Determinants of corporate borrowing: Some evidence from the Indian corporate structure
Bhaduri, Saumitra N
Abstract
In contrast to previous empirical work on capital structure, which is mainly confined to the United States and a few other advanced countries, this paper attempts to study the capital structure choice of developing countries through a case study of the Indian corporate sector. The paper shows that the optimal capital structure choice is influenced by factors such as growth, cash flow, size, and product and industry characteristics. (JEL G32)
Introduction
Ever since Modigliani and Miller (1963) have made their irrelevance proposition, a key theme in corporate finance has been to explain the conditions under which capital structure does affect firm value. However, the existing empirical research on this issue has been largely confined to the United States and a few other developed countries that may have institutional similarities. The issue of capital structure choice in developing countries has, however, received little attention. Only in recent years, a few studies have emphasized these issues. The prevailing view in this context-for example, Mayer (1990) – seems to be that financial decisions in developing countries are somehow different from the issue of capital structure choice in developed economies. To empirically corroborate his view, Mayer (1990) has used an aggregate flow of funds data instead of data from individual firms. This approach, however, poses a problem since it does not control for many firm-specific attributes that can influence individual financial decisions.
On the contrary, some recent empirical studies have attempted to shed light on the capital structure issue within the institutional specifics of developing countries using firm-specific information (Booth et al. 2001; Cobham and Subramaniam 1995; Cherian 1996; Singh 1995). In a cross-country study Booth et al. (2001) have used a sample of 10 developing countries: five former British colonies (India, Pakistan, Thailand, Malaysia, and Zimbabwe) and two Latin American countries (Mexico and Brazil) with a common inflationary experience and three others (Turkey, Jordan, and Korea). In marked contrast to their predecessors, their study reveals that the factors pertaining to the issue of the determination of capital structure in the United States and the European countries are also potent enough in explaining financial decisions in developing countries despite the profound differences in the institutional framework in which they operate.
However, most of these existing studies on capital structures in developing countries have been restricted in scope due to the poor cross-sectional variation in data. The data set used for Singh’s (1995) study constitutes only the top 100 companies, which might not be representative of the corporate sector in the developing countries, while Cherian (1996) and Cobham and Subramaniam (1995) used consolidated balance sheet data provided by the Reserve Bank of India and other financial institutions. Finally, Booth et al. (2001), who use a sample of 100 top companies in 10 developing countries, also suffer from similar limitations in terms of cross-sectional variation as well as sample selection bias. Hence, despite the illuminating contributions of the recent researchers, the existing literature in modern finance, in strict sense, still lacks adequate depth in empirical research on the issue of capital structure choice in developing countries due to poor cross-sectional variation in their samples.
This paper is an attempt to study the issue of capital structure choice in developing countries through a case study of the Indian corporate sector. The data for this analysis are drawn from the Centre for Monitoring Indian Economy (CMIE) CIMM database. The CIMM database reports balance sheet data for a large number of Indian firms. From this data set we selected firms based on the criteria that the firms should have maintained their identity and reported their annual accounts without any gaps for the financial years 1989-90, 1990-91, 1991-92, 1992-93, 1993-94, and 1994-95. Screening for data consistency1 on the basis of this criterion led to the selection of a sample of 363 firms across nine broad industries. Therefore, the study adds value to the existing literature to the extent that it can adequately control for the possible sources of differentiation among firms in their optimization choices and provide more reliable insights into the validity of various mainstream capital structure theories.
Finally, a rapidly changing economic scenario in developing countries, in the recent years, has also provided a substantial need to address the issue of the determination of the capital structure choice in these countries. Developing countries, such as India, have been introducing many market-oriented reforms in their financial sectors since the mid-eighties and early nineties. This has led to a substantial transformation of the institutional setup within which firms have operated, giving more flexibility to the Indian financial manager in choosing the capital structure of the firm. The move toward the free market, coupled with the widening and deepening of various financial markets, including the capital market, has provided the scope for the corporate sectors to optimally determine their capital structure. Such an environment has also encouraged more meaningful research on the capital structure issue. This paper attempts to study the capital structure issue in India in the context of the country’s ongoing economic reforms.
