Pecking Order or Trade-off Hypothesis?
Evidence on the Capital Structure of Chinese Companies
by
Guanqun Tong and Christopher J. Green
Tong: Green:
Doctoral student, Department of Economics, Loughborough University Department of Economics, Loughborough University
Correspondence to:
Guanqun Tong, Department of Economics, Loughborough University, Loughborough, Leicestershire, LE11 3TU, United Kingdom
Tel: +44 (0) 1509 222718; Fax: +44 (0) 1509 223910; E-mail:G.Q.Tong@lboro.ac.uk
June 2004
Preliminary Draft: Not for Quotation
PECKING ORDER OR TRADE-OFF HYPOTHESIS?
EVIDENCE ON THE CAPITAL STRUCTURE OF CHINESE COMPANIES†
by
Guanqun Tong and Christopher J. Green
Tong: Doctoral student, Department of Economics, Loughborough University Green: Department of Economics, Loughborough University
June 2004
Preliminary Draft: Not for Quotation
JEL Classification: G32, O16
Keywords: capital structure, pecking order hypothesis, trade-off hypothesis, China
ABSTRACT
In this paper we test the pecking order and trade-off hypotheses of corporate financing decisions using a cross-section of the largest Chinese listed companies. We build on Allen (1993) and Baskin (1989) to set up three models in which trade-off and pecking order theories give distinctively different predictions: (1) the determinants of leverage, (2) the relationship between leverage and dividends, and (3) the determinants of corporate investment. In model (1), we find a significant negative correlation between leverage and profitability; in model (2) we find a significant positive correlation between current leverage and past dividends. These results broadly support the pecking order hypothesis over trade-off theory. However, model (3) is inconclusive. Overall, the results provide tentative support for the pecking order hypothesis and demonstrate that a conventional model of corporate capital structure can explain the financing behaviour of Chinese companies.
Correspondence to:
Guanqun Tong, Department of Economics, Loughborough University, Loughborough, Leicestershire, LE11 3TU, United Kingdom
Tel: +44 (0) 1509 222718; Fax: +44 (0) 1509 223910; E-mail:G.Q.Tong@lboro.ac.uk
†
This paper is based on and extends Guanqun Tong’s MSc Dissertation at Loughborough University (Tong
2003). We thank participants in the 2004 EEFS Conference in Gdansk for their helpful comments on an earlier draft of this paper.
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
1. Introduction
How is the capital structure of a firm determined? Finance theory offers two broad and competing models: trade-off theory and pecking order theory. Harris and Raviv (1991) provide a comprehensive overview of these models and their implications, but they can be summarized briefly. Trade-off theory states that a value-maximizing firm will pursue an optimal capital structure by considering the marginal costs and benefits of each additional unit of financing, and then choosing the form of financing that equates these marginal costs and benefits. Benefits of debt include its tax advantage and the reduced agency costs of free cash flow; costs include the increased risk of financial distress and increased monitoring and contracting costs associated with higher debt levels. The pecking order hypothesis is based principally on the argument that asymmetric information creates a hierarchy of costs in the use of external financing which is broadly common to all firms. New investments are financed first by retentions, then by low-risk debt followed by hybrids like convertibles, and equities only as a last resort. At each point in time there is an optimal financing decision which depends critically on net cash flows as the factor which determines available funds. However, in contrast to trade-off theory, there is no unique optimal capital structure to which a firm gravitates in the long-run1.
Notwithstanding the conceptual differences between pecking order and trade-off theories, distinguishing between the two in practice is not straightforward. As Fama and French (2002) point out, the two theories share many common predictions about the determinants of leverage and dividends and, in a study of US corporations, they can only identify two predictions on which either theory performs better than the other. Trade-off does better in one case (large equity issues of low-leverage firms) and pecking order in the other (the negative impact of profitability on leverage), thus rendering the verdict on the two inconclusive. Prasad, Green and Murinde (2001a) survey a large volume of empirical literature on company capital structure, and they too conclude that the evidence on trade-off versus pecking order theories remains inconclusive. However, they also observe that the overwhelming bulk of the available empirical research on corporate capital
1
The origins of trade-off theory pre-date Modigliani and Miller (1958), but modern versions, based particularly
on trade-offs among agency costs, were stimulated by the seminal paper of Jensen and Meckling (1976). Pecking order theory was first set out by Myers and Majluf (1984).
1
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
structure is concerned with the major industrial countries, and there has been relatively little work on developing countries or the transition economies.
