Do board characteristics influence Islamic banks’ capital structure decisions? Empirical evidence from a developing country

Abstract

This study investigates the relationship between board characteristics (Sharia boards and boards of directors) and capital structure decisions of Islamic commercial banks (ICBs). The sample consisted of 14 ICBs in Indonesia operated during 2007–2021. The data were analyzed using the random effect model and the feasible generalized least square model. The results reveal that the size and independence of the board of directors and Sharia board expertise have a positive impact on the Indonesian ICBs’ debt-to-total asset ratio decision. Furthermore, board gender diversity encourages ICBs in Indonesia to adopt a lower debt-to-total equity ratio (DTER) while the size and expertise of Sharia boards encourage ICBs to pursue higher DTER. This study reinforces the agency theory’s view regarding the relationship between board characteristics and corporate capital structure decisions.

PUBLIC INTEREST STATEMENT

Corporate financial management aims to maximize shareholders’ wealth. Furthermore, making optimal capital structure decisions is one of the functions of a financial manager, in addition to investment decisions and dividend payout policies, in order to achieve company goals. Besides being related to maximizing shareholders’ wealth, the capital structure decisions of Islamic banks also indicate the banks’ compliance with Sharia rules. Therefore, capital structure decisions in the context of Islamic banks become more important.

Keywords:

• Board characteristics
• Capital structure
• Islamic bank
• Indonesia
• Developing country

Reviewing editor:

• David McMillan
• University of Stirling
• UK

Subjects:

• Economics
• Finance
• Business, Management and Accounting

1. Introduction

Capital structure decision is a critical area that affects the overall level of company operations. Therefore, for more than six decades, it has attracted the attention of numerous studies in order to understand the determining factors that influence the choice between debt and equity which later result in maximizing company value. However, to this date, the optimal decision between these two choices has not reached any satisfactory conclusion despite many research efforts and extensive debates have been carried out. Discussions regarding capital structure tend to separate non-financial companies from financial ones, especially banks. This bank exclusion is due to the argument that banks have different natures of business, resulting from different regulatory frameworks (Khan et al., 2020). Also, in the banking system itself, the capital structure debate has separated the Islamic banking system (IBS) from the conventional banking system (CBS) due to the same argument. IBS and CBS have different regulatory frameworks, including regulations on debt and equity fulfillment (Guizani & Ajmi, 2021; Hoque & Liu, 2021; Khan et al., 2020).

Islamic banks (IBs), as one Islamic financial institution, do not only focus on economic issues but at the same time also spreads spiritual and social aspects when doing their business (Meutia et al., 2019; Ramdani et al., 2023). Thus, managers of Islamic banks must be subject to dual governance; they must comply with Islamic law and the control of the board of directors which intends to ensure the maximal prosperity of shareholders. In addition, to assure IB’s compliance with Islamic rules, an additional board that acts as Islamic corporate governance has been developed, namely the Sharia Board (SB). The Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI) defines SB as an independent body consisting of Fiqh Muamalat scholars who are responsible for ensuring that Islamic financial institutions, including IBs, comply with Sharia principles. SB plays an important role as an internal control mechanism that has the responsibility to oversee IB’s activities and issue fatwas (Ajili & Bouri, 2018; Buallay, 2019; Nomran et al., 2018). IB’s non-compliance with Sharia principles will result in a serious impact on the trust of customers, investors, depositors and all stakeholders (Baklouti, 2020).

Compliance with dual governance may complicate Islamic bank managers in the decision-making process. IB managers may want to invest in profitable but high-risk projects, including regarding the sources of financing, to maximize shareholders’ wealth. However, this investment project is possibly not in accordance with Sharia principles since they prohibit Gharar (excessive risk-taking). Therefore, SB’s duty to abolish those transactions that seem to violate Sharia principles can conflict with the interests of managers and the board of directors (Mollah & Zaman, 2015). In the context of capital structure decisions, Sharia principles limit the use of debt in the capital structure of Islamic financial institutions, especially IBs. According to these principles, Islamic banks may use interest-based capital up to a certain percentage of their total assets. Meanwhile, the agency costs and free cash flow hypothesis (Jensen, 1999) postulates that debt is a governance tool that is able to reduce the conflict of interests between agents and principals. Greater debt can act as a stronger monitoring mechanism for managers so they are imposed to act in the interests of principals (Harford et al., 2008). Thus, in addition to being related to maximizing shareholders’ wealth, capital structure decisions in IBs are also a form of banks’ compliance with Sharia principles. This uniqueness has prompted numerous studies to explore capital structure decisions in this institution.

Over the past few years, scholars have compared the determinants of conventional banks’ capital structure with those of IBs (Guizani & Ajmi, 2021; Hoque & Liu, 2021; Sheikh & Qureshi, 2017). They affirmed that, compared to IBs, conventional banks have greater debt, profitability, size and operating expenses. Other researchers focused on examining the determinants of IBs’ capital structure (Al-Hunnayan 2020; Bukair, 2019; Smaoui et al., 2020). They reported that profitability has no correlation with IBs’ capital structure (Bukair, 2019; Smaoui et al., 2020) while Al-Hunnayan (2020) claimed that there is a negative relationship between these two. Bank size has a positive linkage with IBs’ capital structure (Al-Hunnayan 2020; Bukair, 2019) while Smaoui et al. (2020) reported the opposite result, that the two variables build a negative correlation. Next, bank growth has a positive relationship with IBs’ capital structure (Al-Hunnayan, 2020; Bukair, 2019) while Smaoui et al. (2020) did not find any relation between the two. Considering the previous studies, it appears that the literature regarding the capital structure of IBs is still limited. In particular, those discussing the relationship between board characteristics and banks’ capital structure decisions, especially IBs, are still scarce to find or perhaps have not been done previously. The literature on corporate governance, however, supports the notion that board characteristics are one of the main determinants that influence the capital structure decisions of non-financial companies. Therefore, a study on the relationship between board characteristics and capital structure decisions of IBs is imperative to conduct.

