Corporate characteristics and leverage: evidence from Gulf countries

Abstract Purpose – This study aims to investigate the impact of corporate characteristics on leverage in the Gulf Cooperation Council (GCC) non-financial listed firms. Design/methodology/approach – A sample comprising a balanced panel for eight years from 2009– 2016 for four Gulf countries is used. In total, 85 non-financial listed companies have been selected using a nonprobability sampling technique. Corporate characteristics are represented by return on assets (ROA), return on equity, return on capital employed, market value-added, Tobin-Q, liquidity and firm size. The study used fixed and random effect models to estimate the results. Findings – The findings of the study revealed that both ROA and FSIZE have a significant negative effect on leverage. However, market value-added, return on capital employed and Tobin-Q exhibited a statistically significant positive effect on leverage. Further, the results indicated that Qatar is better than kingdom of Saudi Arabia (KSA), Oman and the UAE. In addition, evidence noted that KSA is better than both UAE and Oman in terms of the overall impact of corporate characteristics on the leverage. However, this effect is not statistically significant. Practical implications – This study provides an open insight for managers, bankers, financial analysts in the GCC countries and some other developing economies by highlighting the relationship between corporate characteristics and leverage in an emergingmarket. Originality/value – The current study provides an important insight into corporate characteristics and leverage. By so doing, it provides an attempt to identify the factors influencing corporate financing behavior taking into consideration different issues such as different proxies of firms’ profitability, market capitalization, market value added and liquidity, which provides original evidence fromGulf countries emergingmarkets. These countries are characterized by low tax rates and high liquidity. High liquidity may reduce the cost of borrowing and debt financing may not be a huge burden on firms’ profits. This makes the investigation of leverage and corporate characteristics, particularly, firms’ profitability and liquidity, very important. Therefore, the study tries to bridge an existing gap in the body of literature of capital structure and debtfinancing in Gulf countries emergingmarkets.


