AN OVERVIEW OF DETERMINANTS OF DEBT LEVERAGE ACROOS SECTORS: EVIDENCE FROM CHINESE A-SHARE LISTED FIMRS

Transcription

1 AN OVERVIEW OF DETERMINANTS OF DEBT LEVERAGE ACROOS SECTORS: EVIDENCE FROM CHINESE A-SHARE LISTED FIMRS Ying Yang University of Malaya, Malaysia Che Hashim Bin Hassan University of Malaya, Malaysia ABSTRACT This study examined the determinants of debt leverage for listed firms in China using a panel of 547 A-share listed firms during The main findings are: (1) total debt is mainly dominated by short-term debt, while long-term debt is lowly financed. The total and long-term debts increased but the short-term debt decreased; (2) the effects of determinants on long-term and short-term debts are differently identified; (3) the effects of determinants are different across sectors, indicating that specific sectorial factors are at work; (4) the debt leverage decision tended to follow a static trade-off theory with consideration for effects of non-debt tax shield and profits. In general, debt leverage decisions (both long-term and short-term) are mostly affected positively by profits at different levels. Firm size is the most important determinant and positively related with long-term debt, while non-debt tax shield contributes most to short-term decision negatively. JEL Classifications: O19, F15, B28 Keywords: Determinants of capital structure, A-share listed firms, China Corresponding Author s Address: fdmyang@yahoo.com or fdmyang@siswa.um.edu.my

2 1. INTRODUCTION In the last several decades, the determinants of debt leverage have been discussed persistently. Yet, considering the fact that the majority of the studies were conducted in developed countries, the findings cannot be applied to situations in developing countries. And in particular in China, where its economy is growing rapidly with many different characteristics from other countries. With reference to a review of related literature, and as far as can be determined, just a few studies have been conducted to examine the determinants of debt leverage for Chinese listed firms. For example, Huang and Song (2006) suggested that debt leverage of Chinese listed firms was negatively related to profitability, growth opportunities, managerial shareholding, and non-debt tax shield, while it was positively related to firm size and fixed assets. Bhabra, Liu, and Tirtiroglu (2008) suggested that listed firms in China were financed with a low long-term debt leverage. They also suggested that long-term debt leverage is negatively related to growth options and profitability, while it is positively related to tangibility and firm size. Qian, Tian, and Wirjanto found a positive correlation between debt leverage and firm size, state ownership, and tangibility in their study in 2009 (Qian, Tian, and Wirjanto, 2009). In contrast, a negative correlation is found to exist between debt leverage and growth, non-debt tax shield, volatility, and profitability. Qian, Tian, and Wirjanto (2009) also found that debt leverage was positively affected by lagged profitability in a small impact. In addition to that, they reminded that the impact of growth was extremely insignificant. The focus of this paper is on the determinants of debt leverage across sectors in China. Although past studies have tried to make more complex research on this topic, an understanding of the basic information is necessary for academic knowledge. In particular, study designed to examine debt leverage decision based on the different sectors in emerging countries is rare. Meanwhile, it is noticed that the second reform of listed firms in China completed at the end of 2007 had converted 97% non-tradable shares into tradable shares in A-share market. (Li, Wang, Cheung and Jiang, 2011). So it is also necessary to conduct a fresh investigation into A-share listed firms the largest segment in the Chinese stock market and a major development of listed firms in China. In contrast to past studies, a panel sample with the latest data from 2008 to 2012 is used in this study. The empirical study is conducted by using pooled ordinary least square (OLS) regression model. Based on a review of related literature, debt leverage of listed firms in China was extremely lowly financed with long-term debt and highly financed by short-term debt (Bhabra, Liu and Tirtiroglu, 2008). As a departure from past studies, total debt, long-term debt, and short-term debt were used as proxies of debt leverage respectively in this study and examined in detail. On the other hand, without disputing the importance of determinants in theory or other studies, this study set a limit to the independent variables to the seven common cited factors: non-debt tax shield, profits, firm size, growth, dividend payment, financial distress cost, and tangibility. Meanwhile, as a priority, employee productivity was used as our eighth variable and detailed explanation is presented in the next section. The regression test was performed on total debt leverage with full sample. Firstly, it is found that the main source of debt leverage is short-term debt amounting to almost 80% of total debt, and the increase in the level of long-term debt is higher than the increase in short-term debt. Secondly, the energy sector has the highest debt level, while the utility sector is found to have the highest long-term debt level. The consumer discretionary sector is ranked highest in terms of short-term debt. The telecommunication service sector is identified as the sector with the lowest long-term and short-term debts. It is noticed that this study does not include the financial sector because the financial sector is operated with a remarkably different financial structure as compared to the other sectors. On the determinants of debt leverage decision, this study shows that non-debt tax shield is the most important factor with a negative relationship with both long-term and short-term debt. The profit, as a factor, is also considered by most of sectors to have a positive relation with both long-term and short-term debts. This

