The determinants of S·uk uk issuance in GCC countries

Purpose – This paper focuses on S uk uk issuance determinants in Gulf Cooperation Council (GCC) countries. Given the dual characteristic of debt and equity ofS uk uk aswell as their unique benefits of social responsibility, the author questions whether the theories of capital structure, the trade-off and the pecking order are able to well explain the S uk uk issuance. Design/methodology/approach – First, the author verifies these theories using capital structure determinants and regresses the S uk uk change on these determinants. Second, the author tests the trade-off theory with the target debt model and third, verifies the pecking order theory using the fund flow deficit model. Findings – The empirical results show that capital structure determinants fail to explain both theories. The author confirms that the S uk uk change is significatively linked to the deviation from a S uk uk target. So, issuing firms balance the marginal costs of S uk uk and their benefits of religiosity and social responsibility toward a target debt. The author finds no evidence of the pecking order theory. Research limitations/implications – This study contributes to corporate finance theory and corporate social responsibility. It verifies if capital structure theories proved in conventional financing can well explain Islamic bonds issuance given their social responsibility benefits. Practical implications –Managers and investors would pay attention to the social factors explaining S uk uk issuance in their finance and investment decisions. They would be enhanced to use this financing tool knowing its social unique benefits. This also should encourage governments to enhance this socially responsible financing. Rating agencies would be motivated to evaluate S uk uk and firms would improve the quality and relevance of disclosure to get the best rating. Social implications – The author highlights the social factors explaining S uk uk issuance and enhances corporate social responsibility (CSR). Originality/value –The author extends the few literature testing capital structure theories for Islamic bonds and highlights the specific social responsible features of S uk uk that would bridge their issuance to capital structure theories. So the author enhances the concept of Islamic CSR. Tying capital structure theories to CSR would also help developing Islamic finance theory as a unique social responsible framework.


Introduction
Financing decision involves decision on the composition between debt and equity and the decision on type of financial securities to be issued. Many studies on corporate finance have dealt with debt-equity choice and associated shareholders' wealth effect. Researchers have focused on the determinants of bond issuance. They have proved theories of capital structure, mainly the pecking order theory and the trade-off theory. The trade-off theory predicts that there is an optimal debt ratio maximizing the value of a firm. This optimal leverage is determined by a trade-off between the marginal costs and benefits. In contrast, the pecking order theory suggests instead a pecking order of financing choice generated by the problem of information asymmetry (Myers and Majluf, 1984;Rajan and Zingales, 1995; 2.1 Capital structure theories Many corporate finance studies have pointed out that trade-off theory and pecking order theory are major determinants of conventional bond issuance. The trade-off theory has contradicted the theorem of Modigliani and Miller (1958) that postulated no leverage impact on firm's value. On the contrary, the trade-off theory predicts that there is an optimal debt to equity ratio maximizing the value of a firm. This optimal leverage is determined by a trade-off between the marginal costs and benefits (Kraus and Litzenberger, 1973;Myers, 2001;Van Binsbergen et al., 2011). In contrast, the pecking order theory does not predict a target debt ratio. It suggests instead a pecking order of financing choice generated by the problem of information asymmetry. The information asymmetry concerns the bigger knowledge of shareholders/managers about the value of the firm assets and future growth prospect. To overcome this problem, shareholders/managers prefer internal financing to external financing. Besides, in case of external financing, they opt for debt prior to equity to reduce information cost (Myers and Majluf, 1984;Rajan and Zingales, 1995;Al-Sakran, 2001;Kayo and Kimura, 2011;Psillaki and Daskalakis, 2009;Vasiliou et al., 2009).
A first part of these researches in this field has tested these theories using determinants related to capital structure, which are mainly profitability, growth opportunities tangibility, non-debt tax shields, volatility and size. Another part assumes that firms target a particular leverage induced by a trade-off between the securities costs and benefits. A third part uses the funds flow deficit model to assume that in case of deficits, the firm will only issue or retire equity as a last resort. The major part of these researches deals with conventional bonds, while very few authors focus on Islamic bonds.

