Ownership Structure and Bank Dividend Policies: New Empirical Evidence from the Dual Banking Systems of MENA Countries

Abstract

This study investigates the relationship between ownership structures and dividend policies for 46 Islamic and 75 conventional banks from 12 MENA and Asian countries between 2012 and 2020. Logit regression is employed to estimate the regression equation, centering on the moderating impacts of the COVID-19 pandemic and national culture. Our findings remain robust as we tackle the endogeneity issue using probit and logistic regression models. Asset growth and GDP growth serve as proxies for investment opportunities. Additionally, dividend per share acts as a proxy for dividend policy. Our findings emphasize how the ownership structure impacts dividend payouts in both banking systems. We observed positive relationships between dividend payouts and foreign ownership, bank size, age, and performance. Conversely, concentration of ownership and leverage negatively influence dividend payouts. The COVID-19 pandemic directly boosts the dividend policy for conventional banks and alters the relationship between foreign ownership and distribution policy in Islamic banks. Specifically, COVID-19 interacts with foreign and state ownership to reduce dividend payouts, but concentration of ownership does not show this effect. This study furnishes evidence affirming the significance of the ownership structure in shaping the dividend payout policy within Islamic and conventional banking. The results maintain their reliability across various estimation approaches. Moreover, this study accounts for the crisis period as a moderating factor influencing dividend payments.

1. Introduction

The dividend payout policy has received much more attention among scholars, corporate strategists, and policymakers than ever. Previous studies do not support any unique set of the determinants of corporate dividend policy (Athari 2020; Bataineh 2020), and hence it remains an issue that deserves attention from academic researchers (Al-Najjar and Kilincarslan 2019; Allen and Michaely 2003; Black 1996; Ed-Dafali et al. 2023). Regarding the “dividend puzzle”, Black (1996, p. 5) states, “The harder we look at the dividend picture, the more it seems like a puzzle, with pieces that just don’t fit together”. Many studies on payout policy have investigated differences between Islamic and conventional banks (Duqi et al. 2020). For instance, Ataullah et al. (2022) found that dividend cuts during the COVID-19 pandemic were even more extensive than during the financial crisis between 2007 and 2008. However, some studies include only Islamic and conventional banks in the Gulf Cooperation Council countries (e.g., Yousef et al. 2021) or those operating in Arab markets (Athari et al. 2016). On the other hand, few studies have quantitatively investigated the effects of shareholders on dividend payout policy in non-US banks (Duqi et al. 2020). Additionally, limited empirical studies analyze how the ownership structure impacts dividend payout strategies during a crisis in Islamic banks (hereafter, IBs) and conventional banks (hereafter, CBs). This article compares the dividend policies of Islamic and conventional institutions in the MENA region based on cross-country data. Several bibliometric and review papers on dividend policy phenomena concluded that the ownership structure influences the dividend policy (Baker and Weigand 2015; Pinto et al. 2020; Booth and Zhou 2017; Ed-Dafali et al. 2023). The agency hypothesis is undoubtedly one of the significant theories among the various arguments for why firms pay dividends (John et al. 2011; Kim et al. 2017; Oded 2020). As a result, managers alter payouts to reduce the sum of the transaction and agency costs (Moh’d et al. 1995; Puleo et al. 2009). Indeed, dividend payments redistribute free cash flows to shareholders, limiting managers’ capacity to make unwise investments (Dhanani 2005). For example, previous evidence suggests that state ownership affects dividend changes (Ben-Nasr 2015). This is consistent with empirical evidence that state-controlled firms use high target payout ratios to reduce managerial agency costs (Gugler 2003). For instance, Naz et al. (2017) show that managers who go from non-sharia compliant to sharia-compliant firms exhibit noticeably different patterns regarding their influences on financial decisions. The present study is focused on the banking industry since banks play a vital role in today’s world economy, specifically in the MENA economies. More importantly, the main goal of this article is to offer practical proof of the influence of the ownership structure on dividend policy in the MENA region’s conventional and Islamic banking sectors. More specifically, we focus on two dual-system models to explain the dividend policy in these banks: Islamic vs. conventional banks. Further, we conduct a comparative analysis of dividends per share for IBs and CBs. Our cross-country study uses agency theory arguments to compare dividend distribution practices in Islamic and conventional banks. This paper emphasizes how crucial the ownership structure is in explaining differences in dividend payouts between the two types of banks. The findings indicate an advantageous relationship between foreign ownership, bank size, bank age, bank performance, and dividend payout. Leverage and ownership concentration, however, have a detrimental impact on dividend distribution. While COVID-19 negatively moderates the relationship between foreign ownership and the distribution policy of Islamic banks, it has a positive and direct impact on the dividend policy for conventional banks. We discovered, in particular, that while the COVID-19 epidemic interacted with foreign and state ownership to diminish dividend payouts (i.e., dividend yield and dividend payment), the concentration of ownership did not display such an effect.

We also report in our robustness tests that national culture may play a stronger role as a moderator in the relationship of the ownership structure with the dividend policy (Ashraf et al. 2016; DasGupta and Roy 2023; Ashraf 2021). By doing so, this study contributes to the literature in several ways. First, unlike previous studies, our study empirically explains the moderating effect of the crisis period on the relationship between the ownership structure and the dividend policy. In doing so, we highlight how the ownership structure (concentration of ownership, state ownership, foreign ownership) interacts with COVID-19 to influence dividend payouts. Second, while previous studies examine the effect of the ownership structure on the dividend policy in IBs or CBs, our paper reports results from an empirical study that compares dividend policies in a dual-banking system. Thirdly, we complement the recent financial literature that reports corporate dividend policy changes during the pandemic (Ali 2022; Krieger et al. 2021; Hoang and Nguyen 2024). Fourth, we add to the corpus of knowledge about many aspects of corporate dividend policy, as well as studies on the effects of COVID-19 on bank dividend policies and strategies (e.g., Cejnek et al. 2021; Casanova et al. 2021; Hardy 2021), including those investigating the impact of the pandemic on IBs and CBs (El-Chaarani et al. 2022; Hassan et al. 2022; Miah et al. 2021; EL-Chaarani et al. 2023; Demir and Danisman 2021; Elnahass et al. 2021). Fifth, we have contributed significantly to the empirical studies of agency theory’s explanation for the cross-sectional distribution of payout strategies in a dual-banking system. The findings offer valuable insights for future research in emerging economies, highlighting the importance of dividend payout strategies in reducing agency problems for policymakers and practitioners in establishing an optimal dividend policy. Sixth, our paper complements the existing studies on the role of culture in explaining distribution policy (Chang et al. 2020; Kim et al. 2020; Khiar and Kooli 2023), with a focus on Hofstede’s seminal work on national work cultures (Hofstede 1980, 2001, 2011; Hofstede et al. 2010). Our paper complements these studies by examining the interaction effects of uncertainty avoidance with ownership structure variables to explain dividend payouts in the MENA region.

