Islamic Securities in Corporate Financial Hierarchy

Transcription

1. Islamic Securities in Corporate Financial Hierarchy Sara AlBalooshi 1 ,2 ∗1 , Maria Chiara Iannino †2 , and Pejman Abedifar ‡3 School of Economics and Finance, University of St Andrews, Castlecliffe St Andrews, KY16 9AL, United Kingdom 1 College of Business Administration, University of Bahrain, Sukhair 32038, Bahrain 3 Tehran Institute for Advanced Studies, Khatam University, Tehran, Iran In this paper, we investigate the place of Islamic investment securities (sukuk) in firms’ financial hierarchy using the modified pecking order theory. We study the external funding preferences of Malaysian firms using quarterly data of 112 firms for the period between 2005 and 2017. We find that when internal funds are exhausted, firms prefer to issue profit-loss sharing sukuk over bonds and fixed income sukuk is placed before equity. The results show that sukuk can widen external finance spectrum which has important implications for policy makers in countries with dual financial systems. Keywords: Corporate Capital Structure, Pecking Order, Islamic Finance, Sukuk JEL: G32, Z12 1 Introduction This study investigates the place of Islamic investment securities (sukuk)1 in firms’ financial hierarchy. The choice of a specific type of external financing instrument is extensively discussed in the literature. It resulted in the prominent capital structure theories; namely, trade-off (Jensen, 1986), pecking order (Myers, 1984), market timing (Baker and Wurgler, 2002), and modified pecking order theories (Leary and Roberts, 2010).2 In this paper, we examine the same question for firms with access to an additional funding tool, which is ”sukuk” using the framework introduced by Leary and Roberts (2010). Sukuk is the sole capital market instrument in Islamic finance, combining the features of debt and equity. The existing literature identifies the debt-equity characteristics of sukuk. Specifically, fixedincome sukuk (FIS) are based on leasing and cost-plus sale transactions. It has pre-determined profit rates and nominal values. Its cash flows are independent of the firms’ operations. Profit-loss sharing (PLS) sukuk are based on partnership arrangements, in which firms and sukuk-holders are partners in a specific project or asset. Sukuk proceeds depend on the performance of the underlying project. Profit is not guaranteed, and losses are shared between the parties. One key question is to place the different types of sukuk in the firm’s financial hierarchy. There are two conflicting views on this issue. According to the first view, sukuk are praised over bonds claiming that they reduce information asymmetry and adverse selection problems (Abdul Halim et al., 2017; Ebrahim et al., 2016; Nagano, 2010). In fact, adhering to Islamic guidelines, sukuk issuance requires transacting parties to specify a tangible asset and to transfer its ownership to a special-purpose-vehicle (SPV). This process protects investors’ rights. On the other hand, the second view claims that the implementation of sukuk leads to higher degrees of adverse selection and information asymmetry because of the complexity of the mechanism. For instance, (Klein and Weill, 2016) show that firms with higher adverse selection and moral hazard problems consider issuing sukuk. ∗ Corresponding author. This paper is part of the PhD thesis by Sara Al Balooshi, written under the supervision of Maria Chiara Iannino and Pejman Abedifar at the University of St Andrews. Email: salbalooshi@uob.edu.bh † Email: mci@st-andrews.ac.uk ‡ Email: p.abedifar@teias.institute 1 It is defined as ”certificates of equal value representing undivided shares in the ownership of tangible assets, usufructs, and services or (in the ownership of) the assets of particular projects or special investment activity” (AAOIFI, 2008) 2 A comprehensive review is provided by Frank and Goyal (2007, 2009); Harris and Raviv (1990) and Graham and Leary (2011).Others include Titman and Tsyplakov (2007), Flannery and Hankins (2013), Adam and Goyal (2008), Fama and French (2005), Rajan (1992), Leary and Roberts (2010), Lemmon et al. (2015), Graham et al. (2015), and Harris and Raviv (1990). 1
2. The extant literature has addressed the question of why and when firms issue sukuk (Abdul Halim et al., 2017; Grassa and Miniaoui, 2017; Mohamed et al., 2014; Nagano, 2010). However, the studies have categorized firms into groups of sukuk issuers versus bonds issuers. The former group can then choose which type of sukuk to issue. Such analyses exclude the possibility of firms using both sukuk and conventional financial instruments such as bond or equity.3 In addition, juxtaposing sukuk with bonds and overlooking the features and functions of PLS sukuk leads to distorted conclusions. Therefore, we view bonds, equity, and the different types of sukuk as a basket of capital market tools and we investigate the firm’ funding choices among them. In this study, we follow Leary and Roberts (2010) and define constraints imposed on firms’ funding deficit. They are represented by two points at the pecking order hierarchy, which are the exhaustion of internal funds and maximum debt capacity. The constraints define the three versions of the pecking order. In the strict specification, a firm is supposed to exhaust all available internal funds (savings equal to 0) before considering issuing bonds. Similarly, the issuance of equity is only justifiable if a firm debt ratio is equal to 1. In the semi-liberal and liberal specifications of the pecking order, the above conditions are relaxed. In the semi-liberal case, firms minimum savings and maximum debt capacity are matched to the industry averages, whereas in the most liberal version, savings and leverage levels are firm-determinant. In each of the above cases, we examine firms’ funding choice at each of the defined points, given a basket of funding instruments, including bonds, equity, and sukuk. The choice indicates the perceived adverse selection and information asymmetry involved in each instrument. The model is appropriate for this study because the thresholds are data-implied. Since there are no established theoretical thresholds for the timing and volume of sukuk issuances, the model facilitates the assessment of sukuk at the different stages of the financing hierarchy. The model also considers a range of pecking order degrees and various thresholds for utilizing funding sources. Conversely, previous Islamic finance capital structure studies use financial ratios solely to proxy for capital structure theories. In addition, we make no prior assumptions about the characteristics of sukuk. In other words, we do not assume that debt-like sukuk are closer to conventional bonds and that equity-like sukuk should resemble shares. We categorize sukuk according to their structure as fixed income sukuk, profit-losssharing sukuk, and zero-coupon sukuk. Each sukuk structure is an independent category.4 We conduct our investigation using a panel dataset comprised of 112 Malaysian firms operating in 10 sectors which had issued bonds, equity and sukuk concurrently over the 48 quarters covering the period between 2005 and 2017. Our results show that firms prefer to issue profit-loss sharing sukuk over bonds and fixed-income sukuk over equity. Meanwhile, bonds and equity are not significant funding alternatives for Malaysian firms. We find that firms perceive the differences among various types of sukuk, conventional debt and equity. This paper contributes to two bodies of literature. First, it adds to the empirical literature on the pecking order theory. For instance, Ebrahim et al. (2014) and Chen (2004) find that the capital structure of firms in Asia-Pacific countries can be explained by a modified pecking order, as firms prefer to raise external funds via bank loans and equity over bonds. We extend this literature by showing that firms prefer PLS sukuk over bonds and FIS over shares. Our empirical findings provide evidence in support of the modified pecking order in an Asia-Pacific country with a dual financial system. Second, we contribute to the literature on Islamic finance, in particular to the literature on regulatory and supervisory requirements of Islamic financial products. Recently, Abedifar et al. (2020) underline that Islamic banks need greater direct supervision compared to conventional banks. El Qorchi (2005) emphasizes the importance of extending Islamic regulations and supervision beyond banks, to capital markets and other sectors. We highlight that debt-like or fixed-income sukuk are comparable to shares, while profit-loss sharing sukuk have more in common with debt. Thus, with respect to information asymmetry, the definitions of Islamic and conventional debt and equity are distinct. For instance, FIS involve more information asymmetry and higher issuance costs compared to PLS. Therefore, FIS and ZCS require more regulation and supervision. In general, our study shows that there is a need for redefining Islamic debt, equity and capital and to restructure the regulatory framework taking into 3 For instance, the Malaysian capital market data show that at least 112 firms used both sukuk and traditional funding tools parallelly over the period 2005-2017. 4 fixed income sukuk(FIS) include: Murabaha, Wakalah, Ijara, Istisna and Salam contracts, profit-loss-sharing sukuk (PLS) include: Musharakah and Mudharabah contracts, and zero-coupon (ZSC) sukuk are short-term and non-tradeable fixed-income sukuk. 2
3. consideration the unique structure of Islamic investment securities . That includes listings, credit ratings, reporting requirements, and Islamic leverage definition. In the following section we set out a brief background of the Malaysian capital market, followed by hypotheses development in Section 3. We present the testing model and econometric approach in Section 4. In Section 5, we describe our data. Sections 6 and 7 include results and discussion, respectively. Section 8 illustrates robustness checks. Finally, Section 9 concludes. 2 Malaysia We focus on the Malaysian market because it is currently the global hub of Islamic capital markets. In 2000, Islamic investment securities were introduced to the Malaysian capital market. The Malaysian government had the vision to become a global Islamic finance hub. In 2011, more than two-thirds of the listed shares on Kula Lumpur Stock Exchange were Sharia-compliant (by market cap).5 Also, the presence of non-Muslim sukuk investors became substantial. In 2017, sukuk issuances were approximately USD 32.34bn against USD 25.734bn worth of conventional bonds issuances. Interestingly, despite the religious motives to use sukuk, one can observe a concurrent pattern of the issuances of both instruments in the period 2013–2017 (Figure 2). Therefore, analysing corporate financing decisions when sukuk is an option become a necessity to understand firms’ capital structure in dual financial system economies. 