There are several reasons why one might expect firms in developing and transition economies (DTEs) to have different financing objectives from their counterparts in the industrial countries. First, many private firms in the DTEs were originally state enterprises and carry different goals and corporate strategies from this heritage. Prasad et al. (2004) point out that corporate strategy is a significant determinant of capital structure, independently of the basic goals of the company. Second, capital markets are less developed in the DTEs, and there is typically a narrower range of financial instruments available and a wider range of constraints on financing decisions than in the industrial countries (Singh and Hamid, 1992). Finally, accounting and auditing standards in DTEs tend to be relatively lax, implying that asymmetric information is more pervasive and problematic than in the major industrial countries (Cobham and Subramaniam, 1998).
Singh and Hamid (1992) and Singh (1995) pioneered research into corporate capital structure in developing countries. They concluded that firms in developing economies rely more heavily on equity than on debt to finance growth relative to their counterparts in the developed economies. Broadly similar results are reported by Booth et al. (2001), who also argue that it is difficult to distinguish between trade-off and pecking order models because variables used in one model are also relevant in the other model. Cobham and Subramaniam (1998) dispute the Singh and Hamid (1992) results, at least for India, where they conclude that, during the 1980s, large Indian and British firms exhibited broadly similar patterns of debt ratios. However, all three research groups emphasize that there are important inter-country differences in debt ratios in developing countries. A limitation of these studies is that they are mainly descriptive, but there have also been a few studies of the determinants of company leverage in DTEs, notably by: Wiwattanakantang (1999) for Thailand, Prasad, Green and Murinde (2001b) for Malaysia and Thailand, and Colombo (2001) for Hungary. However, these authors generally conclude that the determinants of leverage in these countries are comparable with its determinants in the industrial countries. All these studies use explanatory variables which are consistent with trade-off and pecking order theories, and they
2
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
report a range of results. However, none of them explicitly sets out to compare trade-off and pecking order theories in a manner designed to discriminate between them.
In this paper, we study the determinants of capital structure decisions in a sample of listed Chinese companies. China is of interest for several reasons, but most particularly because it is in the almost unique position of being both a developing economy and a transition economy. Most Chinese listed companies were owned by the state before they were listed on the stock market. In the period covered by this study, they retained a substantial proportion of government ownership, and government intervention played (and continues to play) a major role in controlling their decisions. It is therefore of particular interest to investigate how far this leads to different financial decision-making than in other countries.
Huang and Song (2002) studied the leverage decisions of 799 Chinese listed companies up to the year 20002. They argue that leverage rates are generally at the low end of the spectrum typically found in developing countries3. They then follow Rajan and Zingales (1995) and regress different leverage measures on a short list of “plausible” explanatory variables. They find that the correlation between leverage and these characteristics in Chinese state-controlled listed companies is surprisingly similar to what has been found in other countries. They also conclude that the static trade-off model explains the capital structure of Chinese listed companies better than the pecking order hypothesis. Huang and Song draw this conclusion mainly on the basis of a significant impact of institutional shareholdings (ie. ownership) on leverage. However, a key distinctive prediction of pecking-order theory is also accepted by the data: a significant negative impact of profitability on leverage. Therefore, it is really not possible to conclude from their analysis in favour of either trade-off or pecking order theory.
In our study, we focus specifically on the different predictions implied by trade-off and pecking order theories. We follow Baskin (1985, 1989) and Allen (1993) who provide evidence consistent with the pecking order hypothesis in the US (Baskin) and Australia (Allen). 2
3
The sample is an unbalanced panel, but because of the volatility in much of the data, the authors estimate a
cross-section regression of year 2000 leverage on the time-average of each of the explanatory variables, with the averaging done over the available data for each company.
This finding needs to come with a health warning: Green, Murinde, Suppakitjarak (2003) have shown that cross-country comparisons of leverage are even more difficult than Rajan and Zingales (1995) originally suggested.
3
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Specifically, we investigate the interactions among capital structure, cash flows, dividends and investment decisions implied by pecking order or trade-off theory. This enables us to provide a systematic set of tests of these theories and their applicability in China. We also depart from Huang and Song in using a substantially smaller sample of companies for the research. As a rule, “larger samples are better”, but because China is still very much in transition there is considerable variability in the quality of the accounting data provided by companies, and the interpretation of the accounts of all but the largest and best-established companies continues to be fraught with difficulty4. We therefore limit ourselves to a cross-section of the top 50 listed companies in 2002. This may attenuate the generality of our results, but we believe that this is more than compensated for by an increase in reliability.