Introduction
Capital structure indicates to the mix of the various methods of funding sources that a firm maintains (Niu, 2008). G omez et al. (2014) states that the way a firm is financed is very relevant for investors, directors and all other stakeholders. Accordingly, the financing decisions have a direct influence on the firm value. Prior evidences from financial literature showed that a firm's capital structure could be affected by different factors. Capital structure is one of the most broadly debatable areas of research in corporate finance. A central question was so long raised by Modigliani and Miller (1958), which evolve around how a mix of debt and equity in capital structure has an impact on firm value. Further, the determinants that can have an influence on a firm's capital structure are arguable in prior financial studies. Bauer (2004) indicated that a firm's leverage is not influenced by its size. Further, some studies believe that there is a negative correlation between a firm's leverage and its profitability (Pratheepkanth, 2011;Rouf, 2015;Zeitun and Tian, 2007). While Ahmed et al. (2010) found that there is a positive link between leverage and liquidity, Sharif et al. (2012) reported a negative relationship. Ahmed et al. (2010) argue that high liquidity provides firms with greater capability to pay off its debt accordingly, such firms should opt for debt financing as a major contributor to their capital structure.
In another quest, Miguel and Pindado (2001) advocated that capital structure determinants can be explained by business characteristics and the institutional settings of countries. Jong et al. (2008) investigated capital structure in different countries and found that enterprise-level determinants that affect corporate capital structure vary among countries. However, country-level determinants have an indirect influence on capital structure because they have an impact on enterprise-level factors. Li and Islam (2019) indicated that prior studies have made numerous attempts to investigate the corporate characteristics that may have an impact on the capital structure of firms operating within a specific country. However, modern studies have been conducted based on country comparison taking in consideration country-specific determinants to explain sample firms' capital structure choice.
Theoretically, Li and Islam, (2019) indicated that capital structure theories such as the pecking order theory, the market timing theory, the trade-off theory and the agency theory suggest different capital structure determinants. Jaisinghani et al. (2017) stated that the asymmetric information theory and the tradeoff theory are the prominent theories pertaining to optimal capital structure. However, Cappa et al. (2020) argued that there are a number of competing theories that pertain to capital structure formation, which are, for example, trade-off, asymmetric information, pecking order, market timing and agency costs theories. According to Jaisinghani et al. (2017), tradeoff theory postulates that a firm determines its capital structure depending on a trade-off between the costs and benefits of debt. Miller (1977) indicated that there is a relationship between profitability and debt financing, based on the trade-off theory; firms that are more profitable would prefer debt over equity financing. In the same quest, pecking order theory suggests that a firm with higher profitability would prefer to opt for a low debt level due to the information asymmetric element involved in the pecking order theory (Hamid et al., 2015). The asymmetric information theory proposed that a firm may use its capital structure as a signaling instrument to the market (Jaisinghani et al., 2017). Also, agency theory proposed that financing decisions and capital structure mix depend on agency costs. The agency costs may arise from the conflict between shareholders and creditors accordingly, capital structure could be determined by minimizing such cost and balancing the interests of the parties involved. Further, the theory states that firms with high returns and great potential are more likely to prefer equity financing over dept financing (Li and Islam, 2019).
Majority of prior research in this field have concentrated on developed stock markets. Even though there are some studies that have been conducted in some emerging countries. For example, Bhaduri (2002) investigated the same issue in India, Chen and Strange (2005) conducted a study on some firms in China, Correa et al. (2007) provided an evidence from some Brazilian factories, Crnigoj and Mramor (2009) concentrated on firms in Slovenia, Kim et al. (2006) examined this issue in the South Korean context, Fern andez (2005) introduced evidence from Chilean corporates, Mazur (2007) sampled Polish firms, Omet (2006) focused on Jordanian firms and Salawu and Agboola (2008), Shah and Khan (2007) and Vasiliou and Daskalakis (2009) provided evidence from Nigeria, Pakistan and Greek, respectively. However, the results of these studies are greatly varied and widespread. Hence, the present study aims to investigate the corporate characteristics and leverage in some Gulf Cooperation Council (GCC) countries. The current study contributes to finance literature in several ways. First, the study provides a comprehensive investigation of some selected large firms (85 firms), which are covering all sectors of the economy except financial industry from 4 GCC countries over the years spanning from 2009 up to 2016 for 4 Gulf countries (Saudi Arabia, Oman, UAE and Qatar). Second, the study presents different statistical tools to understand the differences among the countries. Prior studies have used either the same or different statistical tools however, the differences among the GCC countries were not adequately provided. The present study used detailed descriptive statistics, and country wise effect comparison estimation. Third, the study contributes to the state of the art of capital structure, cost of finance, debt financing and finance in general especially from the context of developing countries. The study links corporate characteristics to leverage, and thus, it provides an attempt to identify the factors influencing corporate financing behavior taking into consideration, firm specific factors in the GCC. The study also provides evidence from different countries that are homogenous in their culture, economy and banking systems. Moreover, it attempts to provide evidence from oil-based and bank-based economies of the GCC countries. Several studies investigated different issues in debt financing and capital structure in GCC. However, the present study provides an investigation into firm-specific factors on leverage of GCC firms taking into consideration different issues such as different proxies of firms' profitability, market capitalization, market value added (MVA) and liquidity, which makes the investigation of the present study unique and different from prior studies.
The paper proceeds as follows. Section 2 explains the background of GCC. Section 3 discusses the literature review; this is followed by Section 4, which presents a collection of data, techniques and the study models. Section 5 provides empirical analysis and the related discussions. Section 6 comprises the conclusion, recommendations and prospective avenues that future authors can pursue in line of this research.

