO_AVRRS value. It can be seen that the estimated LNCEO_AVRRS coefficient was 2.619 and the t value was 3.02. Therefore, it is statistically and positively significant and supports H1. In addition, the estimated coefficient of the interaction term (FMOWN*CEO_TEN) was −2.146, significant at the 1% level (t = −2.83), and the interaction term (FMOWN*CEO_NAT) was −2.712, significant at the 5% level (t = −2.25), thereby supporting H3a and H3b. The results are in tandem with the results in the main model. The second approach was to re-estimate the primary model with the lagged values of the independent variables to address the potential problem of reversal of causality [98,99]. The results of these re-evaluations are shown in

Table 9. OLS regression results for lagged values of independent variables.

Independent Variables	Model (1)		Model (2)
	Coef.	t-Value	Coef.	t-Value
CEO_TEN	0.277	3.79 ***	0.201	2.75 ***
CEO_NAT	−1.711	−1.07	−3.703	−2.23 **
FMOWN			2.518	3.10 ***
FMOWN*CEO_TEN			−2.411	−3.39 ***
FMOWN*CEO_NAT			−3.111	−2.29 **
TO_IO	0.107	3.23 ***	0.097	2.67 ***
BMET	−0.742	−2.57 **	−0.642	−2.33 **
BSIZ	1.750	1.51	1.073	0.90
BIND	0.001	0.03	0.000	0.01
MVTB	−0.011	−0.01	−0.782	−1.09
LEVERG	0.121	2.55 **	0.146	3.16 ***
ROA	0.031	0.17	0.184	0.99
FSIZE	0.681	0.79	0.662	0.79
Constant	−21.096	−1.30	−11.100	−0.66
Year Dummies	Yes		Yes
N. of Obs	169		169
R-squared	0.3629		0.4343
Prob > F	<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

, which shows that reversal of causality is not likely to be significant, in tandem with the findings in this paper.

6.3. Sub-Sample Family and Non-Family

To measure the impact of family ownership on the relationship between CEO characteristics (tenure and nationality) and CSR, OLS regression for family and non-family subsamples was estimated as shown in

Table 10. OLS regression results for Sub-Sample Family and Non-Family.

Independent Variables	Family		Non-Family
	Coef.	t-Value	Coef.	t-Value
CEO_TEN	0.061	0.32	0.407	5.39 ***
CEO_NAT	−7.925	−1.62	−3.739	−2.06 **
TO_IO	0.110	1.78 *	0.224	4.54 ***
BMET	−0.466	−0.51	−0.713	−3.23 **
BSIZ	4.708	3.09 ***	0.526	0.31
BIND	−0.113	−0.73	−0.001	−0.03
MVTB	−0.393	−0.40	0.153	0.13
LEVERG	0.262	2.67 **	0.090	2.27 **
ROA	−0.545	−1.57	0.035	0.13
FSIZE	−1.911	−0.92	−1.721	−1.47
Constant	1.533	0.03	29.307	1.36
Year Dummies	Yes		Yes
N. of Obs	79		127
R-squared	0.4472		0.4923
Prob > F	<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

. We found significantly positive results for the coefficient of CEO_TEN in regression tests for non-family firms (0.407 with t-value of 5.39). As for family firms, we found that the CEO_TEN-CSR relationship was insignificant, implying that the benefit of longer tenured CEO in promoting CSR is negligible in family firms. Therefore, the argument that family members may make private gains from the firm to the detriment of other shareholders is supported because they have direct access to the firm’s information, giving rise to agency problem Type II and, consequently, to lower CSR performance. Interestingly, among non-family firms, foreign CEOs are more actively involved in CSR pursuit than their Saudi counterparts.

6.4. Hierarchical Regression (Alternative Regression Approach)

Even though we used OLS regression in our main analysis, we also used hierarchical regression to strengthen our findings, as shown in

Table 11. Hierarchical regression results.

Independent Variables	Phase 1		Phase 2		Phase 3		Phase 4
	Coef.	t-Value	Coef.	t-Value	Coef.	t-Value	Coef.	t-Value
TO_IO	0.123	3.19 ***	0.134	4.09 ***	0.144	4.44 ***	0.130	3.78 ***
BMET	−0.895	−3.53 ***	−0.632	−2.48 **	−0.668	−2.67 ***	−0.567	−2.33 **
BSIZ	2.024	1.85 *	2.145	2.03 **	2.154	2.03 **	1.710	1.48
BIND	0.015	0.35	−0.010	−0.23	−0.002	−0.03	−0.009	−0.18
MVTB	0.572	0.63	0.339	0.40	0.250	0.29	−0.166	−0.20
LEVERG	0.096	2.58 **	0.095	2.57 **	0.101	2.60 **	0.111	2.84 ***
ROA	0.121	0.66	−0.072	−0.40	−0.046	−0.24	0.011	0.05
FSIZE	0.782	0.96	0.068	0.09	0.021	0.03	0.050	0.06
CEO_TEN			0.289	4.14 ***	0.284	4.03 ***	0.243	3.37 ***
CEO_NAT			−0.802	−1.22	−0.886	−1.27	−1.245	−1.81 *
FMOWN					0.441	0.89	1.801	2.35 **
FMOWN*CEO_TEN							−1.776	−2.63 ***
FMOWN*CEO_NAT							−1.970	−1.66 *
Constant	−24.423	−2.01 **	−16.447	−1.26	−16.092	−1.24	−11.009	−0.80
Year Dummies	Yes		Yes		Yes		Yes
N. of Obs	206		206		206		206
R-squared	0.2932		0.3568		0.3582		0.3928
Prob > F	<0.000 ***		<0.000 ***		<0.000 ***		<0.000 ***