The rest of the paper is organized as follows. Section 2 sets out the hypotheses to be tested. We use 3 different models to compare trade-off and pecking order theories, and concentrate on formulations where the two theories give a clearly conflicting prediction. We do this so that, where significant coefficients are found, we can have some confidence that the result can be attributed either to trade-off or pecking order decision-making. Section 3 discusses the main data sources and limitations and the empirical methodology. The estimation and testing results are presented and discussed in section 4. Section 5 summarises our main findings. 2. Hypotheses
Fama and French (2002) emphasize that many of the variables held to determine leverage under trade-off or pecking order theories are common to both theories. This makes it difficult for a “horse-race” between two regressions to distinguish adequately between the two theories, notwithstanding that they have very different implications for corporate behaviour. In our tests therefore we examine three related aspects of corporate financing where trade-off and pecking
4
The Ministry of Finance of China announced the results of the ‘Random Check of the Quality of Accounting Information’ of 2001, according to which, 18.8% of 192 randomly chosen listed companies were found to have provided dishonest information in assets, 53.6% of them were dishonest in profits and 11.5% of them prepared “internal accounts” which were different from those disclosed to the market. (Ministry of Finance of People’s Republic of China, 2003). Although the information about which companies were dishonest is not disclosed to the public, we would argue that the largest firms would be more reliable than others based on the sustained
emphasis of the government on the regulation of key large companies due to their significance to the stability of the emergent Chinese capital markets.
4
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
order theories give different predictions: the determinants of leverage, the relationship between leverage and dividends, and the implications of each theory for corporate investment. 2.1 Determinants of Leverage: Profitability, Size and Growth
Static tradeoff theory argues that since less profitable firms provide low shareholder returns, greater leverage in these firms merely increases bankruptcy risk and the cost of borrowing, and will therefore lower shareholder returns still further. Low shareholder returns will also limit equity issues. Therefore, unprofitable firms facing a positive NPV investment opportunity will avoid external finance in general and leverage in particular. There will also be a demand side effect as the market will be reluctant to provide capital to such firms. Thus, tradeoff theory predicts a positive relationship between leverage and profitability. In contrast, pecking order theory predicts that firms will use retentions first, then debt and equity issues as a last resort. Less profitable firms facing a positive NPV investment opportunity will be more willing to use external funds if cash flows are weak. Therefore, there will be a negative relationship between leverage and profitability. Fama and French (2002) and Myers (1984) both document a negative relationship between leverage and profitability and use this result as a basis for rejecting the static tradeoff model. Baskin (1989) regresses current leverage on current profitability and finds a negative coefficient, but Allen (1993) argues that this result is spurious. Because of the balance sheet equation, an increase in current profitability will necessarily result in a decrease in current debt given investment and dividends, and Allen advocates instead using lagged profitability. To check this point, we use current and lagged profitability in our model.
We also control for firm size to help discriminate between trade-off and pecking order theories. Warner (1977) argues that there are economies of scale in bankruptcy, implying that the agency costs of debt will be lower for larger companies, and this is largely confirmed by other researchers (for example, Bradbury and Lloyd, 1994). Thus trade-off theories suggest a positive relation between leverage and firm size. The converse argument is that firm size is a proxy for information asymmetries between the firm and the market: the larger the firm, the more complex its organization, the higher are the costs of information asymmetries and the more difficult it is
5
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
for the firm to raise external finance (Rajan and Zingales, 1995). Therefore, the pecking order is accentuated by size and there will be a negative relationship between leverage and size.
Finally, these relationships are conditional on available investment opportunities which we control for in a standard way using the growth rate of the firm’s assets (Prasad, Green and Murinde, 2001a). According to Baskin (1989), trade-off theory would suggest a negative sign for this variable because higher growth is associated with greater bankruptcy risk. This implies that a positive sign is more consistent with pecking order theory. Thus model (1) to test the tradeoff theory against the pecking order is as follows:
LEV(t)j = a1 + a2 ROA(t)j + a3 ROA(t-1)j + a4 SIZE(t-1)j + a5 GR(t)j + ej …(1)
where:
LEV(t)j = Leverage of firm j in year t
ROA(t) j = Profitability in year t SIZE(t-1) j = Size at the end of year t - 1 GR(t) j
= Growth rate during year t
2.2 Leverage and Dividends
Pecking order theory does not provide a distinctive theory of dividends, but Baskin (1989) argues that the theory can be combined with the Lintner dividend model (Lintner, 1956) to generate specific predictions for leverage. Lintner argued that firms aim for a long-run target dividend pay-out rate relative to earnings but, in the short-run, smooth their dividends from year-to-year, avoiding sharp changes, especially decreases. Therefore, high past dividend pay-outs will normally be maintained, and this will imply that profitable investment projects will have to be financed by a greater proportion of external funds than if the payout rate was lower. Pecking order theory implies that firms with higher past dividends will have less “financial slack” and therefore higher leverage, because they require more external funds5. Thus, Baskin (1989) concludes that a significant positive relationship between the past dividend rate and current leverage supports the pecking order hypothesis. In contrast, trade-off theory would suggest that dividends are high (retentions low) because external financing (including debt) is low, implying a 5
“Financial slack” is the term used by Myers and Majluf (1984) to sum up the availability of internally-generated funds in a firm.