3 Hypotheses Development The pecking order theory demonstrates a clear-cut distinction between debt and equity-based on information asymmetry. Islamic investment securities combine the features of debt and equity; hence, the evaluation of information asymmetry is not straightforward. The literature shows mixed evidence about the distinction between sukuk types. While sukuk distinction is observable in the stock market reaction (Godlewski et al., 2011), it is not as clear in corporate finance theory studies (Grassa and Miniaoui, 2017). Therefore, to place sukuk types in firms’ financial hierarchies we use a two-stage pecking order framework developed by Leary and Roberts (2010). We intend to understand when and under which circumstances firms issue each type of sukuk. We incorporate sukuk types into the framework based on their technical features of being debt-like or equity-like. Each stage is defined by a threshold justifying the use of debt or equity. The first stage is when firms exhaust their internal funds and need external financing. According to Sharia principles and the above discussion, sukuk, in general, promote transparency and entail lower degrees of information asymmetry (Abdul Halim et al., 2017; Ebrahim et al., 2016; Grassa and Miniaoui, 2017; Mohamed et al., 2014). From this perspective, we formulate our first hypothesis in this stage of the framework: H0a1 : Firms with higher information asymmetry are more likely to issue sukuk than conventional bonds. However, differences persist among sukuk structures. To facilitate the discussion, we categorize sukuk into three groups based on the underlying contracts: (1) profit-loss sharing (PLS), (2) fixed-income sukuk (FIS) and (3) zero-coupon sukuk (ZCS). The latter two structures possess more debt features, such as pre-determined periodic payments. They are similar to lease and installments-sale contracts. On the other hand, PLS sukuk have more equity characteristics, and they are deemed to be the most Sharia-compliant. They are partnership arrangements between parties. Scholars assert that PLS contracts involve more adverse selection and information asymmetry compared to fixed-income sukuk. In PLS financing arrangements, the entrepreneur has more information about the project compared to the investor. Further, the disclosure of such information depends on the entrepreneur. This issue of information disclosure is vital at the time of contracting and at later stages of performance reporting. Another problem with a PLS contract is its inability to demand collateral. According to Sharia principles, equity-based contracts cannot require collateral. Hence, our second hypothesis in the first stage of the framework is: 5 Tax incentives are an essential aspect of achieving the Malaysian vision. Tax law treated Islamic and conventional transactions equally. For example, to avoid over-taxation, Islamic profits are assimilated to traditional interest, and profits from sukuk are tax exempted. Also, Islamic partnerships that are similar to venture capital are not considered partnerships from a tax perspective. 3
4. H0a2 : Firms with higher information asymmetry are more likely to issue profit-loss sharing sukuk than fixed-income and zero-coupon sukuk. In the second stage of the pecking order framework, firms exhaust both their internal funds and maximum debt capacity and require external financing. As we find no literature examining the equity characteristics of sukuk, we adopt the argument of Myers and Majluf (1984) that issuing shares involves the highest degree of adverse selection and firms – theoretically – should never issue shares. Our hypotheses at this stage are: H0b1 : Firms with higher information asymmetry are more likely to issue sukuk than equity. H0b2 : Firms with higher information asymmetry are more likely to issue profit-loss sharing sukuk than fixed-income and zero-coupon sukuk. 4 Methodology This paper investigates the place of sukuk in firms’ financial hierarchies given the availability of traditional funding means. We take into account the debt-equity hybrid of sukuk structures. We apply the framework developed by Leary and Roberts (2010).6 The model is based on Myers and Majluf’s (1984) hierarchy and accommodates three versions of the pecking order theory. Each version embodies a decreasing level of strictness of the theory. The specifications allow for interpretations in light of the pecking order and trade-off models. Given its rationale, in this paper we postulate that a modified form of the model can encompass sukuk and test firms’ capital structure. Due to information asymmetry, firms prefer to use their internal funds to finance their new positive investment opportunities (Myers and Majluf, 1984). Once firms have depleted their internal funds, the issuance of conventional bonds is justified (Figure 1 – point C). Likewise, the issuance of equity is only rationalized if the investment size is large, and firms’ maximum debt capacity is unable to cover it (Figure 1 – point D). Leary and Roberts (2010) quantified points C and D to specify two thresholds: αC and αD . The point at which firms’ internal funds (savings) are exhausted and firms rationally start to use debt is αC . Further, αD is the firm’s maximum debt capacity measured by the debt ratio.7 Reaching αD is what determines, “the timing” a firm can rationally issue equity. We demonstrate the model thoroughly in the next section. In the following section, we present the two-stage model pecking order theory with three strictness degrees. The model is modified to account for sukuk as a funding option. 4.1 4.1.1 Model Modified for Sukuk Stage 1 The first stage of the model is defined by the first threshold, αC . In a strict pecking order theory, a firm is supposed to exhaust all available internal funds before it considers issuing external funds. On such a basis, Leary and Roberts (2010) establish a firm’s lower bound of savings. A firm is only going to issue debt if: C Investment − [Internal funds − αit ] > 0, (1) implying that: C(min) C αit > [Internal funds − Investment] = αit (2) C(min) where αit is the lower bound of savings. In essence, using external funds is only justified under the pecking order theory when the saving requirement is greater than the lower bound of savings. Ranging from a strict to a liberal pecking order theory, the definition of the lower bound of saving will change as well.8 6 The authors criticize the static interpretation of corporate capital structure theories and suggest that allowing flexibility in modeling and hypothesis interpretation can enhance the accuracy of empirical analysis. They assert that the pecking order and trade-off theories are complementary and together they constitute a modified pecking order. 7 Debt ratio = (TD/TA) = total debt/total asset. 8 All quantities are scaled by total assets and calculated following Leary and Roberts (2010): • Investment = Capital Expenditure + Change in Investment + Cash Paid for Acquisitions + Sale of Investments and Property Investments + Other Investment Activities • Internal Funds = Cash + Cash Flow − Total Dividends Paid - Change in Working Capital 4
5. • Case 1: C(min) αit = 0 (Firms do not keep any savings; savings = internal funds = 0) • Case 2: C(min) αit = current cash balance + median of cash balance of firms in the same industry-year combination • Case 3: C(min) αit = current cash balance + median of historical cash balance of the same firm Case 1 assumes a strict pecking order specification, where the firm does not keep any savings. It uses all its internal funds before seeking external funds. Case 2 is a semi-liberal version of the pecking order. The firm decides to keep savings that are equal to the industry’s average. In the last case, the firm follows a more liberal pecking order theory, where the savings are firm-determinant. A firm decision on which type of external funds to use when needing to undertake a new investment can be described as: K C(min) Extf undsit = αit +β Xkit−1 + (3) it k=1 C(min) K Where αit is the lower saving bound of the firm in the tested case, while the term k=1 Xkit−1 includes k lagged firm characteristics as determinants of capital structure and it is the error term. Extf undsit is a multinomial dependent variable with the categories of external funding tools, defined as follows:  1 Conventional Bonds     2 PLS Sukuk    3 FI Sukuk (4) Extf undsit =  4 ZC Sukuk      5 Bank Loan    6 Do Nothing Where PLS is profit-loss sharing, FI is fixed income and ZC is zero-coupon. The dependent variable Extf undsit of each firm in each quarter takes a single value designated to the funding source. It would take values 1, 2, 3 or 4 if the firm issued bonds or any type of sukuk. Extf undsit takes a value of 5 if the change in long-term debt is greater than 5%. Finally, the dependent variable takes a value of 6 if a firm’s internal funds are negative and the firm did not use any alternative external funding source. 4.1.2 Stage 2 According to the strict version of the pecking order theory, a firm will never issue equity. However, a less rigorous interpretation would consider equity financing as the last funding resort for a firm. Thus, if a firm exhausted its savings and is only able to borrow through issuing junk bonds, the issuance of equity is justifiable. That takes place at point “D” on the Myers hierarchy (Figure 1). It is quantified by Leary and Roberts (2010) to represent the point at which a firm reaches its upper bound of debt capacity αD . At that point, the issuance of equity is justifiable. That being the case, a firm would issue equity only when: C D Investment − (Internal funds − αit ) + αit − Debtt−1 > 0, (5) which implies that: D(max) D C αit < Investment − [Internal funds + αit + Debtt−1 ] ≡ αit (6) • Cash Flow = Income before Extraordinary Items + Depreciation and Amortization + Extraordinary Items + Deferred Tax + Cash from Other Operating Activities 5
6. D (max) where αit is the upper bound of debt capacity. Similar to the first stage, depending on the degree of strictness of the pecking order theory, capital structure decisions beyond debt capacity are also looked at across three stages: • Case 1: D(max) αit = 1 (debt capacity = debt ratio = 1) • Case 2: D(max) αit = debt capacity is equal to the debt ratio of the industry’s investment-grade companies • Case 3: D(max) αit = debt capacity can vary according to the firm’s needs. Calculated as the annual average debt ratio of the firm The first case indicates that a firm will utilize its maximum debt capacity, that is its total debt to total assets ratio, (TD/TA) = 1. In the second case, a firm would issue debt up to a certain level equal to the debt ratio of the industry’s investment-grade firms. Case 3 is the most liberal version of the pecking order, where the maximum debt capacity is firm-determinant. At this stage, we attempt to explore what type of sukuk a firm would issue when its investment cost exceeds the firms’ debt capacity upper bound. A firm’s decision to issue equity can be described as: K D(max) Extf undsit = αit +β Xkit−1 + uit (7) k=1 D(max) K where αit is the upper debt capacity bound of the firm in the tested case, while the term k=1 Xkit−1 includes k determinants of capital structure and uit is the error term. Extf undsit is a multinomial dependent variable including the categories of external funding tools other than conventional bonds and bank loans, defined as follows:   1 Equity Shares      2 PLS Sukuk (8) Extf undsit = 3 FI Sukuk    4 ZC Sukuk    5 Do Nothing where PLS is profit-loss sharing, FI is fixed-income, and ZC is zero-coupon. The outcome from Stages 1 and 2 would underline the possible ranking of sukuk along the Myers hierarchy. Also, it would highlight the conditions and firms’ characteristics under which Islamic finance funding tools are issued. In the next section, we discuss the firm characteristics that we employ as capital structure determinants. A summary of variables definitions is presented in Table 3. We perform a multinomial logistic regressions to estimate our model (Davidson et al., 1993; Greene, 2012). Logistic models are a popular approach to examine funding decisions in corporate finance literature as in Baskin (1989), Elliott et al. (2008), Denis and Mihov (2003) and Leary and Roberts (2010). We run an Independence of Irrelevance Assumption (IIA) test developed by Long and Freese (2006). It states that categories represented by a multinomial variable should be alternatives and not substitutes. The test shows that the six categories defined in the dependent variable are significantly distinguishable and will not be combined. This finding supports our argument of the importance of differentiating between sukuk types according to their structure. Firms’ perception of such differences is one aspect we are examining in our research. Although the developers of the IIA test criticized its consistency, we run it for diagnostic purposes. Further, the Hausman test and seemingly unrelated estimation-based Hausman test show significant evidence that regressions outputs are independent of each other (Freese et al., 2000; Kleinbaum and Klein, 2010). We report that all explanatory variables significantly affect the funding choice at the 1% significance level. 4.2 Capital Structure Determinants In addition to the fundamental pillars of corporate structure theories, the literature emphasizes the role of firms’ business activities and financial characteristics in capital decisions. Leary and Roberts (2010) 6