The empirical methodology and the data set per se also add value to the existing work on capital structure. Specifically, the study extends the existing empirical work on Indian capital structure in three ways. First, it extends the range of theoretical determinants of capital structure by examining some recently developed theories that have not yet been analyzed with the Indian data set. Second, this paper also accounts for the fact that some of the capital structure theories have different empirical implications with regard to maturity structure of debt instruments. To this end, we introduce separate measures of short-term and long-term borrowings rather than an aggregate measure of total borrowing (as has been the case in many studies). Third, the methodology used in this paper recognizes and tries to mitigate the measurement problem by using an exploratory factor analysis technique.
The remainder of the paper is organized as follows: the second section describes firm-specific attributes suggested by various theories that are important in determining capital structure and lists the proxies that we intend to use in this exercise. The third section develops the methodology used in this paper. The fourth section interprets the results, and the final section concludes.
Determinants of Capital Structure
In this section, we list various firm-specific attributes suggested by capital structure theories and mention the proxies that are used to capture such attributes.
1. Asset structure: The agency cost and asymmetric information theories of capital structure suggest that the composition of assets owned by a firm influence its capital structure choice. According to the agency cost theory, the shareholders of a highly leveraged firm have a tendency to invest sub-optimally (Galai and Masulis 1976; Jensen and Meckling 1976; Myers 1977). However, a firm with collateralized assets can restrict such opportunistic behavior. Hence, we expect a positive association between collateralizable assets and debt.
Another agency problem arises from the tendency of a firm’s manager to consume more than the optimal level of perquisites, which reduces the value of the firm. However, firms with less collateralizable assets are more vulnerable to such agency costs since monitoring the capital outlays is more difficult for such firms (Grossman and Hart 1982). The firm can use a high debt level as a monitoring instrument to mitigate this problem. Therefore, one can expect a negative association between leverage and collateralizable assets. Moreover, in the presence of asymmetric information, firms may find it advantageous to sell secured debt as it reduces the information premium.
We use three proxies for the collateral asset attribute, namely, the ratio of land and building to total assets (LB), the ratio of plant and equipment to total assets (PM) and the ratio of inventories to total assets (INV). Since the values of these collateral assets can depend on the maturity structure of the debt instruments, we include them separately rather than using an aggregate proxy.
2. Non-debt tax shield: DeAngelo and Masulis (1980) in their study argued that non-debt tax shields are substitutes for tax benefit of debt financing. Thus, a firm with large non-debt tax-shield is likely to be less leveraged. However, the identification and measurement of non-debt tax shield is somewhat problematic. Several proxies such as depreciation, tax loss carry forward, and investment tax credits are used to capture this effect. This study has used the ratio of a change in accumulated depreciation to net operating income (DEP) as a proxy for non-debt tax shield of a firm.
3. Size: There is considerable evidence that the size of a firm plays an important role in capital structure choice. Large firms tend to be more diversified and hence likely to be less susceptible to financial distress. Therefore, a positive association is expected between firm’s size and leverage. Many authors have also suggested that direct financial distress cost is inversely related to firm size (Warner 1977; Ang, Chua, and Mcconnel 1982). Moreover, if the capital market is characterized by transaction costs associated with the issue decision, we can expect size to be an important determinant of capital structure. We introduce natural logarithm of total assets (ln(TA)) as a proxy for firm size.
4. Financial Distress: Since debt involves commitment of periodic payment, highly leveraged firms are prone to financial distress costs. Therefore, firms with volatile incomes are likely to be less leveraged. This study introduces two measures of volatility: an indicator variable for financial distress (B-PROB) and the standard deviation of a percentage change in operating income (VOL) multiplied by probability of financial distress (VOL*B-PROB). The second indicator captures the significance of the volatility effect as it depends on the distance from the financial distress threshold.
5. Growth: For growing firms, the agency problems are likely to be more severe since they are more flexible in their choice of future investments. This is indicative of a negative association between long-term debt and future growth of a firm. Further, Myers (1977) suggested, short-term debt could be used to mitigate this problem; growing firms can use short-term debt instead of long-term debt to avoid such agency costs. This paper uses two indicators for the growth attribute: the ratio of capital expenditure over total assets (CE) and growth of total assets (GRTA).