This present study investigates the relationship between board characteristics and the capital structure decisions of Islamic commercial banks (ICBs) in Indonesia. In addition to the board of directors, this study integrates the characteristics of the Sharia Board as a predictor of capital structure decisions. It is because bank regulators (Bapepam LK Number: KEP-208/BL/2012) stipulate that Islamic banks in Indonesia are prohibited from having more than 45 percent of interest-based total debt ratio to total assets. Bapepam LK (The Capital Market and Financial Institution Supervisory Agency) is in charge of guiding, regulating and supervising capital market activities. It also formulates and implements policies as well as technical standardization in the field of financial institutions in accordance with the Minister of Finance regulations and applicable laws. Therefore, the capital structure decisions of ICBs in Indonesia are a form of banks’ compliance with Islamic rules. This study is expected to fill the gaps in the existing literature regarding the relationship between corporate governance and capital structure decisions. It also aims to provide additional insight and understanding for investors, management, depositors and regulators of Islamic banks in Indonesia.

The rest of this paper is arranged as follows. Section 2 explains a review of theoretical and empirical literature as a basic argument to develop hypotheses. Section 3 describes the methodology. Section 4 explains the results of the study as well as its discussion. Finally, section 5 presents conclusions and implications.

2. Literature review

2.1. Theoretical background

Agency theory assumes that managers, as agents, tend to act in their personal interests, which are not always in line with the principals’ interests (Jensen & Meckling, 1976). They favor making investment decisions on low-risk projects in order to ease their work, even when these investments have a negative NPV (Amihud & Lev, 1981). This agency problem results in additional costs incurred by the company to assure that managers work for the needs of principals. The monitoring of managers is realized by introducing the layers of supervision boards that are responsible for monitoring and controlling managers (Fama, 1980; Fama & Jensen, 1983). Furthermore, Daily et al. (2003) affirmed that the supervisory role of the board is attached to certain characteristics that are related to the independence and effectiveness of the board, such as board size, board diversity and board expertise.

Based on agency theory, Jensen (1999) then proposed the agency costs of free cash flow hypothesis, which postulates that debt can be used as a governance tool to reduce the conflict of interests between agents and principals. It serves as a compensatory mechanism to reduce the agency costs of free cash flows caused by managers by forcing them to distribute it to shareholders (Jensen, 1999; Kochhar, 1996). Similarly, Harford et al. (2008) asserted that greater debt can function as a stronger monitoring mechanism for managers, which obligates them to act in the best interests of the principals.

2.1.2. Agency problems in Islamic banks

The profit-sharing principle of Islamic banks brings more complicated agency problems (Mohammed & Muhammed, 2017; Obid & Naysary, 2014). This is due to the fact that issues in Islamic banks are not only limited to the separation of control and ownership but also the separation of control over cash flow and banks’ compliance with Sharia rules. Thus, agency problems in Islamic banks are more complex than those in conventional banks (Farag et al., 2018; Obid & Naysary, 2014; Safieddine, 2009). The Sharia board (SB) functions as a monitoring mechanism to lessen agency problems in Islamic banks (Obid & Naysary, 2014). Its provision of accurate and transparent information regarding Islamic banks’ compliance with Sharia rules can reduce information asymmetry and increase banks’ credibility as well as investors’ trust. Fikri et al. (2017) also affirmed that SB is an effective tool for reducing agency costs (internal costs caused by agency conflict between owners and managers) which result in extra monitoring and bonding expenses. SB reviews all decisions of the board of directors to ensure the implementation of a good fatwa in order to protect investors’ rights (Nathan & Ribière, 2007). Therefore, Islamic banks employ SB to monitor managers’ behavior in order to reduce agency problems and costs (Obid & Naysary, 2014).

2.2. Hypothesis development

2.2.1. Board size and capital structure

Agency theory argues that the relationship between board size and capital structure decisions can be either positive or negative. On the one hand, a bigger board is often associated with more serious agency problems due to the increased conflicts of interest between the company’s owners and managers (Jensen, 1999). More debts are needed to reduce these agency problems. Besides, creditors tend to believe that companies with larger boards have more effective supervision which may reduce the costs of debt capital (Bokpin & Arko, 2009; Jensen, 1999; Rehman et al., 2010). In this case, board size is positively related to company capital structure decisions, and this was supported by a number of recent studies, including (Agyei & Owusu, 2014; Feng et al., 2020; Sheikh, 2019; Zaid et al., 2020).

On the other hand, the theory assumes that a larger board faces more communication and coordination problems which can reduce board effectiveness and efficiency (Jensen & Meckling, 1976). It is difficult for large board members to reach a consensus on decision-making. Consequently, the board cannot effectively deal with CEOs who tend to avoid risk by pursuing lower debt ratios. In addition, management in companies with larger boards faces greater supervisory pressure, which forces them to reduce debt ratios in order to decrease the pressure to improve company performance (Abor & Biekpe, 2006; Berger et al., 1997). In this case, board size is negatively related to company capital structure decisions. This view was supported by a number of more recent studies (Agyei & Owusu, 2014; Hussainey & Aljifri, 2012; Javaid et al., 2021; Sewpersadh, 2019).

Studies on the relationship between board size and capital structure decisions in the context of Islamic banks are scarce to find. However, referring to the existing literature, it is expected that Indonesian ICBs with larger boards will pursue higher debt ratios. It is because Islamic banks face more complicated and more intense agency problems than conventional ones (Mohammed & Muhammed, 2017; Obid & Naysary, 2014). Therefore, the proposed hypothesis is:

H1: Board size has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia.

2.2.2. Board independence

From the perspective of agency theory, the relationship between board independence and company capital structure varies. On the one hand, the theory argues that an independent board can increase its independence and effectiveness in monitoring executive managers. However, tighter monitoring by this independent board causes executive managers to take lower debt ratios in order to avoid the higher risks due to greater debt use (Lin et al., 1998). Higher debts potentially increase the threat of takeover and bankruptcy, and these risks put managers under pressure to use a smaller debt ratio (Grossman & Hart, 1982). In this context, agency theory, supported by Dimitropoulos (2014), predicts a negative relationship between board independence and capital structure decisions.