The rest of this paper is structured as follows: The theoretical context and the testable hypothesis are outlined in Section 2. We describe the methodology used in Section 3. We present and discuss the research findings and further robustness analysis in Section 4. The Section 5 concludes our study, provides research and policy implications, examines the limitations of the present work, and suggests areas for additional investigations.

2. Literature Review

2.1. Theoretical Background

2.1.1. Dividend Policy of Islamic vs. Conventional Banks

Several studies have thoroughly explained why IBs are more resilient during economic downturns than their conventional counterparts (e.g., Hasan and Dridi 2011; Beck et al. 2013; Soedarmono and Yusgiantoro 2023), on the one hand, and that Islamic banks exhibited a larger credit supply compared to their foreign bank peers (Hasan and Dridi 2011; Louhichi et al. 2020), on the other hand. However, IBs have limited financing choices compared to conventional competitors (Duqi et al. 2020). Furthermore, Kabir et al. (2015) found that bank concentration significantly lowers credit risk, as measured by the distance-to-default (DD) score and Z-score. Bank boards typically comprise insiders and outsiders (Fields et al. 2012). In particular, insiders should balance the needs of creditors and minority shareholders while maximizing dividend policy, lowering equity agency expenses, and reconciling these goals (Shao et al. 2013). Notably, IBs exhibit fewer agency conflicts between insiders and outsiders due to their accountability and transparency (Farag et al. 2018; Agbodjo et al. 2021; Toumi and Hamrouni 2023). For example, Hudayati et al. (2023) noted that IBs often consider a wider range of stakeholders when established instead of CBs, which prioritize stockholders. Naz et al. (2017) found that leverage, dividend payouts, and working capital policies differ considerably between Sharia-compliant and non-sharia-compliant companies. According to a study by Athari et al. (2016), profitability and size favorably impact dividend payouts for both IBs and CBs operating in Arab markets, whereas leverage had adverse effects. Al-Kayed (2017) examined the dividend policy of 12 Saudi banks during a four-year period from 2011 to 2014. Her study revealed that CBs provide higher dividends, even though IBs are more profitable. Nevertheless, dividends and dividend yield distributions are insufficient for predicting the market value of emerging-market banks from the MENA region (Budagaga 2020).

Similarly, Hassan et al. (2003) found that Islamic banks exhibited a stable dividend policy. This is consistent with agency theory’s assumptions that firms follow substantial and stable dividend policies (Gugler 2003). This choice may help Islamic banks send positive signals to the market (Al-Kayed 2017). Elnahass et al. (2021) provide evidence that the negative impact of COVID-19 differed between the two banking systems (i.e., conventional and Islamic). For example, Floyd et al. (2015) compared the payoff strategies of US industrials and banks from 1980 to 2012 using free cash flow and signaling theories, while Duqi et al. (2020) used these theories to analyze the distribution of policy of Islamic banks in comparison to their traditional peers, based in 16 countries, for the period 2000–2015. Based on a large sample of dividend policy studies using bibliometric analysis, Ed-Dafali et al. (2023) concluded that the asymmetric information (signaling) hypothesis and agency costs of free cash flow are among the leading hypotheses of the various arguments for why companies pay dividends.

2.1.2. Ownership Structure and Dividend Policy

The dividend payout policy is considered one of the most important and challenging decisions. This decision is influenced by the firm’s ownership structure and ultimately taken along with investment and financing decisions. The theory of the ownership structure for the firm is viewed as a critical driver to establishing appropriate incentives for the manager (Jensen and Meckling 1976; Parvin et al. 2020). Thus, previous evidence suggests that a dividend policy can mitigate the agency problem (Jensen et al. 1992; Rozeff 1982). Previous studies suggested agency problems are more severe among Islamic banks operating in MENA countries (Athari et al. 2016). According to Boshnak (2021), a well-structured ownership framework will potentially significantly affect a company’s dividend payout strategy. Their findings highlighted the significance of the board structure and ownership configuration as determinants of dividend policy differences among Saudi-listed corporations between 2016 and 2019. They notably underlined a favorable correlation between the likelihood to pay dividends and elements, including the frequency of board meetings, company size and profitability, and institutional ownership.

2.2. Hypotheses Development

The board’s efficiency in supervising top management is harmed by the concentration of power and decision-making in the hands of one person (Fama and Jensen 1983). Renneboog and Szilagyi (2020) reported that the probability of a dividend payout is affected by the concentration of control in the hands of the two largest shareholders. Specifically, strong minority shareholders request dividends to mitigate expropriation risk (Renneboog and Szilagyi 2020; Gugler and Yurtoglu 2003). Thus, dividends supplement rather than replace shareholder control in minimizing agency issues (Renneboog and Szilagyi 2020). Cheung et al. (2005) found a favorable association between the dividend yield and higher levels of CEO ownership. Renneboog and Trojanowski (2011) also argued that with more voting power held by executive directors, there is a higher chance of dividend payments. Schooley and Barney (1994) state that higher executive stock ownership reduces agency costs while decreasing the dividend yield. From an agency theory perspective, increased corporate governance directly affects how agency issues can be minimized in the context of banks operating in emerging economies (Grassa et al. 2021). Adjaoud and Ben-Amar (2010) argued that enhanced shareholder rights are linked to larger dividend distributions because good corporate governance controls managers’ opportunistic conduct in payout decisions. The agency cost justification for corporate dividend policy is supported by prior research (Das Mohapatra and Panda 2022; Rozeff 1982; Easterbrook 1984; Bian et al. 2023), providing evidence that managers work to reduce the agency cost/transaction cost structure both inside and across companies over time (Moh’d et al. 1995). Specifically, agency costs tend to be lowered when insider control grows, given that managers carry a greater share of these costs (Farinha 2003). Therefore, minority shareholders may prefer higher dividends because they have little power over companies (Rozeff 1982; Easterbrook 1984). Khalfan and Wendt (2020) show that ownership concentration negatively affects dividends, reinforcing earlier evidence that overly concentrated ownership decreases dividends (Maury and Pajuste 2002). According to Shleifer and Vishny (1986), ownership concentration enables significant shareholders to monitor management, thus resolving the free rider problem associated with dispersed ownership. Meanwhile, Khan (2006) demonstrates that the identity of the largest shareholder impacts dividend payments. While some researchers have supported a positive relationship between ownership concentration and dividend policy (e.g., Briano-Turrent et al. 2020; Gonzalez et al. 2017), other studies have found that ownership concentration was negatively related to dividend policy (e.g., Farinha 2003; Schooley and Barney 1994; Moh’d et al. 1995; Khalfan and Wendt 2020; Sbai et al. 2024). However, other researchers (Hanafi et al. 2013; Busta et al. 2014) conclude that ownership concentration has a non-linear (U or inverse U) relationship with payout. Regarding short- and medium-term investments, asset concentration was accurately explained by the dividend signaling hypothesis investments in the Islamic interest-free banking system (Hassan et al. 2003). We thus hypothesize the following:

H1a. 

There is a significant association between ownership concentration and dividend payments in conventional banks.

H1b. 

There is a significant association between ownership concentration and dividend payments in Islamic banks.

Ben-Nasr (2015) discovers significant and compelling evidence that the dividend level is inversely connected to state ownership. The study reveals that increased state ownership is associated with a lesser proclivity to pay dividends, a lower likelihood of rising payments, and a lower preference for dividend cuts. For instance, the expected ratio of payouts for state-controlled businesses was 54% instead of 41% for businesses under foreign ownership (Gugler 2003). Hasan et al. (2023) found that government ownership hurt dividend payouts in the Bangladesh context. The benefit of stock dividends being linked to the possibility of higher overall cash dividends, as noted by Bechmann and Raaballe (2007), outweighs the disadvantage of the simultaneous growth in share capital. Similarly, Gugler (2003) argued that state-owned enterprises are the most cautious when reducing payouts. Thus, we hypothesize the following:

H2a. 

State ownership is negatively related to dividend payments in conventional banks.

H2b. 

State ownership is negatively related to dividend payments in Islamic banks.

Previous research has yielded conflicting results regarding the association between foreign ownership and dividend distribution policies (Cao et al. 2017; Bataineh 2020; Duqi et al. 2020; Boshnak 2021). Although several studies have reported a positive relationship between dividends and foreign ownership (e.g., Mossadak et al. 2016), other studies have reported reduced payouts for foreign ownership for emerging countries (Bataineh 2020; Dahlquist and Robertsson 2001; Al-Najjar and Kilincarslan 2016; Lam et al. 2012; Moin et al. 2020). Domestic owners in the Gulf Cooperation Council countries own most of the banking sector (Zeitun 2012). For instance, Bashir (2000) demonstrated that foreign-owned banks outperformed domestic competitors in profitability based on a sample of 14 Islamic banks in Middle Eastern nations from 1993 to 1998. As Al-Kayed (2017) noted, Islamic bank managers primarily use dividends to signal profitability to external investors.

Huang et al. (2021) studied the impact of foreign ownership on the payout. They found that, during the global financial crisis of 2007–2009, Chinese companies with foreign ownership, in particular, increased dividend payouts. The previous literature has demonstrated that, due to dividend increases, foreign-controlled firms seemed to exhibit an underinvestment problem (Huang et al. 2021). Furthermore, according to Cao et al. (2017), foreign investors prefer to invest in publicly traded companies with a more generous payout policy. This could be explained by foreign owners engaging in corporate monitoring (Khalfan and Wendt 2020). Hamao and Mei (2001) find that foreign investment positively correlates with long-term fundamentals such as dividend growth rate. According to Duqi et al. (2020), foreign ownership is linked to an increased payout policy in Islamic banks rather than in conventional ones. Therefore, we hypothesize the following:

H3a. 

Foreign ownership is positively related to dividend payments in conventional banks.

H3b. 

Foreign ownership is positively related to dividend payments in Islamic banks.

2.3. The Moderating Effect of the Crisis Period

Many studies have suggested that the COVID-19 pandemic has affected corporate dividend payments (Ali 2022; Jakubik and Teleu 2021), including bank dividend policies (Casanova et al. 2021; Hardy 2021). It is suggested that rent expropriation and the extraction of private benefits of control are exacerbated during a crisis (Johnson et al. 2000). For example, banks worldwide experienced a sudden drawdown of pre-negotiated credit lines (Casanova et al. 2021). Hardy (2021) analyzes dividend payout limits across US banks during the COVID-19 epidemic and concludes that dividend payout restrictions are implemented by regulators to strengthen bank stability and enable stronger growth in bank lending. For example, Cejnek et al. (2021) show that these restrictions are most severe for the financial industry and companies that have received government support to face the COVID-19 pandemic. Comparatively to the 2008 global financial crisis, the economic uncertainty brought on by the COVID-19 pandemic has a bigger impact on sector volatility across all sectors (Choi 2020). Although the distinction between institutional investors classified as international vs. domestic is unimportant to the compensation strategies adopted by businesses in reaction to the COVID-19 pandemic (Ataullah et al. 2022), foreign investors questioned the veracity of financial statements during the COVID-19 crisis, as well as the actual economic impact of the pandemic (Gong et al. 2022). Due to the substantial crisis impact of the COVID-19 epidemic, Alharasis (2023) argues that the relationship between foreign ownership and audit quality has been strengthened. According to Danisman et al. (2021), banking systems with a higher share of Islamic banks are more resilient to the pandemic. More specifically, Ataullah et al. (2022) discovered that firms with significant institutional ownership are more likely to press executives to keep earnings rather than distribute dividends, which may help mitigate the negative effects of the COVID-19 pandemic and enhance their future prospects. Kabir et al. (2015) showed that Islamic banks had a much higher credit risk in most years, but during the crisis, they showed significantly lower Z-scores than CBs. Although there is a difference between IBs and CBs, Elnahass et al. (2021) conclude that COVID-19 adversely affected bank performance and financial stability in 116 countries from the first quarter of 2019 to the second quarter of 2020. Therefore, we hypothesize the following:

H4. 