6. Profitability: If managers of a firm cannot credibly convey inside information to outsiders, they prefer internally generated capital to external financing (Myers and Majluf 1984). In the presence of asymmetric information, a firm would follow a pecking order preference pattern: it would prefer internal finance, but would issue debt if such low cost options were exhausted. The firm’s last option would be to choose equity issue. Since profitable firms are likely to have more retained earnings, we expect a negative association between leverage and past profitability. On the contrary, static trade-off theories envisage a positive relation between profitability and leverage because a firm with high profits would require greater tax shelter and would have more debt– taking capacity. However, the theoretical models envisaging an association between cash flow and maturity structure is complex and possibly non-monotonic (Diamond 1991a,b). Diamond argues that the firms at the lower end of the quality spectrum would depend more on the short-term borrowing and as credit worthiness increases they tend to depend more on the long-term borrowing.2 Two indicators are selected as proxies for profitability: the ratio of cash flow over total assets (CF/TA) and the ratio of cash flow over sales (CF/S).3
7. Age: The existing literature considers age an important determinant of capital structure. Given the fact that young firms are more vulnerable to the problem of asymmetric information, they are likely to use debt and avoid the equity market. In contrast, as Huisman and Hermes (1997) suggested, the industrial policies in developing countries provide the small and young firms an access to cheap credit and hence make debt cheaper than equity. The study introduces a dummy variable that takes the value one if the firm is below the age of 20 years and zero otherwise, as an indicator of age (AGE-D).
8. Signaling: If a firm can credibly signal its quality to the outsiders, it can avoid an information premium and hence may access external sources of funds, particularly the equity market. This paper introduces two proxies to capture the signaling attribute: the ratio of dividend payment to net operating income (DIVI) and a group dummy (GDUM). The group dummy takes value one if the firm belongs to a business group and is zero otherwise. We expect firms belonging to business groups to suffer less from the problem of asymmetric information because of their reputation. We consider dividend ratio because many studies have argued that dividends are used as a costly signal of earnings. A firm with a reputation of dividend payment faces less asymmetric information in accessing the equity market (John and Williams 1985; Miller and Rock 1985).4 On the other hand, if dividend payment represents a signal of better financial health and hence more debt-taking capacity, one would expect a positive association. In addition to signaling models, agency models also draw links between the dividend payment and leverage (Jensen et al. 1992). Specifically, agency models envisage dividend payment and debt issue as a substitute in mitigating agency problems. Therefore, the existing theories envisage an inverse relationship between leverage and dividend payment.
9. Uniqueness: Firms characterized by unique products are likely to be less leveraged (Titman 1988). A firm with a unique product imposes potential costs on their customers, input supplier, and workers when facing liquidation. Such firms also find it difficult to borrow because their specific use of capital reduces the probability of an alternative use in the event of bankruptcy. Firms with unique products are likely to spend more on R&D because their products are less likely to be duplicated by other firms. Moreover, such firms are likely to incur high selling expenses in order to promote their unique product. Therefore, the indicators for uniqueness are the ratio of R&D to sales (R&D) and the ratio of selling expenses to sales (SEXP).
Measurement of Capital Structure
The objective of our study is to analyze various measures of debt, depending on the maturity structure, because many of the capital structure theories have different implications for different types of debt instruments. We have used three measures of leverage in this study. These are total borrowing to asset ratio (TB/TA), long-term borrowing to asset ratio (LTB/TA), and short-term borrowing to asset ratio (STB/TA).
All variables are measured in book values and not in market values because of data limitation. This might introduce some bias because many of the predicted hypotheses will depend on the measure of leverage used. However, some existing studies have pointed out a possibility of high cross-sectional correlation between book value and market value of debt (Bowman 1980). In such a case mis-specification due to the use of book value will be small.