7. use firm characteristics suggested by the well-known work of Adam and Goyal (2008); Frank and Goyal (2009); Jensen (1986). The literature empirically proves that such factors shape firms’ capital structure. We employ comparable factors contingent on data availability (Table 3). The first factor is the firm size. It is measured by the natural logarithm of total assets. The impact of firm size on sukuk issuance is not clear (Abdul Halim et al., 2017; Azmat et al., 2014; Nagano, 2016). Larger firms can effortlessly tap the bond market, while it is difficult for smaller firms, which motivates such firms to issue sukuk. The seconf factor is profitability whoes equally controversial. Klein and Weill (2016) assert that the sukuk issuance is negatively related to profitability. Thus, sukuk issuers are usually low performers. Less profitable firm would find sukuk an attractive mode to raise funds and transfer bad projects outside the balance sheet. We measure profitability as the ratio of earnings before interest, taxes, and depreciation (EBITD) to total assets, and net income to total assets. The third factor is firm growth or investment opportunity. We use the ratio of market to book ratio to proxy for firm growth. Klein and Weill (2016) document that the rise in the market to book ratio increases information asymmetry, and consequently motivates firms to issue sukuk. Tangibility is the fourth capital structure determinant. The ratio of fixed assets to total assets has various economic interpretations, as the pecking order theory suggests. It argues that greater tangibility decreases leverage levels. That is because higher levels of fixed assets reflect low information asymmetry encouraging firms to issue equity. Likewise, a considerable portion of fixed assets decreases adverse selection, making debt easily accessible. As tangible assets are a mandatory requirement for sukuk issuance, firms with high tangibility are expected to turn to the sukuk market. Also, less collateralized firms are considered riskier and might tap the sukuk market. Such firms find it difficult to access other funding markets (Klein and Weill, 2016). We also use the current ratio to measure firms’ financial distress caused by excess levels of debt. Firms with a low current ratio issue sukuk less often because they are not financially distressed. On the contrary, high levels of the current ratio might also result in more sukuk issuances because of the shortage of short-term debt (Klein and Weill, 2016). 5 Sample Malaysia provides an ideal scope for our study. It is home to 60% of the total global corporate sukuk. It hosts a dual financial system that is optimal to examine corporate capital decisions. We build a panel dataset comprised of 112 Malaysian firms over the period 2005–2017.9 Our data has 40 quarters (T = 40). We include all firms which issued sukuk at least once during that period. We draw sukuk and bond data from Bloomberg. We specifically obtain information about sukuk issuance date, the amount issued, the amount outstanding, contract type and maturity. We use data on equity issuances and quarterly financial results from Compustat Global, SNL Financial and Bloomberg. Malaysian firms raised almost USD 43.36bn with more than 1,000 sukuk issuances. Firms in the sample actively issued different types of sukuk over the 12 years. We filter the dataset for financial, governmentowned and private limited companies. That results in our sample of 112 firms and 1,046 issuances. Since a considerable number of firms issued sukuk more than once in a specific quarter, we summed values of issuance and averaged duration. The previous technique is the most appropriate approach to obtain one observation per quarter without forfeiting valuable information. The compressed number of issuances is 431 (Table 1). We exclude observations of multiple funding resources in a single quarter. Firms in the sample operate in ten sectors (Table 2). Our dataset is winsorized at the 10% and 90% levels to control for outliers. Sukuk are classified into three groups: fixed-income, profit-loss sharing and zero-coupon sukuk. Zero-coupon sukuk are the most utilized external financing source, followed by fixed-income sukuk and conventional bonds and, finally, PLS sukuk. 6 Inference and Results This paper seeks to position sukuk among traditional funding instruments in view of capital structure theory. We examine firms’ decisions to raise funds via sukuk given the availability of bonds, equity and bank loans. We find evidence that it is possible to rank the different sukuk contracts among 9 Before 2005, corporate sukuk issuances were modest. Sukuk offerings were mainly sovereign and quasi-sovereign. 7
8. traditional instruments . Our analysis shows that firms prefer to issue PLS sukuk when internal funds are exhausted and issue FIS beyond maximum debt capacity. Moreover, conventional debt and equity are not significant funding alternatives for Malaysian firms. We address our results from two pecking order specifications: semi-liberal and liberal. We do not discuss our results from the strict version of the pecking order. Imposing the strict conditions on the sample results in a random mass of quarter-firm observations which suffers from large periodic gaps and insufficient observations per firm.10 Scholars such as Leary and Roberts (2010) questioned the practical interpretation of the strict pecking order. Our discovery indicates that the strict version of the pecking order theory does not apply to the Malaysian firms. Marginal effects better present multinomial logistic models and makes categorical results interpretation more tangible (Cameron and Trivedi, 2010; Long, 2009; Williams, 2012). We discuss our results subject to the average marginal effects of variables of interest relative to the choice of funding instrument in both versions of the pecking order. 6.1 Semi-Liberal Pecking Order In this version of the pecking order, minimum savings and maximum debt capacity are matched to the industry averages. Thus, debt is adjusted to a target (industry average), which accords to the trade-off model assumptions. 6.1.1 Stage 1 In the first stage the deficit is constrained by the minimum saving level αC(min) , and Malaysian firms are to raise external funds with sukuk, conventional bonds or bank loans. We find that the effect of the deficit on funding choice is significant but relatively small. On average, firms prefer to obtain funds from bank loans (P=4%) and PLS sukuk (P=1%) when deficit size is large. At lower deficit values, the probability of issuing zero-coupon sukuk is 2% (Table 9 and Figure 4). In terms of firm characteristics in this stage, as firms’ size increase, firms prefer to issue fixed-income sukuk (P=1.3%), bonds (P=1%) and PLS (P=0.8%). Smaller firms, however, have a 2% probability of doing nothing. Tangibility refers to the concentration of fixed assets in a firm. Tangible firms are 3.2% more likely to issue conventional bonds, while it is not a significant factor in choosing sukuk or bank loans (Table 12). All firms in our sample have a book to market ratio that is equal to or greater than 1. That indicates that all firms are undervalued, including firms in the technology sector. Hence, their stocks are traded at prices lower than their actual worth. We find that growth opportunity significantly affects fixed-income and zero-coupon sukuk only. Firms with greater growth opportunity prefer to issue zero-coupon sukuk (P=2%). On the contrary, firms with lower growth opportunity or higher book to market ratio issue fixed-income sukuk (P=0.8%). The sample firms appear to have weak credit positions as the Altman Z-scores reach a maximum of 1.5, which is below the minimum threshold of 1.8. However, firms in the upper percentile prefer to issue zerocoupon sukuk and conventional bonds with probabilities equal to 2.2% and 1.6% respectively. Leverage levels significantly and positively impact the issuance of zero-coupon sukuk (P=13%). Profitability does not have a significant impact on funding sources at this stage (Table 12). 6.1.2 Stage 2 The deficit is constrained by the maximum debt capacity αD(max) in this stage. Firms can issue sukuk or shares to acquire needed funds. The impact of the deficit is more pronounced than in the previous stage. Firms are more likely to issue fixed-income sukuk with 25% probability when deficit size is rather small (Table 9 and Figure 4). As we take into consideration the maximum debt capacity in addition to cash exhaustion, the size of the conditioned deficit affects the relationship between firm characteristics and the choice of capital funding change (Table 12). Our findings show that the probability of issuing equity is 6% as firm size 10 A total of 444 observations or 10% of the original sample. 8