Researches using capital structure determinants
Authors in this field have observed the relation between debt and capital structure determinants relating to profitability, growth opportunities, tangibility, non-debt tax shields, volatility and size.
2.2.1 Profitability. Concerning profitability, the trade-off model argues that profitable firms are less likely to be subject to bankruptcy risk because of their increased ability to meet debt repayment obligations. Thus, they will demand more debt to maximize their tax shield at more attractive costs of debt. The pecking order theory predicts the opposite sign suggesting that high profitable firms will be able to generate more funds through retained earnings and then have less leverage. Compared with debt and equity, retained earnings have no adverse selection problem, and hence, they are the cheapest source of finance (Myers and Majluf, 1984;Rajan and Zingales, 1995;Al-Sakran, 2001;Kayo and Kimura, 2011;Psillaki and Daskalakis, 2009;Vasiliou et al., 2009).
2.2.2 Asset tangibility. The trade-off theory predicts that the risk of lending to firms with more tangible assets is expected to be low, given the higher liquidation value of these assets in the event of financial distress or bankruptcy. Therefore, a firm with a higher percentage of fixed assets is expected to borrow more as compared relatively to firms with smaller fixed asset. Thus, we expect a positive relationship between tangibility of assets and debt (Harris and Raviv, 1991;Rajan and Zingales, 1995;Hovakimian and Li, 2011). In contrast, the pecking order theory predicts that firms with few tangible assets are more sensitive to informational asymmetries. Thus, these firms will issue debt rather than equity when they need external financing, which leads to negative relation between asset tangibility and debt (Titman and Wessels, 1988).
2.2.3 Firm size. Under a trade-off framework, larger firms have higher debt capacity and can borrow at more favorable risk-adjusted interest rates than smaller firms. Also, they are more diversified and less susceptible to bankruptcy (Titman and Wessels, 1988). Therefore, we expect a positive relationship between size and debt (Harris and Raviv, 1991; S Á uk uk issuance determinants in GCC countries Rajan and Zingales, 1995;Shyam-Sunder and Myers, 1999). However, according to the pecking order theory, larger firms are more closely observed by the investment community and thus less subject to information asymmetry than small firms (Rajan and Zingales, 1995).Thus, they should be more capable of issuing equity, which is more sensitive to information asymmetry and have lower debt (Rajan and Zingales, 1995). We suggest a negative relation between firm size and leverage.
2.2.4 Growth opportunities. According to the trade-off theory, low-growth firms should use debt because it has a disciplinary role to alleviate the free cash flow problem (Jensen, 1986;Stulz, 1990). Hence, we expect a negative relationship between debt and growth opportunities. Pecking order theory predicts that growth opportunities should be financed with equity instead of debt. In order to mitigate moral hazard, a negative relationship is expected between debt and growth opportunities (Smith and Watts, 1992). However other authors claim that internal funds may be insufficient for highly growing firms, which will tend to issue debt, thus leading to a positive correlation between debt and growth opportunities (Myers, 1977;Titman and Wessels, 1988).
2.2.5 Non-debt tax shield. In the trade-off scheme, firms consider non-debt tax shields, such as depreciation and investment tax credit deductions, as a substitute for the tax shield and will have less incentive to increase leverage for tax considerations. So, non-debt tax shields and debt should have a negative relationship (Titman and Wessels, 1988;Fama and French, 2002;Flannery and Rangan, 2006). On the other hand, pecking order theory does not offer any judgments on the relationship between debt and non-debt tax shield.
2.2.6 Volatility. In the context of volatility, the trade-off theory assumes that firms with high earnings volatility try to accumulate cash during good years to avoid under-investment problems in the future (Myers, 1977). As DeAngelo and Masulis (1980) point out, an adverse selection problem is more severe to firms with highly volatile earnings. To avoid adverse selection problem, firms with financial surpluses should retire debt or invest in cash or marketable securities, to preserve their debt capacity for future financing needs or to avoid issuing equities at higher costs (Myers, 1984). Higher volatility of earnings increases the probability of financial risk and these firms will face the difficulties in debt financing. According to Jensen (1986), the pecking order theory also suggests the negative relationship between leverage and earnings volatility.

Researches using target leverage model
Authors of these papers assume that firms target a particular leverage. If the actual ratio differs from the target, the firm would adjust its debt or equity to achieve the target. Researchers in this field have regressed the long-term debts change scaled by the total asset on the deviation of the debt ratio from its target value (Bradley et al., 1984;Long and Malitz, 1985;Rajan and Zingales, 1995;Titman and Wessels, 1988;Taggart, 1977;Marsh, 1982;Auerbach and King, 1983;Jalilvand and Harris, 1984;Opler and Titman, 1994;Graham and Harvey, 2001;Marsh, 1982, Hovakimian et al., 2001Ozkan, 2001;Fama and French, 2002;Flannery and Rangan, 2006;Lemmon et al., 2008;Huang and Ritter, 2009).

Researches using fund flow deficit model
Researchers in this field regress the firm's net debt issues on its net financing deficit. The financing deficit is defined using the cash flow identity, as the growth in assets less the growth in current liabilities (except the current portion of long-term debt) less the growths in retained earnings. According to this identity, this deficit must be filled by the net sale of new securities. Except for firms at or near their debt capacity, the pecking order predicts that the deficits will be filled entirely with new debt issues. Authors in this field find that the estimated coefficient on the deficit variable is close to one and interpret this result as evidence IES 28,1 supporting the pecking order theory because a shortfall in funds is first met by debt (Shyam-Sunder and Myers, 1999;Frank and Goyal, 2003).