3 study found that firm size has a positive effect on long-term debt, and a negative effect on short-term debt. It is also found that the impact of non-debt tax shield and profits on short-term debt is greater than on long-term debt, while the impact of firm size on long-term debt is greater compared to its impact on short-term debt. A regression test was then performed on the different sectors one by one. The results showed that profits have a positive effect on most sectors, except the consumer discretionary sector. In addition, firm size is positively to debt leverage and non-debt tax shield is negatively to debt leverage. Dividend paid is the most irrelevant variable in debt leverage decision for all of sectors, except the financial sector. The effect of employee productivity varies on debt leverage in the four sectors. The reminder of the paper is organized as follows: Section 2 is a review of related literature. Section 3 describes the data, variable construction, and research method. The discussion of empirical results of the study is presented in Section 4. The conclusion of the study is presented in Section DETERMINANTS OF DEBT LEVERAGE 2.1 Non-debt Tax Shield and Debt Leverage The fundamental theory of optimal capital structure is known as MM theory, established by Modigliani and Miller (1958). It assumes that capital structure is irrelevant to firm value under a perfect market. According to the theory, debt is the first factor involved related to the issue of determinants of capital structure decision. Modigliani and Miller (1963) renewed the MM theory and pointed out that debt would be an effective factor in cost saving of capital structure. Because the interests on debt paid before tax reduces the tax cost of firm profits. Based on the idea, the trade-off theory suggested that debt leverage would create more benefits than equity to firms due to the advantages of tax shield (Kraus and Litzenberger, 1973). This idea was academically supported by DeAngelo and Masulis (1980) and Fama and Franch (2002), and they suggested that debt leverage was negatively related to non-debt tax shield. In China, Huang and Song (2006) arrived at the same results by employing non-debt tax shield measured by depreciation and amortization to total assets. An (2012) suggested that foreign invested enterprises in China tended to raise more debt after the tax rate of income was reduced by the government in With reference to these findings, we employ non-debt tax shield as our first variable of debt leverage decision in this study. 2.2 Profits and Debt Leverage The pecking order theory suggested that firms tended to rely on internal sources initially and then debt to equity because of asymmetric information preference (Myers and Majluf, 1984). According to this theory, profitability is addressed as one important source of internal fund. That means, firms tend to raise less debt if the required fund can be fulfilled via internal source. Therefore the theory sees a negative association between leverage and profitability (e.g. Titman and Wessels, 1988; Rajan and Zingles, 1995; Wald, 1999). On the other hand, static trade-off theory suggests a positive relationship between debt leverage and profits. It argues that high profit firms tend to raise more debt, signaling a good prospect to public. In China, a negative relationship between leverage and profitability was suggested in past studies (Qian, Tian and Wirjanto, 2009; Huang and Song, 2006) Firm Size and Debt Leverage Frank and Goyal (2003) suggested that debt leverage in capital structure decision in large firms were explained by the pecking order theory which was rejected by small firms. This means that the financial behavior was differently explained in terms of firm size. In contrast, Fama and French (2005) suggested that small firms prefer equity to debt finance. Gatchev Spindt, and Tarhan (2009) also suggested that small firms tended to depend on equity financing. In China, a positive relationship between firm size and debt leverage was suggested

4 in past studies (Qian, Tian, and Wirjanto 2009; Tong and Green, 2005; Huang and Song, 2006). 2.4 Growth and Debt Leverage There is a strong correlation between growth, macroeconomic conditions and economic cycle. Under a stable economic situation and in the expansion phase of business cycle, firms grow faster with a prosperous prospect and tend to easily find equity support. On the other hand, it is also interrupted by economic instability in an opposite way (McConnell, Brue, and Flynn, 2009). In corporate finance, Myers and Majluf (1984) suggested that growth was negatively associated with debt leverage. They stated that, based on asymmetric information of growth opportunities, firms had easy access to equity fund. Since they were perceived to have more growth opportunities, and therefore, heading towards a prosperous future. Thus, a negative correlation between growth and debt leverage was detected while equity was positively correlated. In the case of China, a negative correlation between growth and debt leverage of listed firms was found by Huang and Song (2006), and Qian, Tian, and Wirjanto (2009) Dividend Paid and Debt Leverage Based on the effects of asymmetric information, Stiglitz (1981, 1983) suggested that there were two choices for firms in funding when they were faced with new investment opportunities. The first was to make dividend paid continuously by residual cash flow and then raising debt from external sources to make new investment. The second one was to reduce dividend paid so that new investment could be funded by using residual free cash flow without raising debt from outside. That is to say, the debt leverage decision of firms is also affected by the dividend payment. On the other hand, dividend was also discussed in both pecking order and static trade-off theories. With the assumption thatother effects remain constant, there is a positive relationship between profits and dividend paid. Fama and French (2002) attested to this idea and suggested that firm size was positively related to leverage and dividend. In China, a negative relationship between debt leverage and the both current and past dividend paid were suggested by Tong and Green (2005). 2.6 Financial Distress Cost and Debt Leverage Since 2008, the macroeconomic conditions around the world have changed due to the US subprime mortgage crisis. Ye, Liu, and Miao (2012) opined that the subprime crisis contagion has affected all of Asian stock markets including China, Japan, Korea, Hong Kong, except Taiwan. China View (2008) reported that exports and export-oriented firms in China faced a drop in demand and hence performance due to the crisis. Foreign trade growth rate dropped to below 20 percent and economic growth in Chinese main economic city such as Zhe Jiang slowed down significantly. That is to say, there is a potential risk for firms in debt payment due to poor performance. Therefore, the situation may have pushed firms to face greater risk of distress and bankruptcy. In theory, firms may be declared bankrupt for failing to pay back their debts. This also implies that firms with debt financing are facing the potential cost of financial distress (Brealey and Myers, 1984). Besides, the components of financial distress cost are addressed as direct and indirect costs. Direct costs include legal and related administrative expenses. Indirect costs are incurred when customers are unwilling to purchase products from the firms, or suppliers become reluctant to continue the business relationship (Lee, Koh and Kang, 2011). This means that financial distress costs are unavoidable immaterial of whether a firm goes bankrupt. This is because indirect costs may be the result of poor performance (Robichek and Myers, 1966). Lemmon and Zendar (2004) suggested that financial distress cost was a good predictor and well performed for a modified pecking order theory of debt decision.