On the other hand, agency theory believes that a board with more independent members plays a significant role in monitoring executive managers. Independent boards tend to be aligned with the interests of shareholders so they will encourage greater use of debt in the company’s capital structure (Abor & Biekpe, 2006; Jensen, 1999). The higher the independence of the board, the higher the debt ratio because independent boards can help the company gains recognition and trust from creditors, thereby reducing uncertainty and increasing the company’s funding capabilities (Pfeffer & Salancik, 2003). In this context, agency theory predicts a positive relationship between board independence and capital structure decisions, and this was supported by the majority of studies, including Zaid et al. (2020), Boateng et al. (2017), Tarus and Ayabei (2016) and Ali et al. (2014).

Literature on the relationship between board independence and the capital structure decisions of banks, especially Islamic banks, seems limited. However, referring to the existing literature, it is predicted that Indonesian ICBs with a higher proportion of independent boards adopt a higher debt ratio. It is because Islamic banks face more complicated and more intense agency problems than conventional ones (Mohammed & Muhammed, 2017; Obid & Naysary, 2014). Therefore, the proposed hypothesis is:

H2: Board independence has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia.

2.2.3. Board gender diversity

According to agency theory, boards with more diversity will face less information asymmetry. The decrease in information asymmetry is, of which, caused by the presence of women on the board, who contribute to increasing access to external financial resources and improving the proportion of external equity in the company’s capital structure. Thus, this theory believes that more women on the board are associated with lower capital structure decisions (Adusei & Obeng, 2019; Alves et al., 2015; Hernandez-Nicolas et al., 2015). More recently, García and Herrero (2021) affirmed that female boards of directors, as one of the board characteristics, give the most influence on company capital structure decisions. This characteristic is negatively related to leverage, cost of debt and debt maturity for a sample of companies in Europe.

Studies on the relationship between board gender diversity and capital structure decisions in the context of Islamic banks seem to have never been conducted. However, referring to the existing literature, the hypothesis proposed is:

H3: Board gender diversity has a negative relationship with the capital structure decisions of Islamic commercial banks in Indonesia.

2.2.4. Sharia board size

A bigger Sharia board (SB) gives stronger pressure on Islamic bank managers to always comply with Sharia rules, including those regarding the maximum limit on the use of debt ratios. However, the question is whether SB with a larger size can encourage Islamic bank managers to maximize the debt ratio or vice versa. It seems that no studies have examined this issue.

Agency theory (Jensen & Meckling, 1976) assumes that a larger board tends to face more communication and coordination problems which can reduce its effectiveness and efficiency. As a result, the board cannot work optimally in dealing with managers who tend to avoid risk by pursuing a lower debt ratio. In other words, SB with a larger size is often associated with lower debt ratios. In contrast, the agency costs of free cash flow hypothesis (Jensen, 1999) believes that a larger board faces more serious agency problems, especially free cash flows, which cause companies to require more debt financing in order to reduce the problems. Islamic banks, in this case, face more complex and intense agency problems than conventional ones (Farag et al., 2018; Obid & Naysary, 2014), which may cause them to require a bigger debt ratio. Referring to the existing literature, it is predicted that Indonesian ICBs with larger SB will impose managers to use greater debt financing in order to reduce agency problems. Therefore, the proposed hypothesis is:

H4: Sharia board size has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia.

2.2.5. Sharia board expertise

SB expertise refers to those SB members who comprehend business disciplines (economics, finance and accounting) in addition to mastering Fiqh-al-Muamalat law (Rahma & Bukair, 2015). It indicates that besides calling for compliance with Sharia principles, SB expertise is also capable of providing advice on business strategies, including appropriate capital structure decision-making strategies which are essential for improving bank performance. The question is whether SB expertise encourages managers of Islamic banks to maximize debt ratio or vice versa, and it seems that there has been no study to examine this issue.

The agency relationship is defined as a contract between a company’s principals and agents. Agency theory views the potential for conflict of interest arising from the contract since the two parties have different interests (Jensen & Meckling, 1976). Therefore, the theory believes that debt is an effective governance tool to reduce this conflict of interest (Jensen, 1999), which in turn can improve company performance. According to Obid and Naysary (2014) and Mohammed and Muhammed (2017), agency problems in Islamic banks are more complicated and more intense than in conventional banks. In this case, apart from being responsible for monitoring Islamic banks’ compliance with Sharia principles (Nathan & Ribière, 2007), SB also functions to supervise managers’ behaviors in order to reduce agency costs (Obid & Naysary, 2014; Quttainah et al., 2013). Referring to the existing literature, it is predicted that greater SB expertise will further encourage ICB managers in Indonesia to maximize debt ratios in order to reduce agency costs, which in turn can improve bank performance. Therefore, the proposed hypothesis is:

H5: SB expertise has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia.

3. Methodology

3.1. Sample

This study employed a sample of 14 Islamic commercial banks (ICBs) in Indonesia operated during 2007–2021, which established unbalanced panel data with a total of 163 bank-year observations (Table 1). Banks’ data were obtained from the annual reports of each ICB while the data regarding annual real GDP growth and inflation rate were downloaded from the Indonesian Central Bureau of Statistics (BPS) website.

Table 1. Sample distribution for the year period.

Years	ICBs that were operating	ICBs that published annual reports	ICBs as sample
2007	6	2	2
2008	7	3	3
2009	9	5	5
2010	12	11	11
2011	12	11	11
2012	12	11	11
2013	13	13	13
2014	13	13	13
2015	13	13	13
2016	14	14	14
2017	14	14	14
2018	14	14	14
2019	14	14	14
2020	14	14	14
2021	12	12	11
Total	179	164	163

Source: OJK (2023).

3.2. Variables

3.2.1. Dependent variable

This study considered the main capital structure measured by the total debt to total asset ratio (DTAR) as the first dependent variable. This capital structure measure, a maximum of 45%, is a debt ratio regulated and limited by the Indonesian Sharia regulations (Bapepam LK No: KEP-208/BL/2012). This capital structure measure has been used by other previous studies (Bukair, 2019; García & Herrero, 2021; Neves et al., 2020; Wassie, 2020). This study also considered another measure of capital structure, namely the total debt to total equity ratio (DTER) to further strengthen and enrich the discussion. Several previous studies have used DTER as a measure of capital structure, including Wassie (2020); Moradi and Paulet (2019); Chakrabarti and Chakrabarti (2019); and Hussainey and Aljifri (2012).