Capital structure theories for Islamic bonds
To state capital structure theories for s Á uk uk, we begin by analyzing their specific features 2.5.1 Hybrid nature of s Á uk uk. The word s Á uk uk is the plural of Arabic word s Á akk which has the literal meaning of legal instrument/certificate, deed or cheque. The Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI, 2017) defines s Á uk uk as follows: "S Á uk uk are certificates of equal value representing undivided shares in ownership of tangible assets, usufruct and services or (in the ownership of) the assets of particular projects or special investment activity." In other words, s Á uk uk provide ownership of a part of the underlying asset to the holders. These certificates are rewarded with a pre-agreed profitsharing rate and thus avoiding any interest-based transaction.
So, s Á uk uk combine characteristics of conventional bonds and stocks. Like bonds, they have a face value, a maturity date, a remuneration rate and provide a regular stream of cash flows to investors including capital refunding with a margin. However, unlike bonds, the return on the s Á uk uk is generated from an underlying asset, not from the obligation to pay interest. Thus, they share some common features with capital-like instruments as they give the right of a stream of revenue from an investment project (Miller et al., 2007;Nathif and Thomas, 2004;Klein and Weil, 2016;Wilson, 2008). This hybrid nature is influenced by the s Á uk uk types. In fact, some s Á uk uk are more debt-like s Á uk uk as Mur abah Á ah s Á uk uk and Ij arah s Á uk uk, while Mush arakah s Á uk uk and Mud Á arabah s Á uk uk are more partnership-like s Á uk uk. 2.5.2 Benefits of s Á uk uk: religiosity and corporate social responsibility. S Á uk uk are desirable by Shar ıʿah-conscious investors and entrepreneurs for their religious content. Hence, s Á uk uk offer unique benefit of strong adherence to Islamic financial directives. Shafron (2019) and Paltrinieri et al. (2019) explain the effect of this religious benefit on the choice of s Á uk uk investment using the theory of "investor tastes" of Fama and French (2007). According to this theory, "investor tastes" are persistent in nature and exist when certain investors "get direct utility from their holdings of some assets, above and beyond the utility from general consumption that the payoffs on the assets provide". Specifically, investors with a taste for Shar ıʿah-compliant investments achieve a higher utility from investing in Shar ıʿah-compliant investments even with lower expected cash flows, than they would if they had instead held non-Shar ıʿah-compliant investments with higher expected cash flows. The investors' tastes of religiosity should encourage firms to meet these needs by issuing s Á uk uk. Moreover, the Islamic entrepreneurs may themselves have a taste for Shar ıʿah-compliant financing and thus a higher utility from issuing s Á uk uk, than they would if they issue instead non-Shar ıʿahcompliant securities. So, religiosity influences the behavior of stock market investors and issuers. These unique benefits should encourage firms to choose to issue s Á uk uk. Besides, s Á uk uk presents another unique benefit of Islamic CSR. CSR is the recognition on the part of management of an obligation to the society it serves not only for maximum economic performance but for humane and constructive social policies as well (Heald, 1957). The most notable theory inherent to this concept is Freeman's (1984) stakeholder theory. This theory assumes that sharing values with stakeholders is necessarily and explicitly a part of doing business (Freeman, Wicks and Parmar, 2004). Islamic finance implements a variation of the conventional CSR, the Islamic CSR. Indeed, it is based on the ethical principles embodied in the Shar ıʿah (Islamic legal and ethical system), where its underlying objective are generally aimed at realizing overall human wellbeing and social justice (Ullah and Jamali, 2010). One of the most important ethical principles is the ban of interest. Thus, investors of s Á uk uk are paid dividends on the outcome of profit-sharing agreements between issuers and investors instead of fixed interest installment payments as in normal bonds (Siddiqi, 1987). Therefore, s Á uk uk S Á uk uk issuance determinants in GCC countries integrate social concerns building justice between the money holder and the entrepreneur. On the one hand, the s Á uk uk holder is not unfairly assured of a positive return without doing any work or sharing in the risk, while the entrepreneur, in spite of his management and hard work, will bear all the risk to provide guaranteed return to the capital provider. On the other hand, Islamic finance presents schemes of risk management and insurance of s Á uk uk respectively by special purpose vehicle (SPV) and Takaful. The SPV maintains the underlying asset to ensure the returns stream while Takaful alleviates the risk of asset loss. So, Islamic bonds further social benefits beyond financial interest with requests for collective welfare.
Empirically, authors focusing on ethical activities prove the beneficial effect of social responsibility on the raise of the corporate value through an increase in additional equity investment from the external investors. They find significant relation between CSR indicators and measures of financial performance such us ROA, ROE, market-to-book ratio, Tobin's q and cost of equity (Jensen et al., 2002;Heinkel et al., 2001;Graff Zivin and Small, 2005;El Ghoul et al., 2011;Lee and Faff, 2009;Eccles et al., 2013;Dixon-Fowler et al., 2013;Marti et al., 2013;Hu, 2019). Other authors deal with the unique benefits of religiosity and Islamic CSR. Shafron (2019) shows that investors with Islamic beliefs tend to invest more in s Á uk uk than they would without such beliefs. Specifically in Ramadan, Klein et al. (2017) and Bialkowski et al. (2012) assert that religiosity can influence the investor behavior and find that investors react more positively to the s Á uk uk issuance than conventional bonds. Similarly, some authors suggest positive market reaction to s Á uk uk issuance (Nagano, 2010;Mohamed et al., 2017). Although other authors find evidence of negative market reaction to s Á uk uk issuance (Ahmed et al., 2018;Ameer and Othman, 2010;Modirzadehbami and Mansourfar, 2011;Godlewski et al., 2013;Hasib et al., 2017). Mohamed et al. (2017) relate these latter findings to the longer time taken by investors to absorb the information from the s Á uk uk announcement. Indeed, they prove a significantly positive reaction 30 days after the announcement of s Á uk uk issuance. 2.5.3 Capital structure determinants of s Á uk uk. The hybrid nature of s Á uk uk and their unique benefits of Islamic CSR and religiosity address the issue of whether capital structure theories can explain s Á uk uk issuance. Some authors argue that the profit-sharing type of this financing tool depends on greater internal information of the issuers when investors would like to receive maximum dividends. Therefore, the information cost of s Á uk uk issuance is predicted to be between normal debt finance and equity issuance. Thus, the choice of s Á uk uk is accordingly subordinated to normal debt finance but prior to equity issuance according to pecking order theory (Nagano, 2010;Nagano, 2016;Azmat et al., 2014). However, other authors reject the pecking order theory and claim that firms choose to issue s Á uk uk independently of the internal funding and the information cost. So, according to the trade-off theory, the firm opts for a target ratio of s Á uk uk to assets to maximize its value. This optimal leverage is determined by a trade-off between the marginal costs and benefits of the s Á uk uk (Shahida and Saharah, 2013;Mohamed et al., 2015). As s Á uk uk benefits are inherent to religiosity and social responsibility, the trade-off prediction supposes that issuing firms balance the costs and the benefits of religiosity and social responsibility benefits of s Á uk uks.