The crisis period negatively affects dividend payments while moderating the effects of different ownership categories and for both bank types.

3. Data and Method

3.1. Sample

The sample comprises 46 Islamic and 75 CBs from 12 countries in the MENA region and Asia: Bahrain, Bangladesh, Egypt, Jordan, Kuwait, Malaysia, Oman, Pakistan, Qatar, Saudi Arabia, Turkey, and the United Arab Emirates. The study period covers 2012 to 2020. We have selected countries with at least one Islamic and one conventional bank.

Table 1. Distribution of Islamic and conventional banks by country.

	Islamic Banks		Conventional Banks
	Number of Banks	Number of Observations	Number of Banks	Number of Observations
Bahrain	1	9	3	27
Bangladesh	20	180	5	45
Egypt	2	18	2	18
Jordan	10	90	1	9
Kuwait	2	18	5	45
Malaysia	3	27	3	27
Oman	2	18	1	9
Pakistan	13	117	6	54
Qatar	5	45	2	18
Saudi Arabia	1	9	7	63
Turkey	10	90	2	18
United Arab Emirates	6	54	9	81
Total	46	414	75	675

shows a breakdown of IBs and CBs by country. Islamic banks represent 38% of the total sample.

The banks’ ownership structure data were manually collected from their annual reports, accessible on their respective websites between 2012 and 2020. The banks’ financial data were sourced from the Thomson Reuters database.

3.2. Dependent Variable

In this study, we employ two measures for dependent variables. Firstly, akin to Bradford et al. (2013); Byoun et al. (2016), and Saeed and Sameer (2017), we utilize the dividend yield. This is calculated as the ratio of dividend per share to share price. Secondly, we utilize a binary variable that assumes a value of 1 if the bank opts to pay dividends and 0 otherwise (i.e., dividend cuts). This approach aligns with prior studies (e.g., Byoun et al. 2016; Pucheta-Martínez and Bel-Oms 2016; Ye et al. 2019; Almeida et al. 2020).

3.3. Independent Variables

This study employs three independent variables: ownership concentration (Con_k), foreign ownership (FOR), and state ownership (STA). We measure ownership concentration as the equity percentage held by the largest shareholder, following previous studies on conventional and Islamic banks, since this metric is essential for evaluating ownership dispersion in the banking sector (Iannotta et al. 2007; Srairi 2013; Pichard-Stamford 2000; Godard 2005; Sbai and Meghouar 2017). In line with Al-Najjar and Kilincarslan (2016), state ownership is gauged by the proportion owned by the state. Following the approach of Lam et al. (2012); Bataineh (2020), foreign ownership is measured by the proportion held by foreign shareholders.

3.4. The Moderating Effects of COVID-19 Pandemic

The COVID-19 crisis is represented by a binary variable, with a value of 1 if the year under consideration is 2020 and 0 otherwise.

3.5. Control Variables

The model is supplemented by several control variables. These encompass bank characteristics such as profitability, size, age, asset growth, and the presence of Islamic banking. Additionally, it incorporates country characteristics like the GDP growth rate. We incorporate control variables for bank size (Size), quantified by the natural logarithm of total assets (as per Lepetit et al. 2018). Bank age is assessed using the natural logarithm of years since establishment, following Talavera et al. (2018). Moreover, in line Arhinful and Radmehr (2023), the leverage ratio (Leverage), calculated as the ratio of total debt to total assets, will be included as a determinant influencing dividend payout policy. The Islamic banking (IB) variable is a binary variable, assuming the value 1 for Islamic banks and 0 for others. Asset growth is determined by the change in total assets within one year. Similarly, the return on assets (ROA) is computed as the net income ratio to total assets. Additionally, we introduce a macroeconomic variable: the GDP growth rate. These control variables are used in this study because they have been found to have a statistically significant impact on dividends in previous studies.

Table 2. Measures of variables.

Variables	Notation	Definition
Dependent Variables
Dividend yield	DIVY	Dividend per share to price per share
Dividend payment	DIVP	A dummy variable that takes the value 1 if the bank decides to pay dividends and the value 0 otherwise (no dividend payment)
Independent Variables
Concentration of ownership	Con_K	The percentage of capital held by the first shareholder
State ownership	STA	The percentage of capital held by government
Foreign ownership	FOR	The percentage of capital held by foreign
Moderator variable
Crisis due to COVID-19 pandemic	COVID	A dummy variable that takes the value of 1 during year 2020 and 0 otherwise
Control Variables
Islamic bank	IB	The value equal to one for Islamic banks, otherwise zero
Bank size	Size	Natural logarithm of the total assets
Performance	ROA	Net profit over total assets
Leverage	Leverage	Ratio of total debts over total assets
Bank age	AGE	Natural logarithm of bank age
Gowth opprtunities	Growth	Annual change in assets
GDP growth rate	GDP-g	The annual change in GDP as a percentage

contains a description and measurements of the variables used in this study.

3.6. Model Specification

To achieve the objective of this study, the empirical form of the model was formulated as follows:

$${\mathrm{D}\mathrm{I}\mathrm{V}\mathrm{Y}}_{i,t}={\beta }_0+{\beta }_{1}Covid+{\beta }_2\mathrm{C}\mathrm{o}\mathrm{n}\_\mathrm{K}+{\beta }_3\mathrm{S}\mathrm{T}\mathrm{A}+{\beta }_4\mathrm{F}\mathrm{O}\mathrm{R}+{\beta }_5\mathrm{C}\mathrm{O}\mathrm{V}\mathrm{I}\mathrm{D}+\sum {\beta }_{k}controls+{\epsilon }_{i,t}$$

(1)

To estimate Model 1, we employ a Tobit model. Maddala (1987) argues that the simple Tobit model with double censoring is particularly well suited for modeling bank dividend policy. This estimation method has been utilized in several empirical studies on dividend policy (e.g., Bataineh 2020; Briano-Turrent et al. 2020; Saeed and Sameer 2017; Attig et al. 2016; Lam et al. 2012).

$${\mathrm{D}\mathrm{I}\mathrm{V}\mathrm{P}}_{i,t}={\beta }_0+{\beta }_{1}Covid+{\beta }_2\mathrm{C}\mathrm{o}\mathrm{n}\_\mathrm{K}+{\beta }_3\mathrm{S}\mathrm{T}\mathrm{A}+{\beta }_4\mathrm{F}\mathrm{O}\mathrm{R}+{\beta }_5\mathrm{C}\mathrm{O}\mathrm{V}\mathrm{I}\mathrm{D}\times \mathrm{o}\mathrm{w}\mathrm{n}\mathrm{e}\mathrm{r}\mathrm{s}\mathrm{h}\mathrm{i}\mathrm{p}\mathrm{s}\mathrm{t}\mathrm{r}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{e}+\sum {\beta }_{k}controls+{\epsilon }_{i,t}$$

(2)

We employ a logit model to estimate Model 2. This choice aligns with other similar studies (Byoun et al. 2016; Pucheta-Martínez and Bel-Oms 2016; Ye et al. 2019; Almeida et al. 2020) due to the binary nature of the dependent variable. Logistic regression is more suitable for analyzing the relationship between ownership structure and dividend payout.