3.2.2. Independent variables

The independent variables consist of board size, as used by Zaid et al. (2020) and Sheikh (2019); board independence, as used by Feng et al. (2020) and Ali et al. (2014); board gender diversity, as used by García and Herrero (2021) and Usman et al. (2019); and sharia supervisory board size as well as expertise, as used by Baklouti (2020), Nomran et al. (2018) and Nugraheni and Khasanah (2019).

3.2.3. Control variables

This study employed bank-specific factors and macroeconomic factors as the control variables. Bank-specific factors consist of bank profitability and bank growth as suggested by Al-Hunnayan (2020) and Sheikh and Qureshi (2017). Meanwhile, macroeconomic factors consist of annual real GDP growth and inflation rate as used by Le and Phan (2017), Neves et al. (2020) and Smaoui et al. (2020). Table 2 presents the names, notations and measurements of the variables.

Table 2. Variables, acronyms and measurements.

Variables	Acronyms	Measurements
Debt to asset ratio	DTAR	The ratio of total liabilities to total asset
Debt to equity ratio	DTER	The ratio of total liabilities to total equity
Board size	B_SIZE	Total number of board members
Board independence	B_IND	The ratio of the number of independent members to the total members of the board
Board gender diversity	B_GEN	The ratio of the number of female members to the total members of the board
Sharia board size	SB_SIZE	Total number of Sharia Board (SB) members
Sharia board expertise	SB_EXP	The ratio of the number of SB members with economic, finance and accounting expertise to the total members of the SB
Profitability	PROF	The ratio of profit after taxes to total asset
Bank growth	GROWTH	Total asset(t) – total asset(t−1)/total asset(t−1)
GDP growth	GDP	Annual growth of real GDP
Inflation	INFLA	Annual inflation rate

3.3. Model specification

This study used a panel data regression model which provides three alternatives; Common Effect Model, Fixed Effect Model and Random Effect Model. The authors then conducted the Chow test, Hausman test and Lagrange Multipliers (LM) test to determine the most suitable model among the three. The specification of the panel data regression model is: (1) $${\text{DTAR}}_{\text{it}}={\beta }_0+{\beta }_1\mathrm{B}\_{\text{SIZE}}_{\text{it}}+{\beta }_2\mathrm{B}\_{\text{IND}}_{\text{it}}+{\beta }_3\mathrm{B}\_{\text{GEN}}_{\text{it}}+{\beta }_4\text{SB}\_{\text{SIZE}}_{\text{it}}+{\beta }_5\text{SB}\_{\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}}_{\text{it}}+{\beta }_6{\text{PROF}}_{\text{it}}+{\beta }_7{\text{GROWTH}}_{\text{it}}+{\beta }_8{\text{LGDP}}_{\text{it}}+{\beta }_9{\text{INFLA}}_{\text{it}}+{\epsilon }_{\text{1it}}$$(1) (2) $${\text{DTER}}_{\text{it}}={\beta }_0+{\beta }_1\mathrm{B}\_{\text{SIZE}}_{\text{it}}+{\beta }_2\mathrm{B}\_{\text{IND}}_{\text{it}}+{\beta }_3\mathrm{B}\_{\text{GEN}}_{\text{it}}+{\beta }_4\text{SB}\_{\text{SIZE}}_{\text{it}}+{\beta }_5\text{SB}\_{\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}}_{\text{it}}+{\beta }_6{\text{PROF}}_{\text{it}}+{\beta }_7{\text{GROWTH}}_{\text{it}}+{\beta }_8{\text{LGDP}}_{\text{it}}+{\beta }_9{\text{INFLA}}_{\text{it}}+{\epsilon }_{\text{2it}}$$(2)

Description of the model (1) and (2):

• DTAR = Debt to asset ratio

• DTER = Debt to equity ratio

• B_SIZE = Board size, ie the size of the board of directors

• B_IND = Board independence, ie the independence of the board of directors

• B_GEN = Board gender diversity, ie female board of directors

• SB_SIZE = Sharia board size

• SB_EXP = Sharia board expertise

• PROF = Profitability, ie bank’s return on assets

• GROWTH = Bank growth, ie the annual growth of the book value of the bank’s total assets

• GDP = GDP growth, ie the annual growth of Indonesia’s real Gross Domestic Product

• INFLA = Inflation rate, ie Indonesian annual inflation rate

• β₀ = Constant value

• β₁–β₉ = The regression coefficients

• ɛ = The vector of the error term

• i = Indonesian ICBs selected as a sample

• t = time dimension in the year period

4. Results and discussion

This section explains the descriptive statistics, the most suitable model selected, the correlation matrix and the result of the variance inflation factor (VIF) model. The estimation result of the selected model is also presented, followed by a discussion.

4.1. Descriptive analysis

Table 3 presents the descriptive statistics of several selected variables.

Table 3. Descriptive statistics.

Variables	Mean	Std. dev	Minimum	Maximum	Skewness	Kurtosis
DTAR	0.1687	0.0741	0.0116	0.3826	0.1483	2.6343
DTER	2.0679	1.6720	0.0531	8.3839	1.1171	4.3366
B_SIZE	4.1227	1.1588	2	8	1.4309	5.7029
B_IND	0.5349	0.1634	0.3333	0.8000	0.0356	1.5261
B_GEN	0.1767	0.1921	0.0000	0.6667	0.8706	3.1199
SB_SIZE	2.3252	0.4955	1	4	0.7436	2.3661
SB_EXP	0.3451	0.2352	0.0000	0.7500	−0.4926	1.8085
PROF	1.6513	2.8828	−12.1300	12.1600	−0.6029	9.7203
GROWTH	0.0514	0.1667	−0.0394	1.1240	5.4629	13.2005
GDP	4.6198	2.1522	−2.0700	6.3500	−2.5142	8.1842
INFLA	4.3300	2.3920	1.6800	11.0600	1.0297	2.8389

Source: ICBs’ annual report, world bank and Indonesian Central Bureau of Statistics during 2007–2021.