Researches dealing with determinants of Islamic bonds issuance
Very few studies focus on the determinants of s Á uk uk issuance. Islamic corporate finance research has investigated whether capital structure theories explain s Á uk uk issuance. Some authors have performed logit and probit models to analyze the determinants of s Á uk uk and conventional bonds issuance. They have tested if these determinants relate to pecking order, trade-off or timing theory. One of these researchers, Nagano (2010) finds evidence that Malaysian firms choose to issue s Á uk uk prior to bank borrowing and other external IES 28,1 financing tools. He shows that s Á uk uk issuance does not relate to the issuer's internal funds or to the information cost, but that Islamic bank borrowing always does. The author explains the results by the fact that firms issue s Á uk uk to obtain other benefits no matter how large the information cost is. He shows that firms obtain an increase in the corporate value by issuing s Á uk uk, which must be due to its ethical benefits. He concludes that s Á uk uk issuance is preferentially chosen as a funding scheme because it brings unique financial and ethical benefits.
Nagano (2016) did not find any evidence of the pecking order theory in a comparative study concerning Malaysia, Saudi Arabia and the United Arab Emirates. His findings show that the possible determinants of s Á uk uk are firm size and past s Á uk uk issuance. The insignificant relationship with other variables also indicates that s Á uk uk is considered to be chosen prior to the normal bond issuance regardless of the availability of firms' internal funds. In another study concerning Malaysia and Indonesia, Nagano (2017) proves that the pecking order theory explains s Á uk uk issuance decision in case of large funding demand. Indeed, he finds that, under high information asymmetry, a firm with a high stock price and a large demanding fund prefers s Á uk uk issuance to conventional debt. Focusing on specific s Á uk uk type, Azmat et al. (2014) performed probit model on utility function to test Malaysian issuers' choice of Islamic bonds. They show that Islamic joint venture bonds do not align with debt-equity target, while secured against real estate these s Á uk uk do not always represent ownership of the underlying asset. Shahida and Saharah (2013) use OLS, fixed effect and random effect models to prove that s Á uk uk issuance depends on firm size, past s Á uk uk issuance experiences and finally the government tax incentive. These findings are consistent with trade-off theory; however, leverage and profitability remain insignificant for s Á uk uk issuance decisions. Hanifa et al. (2014) perform the partial adjustment model to find the firm specific determinants of target debt ratio. Using s Á uk uk and conventional bond issuance dataset for the period 2000 to 2012, the results of the dynamic panel data estimators provide strong support for trade-off theory. However, when the authors took consideration of bond and s Á uk uk types, they show, on the one hand, that partnership-based s Á uk uk and convertible bonds follow pecking order theory. On the other hand, straight bonds and exchange-based s Á uk uk align toward a target debt.
In GCC countries, Grassa and Miniaoui (2018) use capital structure determinants and find mixed results. Aligning with the pecking order theory's predictions, they document a positive relation between growth opportunity and s Á uk uk issuance and a negative correlation between size and s Á uk uk issuance. However, concerning asset tangibility, their results support the positive sign of the trade-off theory.
These studies provide little evidence that capital structure theory can explain s Á uk uk issuance.