Background to Gulf cooperation council
The GCC region comprises fast-growing economies with government revenues fueled by sizeable oil rents, which are characterized by administrated consumer energy prices. Historically, GCC countries except Oman did not levy corporate taxes on owned domestic firms (Almutairi, 2014). The GCC financial markets are small in size and less developed by international standards and its emerging peers, they account for about 0.8% of the global financial markets (Zeitun and Saleh, 2015). In comparison of the GCC stock markets with the worldwide stock markets, they are considered young and small in size. They are lagging behind Asia and Latin America. They only account for 0.8% of the global financial markets.
Further, in 2010, they accounted for 1.3% of the global equity market capitalization. Furthermore, the GCC stock markets have some barriers such as low number of listed firms; 707 firms in 2013, restricted entry by foreigners and ownership concentration (Woertz, 2012). These characteristics have motivated the current study to investigate this issue to extend the evidence from prior studies to an emerging economy such as the GCC countries. The financial crisis negatively affected the global liquidity, which, in turn, affected banks and other financial instructions' ability to offer credit facilities. Therefore, the cost of borrowing increased tremendously as the accessibility of liquidity decreased and banks created tighter credit policies, which affected firms' performance negatively (Ellaboudy, 2010). While the majority of prior evidence on the relationship between corporate characteristics and leverage comes from capital structure determinants in both developed and developing countries, empirical research in this area did not specifically concentrate on the GCC countries (El-Khatib, 2017;Sbeiti, 2010;Twairesh, 2014). Further, the ambiguous results related to the relationship of corporate characteristics and leverage have motivated the present study to be conducted to investigate this relationship.

Literature review
The relationship between capital structure and firm performance is widely discussed and tested in prior research. However, this field of research still the most perplexing area in the strand literature of corporate finance (Brounen and Eichholtz, 2001). Different studies have been focused on the relationship between a firm's performance and capital structure (Champion, 1999;Ghosh et al., 2000;Hadlock and James, 2002;Jensen and Meckling, 1976;Modigliani and Miller, 1963;Miller, 1977;Myers, 1977;Margaritis and Psillaki, 2010;Titman and Wessels, 1988). These studies contributed significantly to the literature of corporate finance, which started early with the research of (Hirshleifer, 1958;Lintner, 1956;Modigliani and Miller, 1958).
In the GCC context, different studies have highlighted the link between corporate characteristics and capital structure (El-Khatib, 2017;Sbeiti, 2010;Twairesh, 2014). Zeitun and Saleh, (2015) investigated the impact of financial leverage on a firm's performance in the GCC countries taking in consideration the effect of the last financial crisis. The results revealed that firm's performance is significantly influenced by leverage and negatively affected by the financial crisis. In the same context, El-Khatib (2017) studied conventional leverage determinants in some publicly listed firms in many GCC countries, namely; Qatar, Saudi Arabia and the United Arab Emirates over a period from 2005 up to 2014. The study found that conventional leverage is significantly influenced by some determinants such as the tendency to pay dividends, firm's size, tangibility, profitability and age. Both firm's size and Sharīʿah principles were found to have the most significant effect but utilization of Sukuk as a financing vehicle has no significant effect. On the contrary, Sbeiti (2010) investigated the "determinants of capital structure in the context of three GCC countries and the impact of their stock markets' development on the financing choices of firms operating in these markets." The results suggested that stock markets in the sampled countries are regarded as an important option for financing decisions. Further, the results reveal that capital structure in the selected countries is not different from finance models of the developed countries.
Although there are few studies that addressed the relationship between firms' performance and financial leverage in the GCC countries, there are numerous studies that investigated this issue in other countries (Muritala, 2012;Ojo, 2012;Rehman, 2013;Zeitun and Tian, 2007). G omez et al. (2014) investigated capital structure determinants of nonfinancial companies listed in the Stock Exchange of Lima. In the scope of trade-off theory and pecking order theory and with utilization of panel data with random effect model, the results reveal that the level of firms' long-term debt is significantly influenced by their size, profitability, non-debt tax shields and collateral value of assets. Soumadi and Hayajneh (2011) investigated the "effect of capital structure on the performance of the public Jordanian firms listed in Amman stock market." The results indicated that there is a significant and negative relationship between both capital structure and firm performance and this relationship does not change in the case of firms with high or low financial leverage. In the same line, (Pratheepkanth, 2011;Rouf, 2015;San and Heng, 2011;Zeitun and Tian, 2007) found that firms' performance is negatively linked with capital structure. Salim and Yadav (2012) provided evidence on the link between 237 Malaysian firms' performance and capital structure. They advocate a negative relationship between them.
Concerning liquidity, while Ahmed et al. (2010) indicated a positive link between both leverage and liquidity. Firms with high liquidity have a capital structure mix that is characterized by debt. This observation is attributed to the ability of such firms to repay the cost of borrowing with financial ease. Sharif et al. (2012) advocated that the link between leverage and liquidity is negative.
Despite the fact, the prior studies have investigated corporate characteristics and leverage in different countries, however, studies conducted in the GCC countries are unreliable as the stock markets in these countries are not well-developed, and cannot be benchmarked with international standards. Further, the tax system in these countries have a notable impact on capital structure. Therefore, the present study attempts to explore this impact. Referencing Table 1, prior studies have investigated different issues of capital structure and debt financing; the present study is different from prior studies in several ways. For example, El-Khatib (2017) investigated tangibility, market to book ratio, profitability, size, financial deficit and age of the company with relation to equity financing. He further examined firm size, liquidity, profitability, tangibility and growth opportunities and its relation with book leverage and market leverage. None of the studies conducted in GCC have investigated corporate characteristics such as liquidity, different proxies of profitability, Tobin Q and MVA with its relation to leverage. Thus, the present study bridges an existing gap in the debt financing and capital structure. Moreover, it provides a unique and comprehensive examination of the impact of corporate characteristics on leverage in emerging markets especially GCC countries. The study also extends the evidence on capital structure, leverage and debt financing by re-visiting the impact of corporate characteristics on leverage in GCC countries.