Table 3 shows that the mean of LEV is 16.87%, indicating that on average, ICBs in Indonesia have a low total debt-to-total asset ratio. The maximum value of LEV is 38.26%, indicating that those ICBs have complied with Bapepam LK Decree Number: KEP-208/BL/2012 which stipulates that Islamic banks in Indonesia are allowed to have a maximum total debt to total asset ratio of 45%. The mean of DTER is 2.0679 or 206.79%, indicating that on average, ICBs in Indonesia have greater total debt than total equity. Bapepam LK Decree Number: KEP-208/BL/2012 does not regulate the amount of total debt to total equity (DTER) of Islamic banks in Indonesia; thus, those ICBs do not violate the regulation. The mean of B_SIZE is 4.123 with a minimum value of 2 members and a maximum of 8 members. The mean of B_IND is 53.49% with a maximum value of 80.00% and a minimum of 33.33%, indicating that ICBs have complied with Indonesian bank regulations which stipulate that the minimum ratio of independent members to total board members is 30%. The mean of B_GEN is 17.67% with a minimum value of 0.00%, meaning that ICBs do not have female members on their board of directors and the maximum of 66.67%. The mean of SB_SIZE is 2.325 with a minimum value of 1 member and a maximum of 4 members. The mean of SB_EXP is 34.51%, indicating that on average, Sharia board members of ICBs understand the disciplines of economics, finance and accounting in addition to their Fiqh Muamalat expertise. The minimum value of SB_EXP is 0.00% and the maximum is 75.00%. Due to limited space, descriptive statistics of other variables can be seen in Table 3.

4.2. The multicolinearility test

This study also conducted an analysis of the correlation among independent variables and observed the inflation variance factor to detect the multicollinearity symptoms.

4.2.1. Correlation analysis

The estimated correlations of all independent and control variables are presented in Table 4. The multicollinearity issue occurs when the correlation coefficient between variables is bigger than 0.80 (Kennedy, 2008). The result of this analysis shows that the estimated data set is free from multicollinearity symptoms since the correlation coefficient between variables is below 0.80.

Table 4. Matrix correlation of independent variables.

Variables	B_SIZE	B_IND	B_GEN	SSB_SIZE	SSB_EXP	PROF	GROW	GDP	INFLA
B_SIZE	1
B_IND	0.0490	1
B_GEN	−0.0084	−0.3352	1
SSB_SIZE	−0.0325	0.1386	−0.1419	1
SSB_EXP	−0.0735	0.3286	−0.2182	0.2169	1
PROF	0.1159	−0.2730	−0.0657	−0.1043	−0.0023	1
GROW	−0.2143	0.1902	−0.1840	0.1178	0.1748	0.0124	1
GDP	0.0213	0.0113	0.0102	−0.0278	−0.0479	−0.0709	−0.0446	1
INFLA	0.0027	−0.0031	0.0399	0.0298	0.0414	−0.0548	0.0475	0.4433	1

Source: ICBs’ annual report dan Indonesian Central Bureau of Statistics in the period 2007–2021.

4.2.2. Variance inflation factor

In addition to the correlational test, this study carried out the Variance Inflation Factor (VIF) test to detect multicollinearity issues. The result, presented in Table 5, shows that the mean of VIF for all explanatory variables is 4.13. However, it is indicated that the individual VIF value of several variables is greater than 5.00, for example, the VIF value of SSB_SIZE is 9.52, GDP is 6.36, B_IND is 6.27, and INFLA is 5.39. Hence, we refer to Al-Faryan and Alokla (2023) and Gujarati and Porter (2011), that if the individual VIF value of the explanatory variables is smaller or equal to 10.00, the estimated set data can be considered free from serious symptoms of multicollinearity.

Table 5. The results of the variance inflation factor (VIF) model.

Variable	VIF	1/VIF
SB_SIZE	9.52	0.105037
GDP	6.36	0.157110
B_IND	6.27	0.159404
INFLA	5.39	0.185499
SB_EXP	3.53	0.282941
B_GEN	2.17	0.460011
PROF	1.55	0.644261
GROWTH	1.25	0.802752
B_SIZE	1.10	0.913000
Mean VIF	4.13

Source: Self-calculation results.

4.3. Model fit test

Table 6 presents the results of the Chow test, Hausman test and Lagrange Multipliers (LM) test. These three were conducted to select the most suitable regression model among the available ones; Common Effect Model (CEM), Fixed Effect Model (FEM) and Random Effect Model (REM).

Table 6. The results of the model fit test.

Dependent variable	Probability values			Suitable model
	Chow test	Hausman test	LM test
Debt to asset ratio (DTAR)	0.0000	0.9508	0.0000	REM
Debt to equity ratio (DTER)	0.0000	0.0032	0.0000	FEM

Source: Self-calculation results.

For the DTAR variable, Table 6 shows that the Chow test results give a probability value of 0.000, indicating that FEM is more suitable than CEM. The Hausman test results show a probability value of 0.951, indicating that REM is more suitable than FEM. Finally, the LM test results show a probability value of 0.000, indicating that REM is more suitable than CEM. Therefore, the panel data regression equation model of this study was estimated using REM. On the other hand, for the DTER variable, Table 6 shows that the Chow test results give a probability value of 0.000, indicating that FEM is more suitable than CEM. Furthermore, the LM test results give a probability value of 0.000, indicating that REM is more suitable than CEM. Finally, the Hausman test results give a probability value of 0.034, indicating that FEM is more suitable than REM. Therefore, the second panel data regression equation model was estimated using FEM.