Methodology
In this paper, we test first the trade-off theory and the pecking order theory using capital structure determinants. Second, we apply the target debt model to verify the trade-off theory. Third, we use the fund flow deficit model to test the pecking order theory. In the current section, we present the methods of each model. To deal with the problems of heteroscedasticity and serial correlation in the residuals, we use techniques of panel estimation: fixed effects model and random effects model. We also use the instrumental variable technique to resolve the problem of lagged independent variable 3.1 The capital structure model This model aims to verify if s Á uk uk issuance is explained either by trade-off theory or by pecking order theory using capital structure determinants. We will examine if s Á uk uk amount S Á uk uk issuance determinants in GCC countries is influenced by determinants of capital structure, which are profitability, growth opportunities, tangibility, non-debt tax shields, volatility and size. 3.1.1 Hypotheses. Some authors argue that the profit-sharing type of s Á uk uk depends on internal information of the issuers when investors would like to receive maximum dividends. However, the information cost of s Á uk uk issuance is predicted to be inferior to equity issuance. Thus, the choice of s Á uk uk is prior to equity issuance according to pecking order theory. However, according to trade-off theory, the firm opts for a target ratio of s Á uk uk to assets to maximize its value. This optimal leverage is set by a trade-off between the marginal costs and benefits of the s Á uk uk (Nagano, 2010;Nagano, 2016;Shahida and Saharah, 2013;Azmat et al., 2013).
The trade-off theory anticipates a positive relation between leverage and the capital determinants tangibility, size and profitability and a negative relation with growth opportunities, non-debt tax shields and volatility. However, the pecking order theory predicts that leverage depends positively on growth opportunities and negatively on profitability, tangibility, size and volatility.
Hence, we posit the following hypotheses: H1. According to the trade-off theory, s Á uk uk issuance is positively influenced by tangibility, profitability and size and negatively related to growth opportunities, non-debt tax shields and volatility.
H2. According to the pecking order theory, s Á uk uk issuance is positively influenced by growth opportunities and negatively related to size, profitability, tangibility and volatility.

Econometric models.
To test Hypothesis 1, we regress s Á uk uk on these lagged determinants using the following model To test Hypothesis 2, we regress s Á uk uk on these lagged determinants using the following model In the two econometric models, the dependent variable is s Á uk uk, while the independent variables are capital structure determinants, notably profitability, growth opportunities, tangibility, non-debt tax shields, volatility and size.
(1) S Á uk uk S Á uk uk is measured by the ratio of the amount of s Á uk uk divided by total assets (2) Profitability This variable is measured by the ratio of earnings before interest and taxes to the total assets (Following Titman and Wessels, 1988;Rajan and Zingales, 1995).
(3) Growth opportunities Following Rajan and Zingles (1995) and Danbolt (2002, 2004), we use the ratio of market-to-book value as a proxy for growth opportunities.

IES 28,1 (4) Tangibility
We adopt the ratio of fixed assets to the total assets in line with Rajan and Zingales (1995) and Bevan and Danbolt (2004).

(5) Size
As well as Titman and Wessels (1988) and Rajan and Zingales (1995), we employ the natural logarithm of total assets as proxy for the size of the firms.
(6) Non-debt tax shields We calculate it by the ratio of annual depreciation to total assets as done in prior researches (Titman and Wessels, 1988;Ozkan, 2001).
(7) Earning volatility Following Titman and Wessels (1988), we use the standard deviation of return on assets as measure of volatility of earnings, where the return on assets for each year is measured by the ratio of earnings before interest and taxes to the total assets.

The debt target model
The debt target model aims to verify if s Á uk uk issuance is explained by the trade-off theory using the target debt prediction.
3.2.1 Hypothesis. The trade-off theory indicates that a firm aims to achieve an optimal capital structure of debt and equity that is determined by the trade-off between marginal costs and benefits. We suppose that the marginal social responsibility benefits of s Á uk uk issuances also impact a firm's capital structure. So, according to the trade-off theory, the firm opts for a target ratio of s Á uk uk to assets to maximize its value. This optimal leverage is generated by a trade-off between the marginal costs and benefits of the s Á uk uk. The prediction of this model assumes that firms target a particular leverage. If the actual ratio differs from the target, the firm would adjust its s Á uk uk to achieve the target. We will examine if the s Á uk uk change is linked to the deviation from a s Á uk uk target. Thus, we propose the subsequent hypothesis: H3. According to the trade-off theory, s Á uk uk change is significantly linked to the deviation from a s Á uk uk target. 3.2.2 Econometric model. Hypothesis 3 will be verified using the partial adjustment model of debt (Gaud et al., 2005;Drobetz and Wanzenried, 2006;Flannery and Rangan, 2006). This model is specified as follows: Where Debt * it : the target debt level for firm i at time t. We replace debt by S Á uk uk and set the following model Where S Á uk uk * it : the target debt level for firm i at time t. S Á uk uk * it ¼ βX it þ ε i ; X it : Vector of explanatory variables, identified by the capital structure theories. So, we use the same explanatory variables of the precedent models, profitability, growth opportunities, tangibility, non-debt tax shields, volatility and size.
This model measures the change in debt between two periods. The first term on the right side of the equation is the speed of adjustment, γ; the speed by which firms adjust toward their S Á uk uk issuance determinants in GCC countries target s Á uk uk ratio from their s Á uk uk ratio in the previous period. To deal with the endogeneity problem, we use as instrument the sector of activity.

Variable measurement.
The dependent variable is change in s Á uk uk, while the independent variable is the difference between target s Á uk uk and s Á uk uk.
(1) Change in s Á uk uk It is the difference of s Á uk uk in two successive periods. However, for many cases the amount of s Á uk uk of the year before issuance is zero. This would create problems in the measurement of this variable. Therefore, we scaled s Á uk uk by total assets. (2) The difference between target s Á uk uk/assets and s Á uk uk/assets. The target s Á uk uk ratio is measured as: s Á uk uk=Assets Ã ¼ bX it , where X it is a vector of the capital structure determinants used.