4. Empirical Results

4.1. Descriptive Statistics and Correlation

Table 3. Descriptive statistics.

Variables	Mean	Median	S.D.	Minimum	Maximum
DIVY	0.036	0.035	0.03	0.00	0.16
DIVP	0.71	1.00	0.45	0	1
CON	0.36	0.30	0.25	0.02	0.999
STA	0.17	0.01	0.24	0.000	0.991
FOR	0.14	0.00	0.25	0.00	0.999
IB	0.38	0	0.49	0	1
SIZE	446.06	116.36	914.75	1.035	6049
ROA	0.013	0.012	0.01	−0.12	0.05
Leverage	0.17	0.13	0.44	0	14
AGE	3.47	3.62	0.62	0.69	5.05
GROWTH	0.14	0.10	0.57	−0.53	17.91

Notes: This table reports the descriptive statistics for a set of variables used in the present study for the period 2012–2020.

reports the descriptive statistics for a set of variables used in this study, spanning 2012–2020. And

Table 4. Pearson correlation matrix.

Variables	DIVY	DIVP	STAT	CON_K	FOR	IB	GDP	ROA	LEVERAGE	GROWTH	AGE	SIZE	COVID
DIVY	1
DIVP	0.72 ***	1
STA	−0.04	−0.02	1
FOR	0.01	−0.06 *	−0.23 ***	1
CON_K	−0.09 ***	−0.17 ***	0.35 ***	0.45 ***	1
IB	−0.01	0.06 **	−0.04	−0.05 *	−0.04	1
GDP	0.01	−0.04	−0.07 **	−0.02	−0.08 ***	−0.17 ***	1
ROA	0.22 ***	0.33 ***	0.07 **	−0.03	0.06 **	0.00	0.10 ***	1
Leverage	−0.12 ***	−0.03	0.13 ***	0.07 **	0.19 ***	−0.06	−0.05 *	0.11 ***	1
GROWTH	−0.01	0.00	0.05 *	0.05	0.04	−0.001	−0.01	−0.02	0.02	1
AGE	0.09 ***	0.18 ***	0.05	0.15 ***	0.16 ***	−0.23 ***	−0.11 ***	0.19 ***	0.002	−0.07 **	1
SIZE	0.03	0.30 ***	0.15 ***	−0.11 ***	0.06 *	0.25 ***	−0.28 ***	0.29 ***	0.10 ***	−0.001	0.39 ***	1
COVID	−0.01	−0.03	−0.02	0.003	−0.00	0.001	−0.5 ***	−0.15 ***	0.05 *	0.04	0.11 ***	0.13 ***	1

Notes: This table reports the correlation matrix of variables used for this study. *, **, *** represent significance at the 10%, 5%, and 1% levels, respectively.

reports the correlation matrix of variables used for this study. According to these initial results, 38% of the sample are Islamic banks. Concerning financial characteristics, not all banks in our sample paid dividends yearly during the studied period. In addition, the dividend yield varied from 0% to 16%, which means that some banks had kept the same level of the paid dividend for years. Concerning the ownership structure, the percentage of capital the largest shareholder holds varies between 2% and 99.9%. This point reflects how concentrated the ownership of some banks is in the MENA region. For state ownership and foreign ownership, these two elements are not present in the capital of all banks, and when this is the case, these two variables could each represent 99% of the bank’s capital. Finally, for the control variables, we noted that not all banks are indebted; some others achieve a negative ROA, and finally, the growth opportunities variable shows a significant difference between these banks.

4.2. Regression Results

Table 5. Baseline regressions.

Variables	DIVY		DIVP
	All Banks	Interaction Effects	All Banks	Interaction Effects
Constant	0.003 (0.298)	−0.001 (−0.105)	−4.085 *** (−5.755)	−4.424 *** (−6.051)
Con-K	−0.033 *** (−4.785)	−0.035 *** (−4.718)	−2.387 *** (−6.150)	−2.370 *** (−5.706)
STA	0.005 (0.693)	0.009 (1.343)	0.392 (1.006)	0.637 (1.542)
FOR	0.016 ** (2.408)	0.019 *** (2.779)	0.721 * (1.930)	0.923 ** (2.311)
Con-K $\times$ COVID		0.013 (0.638)		−0.384 (−0.319)
STA $\times$ COVID		−0.043 * (−1.914)		−2.465 * (−1.194)
FOR $\times$ COVID		−0.03 (−1.425)		−1.969 * (−1.668)
COVID		0.013 * (1.687)		1.037 ** (2.021)
IB	0.001 (0.343)	0.001 (0.477)	0.133 (0.739)	0.140 (0.772)
Size	0.001 (0.195)	0.0003 (0.297)	0.427 *** (5.116)	0.442 *** (5.264)
ROA	1.532 *** (8.903)	1.564 ** (8.965)	107.87 *** (8.443)	111.009 *** (8.474)
Leverage	−0.034 *** (−3.715)	−0.034 *** (−3.759)	−1.054 ** (−1.976)	−1.08 ** (−2.025)
AGE	0.005 ** (2.055)	0.005 ** (2.009)	0.265 * (1.825)	0.266 * (1.826)
Growth	0.001 (0.29)	0.002 (0.766)	0.107 (0.649)	0.213 (1.196)
GDP-g	−0.0003 (−0.859)	0.0005 (0.103)	−0.019 (−0.729)	−0.003 (−0.079)
Country FEs	Yes	Yes	Yes	Yes
Observations	1089		1089
Pseudo R2	0.18	0.18	0.20	0.21