We attempted to conduct the heteroscedasticity test on Model 1 (dependent variable DTAR), but the facility of such a test for REM was not available in the application. As explained by Wassie (2020), Githaiga (2020) and Sheikh and Qureshi (2017), REM indeed does not require the heteroscedasticity test. We then examined the heteroscedasticity of Model 1 on the Pooling Least Square model with the estat hettest command (Breusch-Pagan/Cook-Weisberg test). Next, using the groupwise method or xttest3 command, we conducted the heteroscedasticity test on Model 2 (dependent variable DTER). The results are presented in Table 7.

Table 7. The results of the heteroscedasticity test.

Observed model	chi2(1)	Prob > chi2	Results interpretation
Model 1: dependent variable DTAR	1.67	0.1962	There is no heteroscedasticity
Model 2: dependent variable DTER	892.77	0.0000	There is heteroscedasticity

Source: Authors’ calculation results.

Table 7 reveals that Model 1 does not show any heteroscedasticity symptoms, or in other words, REM does not need the heteroscedasticity test. Hence, we used the results of REM for Model 1. However, Model 2 shows heteroscedasticity symptoms. Thus, we employed the Feasible Generalized Least Square (FGLS) method to estimate Model 2, as suggested by a number of previous studies (Al-Malkawi & Pillai, 2018; Ghaleb et al., 2020; Owar & Appiah, 2022; Samet & Jarboui, 2017; Tadele et al., 2021).

4.4. Robustness check of results

Checking the robustness of the model focused on our main model, namely the regression model with the dependent variable Debt to asset ratio (DTAR). In this context, we use the System Generalized Method of Moment (SYS-GMM) one-step model which was developed by Arellano and Bover (1995) and has been believed to overcome serial correlation and potential endogeneity problems. Next, a summary of the fitness test results of the SYS-GMM model is presented in Table 8.

Table 8. The summary of GMM model fitness test results.

Lagged dependent variable	Sargan test		Arellano-Bond test		Coefficient
	Chi2	Prob > chi2	z	Prob > z	FEM	GMM	PLS
DTARit−1	122.928	0.068	−0.767	0.4428	0.399***	0.425***	0.479***

Source: Processed ICBs Indonesia’s annual report for the 2007–2021 period. Notes: ***statistically significant at the 1% level.

From Table 8, it is known that the SYS GMM model used looks fit. This is shown by the results of the instrument validity test (Sargan test) and the estimator consistency test (Arellano-Bond test) which give a value of Prob > chi2 = 0.068 and Prob > z = 0.4428, both of these figures are greater than 0.05. The SYS GMM model used also passes the unbiased test. This is indicated by the coefficient value of the influence of DTAR_it−1 on DTAR_it from the SYS-GMM model (0.425) which is greater than the FEM results (0.399) and smaller than the PLS model results (0.479).

Table 9 contains a summary of the REM results from model Equation (1)(1) $${\text{DTAR}}_{\text{it}}={\beta }_0+{\beta }_1\mathrm{B}\_{\text{SIZE}}_{\text{it}}+{\beta }_2\mathrm{B}\_{\text{IND}}_{\text{it}}+{\beta }_3\mathrm{B}\_{\text{GEN}}_{\text{it}}+{\beta }_4\text{SB}\_{\text{SIZE}}_{\text{it}}+{\beta }_5\text{SB}\_{\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}}_{\text{it}}+{\beta }_6{\text{PROF}}_{\text{it}}+{\beta }_7{\text{GROWTH}}_{\text{it}}+{\beta }_8{\text{LGDP}}_{\text{it}}+{\beta }_9{\text{INFLA}}_{\text{it}}+{\epsilon }_{\text{1it}}$$(1) which are compared with the results of the SYS-GMM model to check its robustness. Table 9 also contains a summary of the Feasible Generalized Least Square (FGLS) results from model Equation (2)(2) $${\text{DTER}}_{\text{it}}={\beta }_0+{\beta }_1\mathrm{B}\_{\text{SIZE}}_{\text{it}}+{\beta }_2\mathrm{B}\_{\text{IND}}_{\text{it}}+{\beta }_3\mathrm{B}\_{\text{GEN}}_{\text{it}}+{\beta }_4\text{SB}\_{\text{SIZE}}_{\text{it}}+{\beta }_5\text{SB}\_{\hspace{0.17em}\mathrm{\text{EXP}}\hspace{0.17em}}_{\text{it}}+{\beta }_6{\text{PROF}}_{\text{it}}+{\beta }_7{\text{GROWTH}}_{\text{it}}+{\beta }_8{\text{LGDP}}_{\text{it}}+{\beta }_9{\text{INFLA}}_{\text{it}}+{\epsilon }_{\text{2it}}$$(2) .

Table 9. Summary of REM, SYS-GMM and FGLS estimation results.

Variable/ model	REM			SYS-GMM to check robustness			FGLS
Dependent variable	Debt to asset ratio (DTAR)			Debt to asset ratio (DTARit)			Debt to equty ratio (DTER)
Independent variable	Coefficient	Std. error	z	Coefficient	Std. Error	z	Coefficient	Std. Error	z
DTARit−1	–	–	–	0.4252***	0.0747	5.70	–	–	–
B_SIZE	0.0101***	0.0031	3.24	0.0106***	0.0027	3.91	−0.0676	0.0721	−0.94
B_IND	0.2073***	0.0346	5.98	0.1262***	0.0350	3.61	−0.0998	0.8459	−0.12
B_GEN	−0.0110	0.0243	−0.41	0.0001	0.0249	0.00	−2.9629***	0.6051	−4.90
SB_SIZE	0.0399	0.0436	0.92	0.0059	0.0391	0.15	2.7401***	1.0347	2.65
SB_EXP	0.1232***	0.0222	5.56	0.1123***	0.0209	5.37	1.2328**	0.5971	2.06
PROF	−0.0239	0.0016	−1.52	−0.0052***	0.0015	−3.46	0.0287	0.0401	0.72
GROWTH	0.0607***	0.0229	2.66	0.0965***	0.0234	4.13	0.3503	0.5342	0.66
GDP	0.0231	0.0018	1.31	0.0031**	0.0015	2.06	−0.0450	0.0408	−1.10
INFLA	0.0109	0.0016	0.69	−0.0023	0.0015	−1.53	0.0675**	0.0368	1.93
Constant	0.1594***	0.0228	6.99	0.0917***	0.0234	3.92	0.7315	0.6261	1.17
Adj R-squared	60.93%			–			–
Wald chi2	133.08			137.04			307.89
Probability	0.0000			0.0000			0.0000
Banks	14			14			14
Observations	163			163			163

Note: ***statistically significant at 1%; **statistically significant at 5%.