The fund flow deficit model
This model tests the pecking order theory using the funds flow deficit model.
3.3.1 Hypothesis. This model predicts that the firm will only issue or retire equity as a last resort. It fills its deficit by using only debt. Therefore, the coefficient of the regression of debt change on funds flow deficit would be close to one (Shyam-Sunder and Myers, 1999;Frank and Goyal, 2003). We predict that this model can be applied for s Á uk uk. As s Á uk uk have hybrid nature of debt and equity, they can be appropriated to fill firms deficit, thus letting equity issuing as a last resort. We set the Hypothesis 4.
H4. The coefficient of the regression of debt change on funds flow deficit is close to one 3.3.2 Econometric model. We test Hypothesis 4 using the following model: Where DEF is the funds flow deficit 3.3.3 Variables measurement. The independent variable is the s Á uk uk change, while the dependent variable is the funds flow deficit.
(1) Change in s Á uk uk This variable should be calculated as the variation of the amount of s Á uk uk scaled by total of assets.
(2) The funds flow deficit DEF The funds flow deficit DEF which is measured as follows: Where, DIV t : dividend payments; X t : capital expenditures; ΔW t : net increase in working capital; R t : current portion of long-term debt at start of period; C t : operating cash flows, after interest and taxes. D it : is the amount of debt issued or retired.

Statistical tools
We use techniques of panel estimation: fixed effects model and random effects model to deal with the problems of endogeneity of lagged dependent variable, heteroscedasticity and serial correlation in the residuals. The random effects model can be viewed as a regression model with a random constant term. This model assumes independence between the error term and the explanatory variables. However, the fixed effect model is a regression model with a fixed IES 28,1 constant term. This model assumes correlation between the error term and the explanatory variables and uses deviations from individual averages to eliminate persistent differences between firms. The Hausman test allows us to choose the appropriate model for the sample. For each regression we perform the two methods. Then, we perform the Hausman test to choose the appropriate model for the sample. We also use the Wooldridge autocorrelation to verify if there is a need for autoregressive panels. To deal with the endogeneity problem, we use the instrumental variable technique. We perform these estimations using STATA tool.

Sample and data
The sample of the study includes s Á uk uk issuing firms of GCC countries with available requested data. Thus, the retained countries are KSA, UAE, Oman and Qatar. Three firms were excluded for non-available data. We observe 19 issuing firms from 2004 to 2016. These firms are included only at the year of issuance, so we obtain an unbalanced sample of 36 observations. As demonstrated by Arellano (2003), the results provided by unbalanced panels are as reliable as those based on balanced panels. Furthermore, we believe the sample size is suitable according to Austin and Steyerberg (2015) who proved that the number of subjects per variable required in linear regression analyses for adequate estimation of regression coefficients, standard errors and confidence intervals is only two. So, the minimum required sample size in our case would be 12. Moreover, we reviewed all studies dealing with minimum sample size for panel data using fixed effects and random effects models. There are no studies determining the minimum individual-level sample size. However, researches examining the group-level sample size show that predictors at either level are unbiased with 30 clusters and remain unbiased with as few as 15 clusters (Baldwin, S.A and Fellingham, 2013;Bell et al., 2014;Maas, C andHox, 2004, 2005). By analogy, as we have 19 firms, we believe our results are unbiased. As we include 84% of the firms of the population of issuing firms, then our sample is representative.
We collect data from DATASTREAM. Table 1 presents the list of the issuing firms and the types of s Á uk uk, when available.

Descriptive statistics
We compute in Table 2 the descriptive statistics of the three models of study: the capital structure determinants, the target level and the fund flow financing.
Concerning the capital structure model, we remark that the standard deviation of the part of s Á uk uk in assets, the economic profitability, the ratio of depreciation to assets and the return to equity are inferior to 0.1. This indicates that issuing amount, profitability, nondebt tax shields and earning volatility are relatively heterogeneous among the sample. However, the standard deviation of (ln assets) and the market-to-book ratio are superior to 1. Indeed, the market-to-book ratio varies from 0.230 to 15.730 and (ln assets) varies from 8.020 to 19.577. This indicates that the firms of the sample have different size and growth opportunities.
Regarding the target leverage model, Table 2 shows that change in s Á uk uk varies from À 0.327 to 0.327, with an average of 0.057 and a standard deviation of 0.090. This indicates that the sign of this variable is not the same for all the firms, but it doesn't have a big variation. Concerning the difference between s Á uk uk and the target s Á uk uk, it varies from 0.856 to 2.349, with an average of 1.868 and a standard deviation of 0.789. This indicates that this change has the same sign in the sample and it has a big variation.
The descriptive statistics of the dependent and independent variables of the fund flow deficit model show that s Á uk uk varies from 0 to 0.327, with an average of 0.058 and a standard deviation of 0.071. This indicates that s Á uk uk does not have a big variation in the sample.

S Á uk uk issuance determinants in GCC countries
Concerning the fund flow deficit, it varies from À0.408 to 75,661, with an average of 12,809 and a standard deviation of 0.789. This indicates that this variable does not have the same sign in the sample and it has a big variation, which would have an important effect on the sign of its coefficient.