Notes: ***, **, * Significant at the 1%, 5%, and 10% levels, respectively; all variables are defined in Table 2.

. Hierarchical regression analysis is an alternative comparative process with betas to evaluate the significance of the independent variables. Many authors have suggested this regression as a common technique for the identification of moderating effects [100,101,102]. Four steps were taken in entering variables in the regression equation (see Table 11), as recommended by [100,101]. According to [103], a moderate variable is significant only if a change occurs in R-squared (R2). During the first phase, R2 was 0.2932, indicating that the CSR level could be explained by institutional ownership, board meeting, board size, board independence, leverage, market-to-book price, return on asset, and firm size. In phase 2, R2 increased to 0.3568 with the addition of the independent variables. This means that CEO_TEN and CEO_NAT explain a further increase (0.0636) in CSR. In phase 3, R2 is notably changed by FMOWN to 0.0014, which shows that FMOWN has an impact on CSR. In the final phase, R2 increased from 0.3582 to 0.3928 when the interaction was entered. This difference in R2 (0.0346) is substantial, indicating that FMOWN influences both the CEO_TEN-CSR and CEO_NAT-CSR relationships. The hierarchical regression results corroborate the primary findings of this paper. The findings also support our empirical findings, i.e., the influence of family ownership interacts negatively with CEO_TEN, CEO_NAT, and CSR.

6.5. Alternative Regression Approaches

As previously explained (Section 3.2), our research models employed the OLS regression approach in the main analysis. Following [9,44,55], we further employed Feasible Generalized Least Square (FGLS), Panel Corrected Standard Errors (PCSE), and Driscoll-Kraay standard errors (SCC) regressions to confirm the strength of the main results. The findings in

Table 12. Results of FGLS, PCSE, and SCC regressions.

Independent Variables	FGLS				PCSE				SCC
	Model (1)		Model (2)		Model (1)		Model (2)		Model (1)		Model (2)
	Coef.	z-Value	Coef.	z-Value	Coef.	z-Value	Coef.	t-Value	Coef.	t-Value	Coef.	t-Value
CEO_TEN	0.110	2.32 **	0.112	2.35 **	0.289	4.59 ***	0.243	3.72 ***	0.289	3.80 ***	0.243	2.88 **
CEO_NAT	−0.363	−0.40	−2.635	−2.34 **	−1.921	−1.50	−2.982	−2.25 **	−1.921	−1.72	−2.982	−2.69 **
FMOWN			1.937	3.63 ***			1.801	2.73 ***			1.801	4.97 ***
FMOWN*CEO_TEN			−1.493	−3.89 ***			−1.776	−3.13 ***			−1.776	−3.23 **
FMOWN*CEO_NAT			−2.390	−2.77 ***			−1.970	−1.98 **			−1.970	−3.35 ***
TO_IO	0.077	4.48 ***	0.082	4.05 ***	0.134	4.21 ***	0.130	3.91 ***	0.134	5.82 ***	0.130	5.27 ***
BMET	−0.489	−2.49 **	−0.489	−2.72 ***	−0.632	−2.40 **	−0.567	−2.19 **	−0.632	−5.69 ***	−0.567	−5.83 ***
BSIZ	0.594	1.31	0.080	0.17	2.145	2.12 **	1.710	1.58	2.145	1.57	1.710	1.20
BIND	0.017	0.51	−0.005	−0.14	−0.010	−0.19	−0.009	−0.15	−0.010	−0.20	−0.009	−0.15
MVTB	0.876	2.11 **	0.668	1.66 *	0.339	0.42	−0.166	−0.20	0.339	0.42	−0.166	−0.20
LEVERG	0.078	3.87 ***	0.075	3.34 ***	0.095	2.66 ***	0.111	2.95 ***	0.095	1.87 *	0.111	2.95 **
ROA	0.056	0.57	0.014	0.14	−0.072	−0.41	0.011	0.06	−0.072	−0.78	0.011	0.09
FSIZE	0.911	2.50 **	0.708	1.78 *	0.068	0.09	0.050	0.07	0.068	0.15	0.050	0.10
Constant	−14.122	−1.63	−1.498	−0.16	−14.955	−1.14	−8.693	−0.65	−14.955	−1.40	−8.693	−0.64
Year Dummies	Yes		Yes		Yes		Yes		Yes		Yes
N. of Obs	206		206		206		206		206		206
Wald chi2 (9)/R2	354.61		151.39		121.04		152.84		0.3568		0.3928
Prob > chi2	<0.000 ***		<0.000 ***		<0.000 ***		<0.000 ***		<0.000 ***		<0.000 ***