6
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
negative or insignificant relationship between dividends and leverage. So model (2) involves adding the lagged dividend rate to model (1):
LEV(t)j = a1 + a2 ROA(t)j + a3 ROA(t-1)j + a4 SIZE(t-1)j + a5 GR(t)j + a6 DIV(t-1)j + ej …(2) where:
DIV(t-1) j = Dividend during year t – 1
2.3 Corporate Investment and Financing
According to Myers and Majluf (1984), asymmetric information may cause rejection of profitable investment opportunities because of the costs associated with raising external finance. This implies the existence of direct links between asset growth and financing which are not present according to trade-off theory. Following Baskin (1989), high dividends decrease the availability of the internal funds that are the preferred form of financing: investment outlays will therefore be negatively related to the dividend rate. We also include profitability, size and past leverage in the model. As argued above, the asymmetric information which lies behind pecking-order theory implies that larger firms are less transparent than smaller firms. After controlling for dividends therefore, we would expect a negative relation between investment and size. Profitability should be positively related to investment as it reflects the availability of internal funds. Finally, leverage should be negatively related to investment as a result of possible funding limitations arising from high leverage. (Baskin, 1989) To reduce the problem of reverse causation, we lag all the variables in this regression by one year6. Hence model (3) is:
INVGROW(t)j = a1 + a2 DIV(t-1)j + a3 ROA(t-1)j + a4 SIZE(t-1)j + a5 LEV(t-1)j + ej (3)
where:
INVGROW(t) = growth in invested capital during year t.
3. Data and Methodology
The data were extracted from the published accounts of non-financial companies listed on the Shanghai and Shenzhen stock exchanges for the years 2001 and 20027. As noted at the outset, we confine ourselves to China’s top 50 listed firms, so as to concentrate on a sample in which the 67
This issue is discussed by Chirinko and Singha (2000) in their comment on the pecking order tests of Shyam-Sunder and Myers (1999), and by Allen (1993).
Balance sheet, profit and loss and cash flow statements for these firms are available on the website of the Shenzhen Securities Information Company, Ltd (2003)
7
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
data are relatively more reliable. These consist of the top 50 firms reported by \"Listed Companies\People's Daily and the Shanghai Stock Exchange. The listing is based on total assets, income from main businesses, net profit and market value. Of the top 50 firms, 4 were excluded because they are financial companies and a further two did not report adequate cash flow information for the study. This leaves a sample of 44 non-financial companies for the analysis. Of these, 15 are in manufacturing and the rest are in non-manufacturing8.
__________________________________________________________________________
Table 1 about here
__________________________________________________________________________ The definitions of the variables used in the regression models are listed in table 1. The variables are all measured at book value. There is no widespread agreement on whether book or market values are more appropriate for tests of capital structure theory. For a discussion, see Baskin (1989) and Prasad, Green and Murinde (2001a). However, given the volatility of China’s stock market and the high proportion of shares which are still held by the state (Huang and Song, 2002), we would argue that book values are more reliable, at least for the largest firms which do make up our sample. We employ two measures of leverage (LEV1 and LEV2). The first is a wide measure: the ratio of total liabilities (excluding taxation and provisions) to total assets, and is commonly used in leverage studies (Rajan and Zingales, 1995). However, several authors argue that trade credit given is not a planned source of (negative) external financing on the part of the recipient but a means of promoting sales to others (Baskin, 1989). On this view, trade credit given contributes to the current operations of the firm and is not a means of financing the firm’s investments. This suggests that trade credit given should be deducted from total assets. Correspondingly, it can be argued that trade credit received should be deducted from liabilities. However many Chinese companies actually use trade credit as a means of financing. We argue
8
The names of the individual companies are available from the authors on request. Singh and Hamid (1992)
include only manufacturing companies in their analysis, but more recent studies include all non-financial companies, for example: Prasad, Green, and Murinde (2001b).