Data and methodology 4.1 Sample selection and data collection
This study focuses on examining the associations of corporate characteristics and leverage in GCC non-financial listed firms. We use a sample comprising balanced panel data for eight years from 2009 up to 2016 for four Gulf countries, which is covering all sectors of the economy except the financial industry. In total, 85 non-financial listed companies have been selected using a non-probability sampling technique. As a result, the final sample consists of 23 companies from Saudi Arabia, 19 companies from Oman, 23 companies from the UAE and 20 companies from Qatar. The study uses a sample of large firms from Saudi Arabia, UAE, Qatar and Oman. According to KPMG (2015), these four countries have the highest total banks assets and net profits. Further, according to The World Bank (2015), the gross domestic product annual growth in Saudi Arabia, UAE, Qatar and Oman is the highest among all GCC member states.

Variables description
Corporate characteristics represented by return on assets (ROA), return on equity (ROE), return on capital employed, MVA, Tobin-Q, liquidity and firm size are considered as the independent variables and leverage is the dependent variable. The relationship between corporate characteristics and leverage in the present study is viewed and motivated by different studies (G omez et al., 2014;Kim and Berger, 2008;Rouf, 2015).

Model specification
Consistent with previous literature (Al-ahdal et al., 2020;Almaqtari et al., 2019;Rouf, 2015), we developed the following model to investigate the impact of corporate characteristics on leverage in the GCC non-financial listed firms: Where Y is the dependent variable. b 0 is the constant, b is the coefficient of the explanatory variable (corporate characteristics), fit is the explanatory variable and eit is the error term.
By adopting the economic model as in equation (1) specifically in this study, we can estimate a fixed effects model using equation (2) as follows: Where all variables are as defined in Table 2, a i (i = 1. . .. n) is the unknown intercept for each entity (n entity-specific intercepts) and u it is the error term. Table 3 provides a summary of panel descriptive statistics for the variables used by the current study using xtsum command in Stata. Xtsum generates additional descriptive statistics to the normal one. It provides an analysis for the variables into an overall, between and within descriptive statistics. The overall and within descriptive statistics are calculated over 680 (N = 680) firm-years observations. The between descriptive statistics is calculated over 85 firms (n = 85) and the average number of years a firm was observed in the variable data (T = 8). Overall, descriptive statistics reports common descriptive statistics in which the average value of a variable across all firms for the entire time period is calculated. Besides, minimum, maximum and the standard deviation values are given for the variables. While the minimum values of a variable denote the lowest value of a variable across the sample over the study time period, the maximum values are calculated as the highest values across the sampled firms during the time period.

Descriptive statistics
The average ROA of the sampled firms has a mean value of 8.57, which varies between a minimum of À65.13 and a maximum of 75.23 with a standard deviation (SD) of 9.83. This indicates that there is no high variation in ROA of the sampled companies.   (Figure 1). The value of SD is much higher than the mean value indicating that the mean does not represent the average values of the sampled firms. Finally, LIQ has an average value of 241.89 with a minimum of 0.02, a maximum of 2,1645.40 and SD of 1,377.38. Table 4 reports country wise descriptive statistics. As we have seen in Table 3, there is a variation between the standard deviation and the average values of some variables. This prompts us to conduct country-wise descriptive statistics to know the characteristics of the data of each country. Table 4 presents country-wise descriptive statistics in the form of mean, median, minimum, maximum and standard deviation for all the variables of the study. The results show the variation among the countries in LIQ, MVA and TOBINQ. All other variables have slight variation.