The REM and SYS-GMM columns in Table 9 show that the influence of the main variables from this study, namely, B_SIZE, B_IND, B_GEN, SB_SIZE and SB_EXP on DTAR gives the same results, both in sign and significance. Thus, the results of the model robustness checks performed provide strong evidence that our findings are robust.

The estimation results of REM are fit, with the Adjusted R-squared of 60.93%, the Wald chi2 value of 133.08 and a probability value of 0.000. Meanwhile, the results of the FGLS model show the Wald chi2 value of 307.89 with a probability value of 0.000, indicating that the estimation results of this model are also fit.

4.5. Discussion

This study investigates the relationship between board characteristics (the board of directors and Sharia board) and the capital structure decisions of Islamic commercial banks (ICBs) in Indonesia. The capital structure decisions are measured using Debt to Asset Ratio (DTAR) and Debt to Equity Ratio (DTER). This study employs two estimation models, namely Random Effect Model (REM) and Feasible Generalized Least Square (FGLS). Furthermore, by controlling bank-specific factors (profitability and growth) and macroeconomic factors (GDP and inflation), the discussion is as follows.

First, the result shows that board size has a significant positive relationship with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTAR. It implies that ICBs with a larger board of directors are able to encourage bank managers to make capital structure decisions that lead to the maximum use of the debt ratio allowed by Sharia rules (45%). Larger boards understand that agency problems in Islamic banks are more complex and more intense than those in conventional banks, but these problems can be overcome by adopting a higher debt ratio. This result is in accordance with agency theory which postulates that boards with larger sizes face more serious agency problems, causing companies to require more debt in order to solve the problems (Jensen, 1999). The result showing a positive relationship between board size and capital structure decisions is supported by a number of studies on non-banking institutions (Agyei & Owusu, 2014; Feng et al., 2020; Sheikh, 2019; Zaid et al., 2020). Furthermore, the result of this study reveals that board size does not have any correlation with the capital structure decisions of the Indonesian ICBs when the capital structure decisions are measured using DTER. This result differs from the prediction of agency theory (Jensen, 1999; Jensen & Meckling, 1976). Hence, this study accepts H1: Board size has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia when the capital structure decisions are measured using DTAR.

Second, the result shows that board independence has a significant positive relationship with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTAR. It indicates that bigger board independence is able to encourage bank managers to make capital structure decisions that lead to the maximum use of the debt ratio allowed by Sharia rules (45%). The board understands that agency problems in Islamic banks are more complex and more intense than those in conventional banks, but they can be overcome by adopting a higher debt ratio. This result is in line with agency theory which argues that more independent members on the board of directors strengthen the company’s ability to protect itself from bankruptcy. It is because an independent board can increase transparency which allows companies to attract greater debt capital to finance their investments (Chen & Hsu, 2009). The positive relationship between board independence and capital structure decisions has also been reported by several studies on non-banking companies (Ali et al., 2014; Tarus & Ayabei, 2016; Zaid et al., 2020). Furthermore, the estimation result shows that board independence does not have any significant linkage with the capital structure decisions of ICBs when the capital structure decisions are measured using DTER. This is not in accordance with the prediction of agency theory (Jensen, 1999; Jensen & Meckling, 1976). Therefore, this study accepts H2: Board independence has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia when the capital structure decisions are measured using DTAR.

Third, the result shows that board gender diversity has no correlation with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTAR. This result is supported by Detthamrong et al. (2017) who found that female boards are not related to the capital structure of companies in Thailand and Balqis et al. (2022) in Indonesia. However, it differs from García and Herrero (2021) who reported that board gender diversity has a negative relationship with the capital structure decisions of non-banking companies in Europe. This result implies that the presence of women on ICBs’ board of directors is often associated with risk avoidance of the bank’s core business, namely giving financing in order to avoid higher non-performing financing (NPF). They pay little attention to capital structure decisions whose maximum limit has been set. The result also reveals that board gender diversity has a significant negative relationship with the capital structure decisions of ICBs when the capital structure decisions are measured using DTER. This is in line with agency theory which postulates that the existence of women on the boards is associated with lower capital structure decisions (Alves et al., 2015; Hernandez-Nicolas et al., 2015). This theory also argues that the female board of directors is able to reduce information asymmetry, so the company does not need to pursue higher debt ratios. Hence, this study accepts H3: Board gender diversity has a negative relationship with the capital structure decisions of Islamic commercial banks in Indonesia when the capital structure decisions are measured using DTER.

Fourth, the result shows that Sharia board size has no correlation with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTAR. This result is not in accordance with agency theory (Jensen & Meckling, 1976) and the agency costs of free cash flow hypothesis (Jensen, 1999) which argue that a larger board brings more serious agency problems and requires greater debt financing to overcome them. In addition, the result reveals that Sharia board size has a significant positive relationship with the capital structure decisions of ICBs when the capital structure decisions are measured using DTER. This finding is supported by agency theory (Jensen & Meckling, 1976) and the agency costs of free cash flow hypothesis (Jensen, 1999) which predict that board size has a positive correlation with debt ratios. Thus, this study accepts H4: Sharia board size has a positive relationship with the capital structure decisions of Islamic commercial banks in Indonesia when the capital structure decisions are measured using DTER.

Fifth, the result shows that Sharia board expertise has a significant positive relationship with the capital structure decisions (DTAR and DTER) of ICBs in Indonesia. This result implies that in addition to ensuring Islamic banks’ compliance with Sharia rules, SB expertise is also capable of providing guidance on appropriate capital structure decision-making strategies to improve bank performance. SB expertise encourages ICB managers in Indonesia to make capital structure decisions that lead to the maximum use of the debt ratio allowed by Sharia rules (45%). The board understands that agency problems in Islamic banks are more complicated and more intense than those of conventional banks; thus, adopting a higher debt ratio is important in order to overcome the problems. Therefore, this study accepts H5: Sharia board expertise has a positive relationship with the capital structure decisions (DTAR and DTER) of Islamic commercial banks in Indonesia.