8
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
therefore that liabilities should not be adjusted for trade credit received. Therefore, the second leverage measure differs from the first in that accounts receivable are deducted from total assets. The regressions are based on year 2002 measures of leverage (ie. year t = 2002), with the other variables dated correspondingly, as shown in table 1 and equations (1) – (3). All the regressions were estimated by OLS. In several regressions we found that the same one to four firms produced outliers in the residuals. As a check on the results, we re-estimated the regressions including separate shift dummies for each of these firms. In all three models, the effect of the dummies was to eliminate any heteroskedasticity and non-normality. There were some important changes in the significance of the estimated coefficients in all three models. We therefore report our results without and then with these dummies9.
__________________________________________________________________________
Table 2 about here
__________________________________________________________________________ Descriptive statistics are shown in Table 2. It is difficult to compare these with other countries because, as we have emphasized, there are substantial variations in the ways in which accounting standards are applied in China. Our median leverage rate of about 50% is a relatively high figure by international standards and suggests that Chinese companies may be more highly levered than those in other developing countries (Singh and Hamid, 1995). However, as we employ broad-based definitions of leverage, these ratios may overstate the true leverage rates. Huang and Song (2002) report an almost identical figure for leverage based on the same definition as ours, but for narrower definitions, their reported leverage rates are lower. In addition, it should be borne in mind that these companies still have substantial state participation, so that the actual leverage rates we observe may simply be a bequest from the past as much as a reflection of current financing behaviour. In addition, the top 50 companies are generally given somewhat more favorable borrowing facilities by banks, and this too may encourage higher leverage rates than one might otherwise expect. 9
Also,we found 4 outliers in the data for INVGROW with values of 725.73, 10.01, 155.50 and 53.80, compared to a maximum value of the rest of 3.65, and a mean of 1.38. Therefore we re-estimated the INVGROW regressions after dropping these 4 observations.
9
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
3. Results
In general, model (1) performs well as it explains about 35% of the cross-sectional variation in leverage (table 3)10. Furthermore, the qualitative pattern and significance of the coefficients are essentially the same irrespective of the definition of leverage employed. Considering that this is a relatively simple model these are reassuring results. Profitability has a negative coefficient irrespective of whether it is lagged in the regression. Moreover, when current and lagged profitability are included together, it is the latter which is significant, and this provides relatively robust support for pecking order theory in this context. Asset growth is also significant and has a positive sign contrary to the predictions of trade-off theory. This result has been noted by several authors, for example in the seminal study of Titman and Wessels (1988). Size is signed positive, a finding that is more consistent with trade-off theory, but the coefficients are generally insignificant suggesting that not too much weight should be put on this finding.
The signs of the coefficients on profitability, growth and size are all consistent with those found by Huang and Song (2002). This is reassuring as it suggests on the one hand that our smaller sample is capable of obtaining robust results, and on the other that any unreliability in the accounts of the smaller companies used by Huang and Song may not be too serious. Despite the relatively high absolute levels of leverage which we remarked on earlier, it seems clear that the current financing decisions of Chinese companies can be explained by a relatively compact and plausible group of variables and in a manner which is not inconsistent with those of other countries. Overall, the evidence from model (1) is somewhat in favour of pecking order theory. __________________________________________________________________________
Tables 3, 4, 5 about here
__________________________________________________________________________ Model (2) involves the addition of lagged dividends to model (1) (table 4). Reassuringly, these results are broadly consistent with model (1), as the addition of dividends does not substantively change any of the findings from model (1). Dividends are positively signed as predicted by 10
The proportion of variance explained is obviously higher when the outlier dummies are included.
10
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
pecking order theory but are insignificant or only marginally significant. Model (2) therefore provides further if somewhat weaker support for the pecking order hypothesis.
We turn finally to the interaction between investment and financing (table 5). As for models (1) and (2) the general performance of this model is satisfactory, as it explains about 35% of the variation in investment growth, and the qualitative pattern and significance of the coefficients are generally robust across the two definitions of leverage. However, the coefficient estimates give a more ambiguous picture about whether we should prefer the trade-off or pecking order theory in China. Once the 4 outliers of INVGROW are dropped, the performance of the model decreases dramatically. Dividends and size are both signed negative but generally not significant, which is consistent with pecking order theory. On the other hand, positive signs on leverage and negative signs on profitability are more consistent with trade-off theory, but few of these coefficients are significant. 4. Conclusion
This paper aims to shed light on the empirical debate between the trade-off and pecking order theories of corporate financing, using data on Chinese companies. The main findings are three: first, a significant negative correlation between leverage and profitability; second a significant positive correlation between current leverage and past dividends, although at a lower significance level. These results broadly favour the pecking order hypothesis. Third, the investment model is inconclusive since an insignificant negative correlation is found between the growth of investment and the rate of past dividends. Put together, these results are tilted towards the pecking order theory, but the tilt is by no means a decisive one. Clearly, further work is needed on this, possibly with a larger sample which takes more specific account of variations in the quality of accounting data provided by Chinese companies
Perhaps the most important finding however is that a surprisingly simple and conventional model is capable of explaining a significant proportion of the cross-sectional variation in leverage among Chinese companies. This suggests that further research into Chinese companies along conventional lines will prove fruitful.