Notes: This table presents the results of Tobit regression (Model 1) and logistic regression (Model 2). Definitions of variables are provided in Table 2. *, **, *** represent significance at the 10%, 5%, and 1% levels, respectively.

shows the baseline regressions of our study. Two columns are dedicated to each dependent variable, DIVY and DIVP. The first column reports the results for the full sample, i.e., without differentiating the two types of banks (IBs or CBs). A second column includes the interaction effects between the dividend payouts and COVID-19. The column “All banks” results indicate that the coefficients of ownership and foreign shareholders concentrations are significant at 1%, 5%, and 10%, respectively. The effects of these two variables are negative and positive, respectively, whereas the coefficient of government is not significant. This implies that concentrated ownership negatively affects the decision to pay dividends, and when this is the case, banks pay a lower dividend ratio. This is consistent with H1a and H1b. Although this result is consistent with some previous studies (Khalfan and Wendt 2020), it does not align with others’ evidence, which found that ownership concentration positively influences the likelihood and amount of dividend payments in Nordic public companies (Aleknevičienė and Vilimaitė 2023). On the other hand, the results show that the presence of foreign shareholders in banks’ capital encourages and positively influences dividend payouts. These results align with H3a and H3b, supporting previous research findings (Duqi et al. 2020; Mossadak et al. 2016; Huang et al. 2021). This point does not corroborate Bataineh (2020), who noted that foreign ownership is associated with a lesser likelihood of Jordanian companies paying dividends. The impact of state ownership is insignificant, meaning that governments do not influence the dividend policy of state-owned banks or banks with significant state ownership. This point is not consistent with H2a and H2b. Our results align with those of Bataineh (2020) but differ from those of other previous studies. The later studies found that government ownerships have a negative effect on dividend payouts (Gugler 2003; Hasan et al. 2023).

The COVID dummy is positive and significant, indicating that the COVID-19 crisis affected the dividend policy of banks. This finding confirms previous studies (Hardy 2021; Ali 2022) and could be explained by restrictions implemented by regulators to strengthen bank stability. Regarding the interactive effect, the results show that Conc-K $\times COVID$ is insignificant, whereas $STA\times COVID$ and $FOR\times$ COVID has a significance level of 10%. In both cases, state-owned banks and privately (foreign-) owned banks are negatively linked to dividend payouts in the COVID-19 crisis for both bank types. This result is in line with H4. We observe that these owners implement a more prudent policy aimed at having highly capitalized banks during the crisis. This point is a main priority for governments rather than distributing dividends to reduce agency costs.

Regarding the control variables, dividend payouts are positively influenced by bank performance and the age variable, whereas the dividend payment is only impacted by the bank size. On the other hand, leverage negatively affects dividend payouts.

Previous studies have found that the main differences in terms of the dividend payout between both IBs and CBs could be driven by the bank size. According to the results in

Table 6. Estimations by size.

Variables	Large Banks		Small Banks
	Model 1	Model 2	Model 1	Model 2
Constant	0.058 ** (4.167)	0.754 (0.516)	−0.167 *** (−4.077)	−9.06 *** (−5.09)
Con-k	−0.031 *** (−4.222)	−3.465 *** (−5.28)	−0.026 * (−1.165)	−1.372 ** (−2.345)
STA	0.034 *** (4.716)	2.179 *** (2.843)	−0.02 (−1.36)	−0.492 (−0.877)
FOR	0.014 * (1.832)	0.381 (0.578)	0.019 (1.45)	0.945 * (1.704)
Con-k $\times$ COVID	0.022 (1.044)	1.008 (0.566)	0.005 (0.124)	−1.567 (−0.91)
STA $\times$ COVID	−0.04 * (−1.83)	−3.126 * (−1.708)	−0.046 (−1.02)	−2.153 (−1.115)
FOR $\times$ COVID	−0.008 (−0.369)	−1.657 (−0.977)	−0.05 (−1.18)	−1.937 (−1.124)
COVID	−0.005 (−0.479)	0.527 (0.523)	0.026 ** (1.95)	1.579 ** (2.40)
IB	0.003 (1.37)	0.25 (0.84)	−0.009 (−1.42)	−0.235 (−0.90)
Size	−0.005 *** (−3.246)	0.007 (0.044)	0.022 *** (4.126)	1.02 *** (4.454)
ROA	0.691 *** (3.88)	73.166 *** (3.818)	3.057 *** (8.04)	155.62 *** (7.416)
Leverage	−0.001 (−0.378)	−0.042 (−0.180)	−0.053 *** (−3.02)	−1.497 ** (−2.09)
AGE	0.005 ** (2.123)	0.174 (0.666)	−0.001 (−0.287)	0.099 (0.47)
Growth	−0.011 (−1.422)	−0.852 (−1.285)	0.003 (0.79)	0.495 (0.803)
GDP-g	−0.0002 (−0.305)	0.051 (0.927)	0.001 (0.63)	0.027 (0.58)
Country FEs	Yes	Yes	Yes	Yes
Observations	540		549
Pseudo R2	0.25	0.32	0.22	0.19

Notes: This table presents the results of Tobit regression (Model 1) and logistic regression (Model 2). Definitions of variables are provided in Table 2. *, **, *** represent significance at the 10%, 5%, and 1% levels, respectively.

, the coefficients of concentration, state, and foreign ownership are negative (or positive) and significant at 10%, 5%, or 1%. Indeed, the concentration of ownership negatively affected dividend payouts, specifically in large banks. The effect of state ownership is positive and only in large banks, indicating that the government encourages dividend payouts and may put less pressure on the stability of Islamic and conventional large banks. Concerning foreign ownership, it affected dividend policy less in large banks compared to small banks. On the other hand, COVID-19 negatively moderates the relationship between state ownership and dividend payouts of large banks (STA*COVID is negative and significant at 10%). The COVID dummy is positive and significant at 5%. This observation indicated that the COVID-19 crisis affected the dividend policy of small banks, both Islamic and conventional ones. Finally, regarding the control variables, the findings showed a negative (positive) and significant effect of size on the dividend yield/dividend payouts of large (small) banks. Bank performance is positively associated with dividend policy, more specifically in small banks (ROA is significant at 1%). The age variable positively impacts the dividend yield in large banks, and leverage negatively affects dividend payouts in small banks.