Finally, the result shows that the control variable ‘profitability’ has a negative but statistically insignificant relationship with the capital structure decisions (DTAR) of ICBs in Indonesia. It differs from Al-Hunnayan (2020) who argued that profitability has a negative and significant impact on the leverage of Islamic banks in the Gulf Cooperation Council (GCC) region. Besides, the result shows that the control variable ‘bank growth’ has a significant positive relationship with the capital structure decisions (DTAR) of ICBs in Indonesia. It is supported by Al-Hunnayan (2020) who affirmed that growth has a positive impact on Islamic bank leverage. Next, the analysis of macroeconomic factors shows that GDP growth and inflation do not have any significant relationship with the capital structure decisions of ICBs in Indonesia. This result is not in accordance with Hanousek and Shamshur (2011) who reported that GDP growth has a positive impact on the leverage of companies listed in the database constructed by Bureau Van Dijk. Finally, the result reports that inflation has a significant positive correlation with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTER, but it is not significant when the capital structure decisions are measured using DTAR. This finding is supported by Guizani and Ajmi (2021) who affirmed that inflation has no correlation with the DTAR of Islamic banks in Malaysia.

The results and discussion justify that the characteristics of boards of directors and Sharia boards play a very essential role in the capital structure decisions of Islamic commercial banks (ICBs) in Indonesia. Meanwhile, agency theory believes that the component of debt in the capital structure is important since it is able to reduce agency problems. In fact, agency problems in Islamic banks are far more complicated and more intense than those in conventional banks. These complex agency problems may in turn reduce the performance of Islamic banks. Therefore, this study proposes some suggestions for practitioners, including the government as the regulator, Islamic banks’ investors, management and depositors to control the performance of Islamic banks through capital structure decisions. First, the government as the regulator is expected to make more effective regulations by setting the minimum, in addition to the maximum, debt ratio for Islamic banks. These regulations may prevent banks from avoiding high risks by pursuing lower debt ratios (since they believe in high risk high return, low risk low return). Second, investors should encourage the boards to aim for a higher debt ratio by maximizing the allowable upper limit of this ratio, ie 45% for Islamic banks in Indonesia. They are also expected to bring more board members, board independence and Sharia board expertise in the Islamic banks’ governance, including ICBs in Indonesia, since these three components of corporate governance are able to encourage managers to adopt higher debt to total assets (DTAR) in capital structure. In this case, agency theory believes that debt is one of the governance tools to reduce conflicts of interest between agents and principals, which in turn can improve bank performance. Third, management should realize that they are working in the best interest of shareholders. Islamic bank managers should acknowledge that they face more complex and more intense agency problems than those of conventional banks. Thus, they need to aim for a higher debt ratio to reduce those problems and increase company performance. However, the managers still need to be cautious since a high debt ratio is associated with high financial distress. Fourth, Islamic bank depositors are suggested to choose banks with better performance to ensure the safety of their deposits and the certainty of their expected returns. Existing literature believes and has proven that companies with high debt ratios imply better performance, both in the current and in the future. Therefore, depositors need to pay attention to the debt ratio in the capital structure of Islamic banks, especially ICBs in Indonesia, by choosing the banks that dare to set a high debt ratio.

5. Conclusions

The results of this study reveal that board characteristics (both boards of directors and Sharia boards) play a very significant role in the capital structure decisions of ICBs in Indonesia. In particular, the results show that the size and independence of the board of directors have a significant positive relationship with the capital structure decisions of ICBs in Indonesia when the capital structure decisions are measured using DTAR. Furthermore, Sharia board expertise encourages Islamic banks’ managers to make capital structure decisions that lead to the maximum debt to total asset ratio (DTAR) allowed by Sharia regulations in Indonesia. Finally, board gender diversity encourages ICBs’ managers to adopt lower debt to total equity ratio (DTER) while the size and expertise of the Sharia board encourage managers to pursue higher DTER.

This study contributes to two important implications. First, it fills the gaps in the existing literature regarding the relationship between board characteristics and the capital structure decisions of Islamic banks, especially in the context of ICBs in Indonesia. Second, it provides additional insight for practitioners regarding the influence of the board of directors and Sharia board on the capital structure decisions of ICBs in Indonesia. Existing literature explains that the right capital structure decisions are able to improve company performance. Thus, practitioners can monitor and control ICBs’ performance through their capital structure decisions.

Despite important contributions, this research bears some limitations. The sample is limited to one type of Islamic bank in Indonesia, namely Islamic commercial banks (ICBs). Future studies are suggested to examine the relationship between board characteristics and the capital structure decisions of Islamic rural banks or conventional commercial banks in Indonesia. Board characteristics in this study are measured using board size, board gender diversity, board independence, Sharia board size and Sharia board expertise. Further studies can integrate other variables of board characteristics, such as CEO tenure, CEO age and CEO duality. Future studies are also advised to consider ownership structure and audit committees that can determine the capital structure decisions of ICBs in Indonesia. Finally, they can examine the impact of capital structure decisions on the performance of Islamic commercial banks.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Notes on contributors

Muhamad Umar Mai

Muhamad Umar Mai is an Associate Professor of Corporate Finance in the Applied Masters Program of Islamic Finance and Banking, Accounting Department, Politeknik Negeri Bandung, Bandung, Indonesia. His expertise is in the studies of corporate financial management, financial research methods, and Islamic banking financial performance.

Tjetjep Djuwarsa

Tjetjep Djuwarsa is an Lecturer of Economics and Econometrics in the Applied Study Program of Islamic Finance, Accounting Department, Politeknik Negeri Bandung, Bandung, Indonesia.

Setiawan

Setiawan is an Lecturer and Researcher in the Applied Study Program of Islamic Finance, Accounting Department, Politeknik Negeri Bandung, Bandung, Indonesia.