9. increases . That is followed by PLS sukuk with 3.2% probability. Smaller firms prefer to issue zerocoupon sukuk (P=2%). Tangibility positively and significantly affect the probability of issuing FIS (P=6.5%). Intangible firms prefer to issue zero-coupon (P=9%) and PLS sukuk (P=5.2%). Similar to the first stage, firms with low growth outlook prefer to issue fixed-income sukuk (P=3.14%). While growing firms prefer to issue shares and zero-coupon sukuk (P=4.88% and P=3% respectively). Profitable firms are more likely to issue shares with 2% probability. Troubled firms prefer to issue zero-coupon and fixed-income sukuk (P=36% and 9.27% respectively). Solvent firms continue to prefer zero-coupon sukuk, while insolvent firms prefer to issue PLS sukuk. Unlike the first stage, leveraged firms prefer to issue PLS sukuk (P=5.5%). Figure 6 demonstrates the marginal effects of firm characteristics in this stage. In summary, we observe that in a semi-liberal pecking order specification, the deficit size significantly explains the issuance of the different types of sukuk. We show that beyond the minimum savings thresholds, Malaysian firms prefer to raise funds via zero-coupon sukuk and PLS sukuk, whereas, the issuance of FIS is significant beyond firms’ maximum debt capacity. The implications of firm characteristics are more pronounced than in the second stage of the specification; specifically, when comparing sukuk with equity rather than conventional bonds. We furnish more insights in the discussion section. 6.2 Liberal Pecking Order In the liberal interpretation of the pecking order theory, firms’ cash management and leverage policies are firm-determinant. Therefore, firms keep cash and debt equal to their average historical levels. Figure 5 illustrates the marginal effects of each threshold with respect to funding instruments 6.2.1 Stage 1 The first constrained deficit (by αC(min) ) significantly affects zero-coupon sukuk. When the deficit size is small, the probability of issuing zero-coupon sukuk is 14.1%. Beyond their minimum saving requirement thresholds, large firms are more likely to issue FIS, PLS sukuk and conventional bonds(P=1.35%, 1.32% and 0.88%). Small firms have a greater probability of issuing zero-coupon sukuk and bank loans. Tangibility does not have a significant impact on funding options (Table 13). Firms with future growth outlook issue bonds and zero-coupon sukuk (P=1.94% and 1.27% respectively). Firms with lower growth perspective (higher book to market ratio) are more likely to issue FIS (P=0.7%). Troubled firms are more likely to issue PLS sukuk with 18.1% probability. Firms with better credit stability are more likely to issue PLS sukuk and bank loans (P=1.5% and 8.7% respectively). Leverage levels have a positive and significant effect on issuing conventional bonds (P=6.4%) and zero-coupon sukuk (P=5.2%). Low-leveraged firms prefer to raise funds through bank loans (Table 13). 6.2.2 Stage 2 Beyond the second threshold (αD(max) ) firms fund small financial deficiencies with zero-coupon sukuk (P=31.7%) and fixed-income sukuk (P=24.1%). Further, larger firms prefer to raise funds via PLS and FIS (P=1.34%, P=1.06%) and through issuing sharea with a greater probability (P=4.7%). Being intangible and profitable significantly impacts the probability of issuing equity (P=8% and P=9.4%). Fixed-income sukuk are more likely to be issued by firms with high book to market ratio (P=1.2%). Firms with future growth opportunity prefer to issue zero-coupon sukuk (P=1.9%). Financially stable firms have a higher probability of issuing equity to raise funds. Conversely, insolvent firms prefer to use FIS (P=3.4%). The increase in leverage levels is associated with the probability of issuing ZC, FIS and PLS sukuk (P=35%, 17.4% and 9.7%). Figure 7 exhibits marginal effects of firm characteristics. Our results from this specification give further evidence that the market recognizes the different sukuk structures. Also, sukuk types can be ranked according to a liberal pecking order interpretation. The effect of firm characteristics shows some degree of consistency with the semi-liberal pecking order case. 9
10. 7 Discussion Our findings significantly underline the differences between sukuk types . Hence, our hypotheses that test issuing sukuk as a single instrument against bonds or equity are invalid. In both pecking order specifications, firms prefer to issue PLS sukuk when internal funds are exhausted and prefer to issue FIS beyond the maximum debt capacity. Therefore, it is critical to acknowledge the type of sukuk contract when analysing capital structure decisions. Particularly, when constrained by minimum savings, it is safe to say that to Malaysian firms, PLS sukuk involve less adverse selection problems compared to conventional bonds and FIS. This conclusion is partially in line with the argument of Abdul Halim et al. (2017); Ebrahim et al. (2014); Mohamed et al. (2014) and Grassa and Miniaoui (2017). Similarly, Malaysian firms’ preference to obtain funds via FIS beyond the maximum debt capacity is a critical observation. First, it emphasizes Myers and Majluf’s (1984) claim that equity is the most informationally expensive funding tool. Second, it indicates that a specific type of sukuk is preferred over equity. Third, it provides additional evidence that Malaysian firms perceive the difference between sukuk types. Fourth, it - to some degree - confirms Klein and Weill’s (2016) argument of why sukuk are the choice of firms with high information asymmetry. In short, when internal funds are fully consumed, firms in a dual financial system raise funds via PLS sukuk first, then choose conventional bonds, thus rejecting our first and second hypothesis stating that firms prefer sukuk and PLS instead of bonds when information asymmetry is high. Depleting the borrowing limit drives firms to issue FIS followed by issuing equity. That also leads to rejecting our hypothesis denoting that firms prefer sukuk and PLS sukuk over equity. The discussion highlights that categorization of sukuk based on debt-equity characteristics is theoretically accurate. However, the market perceives otherwise. According to our results, the debt-like sukuk (FIS) is comparable to shares, while the equity-like sukuk (PLS) is comparable to debt. The latter is conceivably due to the contractual requirements of issuing sukuk and the mandatory establishment of the SPV, which reduces adverse selection. On the other hand, FIS contracts are based on leasing and cost-plus sale transactions with pre-determined periodic payments. Such sukuk are usually non-tradeable, inducing higher adverse selection. The characteristics of zero-coupon sukuk are not clear because they are issued under both thresholds. However, the positive relationship between deficit and the issuance of debt is in line with Shyam-Sunder and Myers’s (1999) and Helwege and Liang’s (1996) findings, which acknowledge the debt characteristics of zero-coupon sukuk. Consistent with Chen (2004), Malaysian firms use bank loans before selecting sukuk as predicted for bank-based economies. The previous is with respect to deficit size, which is the key variable in our model. Trying to depict capital structure preferences by only analysing firm characteristics results in mixed evidence, as in Abdul Halim et al. (2017); Mohamed et al. (2014); Nagano (2016) and Grassa and Miniaoui (2017). We find that sukuk debt-equity features are not clear because sukuk are issued in conjunction with shares and conventional bonds. However, results confirm that to Malaysian firms’ sukuk encompass less information asymmetry compared to conventional bonds, but more than bank loans. Equity is the most expensive funding tool, consistent with the Myers and Majluf (1984) pecking order. Issuing equity has the least issuance probability. We find that large firms issue PLS sukuk, FIS and bonds with comparable magnitudes, which is in line with the trade-off model, whereas small firms prefer to issue funds via zero-coupon sukuk. These findings refute Klein and Weill’s (2016) argument; that small firms issue sukuk because large firms are solidly established and can effortlessly tap the bond market. Klein and Weill (2016) state that sukuk issuers are those with greater market to book ratios. Nevertheless, we find firms with growth potential issue fixed-income sukuk, while firms with a high market to book ratio prefer zero-coupon sukuk. That is another confirmation of the different debt-equity characteristics of sukuk. The impact of tangibility and profitability is significant beyond the maximum debt threshold in the comparison between equity and sukuk. Despite the role of tangible assets in sukuk structures, we find that intangible firms prefer to obtain funds via PLS and ZC sukuk. Klein and Weill (2016) explain that less collateralized firms resort to sukuk because they are risky. As proposed by the literature, we find non-profitable firms select FIS and zero-coupon sukuk, while profitable firms go for shares. Issuing sukuk allow firms to transfer unprofitable projects outside the balance sheet to the sukuk-holders. 10
11. For further insight , in Table 14 we compare the expected signs of the association between firm characteristics and the issuance of sukuk, equity and bonds according to the pecking order theory. The combination of trade-off and pecking order theories affirms sukuk characteristics as the predicted signs show. 8 8.1 Robustness Check Modified Pecking Order in Emerging Markets Scholars such as Demirg¨ u¸c-Kunt and Maksimovic (1999), Booth et al. (2001); De Jong et al. (2008) and Fan et al. (2012) suggest that firms in emerging markets maintain a modified pecking order of funding sources where issuing equity is favoured over bonds. We use their argument to examine corporate funding decisions further, adding sukuk as an option. We allow bonds and equity to be possible funding choices for firms in both stages of the pecking order. Therefore, we estimate firm capital decisions using Equations 3 and 7 given the following funding alternatives:  1      2      3 Extf undsit = 4   5      6    7 Conventional Bonds PLS Sukuk FI Sukuk ZC Sukuk Equity Bank Loan Do Nothing (9) The marginal effects confirm our main results. Firms prefer FIS over shares when the debt capacity reaches its maximum. The probability of issuing sukuk is 15% and 10.6% in both specifications, while issuing shares is insignificant. Zero-coupon sukuk remain an option regardless of the threshold. Additionally, at this stage, the probability of writing conventional bonds is significant, with approximately 5% probability. On the contrary, firms are more likely to borrow through issuing shares (P=15%) when firms approach their minimum cash thresholds. The behavior of firms toward bonds and shares is in line with the findings of Demirg¨ uc¸-Kunt and Maksimovic (1999),Booth et al. (2001), De Jong et al. (2008) and Fan et al. (2012).11 8.2 Debt-Equity Characteristics of Sukuk Sukuk are not an innovation (Wilson, 2008). One can identify grey areas between structures of sukuk, bonds and shares. However, that does not impact the fact that such instruments are compliant with the Islamic guidelines and serve purposes other than profit maximization. Scholars find the resemblance between the instruments convenient grounds to classify sukuk. We shall reiterate that fixed-income and zero-coupon sukuk are considered debt-like, while PLS sukuk are deemed to have more equity characteristics. We use this classification to examine the probability of issuing sukuk among conventional instruments. We assume that debt-like sukuk, along with conventional bonds, are used when firms exhaust their internal funds. PLS sukuk and shares the least preferred according to the pecking order, beyond maximum debt. Therefore, external funding alternatives are defined as follows: Stage 2: Deficit constrained by αD(max)  C(min) Stage 1: Deficit constrained by α 1 PLS Sukuk  1      2    3 Extf undsit = 4      5    7 11 Tabulated Extf undsit = Conventional Bonds PLS Sukuk FI Sukuk ZC Sukuk Bank Loan Do Nothing    results are available upon request. 11 2 3 Equity Do Nothing