Data
Data on the proxies for various unobserved attributes discussed in the previous section are collected from CIMM5 database. Our data set comprises of 363 manufacturing firms collected across nine broadly defined industries. The distribution of firms by industries is presented in Table 1. In our sample almost 60 percent of firms come from chemicals, textile, or machinery, indicating the general structure of India’s manufacturing sector. The data on cross-section variables are collected over the 1989 through 1995 time period. Descriptive statistics of variables are presented in Table 2. Since data on the firm-level R&D expenditure are not available, we use industry level data as the second best proxy for it. This implicitly assumes that all the firms in that particular industry incur the same amount of R&D expenditure. Therefore, the resulting bias should be considered while interpreting the results.
In any factor analysis it is important to have a large sample to ensure reliability. Experts have suggested that a sample size of 100-200 is good enough for most purposes, particularly when numbers of variables are not too large. Thus, selection of a sample of 363 firms enables us to do a meaningful factor analysis. It is noteworthy that though the sample covers a wide range of industries and has adequate variation across size, age, and ownership structure, any conclusion that we make should be seen in the context that ours is not a totally random sample.
All explanatory proxies, except the growth rate,6 are averaged over a five-year period (199094) to reduce the measurement error due to random year-to-year fluctuation in variables. We also take into account the simultaneity problem by lagging proxies by one year, that is, the dependent variables are from 1995. This will help us to infer causal links between leverage and various firmspecific attributes rather than a short-term effect.
Methodology
The empirical work undertaken to identify determinants of capital structure has been lagging behind theoretical research, perhaps due to the fact that relevant attributes advanced by various capital structure theories are quite abstract in nature and are not directly observable. Theoretical models in capital structure have emphasized many factors like bankruptcy, agency conflict between various stakeholders, and signaling as important determinants of capital structure. But these factors are unobservable in nature and can only be captured by noting their effect on the observed variable. Therefore, some technique that recognizes and mitigates this measurement problem becomes necessary in this empirical research. Most of the existing empirical work on capital structure has ignored this issue and used simple regression techniques with proxies for the unobservable theoretical attributes to explain the variation in leverage ratios across firms. This approach suffers from a number of shortcomings as pointed out by Titman and Wessels (1988). First, there may be no unique representation of the attribute we wish to measure. Second, it is difficult to find a measure of a particular attribute that is unrelated to the other relevant attributes. Third, since the observed variables are imperfect representations of the attributes they are supposed to measure, their use in regression analysis introduces an error in variable problem. Finally, measurement errors in the proxy variables may be correlated with measurement errors in the dependent variable creating spurious correlation even when the unobserved attribute being measured is unrelated to the dependent variables. In order to validate their argument, they used confirmatory factor analysis (specifically the LISREL technique) to measure unobserved or latent variables, which proved to be a pioneering work beyond doubt. Thus, it provides a strong motivation to use confirmatory factor analysis.
Factor analysis can be used for exploratory as well as for confirmatory analysis, depending on the objective of the study. Factor analysis may be used as an expedient way of ascertaining the minimum number of hypothetical factors that can account for the observed correlation and as a means of exploring the data for possible data reduction. The analysis used for exploring the underlying factor structure without prior specification of number of factors and their loading is called exploratory factor analysis. Factor analysis can also be used as a means of testing a specific hypothesis on the factor loading structure, which is referred to as confirmatory factor analysis.
The measurement model is estimated using a simple factor analytic model. In reality, many variables of theoretical interest are not directly observable. This is the fundamental idea behind the factor analytic model. These unobserved variables are known as latent variables or factors. The information about latent variables can be obtained by noting the impact on observed variables. Factor analysis is a statistical tool to determine a minimum number of unobservable common factors (which are smaller in number than the number of variables) by studying the covariance among a set of observed variables. The methodology adopted in this paper proceeds in two steps. The first step involves extraction of initial factors. There are various methods that could be used for the factor extraction: the Principal Component method, Maximum Likelihood method, Least Square method, Alpha Factoring, Image Factoring, etc. We have employed one of the most widely used methodologies, the Principal Axis with iteration (P.A.A) method, to extract the initial factors, mainly because it is easily comprehensible and requires less stringent conditions (Kim and Muelller 1988). The minimum number of factors is extracted by using the commonly used Kaiser rule of thumb, i.e., the initial eigen value should be greater than or equal to one. We have also used a scree plot to determine the number of factors to retain. The scree diagram plots initial eigen values in a descending order against a number of factors. An elbow on the scree plot indicates the point at which the inclusion of any additional factor does not significantly add to the explanatory power of the model. Therefore, factors that are below the elbow are rejected. However, it is important to note that this procedure involves a certain amount of subjectivity, particularly when there is no clear elbow or many of them in the scree plot.