11
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
References
Allen, D.E., 1993, “The Pecking Order Hypothesis: Australian Evidence”, Applied Financial
Economics, Vol.3, pp.101-112. Baskin, J. B., 1985, “On the Financial Policy of Large Mature Corporations”, Ph D. dissertation,
Department Economics, Harvard University. Baskin, J. B., 1989, “An Empirical Investigation of the Pecking Order Hypothesis”, Financial
Management, Vol.18, pp.26-35. Booth, L., Aivazian, V, Demirguc-Kunt, A., and Maksimovic, V., 2001, \"Capital Structures in
Developing Countries\Journal of Finance, Vol. 56, no. 1, pp.87-130. Bradbury, M. and Lloyd, S. (1994), “An Estimate of the Direct Costs of Bankruptcy in New
Zealand”, Asia-Pacific Journal of Management, Vol. 11, pp. 103- 111. Chirinko, R.S. and Singha, A.R. (2000), “Testing Static Tradeoff against Pecking Order Models
of Capital Structure; A Critical Comment”, Journal of Financial Economics, Vol.58, pp.417-425. Cobham, D. and Subramaniam, R., 1998, “Corporate Finance in Developing Countries: New
Evidence for India”, World Development, Vol.26, No.6, pp.1033-1047. Colombo, E., 2001, “Determinants of Corporate Capital Structure: Evidence from Hungarian
Firms”, Applied Economics, Vol. 33, pp. 1689-1701. Fama, E.F. and French, K.R., 2002, “Testing Trade-off and Pecking Order Predictions about
Dividends and Debt”, The Review of Financial Studies, Vol. 15, No. 1, pp. 1-33. Green, C.J., Murinde, V. and Suppakitjarak. J. 2003, \"Corporate Financial Structures in India\
South Asian Economic Journal Vol. 4, No. 2, July-December pp. 245-274. Harris, M. and Raviv, A., 1991, “The Theory of Capital Structure”, Journal of Finance, Vol.49,
pp. 297-355. Huang, S.G.H. and Song, F.M., 2002, “The Determinants of Capital Structure: Evidence from
China”, Working Paper of Hong Kong University, available at http://www.hiebs.hku.hk/working_papers.asp?ID=60 Jensen, M.C. and Meckling, W.H., 1976, “Theory of the Firm: Managerial Behaviour, Agency
Costs and Ownership Structure”, Journal of Financial Economics, Vol.3, pp. 305-360. Lintner, J., 1956, “Distribution of Incomes of Corporations among Dividends, Retained Earnings,
and Taxes”, American Economic Review, Vol.46, pp. 97-113. Ministry of Finance of People’s Republic of China, 2003, “The Eighth Announcement of the
Random Check of the Quality of Accounting Information”, available at http://www.chinaacc.com/fagui/qyckfg/31013091521.htm
12
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Modigliani, F. and Miller, M.H., 1958, “The Cost of Capital, Corporation Finance and the Theory
of Investment”, American Economic Review, 48, pp.261-297. Myers, S.C., 1984, “The Capital Structure Puzzle”, Journal of Finance, Vol.34, pp. 575-592. 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, Vol. 13, pp.187-221. Prasad, S.J., Green, C.J., and Murinde, V., 2001a, \"Company Financing, Capital Structure, and
Ownership: A Survey, and Implications for Developing Economies\SUERF Studies, No. 12, February. Prasad, S.J., Green, C.J., and Murinde, V., 2001b, \"Corporate Financial Structures in Developing
Economies: Evidence from a Comparative Analysis of Thai and Malay Corporations\Finance and Development Research Programme Working Paper, No. 35, December. Rajan, R. and Zingales, L., 1995, “What Do We Know about Capital Structure? Some Evidence
from International Data”, Journal of Finance, Vol.50, pp.1421-1460. Shyam-Sunder, L. and Myers, S.C., 1999, “Testing static trade-off against pecking order models
of capital structure”, Journal of Financial Economics, Vol. 51, pp. 219-244. Shenzhen Securities Information Company, Ltd, 2003, “Accounts of China’s Top 50 Companies”,
available at: www.cninfo.com.cn. Singh, A., 1995, “Corporate Financial Patterns in Industrialising Economics: A Comparative
Study”, IFC Technical Paper No.2: IFC, Washington DC, USA. Singh, A. and Hamid, J., 1992, “Corporate Financial Structures in Developing Countries”, IFC
Technical Paper No.1, Washington D.C., IFC. Titman, S. and Wessels, R., 1988, “The Determinants of Capital Structure Choice”, Journal of
Finance, Vol.43, pp.1-19. Tong, G, 2003, “The Pecking Order Hypothesis: Evidence from China” Unpublished MSc
Dissertation, Department of Economics, Loughborough University.
