Agency costs are due to the conflicts of interest that probably occur between different stakeholders. In corporate finance, agency costs mainly arise between shareholders and management, between large shareholders and minority shareholders, and between management and debt holders. The firm should minimize agency costs. There are two agency costs about debt, namely underinvestment and overinvestment. As for underinvestment issue, Myers (1977) found that the managers of company, who often serve the interests of shareholders, would probably pass up the positive NPV investment opportunities if they are financed by risky debt. He explained that high risky debt is highly likely to default and shareholders will bear this kind of potential loss. Myers (1977) also suggested that a firm can tackle underinvestment problem by including restrictive covenants in the debt contract and monitoring that covenants, and shortening debt maturity.
For overinvestment, Hart and Moore (1995) analyzed this problem and pointed out the long-term debt level has a negative relation with earnings of new investment opportunities but a positive relation with earnings of companies' existing assets.
2.1.2. Proxies for agency cost theory
Firm size. Barclay and Smith (1995) pointed out firm size would influence the debt maturity because smaller firms would not choose to issue public debt with a huge fixed cost. Kailan, Richard and Yilmaz (2008) also regarded firm size as an explanatory variable in their models. We measure firm size by the natural logarithm of its total asset every year from 2006 to 2009. The firm size is supposed to have a positive relation with debt maturity.
2.2. Signalling effects
2.2.1. Signalling effects theory
Flannery (1986) explored signaling effects of the firm's debt-maturity choice. Under the condition that the market cannot tell good firms from bad ones, it is better for good ones to issue short debts because long-term debt shows relatively more risks to investors and hence would be underpriced. That is to say, good firms would not sell long-term debts at a fair price when good and bad ones issue long-term debts at the same time. Oppositely, bad firms tend to issue relatively overpriced long debts. Besides, Flannery also explained that some firms raise short term debts to finance investment opportunities with longer periods to signal their confidence on the bright prospects.
Diamond (1991) stated that firms with sufficiently good and exceedingly poor ratings will tend to raise short-term debts and those with ratings between the two will prefer to issue relatively long-term debts.
2.2.2. Proxies for signaling effects theory
Firm quality. We employ the ratio of earnings per share to measure the firm quality. Cai, K.,Fairchild, R.,Guney, Y.(2008) argued that 'changes of a firm's future earnings'（引用）can explain the insider's private information. Empirically, the firm quality is negatively correlated with the debt maturity.
2.3. Liquidity
2.3.1. Liquidity theory
Myers and Rajan (1998) explained influence of the liquidity asset on debt financing. Generally, Firms with more liquidity assets tend to have more cash to be able to repay the debt and hence are easier to raise the debt. Myers and Rajan focused more on another side of the impact of liquidity. They pointed out that it is easier for firms with some illiquid assets to get access to long-term debts.
Leland and Toft (1996) examined the relationship between leverage ratio and debt maturity. He concluded that leverage ratio is bigger for firms with debt of longer maturity and this result is empirically consistent with what Barclay and Smith (1995) got. Dennis et al. (2000), however, found that they have an inverse relationship.(这段leverate是ä¸æ˜¯å•æ‹¿å‡ºæ¥)
2.3.2. Proxies for liquidity theory
The ratio of current assets to current liabilities will be used to measure liquidity and a positive relation between the two is expected. Following Kailan, Richard and Yilmaz (2008) , We measure leverage by the ratio of the total debt to total asset. Unclear pictures are shown in the literature about how leverage influences the debt maturity of company.
2.4. Matching theory
2.4.1. Matching theory
Financing cost may be reduced through matching maturity of firm's asset with that of firm's liability. Myers (1977), and Hart and Moore (1994) suggested firms would match the maturity of their assets with that of their liabilities. Myers, specifically, explained such matching principle as 'an attempt to schedule debt repayments to correspond to the decline in future value of assets currently in place' (Myers,1977 p.171). (引用)
2.4.2. Proxies for matching theory
We measure asset maturity as the ratio of fixed asset to the total asset and predict that the asset maturity is positively associated with debt maturity.
2.5. Mortgage assets
2.5.1. Mortgage assets
Whited (1992) argued that mortgage assets would have important influences on the debt maturity. He further pointed out that mortgage assets has positive relation with the firm's long-term debt.
2.5.2. Proxies for the mortgage assets
We measure mortgage assets by the ratio of the sum of inventories and fixed assets to total assets.
2.6. Interest rate
2.6.1. Interest rate
Titman (1992) found that bad firms would transfer from short-term borrowing to long-term borrowing when interest rate risk increases. And
2.6.2. Proxies for interest rate
Following Kailan, Richard and Yilmaz (2008), we measure the term structure of interest rate using difference of the yearly rate on long-term (5 years) and short-term (6 months) government bond Lending rates are used to replace the government bond rates due to data shortages. A direct link between the lending rate difference and debt maturity is expected as the bigger the lending rate difference, the more the shorter debts.
2.7. Other factors specifically to Chinese Firms
In order to better examine the factors to determine the debt maturity in China, 4 independent variables of Chinese distinctive characters are put into the model, including percentage of shares owned by the biggest shareholder, nature of the biggest shareholder, refinance and economic stimulus package.