12. With respect to the above , we re-estimate Equations 3 and 7. We find that in the semi-liberal specification, none of the Islamic certificates are significant. Thus, the alternatives do not match the thresholds. For instance, being not able to select PLS sukuk in the first stage makes other choices inferior. The same applies to issuing FIS in the second stage. However, in the liberal specification, where we relax the deficit constraints, firms amend funding preferences. In particular, rather than issuing PLS sukuk, firms go for FIS in the first stage. Nonetheless, Malaysian corporations prefer sukuk over bonds.12 8.3 Sharia-Compliance and Conventional Debt Levels In order to be listed as firms that adopt Islamic finance principles, “Sharia-compliant” firms must adhere to all Islamic finance restrictions and prohibitions. However, given that formal Islamic finance practices are recent, scholars and practitioners established thresholds to govern Sharia-compliance. For instance, an Islamic firm can utilize conventional debt under the condition that conventional debt levels do not exceed 5% of total debt in the Kuala Lumpur Stock exchange, while it can go up to 33% as per the general Islamic principles. Building on the above, in this section, we assume that if Malaysian firms aim to maintain a “Sharia-compliant” status, the probability of issuing conventional bonds is positive if firms’ leverage ratio is below 5%. Likewise, the probability of issuing sukuk increases if firms’ leverage ratio exceeds 5%, because issuing bonds is no longer an option. We also extend our analysis to the 33% debt ratio threshold. We re-estimate equation 3 in liberal and semi-liberal pecking order specifications. Our results show that as the constrained deficit increase in the semi-liberal pecking order with firms with book leverage less than or equal to 5% firms prefer to either approach banks to obtain funding or do nothing. Beyond the 5% sharia compliance threshold, first prefer to issue zero-coupon sukuk and fixed income sukuk if we ignore the significance level. In the liberal specification, firms with book leverage less than or equal to 5% prefer to use FIS. Nothing changes for those above 5%. Firms prefer to issue zero-coupon sukuk when leverage levels are greater than 5%. The findings suggest that in Malaysia firms’ preference to issue sukuk is not driven by the desire to maintain a “Sharia-compliant” status. That, support our approach of using a modified version of the pecking order taking into account a constrained deficit and firm characteristics. When using a more general conventional debt leverage, our findings show that the deficit levels are positively and significantly associated with the probability of issuing bonds when firms’ book leverage is less than 33% in both pecking order specifications. Where leverage ratio is greater than 33%, Malaysian firms are more likely to issue FIS in the semi-liberal specification compared to PLS and zero-coupon sukuk in the liberal case. Accordingly, a more rigorous debt threshold achieve the “Sharia-compliant” status. We do not refer to bank loans at this stage because it is not possible to differentiate between Islamic and conventional bank loans.13 9 Conclusions Sukuk have distinct legal, financial and regulatory requirements, which necessitate an adequate understanding of its principles, application and possible conventions. Several countries such as Hong Kong, Tunisia and Kazakhstan are undertaking regulatory and legal reforms to accommodate Islamic banking and finance alongside conventional finance. Hence, it is imperative to study the features and functionality of Islamic financial instruments within the context of finance theories in order to establish an appropriate regulatory framework that incorporate the unique structure of Islamic investment securities. This paper contributes to an understanding of the position of the different types of sukuk in Malaysian firms’ financial hierarchy in light of capital structure theory. Our data show that in a dual financial system, Islamic and conventional capital market instruments are not mutually exclusive. All firms in our sample issued both sukuk and conventional financial instruments during the study period, i.e. 20052017. Therefore, the religious motive to utilize sukuk is void because it would result in a sukuk-based capital structure. We use a theoratical framework where we first define two points in the pecking order hierarchy, which are the exhaustion of internal funds and maximum debt capacity. Then, from a basket of funding 12 Tabulated 13 Tabulated results are available upon request. results are available upon request. 12
13. instruments , including bonds, equity, and sukuk, we look at the firms’ funding choices at each of the defined points. The choice indicates the perceived adverse selection and information asymmetry involved in each instrument. We use quarterly financial and accounting data of 112 Malaysian firms from the period between 2005 and 2017. Our findings show that Malaysian firms recognize distinct levels of information asymmetry not only in conventional bonds and equity but also among different classes of sukuk. Our results highlight that in a modified pecking order, sukuk are preferred to bonds when internal funds are fully utilized. Also, beyond maximum debt capacity, firms prefer to issue sukuk over equity. However, zero-coupon sukuk are the last funding option in both thresholds. We highlight that debt-like or fixed-income sukuk are comparable to shares, while profit-loss sharing sukuk have more in common with debt. Our study provides an empirical proof that firms distinguish between Islamic and conventional instruments and the different types of sukuk. Identifying the rank of sukuk in a corporate financial hierarchy among traditional instruments contributes to better understanding of sukuk structure, which has important implications. It enhances Islamic finance reachability to international investors and issuers. It also assists investors and the market to read the signals when firms choose to raise funds via a specific sukuk contract. Finally, empirically verifying sukuk characteristics has also implications for its valuation and pricing mechanisms, risk evaluation and credit rating. 13
14. 10 10 .1 Appendix Tables Table 1: Frequency of external funding sources Frequency Percent Description 85 181 65 7.09 6.93 2.49 Ijarah, Istitna, Murabaha, and Wakalah Sukuk Mudharabah and Musharakah Bank Loan Equity Bonds 1319 721 139 50.54 27.62 5.33 5% increase in long term debt Leary and Roberts (2010) New shares offered New bond issuances Total 5066 100.00 Islamic Investment Certificates Zero-coupon Sukuk Fixed Income Sukuk Profit-loss sharing Sukuk Conventional Instruments The table presents the issuance frequencies of sukuk, and traditional funding instruments; bonds, equity and bank loans over the period 2005-2017. Each issuance takes place in a specific quarter by a certain firm. In case of multiple issuance per quarter, issuances are summed if the same instrument is used. The last column detail issuance contract types. Fixed income sukuk are Islamic securities with more debt characteristics. PLS possess more equity features. Zero-coupon sukuk is a special type of fixed income sukuk which present pure debt. Sukuk and bonds data are obtained from Bloomberg. New equity issuances and bank credit lines are obtained from Compustat. Table 2: Sectors summary Sector GIC Sector code No. of firms Obsv. PLS FIS ZCS Sukuk (total) Bonds Shares Bank loans Industrial Real Estate Consumer Discretionary Energy Materials Consumer Staples Utilities Telecommunications Healthcare Technology Total 20 60 25 10 15 30 50 55 35 45 34 15 14 12 11 9 7 5 9 2 118 1474 713 641 568 495 360 331 243 143 97 5065 28 17 10 1 3 1 3 2 0 0 65 56 23 14 10 12 17 14 27 8 0 181 30 12 41 32 38 18 1 0 1 12 185 114 52 65 43 53 36 18 29 9 12 431 56 11 8 21 4 16 14 8 1 0 139 240 111 33 44 31 51 76 62 68 5 721 307 153 145 158 97 88 50 54 44 19 1115 This table presents the distribution of Malaysian firms over sectors. It shows the issuance frequencies of sukuk, and traditional funding instruments; bonds, equity and bank loans over the period 2005-2017 per sector. In case of multiple issuance per quarter, issuances are summed if the same instrument is used. Sukuk and bonds data are obtained from Bloomberg. New equity issuances and bank credit lines are obtained from Compustat. A 5% increase in long term debt is considered a new bank loan. 14
15. 15 C (min) D(max) Assets Altman Z-score = σROA [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets Total Current Assets/Total Current Liabilities GDP growth GIC Sectors ROA+( Equity ) Natural log of Total Assets Total Fixed Assets/Total Assets (Debt in Current Liabilities + Long-term Debt)/Total Assets Total Assets/Net Income The upper bound of maximum debt capacity. It is defined in each pecking order case as follows: D(max) Case 1: αit =1 D(max) Case 2: αit = debt capacity is equal to the debt ratio of the industry’s investment grade companies. D(max) Case 3: αit = debt capacity can vary according to the firm’s needs. Calculated as the annual average debt ratio of the firm. The lower bound of savings. It is defined in each pecking order case as follows: C(min) Case 1: αit =0 C(min) Case 2: αit = current cash balance + median of cash balance of firms in the same industry-year combination C(min) Case 3: αit = current cash balance + median of historical cash balance of the same firm A dummy variable representing firms’ funding sources for each firm i in quarter i. , It takes variables from 1-5 where: 1 = shares, 2= profit-loss sharing sukuk(PLS), 3= fixed income sukuk (FIS), 4= zero-coupon sukuk (ZCS), 5= do nothing A dummy variable representing firms’ funding sources for each firm i in quarter i. , It takes variables from 1-7 where: 1 = conventional bonds, 2= profit-loss sharing sukuk(PLS), 3= fixed income sukuk (FIS), 4= zero-coupon sukuk (ZCS), 5= bank loans, 6= do nothing Measure This table presents variables definitions. Dependent variables and deficit constraints (αit and αit ) are defined for each pecking order case in each stage. Definitions of firm characteristics used as control variables according to Dahiya et al. (2017); Frank and Goyal (2009). Credit Strength Growth Opportunity Liquidity Economic Condition Sector (B) Firm-specific characteristics k=1 Xit Firm Size Tangibility Book Leverage Profitability D(max) αit αit C(min) 2. Independent Variable (A) Variable of interest (B) Stage 2 Extf undsit 1. Dependent Variable (A) Stage 1 Extf undsit Variable Table 3: Variable Definitions
16. Table 4 : Descriptive statistics for firm characteristics Variable Mean Std. Dev. Min Max Observations Size overall between within 6.3347 1.5127 1.4604 0.4874 1.5490 2.8541 3.8083 10.4446 10.1015 8.3940 N = 4815 n = 108 T-bar= 44.5833 Tangibility overall between within 0.2810 0.2303 0.2112 0.0971 0.0000 0.0000 -0.1826 0.9367 0.7854 0.7934 N = 4815 n = 108 T-bar= 44.5833 Growth Opportunity overall between within 1.2625 1.1983 1.0112 0.5828 0.1020 0.2217 -3.9296 13.2945 7.4355 7.6970 N = 4815 n = 108 T-bar= 44.5833 Profitability overall between within 0.0159 0.0327 0.0161 0.0291 -0.7557 -0.0390 -0.7131 0.6869 0.1001 0.6864 N = 4686 n = 108 T-bar= 43.3889 Solvency overall between within 0.5363 0.7904 0.4572 0.6460 -30.6262 -1.5659 -28.5240 2.0820 1.5888 3.0290 N = 4815 n = 108 T-bar= 44.5833 Leverage overall between within 0.2786 0.1759 0.1459 0.1066 0.0000 0.0187 -0.2497 1.8258 0.7785 1.6635 N = 4815 n = 108 T-bar = 44.5833 GDP Growth overall between within 0.3843 7.3819 0.6352 7.3679 -48.5751 -3.1804 -48.6630 27.1032 1.6635 26.4003 N = 4986 n = 111 T-bar = 44.9189 Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. We generated quarterly financial information of 108 Malaysian firms from Compustat. Table 3 presents variables definitions. 10.2 Figures 16