5 2.7 Tangibility and Debt Leverage Tangibility is another variable widely studied in debt leverage decision by researchers. Jensen and Meckling (1976) predicted that debt financing was positively correlated with level of tangibility based on agency theory, as a high level of tangibility gives managers more chances and incentives to finance debt. On the other hand, studies have confirmed that tangibility increases lender s confidence in firms (Titman and Wessles, 1988; Rajan and Zingales, 1995; ad Frank and Goyal, 2009). The reason is that tangible assets are less risky as collaterals compared to intangible assets. So it is expected that firms with a higher level of tangibility are able to raise more debt, especially long-term debts. In China, a negative relationship between debt leverage and tangibility was suggested by Huang and Song (2006) and Qian, Tian, and Wirjanto (2009). In this study, due consideration is given to the situation of macroeconomic turmoil and the assumption that firms may be threatened by potential risk of failure in debt payment, and therefore, their tangibility is highly valued by lenders. 2.8 Employee Productivity and Debt Leverage Among the research conducted so far, Titman (1984) was the first to argue that the employee productivity is associated to firm leverage because the risk of switching cost incentives shareholders to consider firm s products. He examined the effect of debt leverage for capital structure on firm s liquidation decision and pointed out that firms producing unique product or having a bilateral relationship of custom-suppliers would hold a lower leverage. Kale et al. (2007) suggested that debt served as a bonding mechanism and effected employees productivity, as managers tried to avoid personal lost as a result of failure in debt pay back. Bae, Kang, and Wang (2011) examined stakeholder theory of capital structure by investigating the relationship between the firm and its employees. They identified that the employees were well treated by the firms with low debt ratios. However, in firms holding high debt ratios employees would not be that well treated. Thus, instead of using the standard equation to test as has been done by most of the researchers, and as far as can be determined, this paper would be the first study to include employee productivity in the investigation into debt leverage decision in China. 3. DATA AND METHODOLOGY Our data are supported by data suppliers Bloomberg and during the periods. After deleting missing values, there are a total of 547 firms with 2,735 observations initially tested as full sample. The determinants were then examined based on the different sectors one by one. The sectors are categorized as follows: consumer discretionary, consumer staples, energy, financials, health care, industrials, information technology, materials, telecommunication services, and utilities. Table 1 shows a detailed distribution of listed firms in full sample. It is noticed that telecommunication service sector is not tested by regression model, as there are only 2 firms in the panel and cannot match the sample requirement (Krejcie and Morgan, 1970) TABLE 1. DISTRIBUTION OF LISTED FIRMS IN FULL SAMPLE BY SECTORS

6 Sector No. of firms (Cross-sections included) Included observations Total pool (balanced) observations % of the full sample Consumer Discretionary % Consumer Staples % Energy % Financials % Health Care % Industrials % Information Technology % Materials % Telecommunication Services % Utilities % Full sample % TABLES 2. DEFINITIONS OF VARIABLES Variables Notation Defination Dependent variables: Leverage Independent variables: TD LD SD Total debt Total assets Long-term debt Total assets Short-term debt Total assets Non-debt tax shield NDTS (Deprecation+ amortization) total assets Profitability EBITDA Operating income before depreciation and amortization (EBITDA) Total equity Firm size SIZE Log (total assets at the end of the year) Growth GR Total sales in year (t) Total sales in year (t-1) Dividend payout DIVPD Dividends paid in year (t) Financial distress cost FDC (SD(operating profit) mean (operating profit) total assets) Tangibility TANG Fixed assets total assets Employee productivity EP Sales number of employees The panel data model and pooled ordinary least square (OLS) regression test are used in this study. According to Dougherty (2011), the panel data has the advantage of both cross-section and time series. Table 1 above presents the detailed number of cross-sections included, included observations, and total pool observation for our sample. In general, the basic model postulates that the leverage of a firm depends on firm-related variables. Specifically, let Leverage = ƒ[profitability, non-debt tax shield, firm size, growth, dividend paid, financial distress cost, tangibility, employee productivity] Table 2 shows the notation and how the variables are computed. Therefore, to estimate the determinants of leverage, the following regression model are formulated: TD = a0 + a1ebitdat + a2ndtst + a3sizet + a4grt + a5divpdt + a6fdct + a7tangt + a8ep1t + u1 Eq. 1 LD = a0 + a1ebitdat + a2ndtst + a3sizet + a4grt + a5divpdt + a6fdct + a7tangt + a8ep1t + u1 Eq. 2