Warner, J.B., 1977, “Bankruptcy Costs: Some Evidence”, Journal of Finance, Vol. 32, pp. 337- 347. Wiwattanakantang, Y., 1999, “An Empirical Study on the Determinants of the Capital Structure
of Thai Firms”, Pacific-Basin Finance Journal, 7, 371-403.
13
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Table 1 Definitions of variables used in regressions Variable Description [noncurrent.liabilities+(current.liabilities−taxation−provisions)] LEV1 total.assets[noncurrent.liabilities+(current.liabilities−taxation−provisions)] LEV2 [total.assets−accounts.receivable]profit.from.operations ROA total.assetstotal.assets.in.year.(t) GR total.assets.in.year.(t−1)dividends.paid. DIV total.equityLn (total invested capital at year-end) SIZE totalinvestedcapitalinyear(t) INVGROW totalinvestedcapitalinyear(t−1) Notes:
Year t is 2002; assets and liabilities are measured at the end of each year.
Table 2 Descriptive statistics of variables for Chinese listed firms
Variables Numbers Mean MedianMinimumMaximum
44 0.40 0.37 0.07 0.81 LEV1 (2002)
44 0.50 0.42 0.07 1.81 LEV2 (2002)
44 0.38 0.36 0.09 0.77 LEV1 (2001)
44 0.46 0.42 0.09 1.21 LEV2 (2001)
44 0.08 0.07 0.01 0.26 ROA (2002)
44 0.07 0.07 -0.05 0.18 ROA (2001)
44 0.88 0.91 0.50 1.27 GR (2002)
1.34 0.18 725.73 INVGROW (2002) 44 22.73 44 0.07 0.06 0.00 0.16 DIV (2001)
44 19.72 20.19 13.77 25.84 SIZE (2001)
14
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Table 3. Results of Model (1): leverage, profitability, growth and size
Dependent variable LEV1 LEV2 A B A B -0.32 -0.34 -0.53 -0.44 -0.55 -0.92
CONSTANT
(-1.22) (-1.28) (-1.15) (-1.28) (-1.21) (-2.20)**
-0.85 -0.15 -0.52 ROA(2002)
(-1.04) (-.25) (-1.10)
-1.18 -1.41 -1.26 -1.49 -1.63 -1.14
ROA(2001)
(-2.33)** (-3.01)*** (-1.44) (-2.29)** (-2.03)** (-1.77)* 0.42 0.41 0.78 0.52 0.76 0.83
GR(2002)
(4.03)*** (3.94)*** (4.40)*** (3.58)*** (4.31)*** (3.84)*** 0.02 0.02 0.01 0.02 0.01 0.03 SIZE(2001)
(1.77)* (1.79)* (.75) (1.56) (.77) (1.87)*
Summary Statistics and Diagnostics R-Bar-Squared F-stat.
Functional Form: CHSQ(1) Normality: CHSQ(2) Heteroskedasticity CHSQ(1)
0.38 0.37 0.36 0.65 0.36 0.63 F(4, 39) 7.49***F(3, 40) 9.54***F(4, 39) 7.06***F(6, 37) 14.37F(3, 40) 9.03***F(6, 37) 13.41***
1.11 0.01 0.57 0.00 0.01 0.04 0.77 0.42 15.93*** 0.64 22.10*** 9.13*** 0.26 0.73 12.74*** 0.30 9.18*** 0.30 44 44 44 44 44 44 Number of observations
Notes
A: estimates including outliers; B: estimates with shift dummies for outliers; other results had no outliers.
Numbers in brackets are t statistics. ***Significant at 0.01 level. **Significant at 0.05 level. *Significant at 0.10 level.