17. 17 -0 .3568 overall between within overall between within overall between within overall between within overall between within overall between within overall between within overall between within αC(min) Size Tangibility Growth Opp. Profitability Solvency Leverage GDP Growth 3.6807 0.6690 3.6418 0.1798 0.1533 0.1106 0.9123 0.4854 0.7544 0.0367 0.0184 0.0327 1.3096 1.0883 0.5964 0.2185 0.2026 0.0934 1.3450 1.3984 0.4676 0.8763 0.6344 0.4255 Std. Dev. -48.5751 -1.5363 -46.6425 0.0000 0.0128 -0.2476 -30.6262 -1.6752 -28.4234 -0.7557 -0.0562 -0.7127 0.1255 0.2243 -4.8327 0.0000 0.0000 -0.1657 1.5490 2.8527 3.4398 -9.6716 -5.6358 -4.3925 Min 6.0221 3.5314 8.8376 1.8258 0.8442 1.6554 2.0820 1.6139 3.1296 0.6869 0.1056 0.6866 13.2945 7.4522 7.1928 0.9367 0.6862 0.7683 9.6885 9.5018 8.1211 -0.0003 -0.0161 3.0550 Max N = 3377 n = 106 T-bar = 31.8585 N = 3377 n = 106 T-bar = 31.8585 N = 3377 n = 106 T-bar= 31.8585 N = 3366 n = 106 T-bar = 31.7547 N = 3377 n = 106 T-bar= 31.8585 N = 3377 n = 106 T-bar = 31.8585 N = 3377 n = 106 T-bar = 31.8585 N = 3377 n = 106 bar = 31.8585 Observations GDP Growth Leverage Solvency Profitability Growth Opp. Tangibility Size αD(max) Variable overall between within overall between within overall between within overall between within overall between within overall between within overall between within overall between within 1.3384 0.3908 0.4086 0.0122 1.1873 0.2963 6.3716 -0.5639 Mean 3.2271 0.8532 3.2055 0.1493 0.1312 0.0896 0.9553 0.4901 0.8458 0.0374 0.0174 0.0348 0.9370 1.0852 0.3693 0.2391 0.2237 0.0913 1.5099 1.4514 0.4160 0.9493 0.6774 0.6148 Std. Dev. -7.8733 -4.8406 -8.1447 0.0000 0.0000 -0.1768 -30.6262 -1.9895 -28.4692 -0.7557 -0.0429 -0.7140 0.1700 0.2175 -2.1165 0.0000 0.0000 -0.1767 1.5490 3.6069 3.8606 -20.3795 -5.8382 -18.7014 Min 6.0221 3.2190 7.1833 1.8258 0.7998 1.7628 1.8428 1.3769 3.0838 0.6869 0.0932 0.6833 10.5061 7.3415 4.5709 0.9115 0.7941 0.6490 10.4446 10.3209 8.3876 -0.0002 -0.0525 3.4202 Max N = 2367 n = 102 T-bar = 23.2059 N = 2367 n = 102 T-bar = 23.2059 N = 2367 n = 102 T-bar = 23.2059 N = 2356 n = 99 T-bar = 23.798 N = 2367 n = 102 T-bar = 23.2059 N = 2367 n = 102 T-bar = 23.2059 N = 2367 n = 102 T-bar = 23.2059 N = 4383 n = 108 bar = 40.5833 Observations This table presents descriptive statistics for firms in the semi-liberal pecking order specification. In this pecking order version, deficit is constrained by firms savings and leverage levels are held at the industry averages. In the first panel we show firm characteristics when internal funds are exhausted and firms can choose between bonds and sukuk to obtain funds externally. The second panel shows firm characteristics when deficit is constrained by maximum debt capacity, and firms can raise funds by issuing sukuk or equity. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. We generated quarterly financial information of 108 Malaysian firms from Compustat. 1.2792 0.2792 0.5277 0.0150 1.3123 0.2567 5.9582 Mean Variable Semi-liberal Pecking order Table 5: Descriptive statistics of firms in the semi-liberal pecking order specification
18. 18 -0 .1274 overall between within overall between within overall between within overall between within overall between within overall between within overall between within overall between within αC(min) Size Tangibility Growth Opportunity Profitability Solvency Leverage GDP Growth 3.7090 0.5761 3.6797 0.1779 0.1524 0.1010 0.6282 0.4392 0.4337 0.0298 0.0168 0.0265 1.1556 0.9940 0.5180 0.2193 0.2068 0.0903 1.4982 1.4584 0.4550 0.1222 0.0832 0.0878 Std. Dev. -48.5751 -1.7785 -47.5494 0.0000 0.0061 -0.2710 -15.2623 -0.8384 -14.2406 -0.3764 -0.0518 -0.3642 0.1128 0.2190 -3.2854 0.0000 0.0000 -0.1874 1.7410 2.8527 3.9671 -1.0330 -0.7131 -0.7611 Min 6.0221 2.7602 7.8344 1.8258 0.8272 1.7033 2.0820 1.6147 2.4895 0.6869 0.0844 0.6829 13.2945 7.4141 8.3002 0.9367 0.7407 0.7460 10.4446 10.2986 8.4590 -5.22E-06 -0.0216 0.5776 Max N = 3516 n = 108 T-bar = 32.5556 N = 3516 n = 108 T-bar = 32.5556 N = 3516 n = 108 T-bar = 32.5556 N = 3505 n = 108 T-bar = 32.4537 N = 3516 n = 108 T-bar = 32.5556 N = 3516 n = 108 T-bar = 32.5556 N = 3516 n = 108 T-bar = 32.5556 N = 3516 n = 108 bar = 32.5556 Observations GDP Growth Leverage Solvency Profitability Growth Opportunity Tangibility Size αD(max) Variable overall between within overall between within overall between within overall between within overall between within overall between within overall between within overall between within 1.2688 0.2836 0.5381 0.0153 1.2667 0.2814 6.3679 -0.3792 Mean 3.6343 0.2961 3.6238 0.1755 0.1484 0.1035 0.8111 0.4707 0.6626 0.0330 0.0166 0.0293 1.2047 1.0215 0.5664 0.2298 0.2119 0.0954 1.5236 1.4656 0.4606 0.2639 0.167 0.2142 Std. Dev. -48.5751 -0.2721 -47.5521 0.0000 0.0224 -0.2839 -30.6262 -1.7484 -28.3397 -0.7557 -0.0429 -0.7109 0.1128 0.2190 -4.0835 0.0000 0.0000 -0.2045 1.5490 2.8527 3.8569 -7.2945 -1.0376 -6.8044 Min 6.0221 1.8845 7.1148 1.8258 0.7998 1.6557 2.0820 1.6147 3.2133 0.6869 0.0997 0.6864 13.2945 7.4141 7.5431 0.9367 0.7825 0.8031 10.4446 10.1168 8.4343 -0.0023 -0.0690 0.2639 Max N = 4416 n = 108 T-bar = 40.8889 N = 4416 n = 108 T-bar = 40.8889 N = 4416 n = 108 T-bar = 40.8889 N = 4405 n = 108 bar = 40.787 N = 4416 n = 108 T-bar = 40.8889 N = 4416 n = 108 T-bar = 40.8889 N = 4416 n = 108 T-bar = 40.8889 N = 4416 n = 108 bar = 40.8889 Observations This table presents descriptive statistics for firms in the liberal pecking order specification. In this pecking order version, saving and leverage levels are determined by firms individually . In the first panel we show firm characteristics when internal funds are exhausted and firms can choose between bonds and sukuk to obtain funds externally. The second panel shows firm characteristics when deficit is constrained by maximum debt capacity, and firms can raise funds by issuing sukuk or equity. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. We generated quarterly financial information of 108 Malaysian firms from Compustat. 1.2715 0.2745 0.5596 0.0151 1.2727 0.2617 6.4114 Mean Variable Liberal Pecking order Table 6: Descriptive statistics of firms in the liberal pecking order specification
19. 19 0 .0388 0.4236 0.0218 -0.0273 0.1296 0.1752 0.0673 0.0017 1 0.048 0.3671 0.0085 -0.0353 0.2404 0.4462 -0.1829 0.0069 1 SemiliberalαD(max) 0.0816 0.058 0.1021 -0.0224 0.0794 -0.0461 0.1188 0.012 1 Liberal αC(min) 1 0.0761 -0.0046 0.0164 -0.0155 0.2544 0.3254 -0.4958 0.0222 Liberal αD(max) 1 0.053 -0.3187 -0.0209 0.0609 -0.1034 -0.0599 0.0115 Sector 1 0.2173 0.019 0.0849 0.0183 0.0792 0.0048 L.Size 1 0.0923 0.0121 -0.0066 0.0932 -0.0032 L.Tangibility 1 0.2247 0.0099 -0.0228 0.0075 L.Growth Opp. 1 0.5172 -0.2044 0.0378 L.Profit. 1 -0.3879 0.0067 L.Solvency 1 0.0089 L.Leverage 1 GDP Growth This table shows the correlation between firm characteristics and constrained deficit of each stage of the two versions of the pecking order. Variables definitions are presented in table 3. Semi-liberal αC(min) = current cash balance + median of cash balance of firms in the same industry-year combination, Semi-liberal αD(max) = debt capacity is equal to the debt ratio of the industry’s investment grade companies, Liberal αC(min) = current cash balance + median of historical cash balance of the same firm, and Liberal αD(max) = the annual average debt ratio of the firm. Sector is a dummy variables taking values from 1 to 10 representing GIC sectors. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. Table 3 presents variables definitions. Semi-liberal αC(min) Semi-liberal αD(max) Liberal αC(min) Liberal αD(max) Sector Size Tangibility Growth Opp. Profitability Solvency Leverage GDP Growth Semi-liberal αC(min) Table 7: Correlation matrix
20. Table 8 : Marginal effects of constrained deficit (semi-liberal specification - stage 1) Instrument Bonds FIS PLS ZCS Bank Loans Do Nothing Observations (1) (2) αC(min) αC(min) + Firm Characteristics Pooled M.logit S.E cluster-firm S.E cluster-sector Pooled S.E cluster-firm S.E cluster-firm/ sector charact. S.E cluster-sector 0.0148 (0.00936) 0.00937 (0.00747) 0.00708 (0.00667) 0.0295** (0.0131) -0.0224** (0.00983) -0.0384*** (0.0145) 2,763 0.0148 (0.0239) 0.00937 (0.0186) 0.00708 (0.0101) 0.0295 (0.0187) -0.0224 (0.0203) -0.0384 (0.0371) 2,763 0.0148 (0.0114) 0.00937 (0.0113) 0.00708 (0.0167) 0.0295 (0.0203) -0.0224 (0.0156) -0.0384 (0.0259) 2,763 0.00760 (0.0104) 0.00196 (0.00741) -0.00563 (0.00389) 0.0284 (0.0198) -0.0436*** (0.0137) 0.0112 (0.0190) 2,759 -0.000924 (0.0130) -0.00101 (0.00983) -0.00662* (0.00352) 0.0556** (0.0259) -0.0394** (0.0180) -0.00767 (0.0312) 2,759 0.00760 (0.0102) 0.00196 (0.0121) -0.00563* (0.00337) 0.0284* (0.0168) -0.0436*** (0.0132) 0.0112 (0.0188) 2,759 0.00760 (0.0162) 0.00196 (0.00874) -0.00563 (0.00457) 0.0284 (0.0282) -0.0436*** (0.0133) 0.0112 (0.0329) 2,759 This table reports the impact of constrained deficit (αC(min) ) on the probability of issuing each external funding source. It reports marginal effects of the minimum cash holding condition under semi-liberal pecking order. In this stage, firms hold savings equal to the industry average. Probabilities are obtained from estimating Equation (3). Variables definitions are shown in table 3. Funding tools are: bonds, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), bank loans, and doing nothing. In (1) constrained deficit is the only explanatory variable. In (2) we add firm characteristics as explanatory variables. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. Table 9: Marginal effects of constrained deficit (semi-liberal specification - stage 2) Instrument Equity FIS PLS ZCS Do Nothing Observations (1) (2) αD(max) αD(max) + Firm Characteristics Pooled M.logit S.E cluster-firm S.E cluster-sector Pooled S.E cluster-firm S.E cluster-firm/ sector charact. S.E cluster-sector 0.251*** (0.0422) 0.121*** (0.0314) 0.0257 (0.0162) 0.0101 (0.0189) -0.408*** (0.0416) 1,520 0.251 (0.174) 0.121* (0.0624) 0.0257 (0.0391) 0.0101 (0.0344) -0.408*** (0.153) 1,520 0.251* (0.136) 0.121*** (0.0412) 0.0257 (0.0423) 0.0101 (0.0231) -0.408*** (0.121) 1,520 -0.0387 (0.0477) 0.252*** (0.0585) -0.0175** (0.00840) 0.0360 (0.0284) -0.232*** (0.0534) 1,517 -0.0387 (0.126) 0.252** (0.125) -0.0175 (0.0133) 0.0360 (0.0399) -0.232* (0.123) 1,517 -0.0387 (0.0652) 0.252*** (0.0877) -0.0175 (0.0112) 0.0360 (0.0284) -0.232*** (0.0718) 1,517 -0.0387 (0.0757) 0.252*** (0.0838) -0.0175 (0.0157) 0.0360 (0.0564) -0.232* (0.138) 1,517 This table reports the impact of constrained deficit (αD(max) ) on the probability of issuing each external funding source. It reports marginal effects of the maximum debt capacity condition under semi-liberal pecking order. In this stage, firm maintain debt ratio equal to the industry average. Probabilities are obtained from estimating Equation (7).Variables definitions are shown in table 3. Funding tools are: equity, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), and doing nothing. In (1) constrained deficit is the only explanatory variable. In (2) we add firm characteristics as explanatory variables. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. Table 10: Marginal effects of constrained deficit (liberal specification - stage 1) VARIABLES Bonds Pooled M.logit -0.00596 (0.0339) FIS 0.0691 (0.0458) PLS -0.00157 (0.0241) ZCS 0.176*** (0.0538) Bank Loans -0.359*** (0.0793) Do Nothing 0.121 (0.0953) Observations 2,814 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 (1) (2) αC(min) αC(min) + Firm Characteristics S.E cluster-firm S.E cluster-sector Pooled S.E cluster-firm S.E cluster-firm/ sector charact. S.E cluster-sector -0.00596 (0.0603) 0.0691 (0.0586) -0.00157 (0.0319) 0.176** (0.0724) -0.359*** (0.120) 0.121 (0.163) 2,814 -0.00596 (0.0821) 0.0691 (0.0547) -0.00157 (0.0426) 0.176*** (0.0510) -0.359*** (0.0668) 0.121 (0.0974) 2,814 0.0274 (0.0394) 0.0494 (0.0497) -0.0119 (0.0279) 0.141*** (0.0493) -0.494*** (0.0879) 0.288*** (0.102) 2,808 0.0274 (0.0527) 0.0494 (0.0644) -0.0119 (0.0430) 0.141** (0.0613) -0.494*** (0.141) 0.288* (0.172) 2,808 0.0274 (0.0306) 0.0494 (0.0528) -0.0119 (0.0248) 0.141*** (0.0441) -0.494*** (0.0884) 0.288*** (0.101) 2,808 0.0274 (0.0552) 0.0494 (0.0470) -0.0119 (0.0526) 0.141*** (0.0436) -0.494*** (0.105) 0.288* (0.154) 2,808 This table reports the impact of constrained deficit (αC(min) ) on the probability of issuing each external funding source. It reports marginal effects of the minimum cash holding condition under liberal pecking order. In this stage, firms hold savings equal to the firms’ historical averages. Probabilities are obtained from estimating Equation (3). Variables definitions are shown in table 3. Funding tools are: bonds, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), bank loans, and doing nothing. In (1) constrained deficit is the only explanatory variable. In (2) we add firm characteristics as explanatory variables. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. 20