7 SD = a0 + a1ebitdat + a2ndtst + a3sizet + a4grt + a5divpdt + a6fdct + a7tangt + a8ep1t + u1 Eq EMPIRICAL RESULTS 4.1 Summary Statistics Table 3 reports basic descriptive statistics. It shows that the average total debt leverage of full sample is 47% and the average short-term debt of full sample is 30%. The average long-term debt of full sample is 10% - close to Brazil s figure of 9.7% (Booth, et al., 2001) and far from the G-7 countries of 28% to 54% (Rajan and Zingales, 1995). With reference to Table 3, the rate of long-term debt to total debt is 20.34% while the contribution of short-term debt to total debt is 79.66%. This means that the long-term debt is very small in China. This result is consistent with findings reported by Bhabra, Liu, and Tirtiroglu (2008) and Li, Yue, and Zhao (2009). The results is also consistent with Booth et al. (2001) who suggested that firms in emerging markets tend to raise more short-term debt. TABLE 3. DESCRIPTIVE STATISTICS Mean Median Max Min Std. Dev. Observations Cross sections TD , LD , SD , EBITDA , NDTS , SIZE , GROW , DIVPD , , , FDC , TANG 14,000 3,380 2,140,000-80,900 2, EP 260,874 72,342 70,787, ,411 1,541,071 2, Notes: DIVPD and TAND are presented by Chinese RMB in million, EP is presented by Chinese RMB in Thousand. Table 4 and Figure 1 present the ranking of debt level for sectors. It shows that the financial sector has the highest total debt leverage (62.18%) while the lowest is telecommunication services (26.90%). Meanwhile, there are only 2 firms in telecommunication services sector in this sample due to a panel data structure. On the other hand, ranking for long-term debt is different from that of total debt. Among the sectors, the long-term debt of utility sector is the highest (23.71%) while that of consumer staples is the lowest (3.68%). In general, the finding is consistent with earlier studies which suggested that firms in the utility sectors used more long-term debt than firms in the other sectors. In contrast, the highest average short-term debt is financial sector (44.95%) and the lowest is telecommunication services sector (24.70%). TABLE 4. THE OVERVIEW OF AVERGAGE DEBT LEVEL OVER

8 TD Rank LD Rank SD Rank Full Sample 46.96% 9.55% 37.41% Consumer Dis % (4) 6.39% (6) 41.40% (2) Consumer Sta % (7) 3.68% (9) 38.11% (4) Energy 44.69% (6) 13.54% (3) 31.10% (7) Financials 62.18% (1) 17.26% (2) 44.95% (1) Health Care 32.97% (9) 4.26% (8) 28.74% (8) Industrials 50.89% (2) 9.97% (5) 40.93% (3) Information Tec % (8) 4.77% (7) 32.44% (6) Materials 45.71% (5) 10.91% (4) 34.82% (5) Tel. Services 26.90% (10) 2.40% (10) 24.70% (10) Utilities 50.42% (3) 23.71% (1) 26.75% (9) FIGURE 1. THE OVERVIEW OF AVERAGE DEBT LEVEL OVER % 62.18% 60% 46.96% 47.81% 50.89% 50.42% 41.82% 44.69% 45.71% 50% 44.95% 40.93% 37.41% 37.22% 40% 41.40% 38.11% 32.97% 34.82% 31.10% 28.74% 32.44% 26.90% 26.75% 30% 24.70% 23.71% 17.26% 20% 13.54% 9.55% 9.97% 10.91% 6.39% 10% 3.68% 4.26% 4.77% 2.40% 0% TD LD SD According to Pessarossi and Weill (2013), state-owned firms tend to have easier access to debt and loan fund compared to non-state-owned firms. On the other hand, it is known that ownership of firms in fundamental sectors such as energy and mining were dominated by the state in China (Stephen, 2006). Thus, based on the results shown above, it is evident that long-term debt is highly financed in the energy and materials sectors. Specifically, the levels of all the proxies of debt leverage in the financial sector are high compared to other sectors. According to Kim Ho, and Giles (2003), 86% of all loans was held by banks and only 2.8% was held by public firms. Our results may confirm this finding and suggest that firms in the financial sector have easier access to debt financing comparing to other public firms in China. Table 5 presents a detailed report of debt level for every sector during the periods. It shows that average debt leverage over the five year period are 42.70%, 43.63%, 44.28%, 44.46%, and 45.37%, indicating an increasing trend. In detail, it is found that the average long-term debt increases form 7.94% (1,977 million Yuan) in 2008 to 10.42% (4,324 million Yuan) in It increases by 2.48% representing a 31.23% of average long-term debt for In particular, the highest is 2011 when it reaches 4,925 million Yuan. On the other hand, the average short-term debt increases from 34.63% (9,774 million Yuan) to 34.94% (22,708 million Yuan). Compared to long-term debt, the degree of increase in short-term debt is lower than that of long-term debt, although the amount increases by % (12,934 million). Taken together, it shows that debt leverage has been increasing in China, in particular the increase in long-term debt during the periods compared to past years. TABLE 5. DEBT LEVEL OF DIFFERENT SETORS OVER