Unit root analysis
Unit root analysis is applied to confirm the stationarity of the variables as a prerequisite requirement for analysis of the models of the study. Three tests were conducted to check the stationarity of the variables; Levin, Lin and Chu test, Im Pesaran and Shin test and Augmented Dickey-Fuller test. Each variable was tested individually at the first level. As shown in Table 5, all variables used in the models are found to be stationary at level in all the tests applied. This leads to reject the null hypothesis of the unit root test at the level. Table 6 shows the results of Pearson correlation matrix for the variables of the present study. The results of Pearson correlation demonstrate that there is a positive and negative relationship between dependent and independent variables. While LEV has a significant Average values of leverage across the countries positive correlation at 1% level of significance with TQ, it has a significant negative correlation with all other independent variables. All independent variables have a maximum correlation coefficient of 0.58 (0.58 < 0.70), which indicates that multicollinearity issues in this study does not exist. Table 7 presents an estimation of the results of the models of the current study. An estimation of fixed effect model is opted. At the initial step of the analysis, a redundant fixed Notes: 1-LEV is leverage, ROE is return on equity, TQ is Tobin-Q, LIQ is liquidity, ROCE is return on capital employed, ROA is return on assets, FSIZE is firm size and MVA is market value added Notes: 1-LEV is leverage, ROE is return on equity, TQ is Tobin-Q, LIQ is liquidity, ROCE is return on capital employed, ROA is return on assets, FSIZE is firm size and MVA is market value added effect estimation is made to choose between the panel set or pooled regression analysis. The results of the redundant fixed effect test shows that the panel set is appropriate than the pooled set. Furthermore, the results of the redundant fixed effect shows that the estimation should be a one-way cross-section fixed effect model. In addition, an analysis is conducted to compare the relationship between fixed or random effect models. Accordingly, Husman test is conducted for this purpose and the yielded result of Husman test was in favor of fixed effect model (P-value < 0.00), which means a rejection of random effect model and the acceptance of fixed effect model. The results of regression analysis of fixed effect model shows that ROA has a significant negative effect on leverage at the level of 1% level of significance (b = À5.07, P-value = 0.000 < 0.01). This negative effect indicates that LEV is significantly influenced by ROA, this result is consistent with (Bandyopadhyay and Barua, 2016;El-Khatib, 2017;G omez et al., 2014;Güner, 2016;Kyissima et al., 2019;Matemilola et al., 2013;Rouf, 2015;Sbeiti, 2010), who found a negative and significant impact of profitability on leverage. The results contradict those of (Alipour et al., 2015;Chadha and Sharma, 2015), who documented a positive impact of profitability on leverage. Notes: 1-LEV is leverage, ROE is return on equity, TQ is Tobin-Q, LIQ is liquidity, ROCE is return on capital employed, ROA is return on assets, FSIZE is firm size and MVA is market value added; ***, ** and *indicate at 1, 5 and 10% level of significance respectively Notes: 1-LEV is leverage, ROE is return on equity, TQ is Tobin-Q, LIQ is liquidity, ROCE is return on capital employed, ROA is return on assets, FSIZE is firm size and MVA is market value added FSIZE has statistically significant negative effect on LEV. This effect is statistically significant at the level of 1% (b = À28.13, P-value = 0.000 < 0.01). This result is in line with (Bandyopadhyay and Barua, 2016;Benkraiem and Gurau, 2013;Güner, 2016;Rouf, 2015), who found a negative and significant impact of size on firms' leverage. These results do not agree with the findings of (Alipour et al., 2015;Chadha and Sharma, 2015;El-Khatib, 2017;G omez et al., 2014;Kyissima et al., 2019;Matemilola et al., 2013;Morri, and Cristanziani, 2009;Sbeiti, 2010), who advocated that size of a firm has a positive and significant impact on firm's leverage.