21. Table 11 : Marginal effects of constrained deficit (liberal specification - stage 2) VARIABLES Equity FIS PLS ZCS Do Nothing Observations (1) (2) αD(max) αD(max) + Firm Characteristics Pooled M.logit S.E cluster-firm S.E cluster-sector Pooled S.E cluster-firm S.E cluster-firm/ sector charact. S.E cluster-sector 0.0242 (0.0386) 0.0900*** (0.0272) 0.0146 (0.0153) 0.0616** (0.0263) -0.190*** (0.0460) 2,675 0.0242 (0.0870) 0.0900** (0.0405) 0.0146 (0.0244) 0.0616* (0.0370) -0.190** (0.0890) 2,675 0.0242 (0.0880) 0.0900*** (0.0318) 0.0146 (0.0256) 0.0616* (0.0369) -0.190** (0.0820) 2,675 -0.00197 (0.0658) 0.241*** (0.0529) 0.0784** (0.0330) 0.317*** (0.0501) -0.635*** (0.0829) 2,668 -0.00197 (0.114) 0.241** (0.0951) 0.0784 (0.0526) 0.317*** (0.0714) -0.635*** (0.134) 2,668 -0.00197 (0.0808) 0.241*** (0.0879) 0.0784** (0.0379) 0.317*** (0.0550) -0.635*** (0.105) 2,668 -0.00197 (0.124) 0.241** (0.112) 0.0784 (0.0755) 0.317*** (0.0825) -0.635*** (0.0867) 2,668 This table reports the impact of constrained deficit (αD(max) ) on the probability of issuing each external funding source. It reports marginal effects of the maximum debt capacity condition under liberal pecking order. In this stage, firms maintain debt capacity equal to firms’ historical averages. Probabilities are obtained from estimating Equation( 7).Variables definitions are shown in table 3. Funding tools are: equity, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), and doing nothing. In (1) constrained deficit is the only explanatory variable. In (2) we add firm characteristics as explanatory variables. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. Figure 1: Leary and Roberts (2010) illustration of the pecking order theory funding resources hierarchy 21
22. Table 12 : Marginal effects of firm characteristics in the semi-liberal specification Stage 1 Instrument Bonds FIS PLS ZCS Bank Loans Do Nothing Observations Size Tangibility Growth Profitability Solvency Leverage Liquidity GDP 0.00957*** (0.00261) 0.0130*** (0.00360) 0.00835*** (0.00213) -0.00492 (0.00479) 0.00308 (0.00859) -0.0290*** (0.00952) 2,759 0.0337* (0.0187) -0.00760 (0.0209) -0.0141 (0.0140) -0.0296 (0.0287) 0.00999 (0.0514) 0.00756 (0.0562) 2,759 -0.00372 (0.00337) 0.00480** (0.00243) -0.00150 (0.00380) -0.0176** (0.00683) 0.0109 (0.00826) 0.00708 (0.00955) 2,759 -0.0305 (0.142) -0.158* (0.0917) -0.0287 (0.0983) -0.122 (0.107) 0.129 (0.345) 0.210 (0.343) 2,759 0.00547 (0.00672) -0.00787** (0.00378) 0.0175** (0.00847) 0.0131 (0.0118) 0.0912*** (0.0289) -0.119*** (0.0275) 2,759 0.0290 (0.0198) -0.0143 (0.0248) 8.08e-05 (0.0136) 0.130*** (0.0258) -0.325*** (0.0620) 0.180*** (0.0657) 2,759 0.00467*** (0.00116) -0.00568 (0.00423) -0.00818** (0.00416) 0.00442** (0.00220) -0.0434*** (0.0101) 0.0481*** (0.00956) 2,759 -1.53e-06** (6.75e-07) 3.83e-07 (2.26e-06) -1.99e-07 (4.13e-07) -2.60e-06** (1.24e-06) -2.28e-06 (2.80e-06) 6.23e-06* (3.20e-06) 2,759 Size Tangibility Growth Profitability Solvency Leverage Liquidity GDP 0.0689*** (0.0119) -0.00395 (0.00789) 0.0334*** (0.00767) -0.0201*** (0.00682) -0.0783*** (0.0137) 1,517 0.0479 (0.0522) 0.0650** (0.0276) -0.0507** (0.0219) 0.0908** (0.0405) -0.153** (0.0641) 1,517 -0.0466*** (0.0120) 0.0302*** (0.00489) 0.000581 (0.00417) -0.0257*** (0.00992) 0.0415*** (0.0149) 1,517 2.093*** (0.637) -1.067*** (0.262) 0.279 (0.188) -0.407* (0.215) -0.898 (0.619) 1,517 -0.0301 (0.0365) 0.00854 (0.0235) -0.0351*** (0.0102) 0.0575*** (0.0201) -0.000777 (0.0389) 1,517 -0.187* (0.108) 0.105 (0.114) 0.0303 (0.0306) 0.0661 (0.0592) -0.0137 (0.124) 1,517 0.0146* (0.00882) -0.0104 (0.0104) 0.00698*** (0.00202) -0.00249 (0.00359) -0.00872 (0.0107) 1,517 -8.10e-06** (3.96e-06) 3.18e-06 (2.51e-06) -3.92e-08 (1.29e-06) -6.72e-06*** (2.59e-06) 1.17e-05** (4.81e-06) 1,517 Stage 2 Instrument Equity FIS PLS ZCS Do Nothing Observations This table reports the impact of firm characteristics and macroeconomic factors on the probability of issuing each external funding source under the semi-liberal pecking order specification. In this specification, firms hold savings and debt capcity equal to the industry average. In stage 1 probabilities are obtained from estimating Equation (3). Variables definitions are shown in table 3. Funding tools are: bonds, Fixed income sukuk (FIS), profit-loss-sharing (PLS) sukuk, zero-coupon sukuk (ZCS), bank loans, and doing nothing. In stage 2 probabilities are obtained from estimating Equation (7). Variables definitions are shown in Table 3. Funding tools are: equity, Fixed income sukuk (FIS), profit-loss-sharing (PLS) sukuk, zero-coupon sukuk (ZCS), and doing nothing. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. 22
23. Table 13 : Marginal effects of firm characteristics in the liberal specification Stage 1 Instrument Bonds FIS PLS ZCS Bank Loans Observations Size Tangibility Growth Profitability Solvency Leverage Liquidity GDP 0.00921*** (0.00257) 0.0135*** (0.00273) 0.0132*** (0.00270) -0.00623** (0.00299) -0.0360*** (0.00631) 2,808 5.92e-06 (0.0200) 0.00543 (0.0225) -0.0236 (0.0191) -0.0165 (0.0247) 0.0686 (0.0502) 2,808 -0.0194*** (0.00514) 0.00757*** (0.00289) 0.00444 (0.00271) -0.0127** (0.00572) -0.00130 (0.00918) 2,808 -0.162 (0.139) -0.172 (0.122) -0.181** (0.0884) -0.0659 (0.0868) 0.470 (0.333) 2,808 0.0101 (0.00911) -0.00806* (0.00432) 0.0155* (0.00918) -0.00238 (0.00635) 0.0873*** (0.0271) 2,808 0.0643*** (0.0213) -0.0294 (0.0259) 0.0168 (0.0209) 0.0521*** (0.0195) -0.246*** (0.0590) 2,808 0.00425*** (0.00117) -0.00279 (0.00465) -0.00199 (0.00400) 0.00511*** (0.00165) -0.0575*** (0.0105) 2,808 -1.66e-06*** (6.33e-07) 8.31e-07 (2.52e-06) -1.26e-07 (8.57e-07) -2.16e-06* (1.14e-06) -6.19e-07 (2.85e-06) 2,808 Stage 2 Instrument Equity FIS PLS ZCS Do Nothing Do Nothing Observations Size Tangibility Growth Profitability Solvency Leverage Liquidity GDP 0.0472*** (0.00583) 0.0103*** (0.00303) 0.0135*** (0.00314) -0.0197*** (0.00346) -0.0513*** (0.00672) 0.00634 (0.00708) 2,668 -0.0819* (0.0436) 0.0338 (0.0231) -0.0141 (0.0157) 0.00626 (0.0273) 0.0559 (0.0504) -0.0340 (0.0557) 2,668 -0.00470 (0.00636) 0.0118*** (0.00331) 0.00209 (0.00456) -0.0158** (0.00797) 0.00655 (0.0105) 0.0214** (0.0106) 2,668 1.026** (0.494) -0.287 (0.180) -0.0368 (0.167) -0.0837 (0.173) -0.619 (0.508) 0.110 (0.347) 2,668 0.0365 (0.0241) -0.0243* (0.0130) 0.000705 (0.0113) 0.0110 (0.0145) -0.0239 (0.0279) -0.103*** (0.0256) 2,668 0.101 (0.0931) 0.167* (0.0897) 0.0926** (0.0393) 0.354*** (0.0521) -0.714*** (0.113) 0.142** (0.0637) 2,668 0.00779* (0.00456) -0.00819 (0.00716) 0.00259 (0.00316) -0.00371 (0.00286) 0.00152 (0.00634) 0.0530*** (0.01000) 2,668 -4.28e-06* (2.59e-06) 7.09e-07 (2.69e-06) -4.23e-07 (8.31e-07) -3.32e-06 (2.42e-06) 7.32e-06** (3.64e-06) 3.74e-06 (3.44e-06) 2,668 This table reports the impact of firm characteristics and macroeconomic factors on the probability of issuing each external funding source under the liberal pecking order specification. In this version, firms savings and debt capacity are firm determinants. They are equal to individuals firms’ historical savings and leverage averages. In stage 1 probabilities are obtained from estimating Equation (3). Variables definitions are shown in Table 3. Funding tools are: bonds, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), bank loans, and doing nothing. In stage 2 probabilities are obtained from estimating Equation (7). Variables definitions are shown in table 3. Funding tools are: equity, Fixed income sukuk (FIS), profit-loss-sharing(PLS) sukuk, zero-coupon sukuk (ZCS), and doing nothing. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. Standard errors in parentheses. Significance levels are represented by: *** p<0.01, ** p<0.05, * p<0.1. 23
24. Table 14 : Expected and observed signs Stage 1 Expected Signs Semi-liberal pecking order Liberal pecking order Determinants Sukuk Bonds FIS PLS ZC Bonds FIS PLS ZC Bonds Size Tangibility Growth Profitability Solvency Leverage Liquidity (-) (-/+) (-) (-) (-) (-) (-/+) (+) (+) (+) (+) (+) (+) (-/+) (+) (+) (+) (-) (-) (-) (-) (+) (+) (+) (-) (+) (-) (-) (-) (-) (-) (-) (+) (+) (+) (+) (+) (-) (-) (+) (+) (+) (+) (+) (+) (-) (-) (-) (-) (+) (-) (+) (-) (+) (-) (-) (-) (-) (-) (-) (-) (+) (+) (+) (-) (-) (-) (+) (+) (+) Stage 2 Size Tangibility Growth Profitability Solvency Leverage Liquidity Expected Signs Semi-liberal pecking order Liberal pecking order Sukuk Equity FIS PLS ZC Equity FIS PLS ZC Equity (+) (+) (+) (+) (+) (+) (-/+ ) (-) (-/+) (-) (-) (-) (-) (-/+) (+) (+) (+) (-) (+) (+) (-) (+) (-) (+) (+) (-) (+) (+) (-) (+) (-) (-) (+) (+) (-) (+) (+) (-) (+) (-) (-) (+) (+) (+) (+) (-) (-) (+) (-) (+) (-) (+) (+) (-) (+) (+) (-) (-) (-) (-) (+) (+) (-) (+) (-) (-) (+) (+) (+) (+) In this table we present the expected and obtained signs of the relationship between source of funding and firm characteristics. In stage 1, firms exhaust internal funds and are to choose between the issuances of sukuk types and conventional bonds. In stage 2, firms reach maximum debt capacity and can raise funds by issuing sukuk or equity. In the semi-liberal specification, deficit is constrained by the industrial average of savings and leverage. In the liberal version, firms savings and leverage are firm determinant and calculated as annual historical averages. In the first panel, we show and compare the expected direction of the affect of firm characteristics on the issuance of sukuk types and conventional bonds. In the second panel, we show and compare the expected direction of the affect of firm characteristics on the issuance of sukuk types and shares. Firm size is measured by the natural logarithm of total assets. Tangibility is measured by the ratio of total fixed assets to total assets. We measure growth opportunity as [Total Assets +Book Equity + (No of Shares outstanding * Share price)]/Total Assets. The return on assets ratio (ROA) is used to measure firm profitability. We measure solvency or credit solvency by Altman z-score. Leverage is the total debt to total asset ratio. We use GDP growth as a macroeconomic control variable. 24
25. Figure 2 : Average frequency and volume of external funding sources over sample period 25
26. Figure 3 : Average issuance frequency and volume of each sukuk type. FIS = Fixed Income Sukuk, PLS = Profit-Loss-Sharing Sukuk, ZC = Zero-Coupon Sukuk 26
27. Figure 4 : Marginal effects of deficit in the semi-liberal pecking order theory. αC(min) = lower bound of savings, αD(max) = upper bound of debt capacity , FIS = Fixed Income Sukuk, PLS = Profit-Loss-Sharing Sukuk, ZC = Zero-Coupon Sukuk 27
28. Figure 5 : Marginal effects of deficit in the liberal pecking order theory. αC(min) = lower bound of savings, αD(max) = upper bound of debt capacity , FIS = Fixed Income Sukuk, PLS = Profit-Loss-Sharing Sukuk, ZC = Zero-Coupon Sukuk 28
29. 29 αC(min) = lower bound of savings, αD(max) = upper bound of debt capacity , FIS = Fixed Income Sukuk, PLS = Profit-Loss-Sharing Sukuk, ZC = Zero-Coupon Sukuk Figure 6: Average marginal effects of firm characteristics in the semi-liberal pecking order.