9 Con. Dis. Con. Sta. Energy Financials Health Care Industrials Info. Tec. Materials Tel. Services Utilities Yearly average Average TD(mil) ,039 1,413 24,140 4, , ,975 70,872 3,852 11, ,697 1,749 35,286 5, ,480 1,046 3, ,197 7,423 17, ,645 2,494 41,786 8,621 1,286 9,802 1,258 4, ,648 7,708 19, ,604 3,095 53,729 10,225 1,586 11,992 1,503 5, ,962 8,454 22, ,183 3,377 63,719 12,256 1,887 13,844 1,663 6, ,070 8,731 27,031 Average TD(%) % 42.37% 41.45% 59.17% 33.70% 50.27% 36.11% 46.20% 21.19% 48.95% 42.70% % 42.37% 44.35% 60.87% 32.24% 49.99% 36.67% 46.64% 25.98% 49.71% 43.63% % 41.94% 44.28% 64.31% 32.51% 50.96% 37.91% 43.89% 27.83% 51.27% 44.28% % 41.39% 46.20% 62.98% 32.47% 51.49% 37.31% 45.37% 28.36% 50.69% 44.46% % 42.18% 47.16% 63.66% 34.06% 51.74% 38.05% 46.45% 31.10% 51.45% 45.37% Average LD(mil) ,750 1, , ,558 1,806 1, ,374 2, , ,387 4,630 2, ,479 3, , ,261 18,696 3,281 4, ,690 3, , ,495 18,212 3,726 4, ,468 3, , ,039 2,156 4,047 4,324 Average LD(%) % 4.06% 10.69% 14.05% 2.55% 7.96% 2.93% 8.71% 1.86% 22.01% 7.94% % 4.04% 14.03% 18.21% 3.91% 8.92% 4.72% 10.91% 1.29% 23.65% 9.58% % 5.04% 14.63% 19.64% 4.66% 9.28% 5.42% 11.59% 4.21% 23.87% 10.46% % 2.53% 14.28% 17.02% 4.30% 10.96% 5.17% 11.22% 3.97% 23.83% 10.05% % 2.85% 14.29% 17.23% 6.00% 12.69% 5.74% 12.12% 0.42% 25.11% 10.42% Average SD(mil) ,844 1,323 16,393 2, , ,196 64,313 2,046 9, ,396 1,634 22,912 3, , ,497 99,810 2,792 14, ,126 2,298 26,306 5,511 1,109 7,791 1,057 3,140 98,952 4,427 15, ,910 2,942 35,038 7,142 1,342 9,247 1,295 4, ,750 4,728 17, ,403 3,217 37,251 8,627 1,545 10,485 1,408 4, ,914 4,684 22,708 Average SD(%) % 37.02% 30.76% 45.12% 31.15% 42.31% 33.19% 37.50% 19.33% 26.94% 34.63% % 38.33% 30.33% 42.65% 28.33% 41.07% 31.95% 35.74% 24.70% 26.06% 34.05% % 36.90% 29.65% 44.67% 27.85% 41.68% 32.49% 32.30% 23.62% 27.40% 33.82% % 38.87% 31.92% 45.96% 28.17% 40.53% 32.14% 34.15% 24.39% 26.87% 34.40% % 39.33% 32.86% 46.43% 28.05% 39.05% 32.31% 34.33% 30.68% 26.34% 34.94%

10 Table 6 shows the results of a comparative analysis of the average value of determinants in different sectors. It is reported that the energy sector has the highest average profits (30%) and dividend paid (3,230 million Yuan), biggest average size (10.14), and the highest average finance distress cost (4.73). The non-debt tax shield (0.05) and tangibility (104,000 million Yuan) in energy sector are also high and ranked as the second highest among all the sectors. Based on the results, it is obvious that natural sources of energy are important contributors to the Chinese economy. At the same time, it is also interesting to note that the energy sector is the biggest sector in terms of assets and also pay the highest dividend it is also the most profitable sector. Whether there is a cause and effect relationship between profits and dividend payment and firm size should be an interesting question calling for further analysis in China. TABLE 6. AVERAGE VALUE OF DETERMINANTS FOR DIFFERENT SECTORS OVER NDTS EBITDA SIZE GROW DIVPD FDC TANG EP Rank Rank Rank Rank Rank Rank Rank Rank Full Sample % % , ,874 Consumer Dis (3) 20% (4) 9.51 (6) 20% (7) -83 (7) 0.34 (4) 6,330 (7) 134,026 (7) Consumer Sta (4) 23% (2) 9.48 (7) 28% (3) -173 (4) 0.34 (4) 5,680 (8) 136,260 (6) Energy: 0.05 (2) 30% (1) (1) 25% (5) -3,230 (1) 4.73 (1) 104,000 (2) 339,234 (4) Financials: 0.01 (5) 17% (7) 9.78 (2) 31% (2) -60 (9) 0.32 (6) 11,700 (5) 910,893 (1) Health Care: 0.02 (4) 19% (5) 9.34 (8) 19% (8) -70 (8) 0.06 (9) 3,420 (9) 366,328 (3) Industrials: 0.02 (4) 18% (6) 9.73 (4) 26% (4) -157 (5) 0.39 (3) 14,600 (3) 178,578 (5) Information Tec (4) 14% (8) 9.31 (9) 18% (9) -41 (10) 0.14 (8) 2,870 (10) 101,838 (8) Materials: 0.03 (3) 22% (3) 9.61 (5) 21% (6) -115 (6) 0.26 (7) 8,400 (6) 96,613 (9) Tele. Services: 0.07 (1) 20% (4) (1) 14% (10) -556 (5) 3.02 (2) 217,000 (1) 51,292 (10) Utilities: 0.03 (3) 17% (7) 9.77 (3) 94% (1) -195 (3) 0.33 (5) 13,100 (4) 537,176 (2) Notes: TAND and DIVPD are presented by Chinese RMB in million, EP is presented by Chinese RMB in Yuan. As follows, the highest non-debt tax shied is 7% for the telecommunication service sector and the lowest is 1% for the financial sector. The most profitable sector measured by using EBITDA is the energy sector reported to be 30% while the lowest is 14% for the information technology sector. The biggest size in terms of total assets is in both the energy and telecommunication service sectors. The highest growth sector is utility with an average of 94%, and the sector with the lowest growth is the telecommunication service. The information technology sector is found to have the lowest average dividend payment (40.89 million Yuan) while the energy sector is highest with an average of 3,230 million Yuan. The sector with the highest financial distress cost is the energy sector (4.73) and the lowest is only 0.06 in the health care sector. The highest tangibility is 217,000 million Yuan for the telecommunication service sector and the lowest is 2,870 million Yuan for the of information technology sector. The financial sector is found to have the highest level of employee productivity with 910,893 Yuan while the telecommunication service sector is ranked lowest at 51,292 Yuan. 4.2 Pooled Panel Regression Results Pooled Panel Regression Results for Full Sample with Eq.1, Eq.2, and Eq.3 Table 7 reports the regression results for model specification as given in Eq. (1), Eq. (2), and Eq. (3) for full sample. It shows that the non-debt tax shield, profits, and firm size are the determinants for all 10