引用ä¸å›½å¦è€… XIAO ZUOPINGçš„-ç«
3. The data and sample description
3.1 The data
The data are found on Shanghai Stock Exchange from Thomas Reuters and RESSET databases as well as statistics published by People's Bank of China from 2005 to 2009. These two databases include 858 companies. We only use the book equity from 2005 to 2009 to calculate the percentage changes from 2006 and all other kinds of data started from 2006 onwards. Specifically, the data of percentage of the biggest shareholder and nature of the shareholder are collected from RESSET database. Long-term and short-term government bond rates are replaced by lending rates respectively shown in People's Bank of China. The rest of data hence are sought from Thomas Reuters database. We ignore the companies from banking industry. Two methods are used to deal with firms that have incomplete data. For data in the field of percentage of the biggest shareholder and nature of the shareholder, only about 230 firms have complete data, thus we take averages of these data and put the average for other firms. For incomplete data that are in the other areas, like earnings per share, ratio of total debt to total assets and etc., these kind of firms are ignored. We are left with 768 firms.
3.2. Sample description
The table 1 shows the descriptive statistics of all the variables in the general regression equation.
Figure 1 : Average debt maturity of Chinese listed companies from 2006 to 2009
Table 1 : Descriptive statistics
The mean value of ltdtd (long-term debt to total debt) is 0.13. In figure 1, specific picture is shown about yearly average debt maturity and it has an increasing trend but the value is around 0.13. This implies Chinese listed firms tend to issue short-term (less than 1 year) debt much more than long term debt. This value decreases a lot compared with 0.23 from 1999 to 2004 indicated by Cai,Fairchild and Guney (2008). This figure is even lower compared with situations in develop economics, which often ranges from 0.6 to 0.8. The reasons behind may include that Chinese listed companies tend to issue equity or stock shares to get financing and the debt market is not mature yet. Besides, many listed companies rest their financing on bank loans and the banks place very strict requirements on long-term lending to companies. The average of ref amounts to 0.53, which shows preference of Chinese firms to issue stock shares to refinance instead of issuing debt. This relatively high value is a good explanation about low value in the ratio of long-term debt to total debt.
Three other variables can tell us some special features about Chinese firms, namely perbs (percentage of the biggest shareholder), natbs (nature of the biggest shareholder) and ref (refinance of Chinese listed firms during periods of 2006 to 2009). The mean value of perbs is 0.35, demonstrating that the biggest shareholder in the Chinese firms often have a big saying. For natbs, the average is 0.62 and it shows the biggest shareholder in 62% firms relates to state-owned holder. These two values express distinct characters of Chinese companies, characters that the biggest shareholder covers a very big portion of shares and many Chinese firms controlled more or less by the nation. Besides, mean value of ref tells us that over a half companies did refinancing by issuing shares during those periods.
3.3 Variable correlation
Table 2 : Variable correlation
From correlation table, relation signs between the ratio of long-term debt to total debt (ltdtd) and most explanatory variables are indicated as we expected. For common explanatory variables, the exception lies in earnings per share, which has a positive relationship with ltdtd. The reason is probably that companies of good qualities in China tend to issue long-term debt for banks can often set preferable interests on long-term debt for them and hence it makes sense for companies to borrow for a long period in terms of cost savings.
Looking at four explanatory variables of Chinese distinctive features, they all have positive relationship with the ltdtd. For perbs, if the biggest shareholder has more shares in the company, it has bigger motivation to monitor the company's operation and hence reduce the agency cost. Therefore, it is common not to use too much short debts to tackle the agency issue, consistent with the statement by Myers (1977).
The nature of the biggest shareholder (natbs) is positively correlated with the ltdtd for the state-owned companies have closer relations with the government and are more probable to be controlled by the government and so probably are easier to get more long-term debt from banks in China.
The variable of ref has a positive relation with ltdtd. It may be that many Chinese listed companies choose to refinance themselves by issuing shares again, then debt decreases in their capital structure and accordingly the ratio of long-term debt to long-term debt increases.
Medium or long-term debt increased from 2008 to 2009 since the expansion policy has been launched after the end of 2008 and it is reasonable to see a positive sign between ltdtd and esp.
4. Methodology
We explore the debt maturity of Chinese listed companies using panel data and incorporate eleven explanatory variables in our general model. A variety of regression methods and tests are employed to further study the influence power of different factors.
4.1. The regression model
Following previous approaches by many researchers, panel data is used to run regressions and deal with the problem. The dependent variable in the general model is the ratio of long-term debt (more than one year) to total debt. Eleven independent variables are used in the general model, which are highlighted in the literature review. Of these independent variables, three dummy variables are employed, namely nature of the biggest shareholder (NatBS), refinance (Ref) and economic stimulus package (ESP).For nature of the biggest shareholder, it equals to 1 if state-owned company is the biggest shareholder and 0 otherwise. For refinance, the percentage changes of book equity from 2006 to 2009 are calculated to examine whether the company refinanced by issuing shares during this period. This dummy equals to 1 if the percentage change of the book equity exceeds 5% and 0 otherwise. Finally, the dummy variable of the economic stimulus package is 1 in 2009 and 0 otherwise because the chinese government launched 4 trillion yuan investment plan at the end of 2008. The general model hence is as follows:
4.2. Regression methods
Fixed-effects, random-effects and mixed-effects methods
å¯ä»¥å€Ÿé‰´ä¹¦ä¸Šçš„è¯ï¼š 与RANDOM-EFFECTS相比，FIX-EFFECTS一般都ä¸ä¼šINCONSISTENT用TESTS区分用哪ç§-法
4.3. Tests
General test
Year-by-year还åš-？
Industry-by-industry
5. Results and discussion
有一部分é‡ç‚¹åˆ†æžåˆ†æžä¸å›½ç‰¹è‰²