Implementation of the first step gives us a factor structure matrix: a matrix of coefficients where the coefficients refer to the correlation between factors and variables, known as factor loading. In the next step we try to identify the factor by observing the factor-loading pattern after an orthogonal rotation, as well as by using the knowledge of various indicator variables used as a proxy for unobservable attributes.
In the second step, we estimate the factor scores and then run a regression of various factor scores on various measures of leverage ratio and test the statistical significance of the extracted factors to explain the variation in the leverage ratio across firms. Since the factor scores are generated through an orthogonal transformation, they do not pose any multi-collinearity problem at the regression equation. At the regression stage, we also include an industry dummy along with factors to explain variation in the observed leverage. The inclusion of the industry dummy is motivated by our observations in Paper 3 that the capital structure choice in India varies across industries. Moreover, the inclusion of the industry dummy is also justified by considerable literature that has predicted industry effects. The theoretical models by Titman (1984), Maksimovik (1988), and others also predict industry effects (Chevalier 1993). The reason for including industry dummies at this stage and not in the initial stage stems from the consideration that an industry dummy usually represents a unique factor rather than a common factor.
Identification of Factors
A factor loading of 0.30 indicates at least a 9 percent overlap in variance between the variable and the factor. Therefore, as a rule of thumb, only loadings that are more than 0.30 are eligible for interpretation. The greater the overlap is between a variable and a factor, the higher the likelihood that the variable is a pure measure of the factor. As a rule of thumb, a loading of more than 0.71 (50 percent overlap) is considered excellent, 0.63 (40 percent) very good, 0.55 (30 percent) good, 0.45 (20 percent) fair, and below 0.32 (less than 10 percent of overlap) poor. In this study we have used 0.32 (10 percent overlap) as the cutoff for interpretation. However, the results are fairly robust to the choice of cutoff.
* Factor 1: The loading pattern given in Table 4 shows that factor 1 is highly loaded in favor of variables LB, PM, and INV. As we have already mentioned that LB, PM, and INV are used as proxies to asset structures, this can be identified as collateral value factor.
* Factor 2: Factor 2 can be identified as a growth factor, as it is loaded in favor of CE and GR (TA), which are used as proxy to growth.
* Factor 3: This factor is a little ambiguous, as it is loaded in favor of many variables. It is positively loaded in favor of size and group dummy but negatively loaded in favor of volatility. We identify this factor as the size factor. The positive loading in favor of group dummy and negative loading on the volatility indicates the possibility that the large firms are often more prone to financial distress and also affiliated to business groups.
* Factor-4: From the loading pattern we can identify this factor to be the cash flow factor, as it is relatively loaded in favor of the variable that measures cash flow of the firm. The loading on capital expenditure (CE) indicates that the profitable firms are also the ones to invest more. The fact that this factor assigns a positive load, though insignificant, in favor of dividend also supports our identification of the factor. This is because dividend payment often represents a signal of better financial health.
* Factor-5: Factor five can be identified as a uniqueness factor, as it is loaded in favor of R&D and SEXP, which are used as proxies to uniqueness.
Goodness of Fit
Our five-factor model explains 58.8 percent of variations in the data. There is a wide range of variation in the proportion being explained by each factor. In contrast to the first factor that explains 22.5 percent of variation, the fifth factor only explains 5.6 percent of the variation. Various measures of goodness of fit discussed in an earlier section show that our five-factor model moderately explains an observed correlation pattern. In terms of residuals, there are only 3.0 percent residuals with absolute values greater than 0.05. The Kaiser’s measure of sampling adequacy for the model is 0.65, implying a “mediocre” fit.