11 of the proxies of debt leverage we identified. Among them, the effect of non-debt tax shield is negative while profits and firm size are positive. Growth and tangibility have positive effects on the total and short-term debts, while dividend payment is positively related to long-term debt. Both financial distress cost and employee productivity show no effect on all the proxies for debt leverage. Among the variables, non-debt tax shield is the most significant variable followed by profits, firm size, and tangibility. TABLE 7. POOLED PANEL REGRESSION RESULTS FOR FULL SAMPLE WITH Eq. 1, Eq.2, and Eq. 3 NDTS Profits SIZE GROW DIVPD FDC TANG EP Obs Adj-R2 Full Sample: TD Coeff Prob. (000)*** (000)*** (000)*** (000)*** (0.002)*** LD Coeff Prob. (0.015)*** (000)*** (000)*** 0.601(0.048)*** SD Coeff Prob. (000)*** (000)***(0.001)*** (000)*** (0.051)** Note: *** at the significant level, ** at the 0.05 significant level. There are remarkable differences and similarities between past studies and this study in terms of results arrived at. Firstly, the profits were found to be negatively associated with debt leverage by past studies (e.g see Huang and Song, 2006; Qian, Tian, and Wirjanto, 2009). In this study, the profits are positive related with debt leverage and statistically significant. It is noticed that only Qian, Tian, and Wirjanto (2009) found a small positive relationship between debt leverage and past profits. Secondly, growth is found to have a positive relationship with debt leverage, while it was found to be negative in past studies in China (Tong and Green, 2005; Huang and Song, 2006; Qian, Tian, and Wirjanto, 2009). The findings for firm size are consistent with past studies. According to Chen, Jian, and Xu (2009), capital market for small and medium firms was limited. The finding for tangibility is also consistent with past studies (Huang and Song, 2006; Qian, Tian, and Wirjanto, 2009). Putting all the results in perspective, particularly with reference to the impact of non-debt tax shield and profits, the conclusion arrived at shows that debt leverage decision in China is well explained by static trade-off theory. This finding throws new lights on the studies of debt leverage decision in China as most of the studies suggested a pecking order theory for debt leverage in China in past years. Analysis of this study shows that the financial behaviors of listed firms in China changed over different periods of time. As a result, we would like to suggest that the main determinants of debt leverage in China have changed since Pooled Panel Regression Results for Sectors with Eq.1 The pooled panel regression results for different sectors with Eq. 1 are reported in Table 8. The first row shows that the non-debt tax shield is negatively associated with total debt leverage in terms of consumer discretionary, energy, health care, industrial, material, and utility sectors. Among them, the most significant effect of non-debt tax shield is on consumer staples sector while the lowest is on health care sector. The profits are positively related to debt leverage in all the sectors under study, except consumer discretionary sector. This may show that profits is the most preferred variable in most firms in China, when they make debt leverage decision. With regard to firm size, it is negative in the consumer discretionary sector and positive in the energy, financial, health care, industrials, information technology, materials, and utility sectors. The 11

12 results of both full sample and separated sectors show that firm size is one of most important determinants for debt leverage in China. This is consistent with findings in past studies and evident that TABLE 8. POOLED PANEL REGRESSION RESULTS FOR SECTORS WITH Eq.1 Sector NDTS EBITDA SIZE GROW DIVPD FDC TANG EP Obs Adj-R2 Consumer Discretionary: Coeff Prob. (000)*** )***0.009)*** )*** Consumer Staples: Energy: Financials: Health Care: Coeff Prob (0.001)*** (0.064)** Coeff Prob (000)*** (000)*** )*** )** Coeff Prob (000)*** (000)*** )***0.007)*** (0.074)**001)*** Coeff Prob. (0.071)** (000)*** (000)*** (000)*** (0.075)** Industrials: Coeff Prob. (000)*** (000)*** (000)***0.001)*** (000)*** )*** Information Technology: Coeff Prob (000)*** (0.069)** Materials: Coeff Prob..002)*** (000)*** 0.006)*** )*** )*** Utilities: Coeff Prob. (000)*** (000)*** 0.040)*** (000)*** )*** Notes: *** at the significant level, ** at the 0.05 significant level. the effect of firm size on debt leverage clearly exists for listed firms in China (Qian, Tian, and Wirjanto, 2009; Tong and Green, 2005; Huang and Song, 2006). Growth is found to be positively related to debt leverage in the consumer discretionary, consumer staples, industries, and utility sectors while it is negatively related to debt leverage in the information technology sector. Financial distress cost is positively related to debt leverage in the energy, health care, industries, and materials sectors while it is negatively related to debt leverage in the financial sector. Tangibility is found to be positively related to debt leverage in the consumer discretionary, financial, and utility sectors while it is negatively related to debt leverage in the health care sector. Employee productivity is found to be negatively related to debt leverage in the energy, financial, industries, and utility sectors. Finally, the impact of dividend paid is positively related to debt leverage in the financial sector only. The results show that the specific effects of different sectors are at work in affecting debt leverage Pooled Panel Regression Results for Sectors with Eq.2 and Eq.3 12