Estimation results
We present first the estimation results relative to capital structure determinants model. Then we report the results of the target leverage model and finally we show the results of the fund flow deficit model.  Table 3 verifies if s Á uk uk issuance is explained by capital structure determinants. Specification 1 presents the results relative to the trade-off theory while specification 2 reports the results relative to the pecking order theory.
The results of specification 1 show that Wooldridge autocorrelation test rejects serial collinearity, so there is no need to perform autoregressive panel. Besides, the Hausman test recommends using fixed effect models. Two variables are significant to the 5% level;  Table 3. Determinants of s Á uk uk issuance Model 1: S Á uk uk Ã ¼ α iþ β i X it−1 þ ε i S Á uk uk issuance determinants in GCC countries profitability measured by the ratio (EBIT/assets), earnings volatility proxied by earnings standard deviation and size measured by (ln assets). However, the signs of these variables do not confirm the trade-off theory. In fact, the sign of profitability is negative, which is contrary to the trade-off theory. Besides, the sign of volatility is positive, which is contrary to the predicted sign. Hypothesis 1 is rejected, s Á uk uk issuance is not positively influenced by tangibility, profitability and size and negatively related to growth opportunities, non-debt tax shields and volatility.
Specification 2 shows the results concerning the capital structure determinants of the pecking order theory. Wooldridge autocorrelation test rejects serial collinearity, so there is no need to perform autoregressive panel. Besides, the Hausman test recommends using fixed effect models. The same variables of specification 2 are significant to the 5%; profitability measured by the ratio (EBIT/assets) and earnings volatility proxied by earnings standard deviation and size measured by (ln assets). Also, the signs of these variables do not all confirm the pecking order theory. In fact, the sign of profitability is negative, which is confirming to the pecking order theory. Nevertheless, the signs of volatility is positive, which is contrary to the predicted sign. Hypothesis 2 is rejected, and s Á uk uk issuance is not positively influenced by growth opportunities and negatively related to size, tangibility and volatility.

The results of the leverage target model
We present the results in Table 4. Wooldridge autocorrelation test rejects serial collinearity, so there is no need to perform autoregressive panel. Besides, the Hausman test is significant, thus recommending fixed effect model. The variable (S Á uk uk* it -S Á uk uk it-1 ) is significant to the 1% level. Hypothesis 3 is confirmed, s Á uk uk change is significantly linked to the deviation from a s Á uk uk target. So, the amount of s Á uk uk converges to a target level following a trade-off between marginal costs and benefits of s Á uk uk. Therefore, we find evidence of trade-off theory.

The results of the funds flow deficit model
These results are reported in Table 5 We notice in Table 5 that the Hausman test is not significant, which recommends random effect model. Wooldridge autocorrelation test rejects serial collinearity, so there is no need to perform autoregressive panel. The coefficient of the variable (funds flow deficit) is positive but not close to one. It is also not significant. Hypothesis 4 is rejected, the coefficient of the regression of debt change on funds flow deficit is not close to one. So, funds deficit is not filled by using only debt. Thus, the pecking order theory is rejected.

Robustness check
We check the robustness of our results in many ways. First, we use techniques of panel estimation: fixed effects model and random effects model to deal with the problems of endogeneity of lagged dependent variable, heteroscedasticity and serial correlation in the residuals. For each regression we perform the two methods. Then, we perform the Hausman test to choose the appropriate model for the sample. We also use the Wooldridge autocorrelation to verify if there is a need for autoregressive panels.
Second, we test capital structure theories using three models to corroborate our findings. The model of capital structure determinants tests if s Á uk uk change depends on capital structure determinants. The debt target model verifies if the firm targets a ratio of s Á uk uk to assets determined by a trade-off between the marginal costs and benefits of the s Á uk uk. The fund flow deficit model predicts that the firm will fill its deficit by using only debt only letting issuing equity as a last resort. We deal with the endogeneity problem using the instrumental variable technique. The capital structure model fails to prove either the trade-off theory or the pecking order theory. The debt target model proves the trade-off theory, while the funds flows deficit model rejects the pecking order theory. Thus, our findings are robust.
However, we do not use alternative measures of the significant independent variables. Indeed, the variables measures are chosen according to the review of previous researches, which used specific measures.