Regression estimation
MVA and TQ exhibit statistically significant effect on LEV at the level of 1% (P-value = 0.000 < 0.01). This effect is positive indicated by (b = 0.01 and 699.48, respectively). Further, the results reveal that ROCE has a significant positive impact (b = 0.32). However, this effect is statistically significant at the level 10% (P-value = 0.08 < 0.10).
Overall, the model is fit with a significance level of less than 1% (P-value = 0.000 < 0.01). The adjusted R 2 is 13% indicating that the variables included in the model contribute about 13% of the variability of LEV. Table 8 illustrates a comparison on the overall impact of all explanatory variables included in the models on the dependent variable. A country wise effect comparison estimation is made based on kingdom of Saudi Arabia (KSA). KSA is considered as the basis of the comparison and other countries are compared with KSA. The results show that both Oman and the UAE have a negative difference from KSA (b = À18.24, À17.88, respectively) however, Qatar has a positive difference from KSA (b = 2.82). This means that Qatar is better than KSA, Oman and the UAE but KSA is better than both UAE and Oman in terms of the overall impact of all explanatory variables on the dependent variable. However, the Pvalue of the three countries (Oman, the UAE and Qatar) are insignificant (P-value > 0.10) indicating that there are no significant differences among the countries in terms of the overall impact of all explanatory variables included in the models on the dependent variable.  Table 9 presents the results of robust regression. The results show consistent outputs to the results of the fixed effect model except in the case of liquidity. Coefficient estimates of both fixed and robust regressions are not highly deviated. All variables exhibit same effect except in case of ROE, which was insignificant in case of the random effect model but it is significant in the robust regression. This signifies a proper estimation of the regression assumptions. Further, the outputs of robust regression signify that the data is not contaminated with outliers.

Conclusion and policy implications
The main objective of the present research was to investigate the impact of corporate characteristics on leverage. These characteristics include ROE, Tobin-Q (TQ), Liquidity (LIQ), ROCE, ROA, firm size (FSIZE) and MVA. The finding of the study revealed that both ROA and FSIZE have a significant negative effect on leverage. However, MVA, ROCE and TQ exhibited statistically significant positive effect on leverage. Further, the results indicated that Qatar is better than KSA, Oman and the UAE but KSA is better than both UAE and Oman in terms of the overall impact of corporate characteristics on the leverage. However, this effect in not statistically significant. Despite the effort made by prior studies to investigate the issue of corporate characteristics and its relationship with leverage, this study focused on the GCC context, which open an insight of different research implications. The study realizes that the GCC countries have some characteristics such as immature stock markets, tax systems and Sharīʿah financing, which may have a significant impact on financing decisions of firms. This study calls upon managers, bankers, inventors, financial analysts and regulators in the GCC stock markets to reconsider capital structure components and try to rebalance between debts and equities depending on corporate characteristics. In addition, future research in the GCC countries should investigate some other determinants such as country-level factors, stock market characteristics, Sharīʿah financing and some areas of capital structure mix. This study contributes to knowledge by providing new insights into the corporate characteristics and leverage. By so doing, it provides an attempt to identify the factors influencing corporate financing behavior taking into consideration Notes: 1-LEV is leverage, ROE is return on equity, TQ is Tobin-Q, LIQ is liquidity, ROCE is return on capital employed, ROA is return on assets, FSIZE is firm size and MVA is market value added different issues such as different proxies of firms' profitability, market capitalization, MVA and liquidity, which provides original evidence from Gulf countries emerging markets. These countries are characterized by low tax rates and high liquidity. High liquidity may reduce the cost of borrowing and debt financing may not be a huge burden on firms' profits. This makes the investigation of leverage and corporate characteristics, particularly, firms' profitability and liquidity, very important. Therefore, the study tries to bridge an existing gap in the body of literature of capital structure and debt financing in Gulf countries emerging markets. The findings of this study will enable regulatory agencies to aim toward greater compliance with the local and international standards and will also enable them to enforce penalties for non-compliance. The limitation of the study is that only non-financial companies have been used as a sample. Hence, the results may not extend across all listed companies in Gulf countries.