15
Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Table 4. Results of model 2: leverage, profitability, growth, previous dividend and size
Dependent variable CONSTANT ROA(2002) ROA(2001) GR(2002) DIV(2001) SIZE(2001)
LEV1 LEV2 A B A B -.46 (-1.71)* -.74 (-1.55) -1.19 (-2.41)** .44 (4.35)*** .94 (1.78)* .02 (2.12)**
-.45 (-1.65) -1.49 (-3.20)*** .42 (4.11)*** .73 (1.41) .02 (2.03)**
-.68 (-1.41) -1.08 (-1.29) -1.27 (-1.46) .81 (4.51)*** 1.00 (1.07) .02 (.94)
-.57 (-1.51) -.53 (-.79) -1.59 (-2.31)** .51 (3.29)*** 1.28 (1.73)* .03 (1.82)*
-.66 (-1.37) -1.71 (-2.10)** .78 (4.34)*** .69 (.76) .02 (.89)
-.53 (-1.53) -1.64 (-2.76)***
.51 (3.65)*** .67 (.97) .02 (1.75)*
Summary Statistics and Diagnostics
R-Bar-Squared F-stat.
Functional Form: CHSQ(1) Normality: CHSQ(2) Heteroskedasticity CHSQ(1) Number of observations
.41 .39 .36 .61 .35 .66 F(5, 38) 6.96***
F(4, 39) 7.82***
F(5, 38) 5.90***
F(6, 37) 12.02***
F(4, 39) 6.84***
F(6, 37) 14.86***
1.06 .44 .41 .03 .08 .02 1.57 .70 19.26*** .39 26.19*** .84 .12 1.24 12.62*** .20 8.81*** .39 44 44 44 44 44 44
Notes
A: estimates including outliers; B: estimates with shift dummies for outliers; other results had no outliers.
Numbers in brackets are t statistics. ***Significant at 0.01 level. **Significant at 0.05 level. *Significant at 0.10 level.
16Pecking Order or Trade-off Hypothesis-CT-ep.doc
© Guanqun Tong and Christopher J. Green 2004
All Rights Reserved
Table 5. Results of model 3: growth of investment, previous profitability, size, previous leverage and previous dividends level
Dependent
INVGROW
variable
A B 399.24 423.60 393.49 394.79 3.58 3.79 3.82 3.90 CONSTANT
(2.77)*** (2.92)*** (2.82)*** (2.88)*** (2.45)** (2.50)** (2.72)*** (2.83)*** -66.20 -155.81 -28.23 -.38 -1.42 -1.50
ROA(2001)
(-.17) (-.43) (-.09) (-.11) (-.45) (-.53) -21.12 -21.25 -21.29 -21.44 -.11 -.11 -.11 -.12
SIZE(2001)
(-2.98)*** (-3.05)*** (-3.07)*** (-3.22)*** (-1.53) (-1.48) (-1.52) (-1.70)*
.61 -20.72 LEV1(2001)
(.64) (-.20)
-55.64 .03 LEV2(2001)
(-.81) (.05)
790.23 820.34 768.63 764.28 -3.66 -3.00 -2.96 -3.22
DIV(2001)
(2.07)** (2.23)** (2.13)** (2.16)** (-1.03) (-.86) (-.88) (-.98)
Summary Statistics and Diagnostics
R-Bar-Squared F-stat.
.22 .24 .24 .26 .01 .34 .03 .05 F(4, 39) 4.10***
F(4, 39) 4.32***
F(3, 40) 5.59***
F(2, 41) 8.59***
F(4, 35) 1.12
F(4, 35) 1.00
F(3, 36) 1.38
F(2, 37) 1.96
Functional Form:
26.54*** 26.16*** 25.13*** 24.66*** 0.01 0.01 0.01 0.02
CHSQ(1) Normality: 750.41*** 733.97*** 746.93*** 749.01*** 5.23* 5.74* 5.76* 4.92*
CHSQ(2)
Heteroskedasticity
32.26*** 31.73*** 32.35*** 32.36*** 0.36 0.11 0.10 0.23
CHSQ(1) Number of
44 44 44 44 40 40 40 40 observations
Notes
A: estimates including outliers; B: estimates after dropping 4 observations.
Numbers in brackets are t statistics. ***Significant at 0.01 level. **Significant at 0.05 level. *Significant at 0.10 level.
17Pecking Order or Trade-off Hypothesis-CT-ep.doc
因篇幅问题不能全部显示,请点此查看更多更全内容