30. 30 αC(min) = lower bound of savings, αD(max) = upper bound of debt capacity , FIS = Fixed Income Sukuk, PLS = Profit-Loss-Sharing Sukuk, ZC = Zero-Coupon Sukuk Figure 7: Average marginal effects of firm characteristics in the liberal pecking order.
31. References AAOIFI (2008), Accounting, Auditing and Governance Standards for Islamic Financial Institutions, The Accounting and Auditing Organization for Islamic Financial Institutions (AAOIFI), Manama, Bahrain. Abdul Halim, Z., How, J. and Verhoeven, P. (2017), ‘Agency costs and corporate sukuk issuance’, Pacific Basin Finance Journal 42, 83–95. Abedifar, P., Bouslah, K., Hashem, S. Q. and Song, L. (2020), ‘How informative are stock prices of islamic banks?’, Journal of International Financial Markets, Institutions and Money p. 101203. Adam, T. and Goyal, V. K. (2008), ‘The investment opportunity set and its proxy variables’, Journal of Financial Research 31(1), 41–63. Azmat, S., Skully, M. and Brown, K. (2014), ‘Issuer’s choice of Islamic bond type’, Pacific Basin Finance Journal 28, 122–135. Baker, M. and Wurgler, J. (2002), ‘Market timing and capital structure’, The journal of finance 57(1), 1–32. Baskin, J. (1989), ‘An empirical investigation of the pecking order hypothesis’, Financial management 18(1), 26–35. Booth, L., Aivazian, V., Demirguc-Kunt, A. and Maksimovic, V. (2001), ‘Capital structures in developing countries’, The journal of finance 56(1), 87–130. Cameron, A. C. and Trivedi, P. K. (2010), Microeconometrics using stata, Vol. 2, Stata press College Station, TX. Chen, J. J. (2004), ‘Determinants of capital structure of chinese-listed companies’, Journal of Business research 57(12), 1341–1351. Dahiya, S., Iannotta, G. and Navone, M. (2017), ‘Firm Opacity Lies in the Eye of the Beholder’, Financial Management 46(3), 553–592. Davidson, R., MacKinnon, J. G. et al. (1993), Estimation and inference in econometrics, Oxford University Press. De Jong, A., Kabir, R. and Nguyen, T. T. (2008), ‘Capital structure around the world: The roles of firm-and countryspecific determinants’, Journal of Banking & Finance 32(9), 1954–1969. Demirg¨ uc¸-Kunt, A. and Maksimovic, V. (1999), ‘Institutions, financial markets, and firm debt maturity’, Journal of Financial Economics 54(3), 295–336. Denis, D. J. and Mihov, V. T. (2003), ‘The choice among bank debt, non-bank private debt, and public debt: evidence from new corporate borrowings’, Journal of financial Economics 70(1), 3–28. Ebrahim, M. S., Girma, S., Shah, M. E. and Williams, J. (2014), ‘Dynamic capital structure and political patronage: The case of malaysia’, International Review of Financial Analysis 31, 117–128. Ebrahim, M. S., Jaafar, A., Omar, F. A. and Salleh, M. O. (2016), ‘Can Islamic injunctions indemnify the structural flaws of securitized debt?’, Journal of Corporate Finance 37, 271–286. El Qorchi, M. (2005), ‘Islamic finance gears up’, Finance and Development A quarterly magazine of the IMF 42(4), 46–52. Elliott, W. B., Ko¨ eter-Kant, J. and Warr, R. S. (2008), ‘Market timing and the debt–equity choice’, Journal of Financial Intermediation 17(2), 175–197. Fama, E. F. and French, K. R. (2005), ‘Financing decisions: Who issues stock?’, Journal of Financial Economics 76(3), 549–582. Fan, J. P., Titman, S. and Twite, G. (2012), ‘An international comparison of capital structure and debt maturity choices’, Journal of Financial and quantitative Analysis 47(1), 23–56. Flannery, M. J. and Hankins, K. W. (2013), ‘Estimating dynamic panel models in corporate finance’, Journal of Corporate Finance 19(1), 1–19. Frank, M. Z. and Goyal, V. K. (2007), ‘Trade-Off and Pecking Order Theories of Debt’, Handbook of Empirical Corporate Finance SET 1(06), 135–202. Frank, M. Z. and Goyal, V. K. (2009), ‘Capital structure decision: Which factors are reliably important’, Financial Management 38(1), 1–37. Freese, J., Long, J. S. et al. (2000), ‘Tests for the multinomial logit model’, Stata Technical Bulletin 10(58), 1–10. Godlewski, C. J., Turk-Ariss, R. and Weill, L. (2011), ‘Do markets perceive sukuk and conventional bonds as different financing instruments?’, Discussion Paper . Graham, J. R. and Leary, M. T. (2011), ‘A review of empirical capital structure research and directions for the future’, Annu. Rev. Financ. Econ. 3(1), 309–345. Graham, J. R., Leary, M. T. and Roberts, M. R. (2015), ‘A century of capital structure: The leveraging of corporate America’, Journal of Financial Economics 118(3), 658–683. 31
32. Grassa , R. and Miniaoui, H. (2017), ‘Corporate choice between conventional bond and Sukuk issuance evidence from GCC countries’, Research in International Business and Finance (October 2016), 1–13. Greene, W. H. (2012), Econometric analysis, internationl edition,7th edn, Pearson. Harris, M. and Raviv, A. (1990), ‘Capital structure and the informational role of debt’, The Journal of Finance 45(2), 321– 349. Helwege, J. and Liang, N. (1996), ‘Is there a pecking order? evidence from a panel of IPO firms’, Journal of financial economics 40(3), 429–458. Jensen, M. C. (1986), ‘Agency costs of free cash flow, corporate finance, and takeovers’, The American economic review 76(2), 323–329. Klein, P. O. and Weill, L. (2016), ‘Why do companies issue sukuk?’, Review of Financial Economics 31, 26–33. Kleinbaum, D. G. and Klein, M. (2010), Logistic Regression: A Self-Learning Text, Springer Science & Business Media. Leary, M. T. and Roberts, M. R. (2010), ‘The pecking order, debt capacity, and information asymmetry’, Journal of Financial Economics 95(3), 332–355. Lemmon, M. L., Roberts, M. R., Zender, J. F., Lemmon, L., Roberts, M. R. and Zender, J. F. (2015), ‘American Finance Association Back to the Beginning : Persistence and the Cross-Section of Corporate Capital Structure All use subject to JSTOR Terms and Conditions Persistence Back to the Beginning : of Corporate Capital and the Structure’, 63(4), 1575–1608. Long, J. S. (2009), ‘Group comparisons in logit and probit using predicted probabilities’, Department of Sociology, University of Indiana. Unpublished manuscript . Long, J. S. and Freese, J. (2006), Regression models for categorical dependent variables using Stata, Stata press. Mohamed, H. H., Masih, M. and Bacha, O. I. (2014), ‘Why do issuers issue Sukuk or conventional bond? Evidence from Malaysian listed firms using partial adjustment models’, Pacific Basin Finance Journal 34, 233–252. Myers, S. C. (1984), ‘The capital structure puzzle’, The journal of finance 39(3), 574–592. Myers, S. C. and Majluf, N. S. (1984), ‘Corporate financing and investment decisions when firms have information that investors do not have’, Journal of Financial Economics 13(2), 187–221. Nagano, M. (2010), ‘Islamic Finance and the Theory of Capital Structure’, Munich Personal RePEc Archive (24567). Nagano, M. (2016), ‘Who issues Sukuk and when?: An analysis of the determinants of Islamic bond issuance’, Review of Financial Economics 31, 45–55. Rajan, R. G. (1992), ‘Insiders and Outsiders: The Choice between Informed andh’s-Eength Debt’, The Journal of Finance XLVII(4), 1367–1400. Shyam-Sunder, L. and Myers, S. C. (1999), ‘Testing static tradeoff against pecking order models of capital structure’, Journal of financial economics 51(2), 219–244. Titman, S. and Tsyplakov, S. (2007), ‘A dynamic model of optimal capital structure’, Review of Finance 11(3), 401–451. Williams, R. (2012), ‘Using the margins command to estimate and interpret adjusted predictions and marginal effects’, The Stata Journal 12(2), 308–331. Wilson, R. (2008), ‘Innovation in the structuring of islamic sukuk securities’, Humanomics 24(3), 170–181. 32 View publication stats