13 The followings are the pooled panel regression results for different sectors with Eq.2 and Eq.3. The analysis suggests that long-term debt is negatively affected by non-debt tax shield in the energy, materials, and utility sectors, while the short-term debt is negatively affected by it in the consumer discretionary and industrials sectors. The second row in Table 9 shows the profits effects. It finds that the effects of profits are positive TABLE 9. POOLED PANEL REGRESSION RESULTS FOR SECTORS WITH Eq.2 & Eq.3 Sector NDTS EBITDA SIZE GROW DIVPD FDC TANG EP Obs Adj-R2 Full Sample: LD Coef f Prob. (0.015)*** (000)*** (000)*** (0.048)*** SD Coef f Prob. (000)*** (000)*** (0.001)*** (000)*** (0.051)** Consumer Discretionary: LD Coef f Prob (000)*** (0.095)** SD Coef f Prob. (000)*** (0.080)** (000)*** (0.011)*** (0.042)*** Consumer Staples: LD Coef f Prob (0.026)*** SD Coef f Prob (0.003)*** Energy: LD Coef f Prob. (0.003)*** (0.004)*** (000)*** (0.036)*** (0.013)*** SD Coef f Prob (0.010)*** (0.030)*** Financials: LD Coef f Prob (0.004)*** SD Coef f Prob (0.023)*** (0.064)** (0.007)*** (0.045)*** Health Care: LD Coef f Prob (0.002)*** (000)*** (0.085)** SD Coef f Prob (000)*** (0.098)** (0.065)** (0.002)*** (0.081)** Industrials: LD Coef f Prob (000)*** (0.063)** (0.001)*** SD Coef f Prob. (000)*** (000)*** (000)*** (000)*** Information Technology: LD Coef f Prob (0.003)*** SD Coef f Prob (000)*** (0.081)** Materials: LD Coef f Prob. (0.006)*** (000)*** (000)*** (0.011)*** (0.019)*** (0.002)*** SD Coef f Prob (000)*** (0.075)** (0.058)** (0.010)*** Utilities: LD Coef f Prob. (0.025)*** (000)*** (0.002)*** (0.009)*** (0.001)*** SD Coef f Prob (0.018)*** (0.025)*** (0.048)*** Notes: *** at the significant level, ** at the 0.05 significant level. to both long-term and short-term debt in the energy, health care, materials, and utility sectors. In contrast, the effects of profits are only positive to short-term debt in the consumer discretionary, 13

14 consumer staples, financial, industrials, and information technology sectors. According to Bhabra, Liu, and Tirtiroglu (2008), long-term debt is negatively affected by profits in China. The result of this study is totally different from earlier findings with regard to the financial behavior of listed firms in China. The third row shows that firm size is positively related to long-termdebt in the financial, industries, materials, and utility sectors when it is negatively related to long term debt in consumer staples. The result is consistent with the findings in study conducted by Bhabra, Liu, and Tirtiroglu (2008). Whereas it is also positively related to short-term debt in the consumer discretionary, energy, and health care sectors. Growth as shown in the fourth row of Table 9 has different effects on debt leverage. Firstly, it shows that both long-term and short-term debts are not affected by growth in the consumer staples, energy, financial, and information technology sectors. In contrast, it shows that both long-term and short-term debts are affected by growth in the health care and industries sectors. Among them, long-term debt is negatively related to growth and short-term debt is positively related to growth. Furthermore, it shows a positive relationship between growth and short-term debt in the consumer discretionary sector. In addition to that, there is a negative relationship between growth and long term debt in the martials sector, while a positive relationship between growth and long term debt in the utility sector. The dividend paid as shown in fifth row shows no significant relationship to both long-term and short-term debts in the consumer discretionary and staples, health care, industrials, and information technology sectors. On the other hand, it is found that dividend paid is significant to both long-term debt (- ve) and short-term debt (- ve) in the utility sector. As follows, a positive relationship between short-term debt and dividend paid is found in financial and materials sectors, while a positive relation between long-term debt and dividend paid in the energy sector. The sixth row shows the effect of financial distress cost. It shows that the effect of financial distress cost on consumer discretionary and staples, financial, information technology, and utility sectors is not statistically significant. As follows, a positive relationship between financial distress cost and long-term debt is found in the industrials, energy, and materials sectors. Finally, there is only a positive relationship between financial distress cost and short-term debt is found in the health care sector. The seventh row reports results for tangibility. It shows that the long-term debt is positively affected by tangibility in the energy, industrials, and materials sectors while the short-term debt is positively affected only in the health care sector. There is no any significant relationship found between tangibility and both long-term and short-term debt in the consumer discretionary and staples, financial, information technology, and utility sectors. The eighth row in Table 9 reports the results for employee productivity. It shows a negative relationship between short-term debt and employee productivity in the financial and industrials sectors. On the other hand, it shows a significant relationship between employee productivity and both long-term and short-term debt in the materials sector, while a positive relationship is found between employee productivity and long-term debt and a negative relationship between employee productivity and short-term debt in the information technology sector. There is no any significant impact of employee productivity on both long-term and short-term debt in the consumer discretionary and staples, energy, health care, and utility sectors. With reference to the results as shown above, the effects of determinants on long-term and 14