Note(s): ***, **, * Significant at the 1%, 5%, and 10% levels, respectively.

demonstrated CEO_TEN has a significantly positive relationship with CSR and CEO_NAT has an insignificant relationship with CSR. Furthermore, FMOWN had a significantly negative moderating effect on these relationships. These results confirm the main results shown in Table 6, suggesting that Saudi CEOs are concerned with CSR, but that family ownership limits their ability to address CSR.

7. Conclusions

The study examined the association between CEO tenure and CEO nationality and CSR reporting. We further investigated whether or not family ownership affects that relationship, given that family firms are ubiquitous in Saudi Arabia. This study is critical for enhancing our understanding of family businesses’ CSR behavior due to their unique ownership structure and agency issues. OLS regression and descriptive statistics were used to investigate 206 firm-year observations from 2010 to 2019. The findings revealed that CEO_TEN and CSR reporting have a positive relationship, consistent with evidence in Pakistan [11]. Meanwhile, CEO_NAT had no effect on CSR reporting, which suggests the mechanism through which CEO_NAT may create CSR transparency is much more complex that the simple dichotomy of local versus foreign CEOs. Moreover, the findings revealed that there is a negatively moderating effect of FMOWN on the association between CEO characteristics (tenure and nationality) and CSR reporting, suggesting that the incentives for family-controlled firms to engage in CSR and provide more information are lower, and the information asymmetry in such firms cannot be adequately reduced. As a result, this finding endorses the notion that family firms face more agency issues, especially arising from corporate control entrenchment. Our study advances the literature on CSR reporting in Saudi Arabia by investigating the leadership role of CEO and using more recent and extended CSR data from Bloomberg, spanning the period 2010–2019. Prior studies on the determinants of CSR reporting in Saudi Arabia by [1,37,96] used self-constructed CSR disclosure index for the periods 2007–2011, 2013–2014, and 2016–2018 respectively, and focused more on the effects of audit committee and board characteristics. This study also adds to the nascent CSR literature on how the CEO profiles interact with family ownership to influence CSR outcomes.

The findings of this study provide valuable insights to policymakers and practitioners to legislate effective regulations on CEO and conduct supervision on CSR reporting and disclosure, while considering the different motivations that influence family firms towards CSR reporting. To investors, our results may encourage them to invest in firms where the CEO has been in office for a long time, because a longer-tenured CEO leads to more positive CSR results. To the stakeholders, our findings suggest that firms with long-tenured CEOs tend to be more socially responsible. Further, the results of this study indicated that interactions between CEO tenure and CEO nationality with family ownership are negative and significant, suggesting that increased family ownership in firms weaken the positive association between CEO tenure and CSR disclosures, consistent with the entrenchment hypothesis. Finally, the findings of this study should assist in assessing the potential contribution of appointing expatriate CEOs to lead Saudi firms towards the accomplishment of Saudi Vision 2030.

Like other studies, this study is not without limitations that pave the way for future empirical studies. While the percentage of foreign directors in Saudi firms has been increasing, more research is required to validate the influence of foreign CEOs. Our study did not find a clear relationship between CEO nationality and authentic CSR activities. This suggests the heterogeneity in the effects of the local/foreign CEO on CSR. The channel through which the nationality of the CEO may create authentic CSR is much more complex than the simple dichotomy of Saudi versus non-Saudi CEO. Future research may consider whether Saudi CEOs’ foreign exposure, or non-Saudi CEOs’ prior work experience in the region, subject to data availability, plays any role in influencing CSR transparency. Another limitation is the small sample size, because the required data from 2010 to 2019 for most listed firms were unavailable, which means the results may not be generalizable to other Saudi firms, or other countries. The difficulty and complexity in gathering data on the profile of CEOs restrict the opportunity to investigate additional behavioral biases and demographic factors. Therefore, research carried out in the future could be enriched by investigating the impact of the relationship between different CEO characteristics and CSR, as well as incorporating additional research methods, such as CEO interviews, surveys, and in-depth case studies, to ascertain corporate management’s perceptions of CSR participation. Despite the inherent limitations associated with data availability and small sample size that beset most corporate governance studies in the MENA region [104], this study paves the way for future investigations on the heterogeneity of family businesses [105] and other contingency factors that drive the CEO profiles and CSR relationship in the MENA region and elsewhere.