4.3. Robustness Tests

4.3.1. Islamic Banks vs. Conventional Banks

The additional robustness tests (

Table 7. Difference Means of Islamic banks and conventional banks.

	Islamic Banks (N = 414)	Conventional Banks (N = 675)	Student Statistic t-Value
DIVY	0.036	0.037	−0.001
DIVP	0.749	0.69	0.059 *
Con-k	0.344	0.367	−0.023
STA	0.158	0.178	−0.02
FOR	0.121	0.149	−0.028
Size	9.325	8.634	0.691 ***
ROA	0.013	0.013	0.00
Leverage	0.185	0.163	0.022 **
Growth	0.135	0.141	−0.006 *
AGE	3.28	3.579	−0.299 ***

*, **, *** represent significance at the 10%, 5%, and 1% levels, respectively.

and

Table 8. The effect of ownership structure and dividend policy: Islamic banks vs. conventional banks.

Variables	Islamic Banks		Conventional Banks
	Model 1	Model 2	Model 1	Model 2
Constant	−0.039 ** (−2.429)	−8.545 *** (−5.682)	0.021 (1.529)	−2.641 *** (−3.275)
Con-k	−0.011 (−1.169)	−1.580 ** (−2.354)	−0.052 *** (−4.727)	−3.284 *** (−5.484)
STA	−0.016 (−1.526)	−1.473 * (−1.816)	0.025 ** (2.434)	1.484 *** (2.714)
FOR	0.001 (0.089)	0.449 (0.684)	0.036 *** (3.616)	1.652 *** (2.995)
Con-k $\times$ COVID	0.016 (0.536)	−0.625 (−0.319)	0.011 (0.391)	−0.342 (−0.211)
STA $\times$ COVID	−0.05 (−1.507)	−3.515 (−1.479)	−0.038 (−1.257)	−2.497 (−1.524)
FOR $\times$ COVID	−0.067 ** (−2.05)	−3.694 * (−1.808)	−0.025 (−0.851)	−1.594 (−1.011)
COVID	0.012 (0.899)	0.448 (0.461)	0.023 *** (2.591)	1.62 *** (2.669)
Size	0.001 (0.571)	0.672 *** (3.873)	−0.001 (−0.464)	0.320 *** (3.392)
ROA	1.391 *** (4.883)	132.941 *** (4.656)	1.87 *** (8.278)	113.64 *** (7.295)
Leverage	0.001 (0.352)	0.511 (0.541)	−0.053 *** (−4.027)	−1.499 ** (−2.060)
AGE	0.013 *** (3.389)	0.927 *** (3.111)	0.014 * (3.11)	0.850 * (2.567)
Growth	−0.005 (−0.565)	−0.506 (−0.612)	0.001 (0.505)	0.256 (0.739)
GDP-g	0.001 * (1.782)	−0.007 (−0.118)	0.001 (0.319)	0.059 (0.316)
Country FEs	Yes	Yes	Yes
Observations	414		675
Pseudo R2	0.25	0.32	0.22	0.19

Notes: This table presents the results of Tobit regression (Model 1) and logistic regression (Model 2). Table 2 provides variable definitions. *, **, and *** represent significance at the 10%, 5%, and 1% levels, respectively.

) indicate that, on the one hand, the effects of the ownership structure (Con-k, STA, and FOR) could be different depending on bank types (IBs or CBs). More specifically, we noted that state ownership negatively (positively) affected the dividend policy of Islamic (conventional) banks, and foreign shareholders positively impacted the dividend payouts of CBs. Finally, the concentration of ownership negatively influenced the dividend policy of both IBs and CBs. On the other hand, the COVID-19 crisis negatively moderated the relationship between foreign ownership and the dividend payouts of Islamic banks and positively impacted the dividend payouts of CBs.

In connection with the control variables, the performance and age variables positively affected the dividend payouts of both IBs and CBs. The leverage variable negatively influenced the dividend policy of CBs. Finally, the GDP growth rate could have a positive effect on the dividend yield of Islamic banks.

4.3.2. National Culture and Dividend Payout

To examine the impact of national culture on the dividend payout, we used the uncertainty avoidance dimension as one of four of Hofstede’s cultural dimensions (Hofstede 1980, 2001, 2011; Hofstede et al. 2010). Since we focus on the moderating effects of culture on the relationship between the ownership structure and dividend payout, we interact with the uncertainty avoidance dimension with the concentration of capital, state ownership, and foreign ownership.

Table 9. Ownership structure and dividend policy: the moderating effect of culture.

Variables	DIVY				DIVP
	(1)	(2)	(3)	(4)	(1)	(2)	(3)	(4)
Constant	0.035 *** (2.601)	−0.05 ** (−2.371)	0.018 (1.268)	0.025 * (1.703)	−0.689 (−0.792)	−5.431 *** (−3.692)	−2.714 *** (−2.677)	−1.56 (−1.597)
UAV	−0.001 *** (−4.538)	0.001 ** (2.269)	−0.0004 *** (−2.712)	−0.0005 *** (−2.907)	−0.073 *** (−6.017)	−0.005 (−0.225)	−0.048 *** (−3.636)	−0.061 *** (−4.406)
Con-K	−0.028 *** (−4.087)	0.204 *** (4.495)	−0.025 *** (−3.631)	−0.029 *** (−4.217)	−1.91 *** (−4.849)	9.473 *** (3.354)	−1.742 *** (−4.393)	−2.048 *** (−5.087)
STA	0.006 (0.824)	−0.001 (−0.152)	0.001 (0.199)	0.065 * (1.703)	0.396 (1.005)	0.048 (0.119)	0.103 (0.257)	5.73 ** (2.073)
FOR	0.016 ** (2.439)	0.016 ** (2.348)	0.139 *** (3.486)	0.015 ** (2.220)	0.652 * (1.704)	0.681 * (1.746)	13.00 *** (4.276)	0.566 (1.46)
UAV $\times$ Con-K		−0.003 *** (−5.167)				−0.167 *** (−4.058)
UAV $\times$ STA				−0.0009 (−1.582)				−0.077 * (−1.948)
UAV $\times$ FOR			−0.001 *** (−3.118)				−0.175 *** (−4.103)
COVID	0.005 (0.935)	0.009 * (1.821)	0.006 (1.253)	0.005 (1.016)	−0.133 (−0.427)	0.187 (0.574)	0.109 (0.339)	−0.083 (−0.266)
IB	0.0002 (0.07)	−0.002 (−0.695)	−0.004 (−0.144)	0.0002 (0.064)	0.124 (0.675)	0.028 (0.151)	0.005 (0.026)	0.131 (0.711)
Size	0.001 (0.556)	0.001 (0.642)	0.001 (1.031)	0.0005 (0.383)	0.572 *** (6.131)	0.565 *** (6.002)	0.643 *** (6.706)	0.549 *** (5.771)
ROA	1.634 *** (9.359)	1.734 *** (9.925)	1.663 *** (9.056)	1.645 *** (9.419)	116.846 *** (8.64)	127.21 *** (8.935)	120.11 *** (8.70)	121.52 *** (8.679)
Leverage	−0.019 ** (−2.04)	−0.010 (−1.062)	−0.023 ** (−2.327)	−0.017 * (−1.698)	0.263 (0.432)	0.744 (1.116)	−0.128 (−0.199)	0.658 (0.992)
AGE	0.005 ** (1.795)	0.005 * (1.958)	0.004 (1.53)	0.006 ** (2.181)	0.27 * (1.835)	0.248 * (1.672)	0.138 (0.921)	0.307 ** (2.063)
Growth	0.0004 (0.179)	0.0002 (0.109)	0.0002 (0.100)	0.0004 (0.193)	0.109 (0.531)	0.114 (0.527)	0.08 (0.40)	0.116 (0.530)
GDP-g	−0.0003 (−0.675)	0.0003 (0.537)	−0.0001 (−0.141)	−0.0003 (−0.571)	−0.06 * (−1.868)	−0.013 (−0.357)	−0.018 (−0.505)	−0.056 * (−1.677)
Observations	1089