15 short-term debts are differently reflected. Although some results are mixed, the profits and firm size are identified as the most activity variables affecting the debt leverage decision of both long-term and short-term. On the other hand, non-debt tax shield is identified as one variable that has a high negative impact on the both long-term and short-term debt. 5. CONCLUSION This study has examined determinants of leverage in China. As can be seen, the models have been subjected to detail empirical testing, although the majority of past studies have discussed total debt leverage. In particular, less past studies have been undertaken to make a comparative study for long-term and short-term debts. As far as can be determined, this paper is possibly the first one to examine the determinants of debt leverage in terms of different sectors. When almost all of the studies examined all the sectors holistically, this study has adopted a design focusing on an examination of the financial behaviour in terms of debt leverage of Chinese listed firms in the different sectors. Besides, this study used a joint model to test the interaction between employees productivity and debt leverage. Therefore, this study provides a new perspective in terms of an additional tool to investigate firm leverage. REFERENCE An. Z. Y., Taxation and capital structure: Empirical evidence from a quasi-experiment in China, Journal of Corporate Finance, 2012, Vol. 18, pp Bae, K., Kang, J. and Wang, J., Employee treatment and firm leverage ratio: A test of the stakeholder theory of capital structure, Journal of Financial Economics, 2011, Vol. 100(1), pp Bhabra, H. S., Liu, T., and Tirtiroglu, D. Capital structure choice in a nascent market: Evidence from listed companies in China, Financial Management, 2008, Vol. 37, pp Booth, L., Aivazian, V., Demirguc-Kunt, A., and Maksimovic, V., Capital structures in developing countries, Journal of Finance, 2001, Vol. 56 (1), pp Brealey, R. A. and Myers S. C., Principles of corporate finance (2 nd ed.), 1984, New York: McGraw-Hill Press. Chen, D. H., Jian, M. and Xu, M, The role of dividends in a regulated economy: the case of China, Pacific-Basin Finance Journal, 2009, Vol. 17, pp China View, Special Report: NPC, CPPCC Annual Sessions 2008, [online]. Available: [March 8th, 2008] DeAngelo, H., and Masulis R.W., Optimal capital structure under corporate and personal taxation, Journal of Financial Economics, 1980, 8, pp Dougherty, C, Introduction to Econometrics (4 th ed.), 2011, U.S.A.: Oxford University Press. Fama, E. F. and French K. R., The anatomy of value and growth returns, Working Paper, 2005, University of Chicago. Fama, E. F., and French K. R., Testing trade-off and pecking order predictions about dividends and debt, Review of Financial Studies, 2002, 15, pp Frank, M. Z., and Goyal V. K., Trade-Off and pecking order theories of debt, Handbook of 15

16 Empirical Corporate Finance, 2009, Vol. 2, pp Frank, M., Goyal, V, Testing the pecking order theory of capital structure, Journal of Financial Economics, 2003, Vol. 67, pp Gatchev, V. A., Spindt, P. A. and Tarhan, V., How do firms finance their investments? The relative importance of equity issuance and debt contracting costs, Journal of Corporate Finance, 2009, Vol. 15, pp Huang, G. and Song, F. M., The determinants of capital structure: Evidence from China, China Economic Review, 2006, Vol. 17, pp Jensen, M. and Meckling, W., Theory of the firm: managerial behavior, agency costs and ownership structure, Journal of Financial Economics, 1976, Vol. 3(4), pp Kale, J., Ryan Jr, H. and Wang, L., Debt as a bonding mechanism: Evidence from the relations between employee productivity, capital structure, and outside employment opportunities, 2007, 18th Annual Conference on Financial Economics and Accounting, NYU, Kim Y., I.S.M. Ho, and Giles, M.S., Developing institutional investors in People's Republic of China, The World Bank Country Study Paper. Krejcie, R. V. and Morgan, D. W., Determining sample size for research activities, Educational and Psychological Measurement, 1970, Vol. 30, pp Kraus, A. and Litzenberger, R. H., A state-preference model of optimal financial leverage, Journal of Finance, 1973, Vol. 9, pp Lee, S., Koh, Y. and Kang, K. H., Moderating effect of capital intensity on the relationship between leverage and financial distress in the U.S. restaurant industry, International Journal of Hospitality Management, 2011, Vol. 30, pp Lemmon, M. L. and Zender, J. F., Debt capacity and tests of capital structure theories, Working Paper, 2004, University of Utah. Li, K., Wang, J., Cheung, Y. L., and Jiang, P., Privatization and risk sharing: Evidence from the split share structure reform in China, The Review of Financial Studies, 2011, Vol. 24(7), pp Li, K., Yue, H. and Zhao, L., Ownership, institutions, and capital structure: Evidence from China, Journal of Comparative Economics, 2009, Vol. 37, pp McConnell, C. R., Brue, S. L. and Flynn, S. M., Economics: Principles, Problems, and Policies (19th ed), 2009, U.S.A.: Mc Graw Hill Press. Modigliani, F. and Miller, M. H., The cost of capital, corporation finance and the theory of investment, The American Economic Review, 1958, Vol. 48, pp Modigliani, F. and Miller, M. H., Corporate income taxes and the cost of capital: A correction, The American Economic Review, 1963, Vol. 53, pp Myers, S.C. and Majluf, N.S., Corporate financing and investment decisions when firms have information that investors do not have, Journal of Financial Economics, 1984, Vol. 13, pp Paul, A. S. and William, D. N., Economics (19 th ed.), 2010, New York: MC Graw Hill Press. Pessarossi, P. and Weill, L., Choice of corporate debt in China: The role of state ownership, China Economic Review, 2013, Vol. 26, pp Qian, Y., Yao, T. and Wirjanto, T. S., Do Chinese publicly listed companies adjust their capital structure toward a target level?, China Economic Review, 2009, Vol. 20, pp