Discussion of results
We test if capital structure theories can explain s Á uk uk issuance using three models; the capital structure determinants model, the debt target model and the fund flows deficit model. The model of capital structure determinants fails to confirm either trade-off theory or pecking order theory. We find that some of the coefficients of the variables measuring these determinants present signs conform to the predicted signs while other coefficient have signs contrary to the predicted signs. These mixed results are in line with those of Grassa and Miniaoui (2018). In fact, the authors document a positive relation between growth opportunity and s Á uk uk issuance and a negative correlation between size and s Á uk uk issuance, which confirm the pecking order theory's predictions. However, they report a positive sign of asset tangibility, which verify the trade-off theory while leverage and profitability remain insignificant for s Á uk uk issuance decisions. Our results conform also those of Nagano (2010) and Nagano (2016) that s Á uk uk issuance is not related to the issuer's internal funds or the information cost and that size is a possible determinant of s Á uk uk. Though, our findings differ from the one of Nagano (2016) that, under high information asymmetry, a firm with a high stock price and a large demanding fund prefers s Á uk uk issuance to equity, thus proving the pecking order theory. Our findings are also different from those of Shahida and Saharah (2013)   To deal with our mixed results, we perform the leverage target model and then the fund flow deficit model. The results of the target s Á uk uk model show that the amount of s Á uk uk converges to a target level confirming the trade-off theory. This evidence implies that s Á uk uk is a desirable financing tool and the firm aims to have a mixed financial structure of equity and s Á uk uk. We explain this attraction by the unique benefits of s Á uk uk. In fact, the hybrid nature of s Á uk uk and its interest-free scheme of outcomes made them an Islamic CSR way to rise funds. This suggestion aligns the findings of authors focusing on ethical activities, which have proven that social responsible activities not only improve the consumer's credibility, but also increase corporate value through an increase in additional equity investment from the external investors (Jensen et al., 2002;Heinkel et al., 2001 andGraff Zivin andSmall, 2005;El Ghoul et al., 2011;Lee and Faff, 2009;Eccles et al., 2013;Dixon-Fowler et al., 2013;Marti et al., 2013;Hu, 2019). Nagano (2010) and Mohamed et al. (2017) have also suggested that s Á uk uk brings unique benefits by increasing issuer's stock returns.
However, the results of the fund flows deficit model reject that funds deficit is filled by using only debt. So, pecking order theory is rejected. These findings are contrary to those of Shyam-Sunder and Myers (1999) and Frank and Goyal (2003) that fund deficit is filled by the net sale of new conventional debt securities.
As our results confirm the trade-off theory and reject the pecking order theory, we assume that firms do not choose to issue s Á uk uk because of asymmetric information or its cost, but for their social unique benefits that other external financing don't afford.

Conclusion
In this paper, we verify if s Á uk uk issuance is explained by theories of capital structure. We extend the literature testing these theories for Islamic bonds. Previous research failed to find evidence of any capital structure theory outlining s Á uk uk issuance in GCC countries. Our study further tests these theories and adds theoretical and empirical contributions. Theoretically, we highlight the specific features of s Á uk uk that would bridge their issuance to capital structure theories. S Á uk uk are couched in the ethical principles embodied in the Shar ıʿah (Islamic legal and ethical system). The underlying objectives of Shar ıʿah are generally aimed at realizing overall human wellbeing and social justice. Indeed, s Á uk uk conform to the principle of no interest and risk sharing. This principle promotes social justice between s Á uk uk holders and issuing firms. Indeed, investors are not allowed to realize financial gains without being exposed to the risk of potential loss. So, in case of profits, they are paid dividends on the outcome of profit-sharing agreements. Therefore, Islamic bonds (s Á uk uk) have hybrid nature between debt and equity. Besides, they offer unique benefits of religiosity and socially responsible financing. This hybrid nature as well as the social and religious benefits are the specific features linking s Á uk uk issuance to capital structure theories. Methodologically, this study adds empirical evidence by using three models, in contrast to previous studies dealing with only one model. Using the model of capital structure determinants, our results show that s Á uk uk issuance is negatively and significantly linked to profitability. This sign confirms the pecking order theory. Nevertheless, s Á uk uk issuance is positively and significantly linked to earnings volatility, which is contrary to both trade-off theory and pecking order theory. Thus, the model of capital structure determinants does not permit to confirm or reject capital theories. So, we used the debt target model to test the tradeoff theory and the fund flow deficit model to test the pecking order theory. Our results show that s Á uk uk converge to a target level determined by a trade-off between the cost and the social responsibility benefits of s Á uk uk. These findings are consistent with the trade-off theory. In addition, the test of the fund flow deficit shows that funds deficit is not filled by using only IES 28,1 debt, thus rejecting the pecking order theory. Overall, we find evidence of the trade-off theory. We suggest that firms aim to have a target level of s Á uk uk in their financial structure due to their unique benefits of religiosity and Islamic CSR.
Our findings present a number of implications for theory and practice. From the theoretical side, this paper contributes to the corporate finance theory and CSR. It highlights the important contribution of corporate Islamic finance to the development of CSR. Indeed, Islamic finance is embedded in ethical and social principles. One important principle is the ban on interest in financing and its replacement by profit-and-loss-sharing. The adoption of this principle in s Á uk uk induces unique social benefits with claims of social justice between s Á uk uk holders and issuing firms. These unique benefits, that other financing schemes do not give, link capital structure theories to CSR. Our research enhances the concept of Islamic CSR. Tying the capital structure theories to CSR would also help developing Islamic finance theory as a unique socially responsible framework. The socially responsible aspect is obvious as the unfair features such as the interest and risk bearing are replaced by the ethical principles of no interest and risk sharing. Therefore, the core of Islamic finance theory is to tailor conventional finance to socially responsible aims.
The main practical implications relate to the actors intervening in the financing process. One important outcome is to encourage managers and investors to further contribute to promote this Islamic financing tool for its unique social and Shar ıʿahcompliance benefits. Our results would encourage governments to enhance firms to adopt this socially responsible financing. Moreover, it would motivate them to issue sovereign s Á uk uk, which constitutes a pricing benchmark and an anchor security for portfolio management and secondary trading. Furthermore, rating agencies would be motivated to evaluate s Á uk uk and ascertain the quality of issuance and subsequently attract more investors. To get the best rating, firms would improve the quality of disclosure and the relevance of their accounting information. This would reasonably lead to socially responsible financing development.