Notes: This table reports the results for Tobit regressions where the dependent variables are the dividends per share to price per share (DIVY) and dividend payment (DIVP). Definitions of variables are provided in Table 2. ***, **, and * denote the significance of the test at the 1%, 5%, and 10% levels, respectively.

shows the results of the moderating effect of culture. Uncertainty avoidance has a negative and significant coefficient in all models for the dependent variable DIVP and in models 1, 3, and 4 for the dependent variable DIVY. We document a negative and significant relationship between uncertainty avoidance and dividend payout. This suggests that banks in countries with high uncertainty avoidance pay less dividends. The interaction terms (UAV*Con-K, UAV*STA, UAV*FOR), our variables of interest, are negative and significant, indicating that uncertainty avoidance reinforces the negative effect of the concentration of capital, state ownership, and foreign ownership on the dividend payout policy of banks in MENA countries. State-owned and privately (foreign-) owned banks are negatively linked to dividend payouts in countries with high levels of uncertainty avoidance for both bank types. These owners practice a more prudent dividend policy aimed at having highly capitalized banks in these countries.

5. Conclusions

This article investigates the effect of the ownership structure on bank dividend policy in MENA countries, spanning 2012–2020. The empirical sample includes the two types of banks (IBs and CBs) and considers the moderating effect of both the COVID-19 pandemic and the cultural dimension of MENA countries. To address the endogeneity problem, we used probit and logistic regression models. More importantly, testing the moderating effect of the crisis period using an interaction variable on the relationship between the ownership structure and dividend policy is a significant addition, similar to studies by EL-Chaarani et al. (2023); El Ammari (2021); Diab et al. (2024).

The baseline regression results showed the relevance of the ownership structure to explain the dividend policy variation in both banks (IBs and CBs). Indeed, concentrated ownership negatively affects the decision to pay dividends, more specifically in large banks. The presence of foreign shareholders in banks’ capital encourages and positively influences dividend payouts. Moreover, state-owned banks and privately (foreign-) owned banks are negatively linked to dividend payouts in the COVID-19 crisis for both bank types. These owners implement a more prudent policy to have highly capitalized banks in the pandemic crisis. Related to bank size, the results deduce that the effect of state ownership is positive in large banks, and foreign ownership less affected the dividend policy of large banks than small banks. In addition, COVID-19 negatively moderates the relationship between state ownership and the dividend policy of large banks. On the other hand, the robustness checks confirm that state ownership negatively (positively) affected the dividend policy of Islamic (conventional) banks, foreign shareholders positively impacted the dividend payouts of CBs, and finally, concentration of ownership negatively influenced the dividend policy of both IBs and CBs. The COVID-19 crisis negatively moderated the relationship between foreign ownership and the dividend payouts of Islamic banks and positively impacted the dividend payouts of CBs. Finally, we examined the moderating effect of national culture on the dividend payout, using uncertainty avoidance as the cultural dimension. It appeared that uncertainty avoidance reinforces the negative influence of the concentration of capital, state ownership, and foreign ownership on the dividend payout policy of banks in the MENA region and Asia.

Our research has important implications for both managers and policymakers. Managers could take further steps to improve the efficiency of dividend policy in different contexts. Policymakers may find important findings for formulating policies related to the banking system. In a future research avenue, we could use other relevant variables to explore the relationship between dividend policy and ownership structure.

Our study had some limitations. Firstly, the data we analyzed were limited to the MENA region and Asia. As a result, data from other countries where Islamic banks operate could produce different results and conclusions. Secondly, we used uncertainty avoidance as a proxy for national culture rather than the various components of Hofstede’s cultural framework at the national or individual levels (Hofstede et al. 2010; Heydari et al. 2021).

Thus, by acknowledging the limitations of this study, we can outline multiple lines of future research in this area. Firstly, future research should look into the effects of female directors on bank dividends and include ownership characteristics in dividend smoothing assumptions. Secondly, studies should focus on optimal dividend approaches for banks with broad income distributions and the interaction of behavioral and cultural influences. Thirdly, future research should look at how investors react to bank dividend announcements and the interconnections between dividend policies and competitive dynamics during high Economic Policy Uncertainty (EPU). Finally, future research should examine bank dividend policies in emerging nations to explain the decline in dividend inclination among banks during their financial life cycle (see a recent review by Ed-Dafali et al. (2023)). As a result, the question of whether Islamic banks have more or less optimal dividend policies than their conventional counterparts merits additional examination. Further studies could examine integrating more potential explanatory variables and conducting cross-country regressions with data from different regions to offer a more comprehensive analysis.