Regression Results
In the second step we regress the factor scores of our five-factor model on various measures of leverage, and the results are presented in Table 5. Except the growth factor, other statistically significant coefficients on factors are consistent with our predicted hypotheses. Moreover, the statistically significant results are consistent across various measures of leverage.
The coefficients on the size factor are consistent in sign with our hypotheses. The coefficients are significant for long-term borrowing and short-term borrowing models, but not for total borrowing. Our regression results show that firms with large size depend more on the long-term borrowing, while the small firms depend more on short-term borrowing. A possible explanation for this is the fact that small firms might face high transaction costs in raising long-term borrowing. It is important to note that the size factor, which is highly and negatively loaded in favor of volatility, is also likely to capture the financial distress effect. Therefore, a significant coefficient on size factor also indicates the fact that small firms are more dependent on the shortterm borrowings, probably because firms with poor health solve the risk premium problem by issuing short-term borrowing, as it involves less risk for the creditor.
The coefficients on the cash flow factor are significant for the short-term and total borrowings. However, it is not significant for long-term borrowing. The theoretical models envisaging an association between cash flow and maturity structure are complex and possibly nonmonotonic (Diamond 1991 a,b). Diamond argues that the firms at the lower end of the quality spectrum would depend more on the short-term borrowing, and as credit worthiness increases they tend to depend more on the long-term borrowing.8 If cash flow can be treated as an indicator of a firm’s quality and credit worthiness, then our finding of a negative coefficient on the cash flow variable in the short-term model therefore lends partial support to this hypothesis. However, a negative and insignificant coefficient on the long-term borrowing does not have any obvious interpretation.
The coefficients on the growth factors are significant but inconsistent with our agency cost hypothesis. However, one can interpret these positive coefficients as indicative of the fact that growth opportunities add value to the firm and hence increase long-term debt-taking capacity. Moreover, as growing firms require more finance to support their planned capital expenditure, they are likely to be more leveraged, particularly if the equity market imposes more transaction and information cost than the debt-a result consistent with the pecking order theory. However, this is not significant in the short-term model.
The uniqueness factor in the regression model turned out to be consistent and significant. The significant negative coefficients on uniqueness indicate that the firms with high selling expense (SEXP) and research and development expenditure (RND) are the ones with less leverage. Firms with unique products find it difficult to borrow because of their specific use of capital and less tangible assets.
Surprisingly, the collateral value factor turns out to be insignificant in all cases. We do not get any significant association between the share of long-term borrowing and the share of fixed assets in the portfolio. This indicates a possibility that the term loans are not always used to finance longer-lived assets.9 We also find strong evidence in favor of industry effects in explaining observed variation in capital structure across firms. Now turning to the diagnostic of our regression model, the adjusted R^sup 2^ is satisfactory in all cases. The adjusted R^sup 2^ is 0.24 for the aggregate borrowing model, while it is 0.27 and 0.16 for the long- and short-term models, respectively. The F-values are also significant in all three cases. Further, the explanatory power of our model is higher when compared to Titman and Wessels’ (1988) study. We also check the robustness of our results by redefining the dependent variable as an average over three years (1993-95) and correcting for unknown form heteroskedasticity (White 1980). Our findings remain invariant to such robustness checks.
Finally, it is important to contrast our results with that of Boot et al. (2001). Notwithstanding the limitation of IFC data used by Boot et al. (2001), both in terms of company coverage and lack of sufficient information to construct many variables10 that are known to be relevant from studies of firms in developed countries, their results show trends very similar to ours in terms of statistically significant factors influencing the optimal capital structure choice. Similar to Boot et al. (2001), our paper also provides evidence that the optimal capital structure choice in developing countries is strongly influenced by factors such as size, asset structure, profitability, and shortterm financial distress cost. To the extent that our study can adequately control for the possible sources of differentiation among firms in their optimization choices and incorporate many relevant variables that were missing in the earlier studies, it provides more reliable insight into the validity of various mainstream capital structure theories.
Conclusion
The study presents a simple factor analytic model to explain the observed variation in capital structure in terms of five unobserved factors: growth, cash flow, size, uniqueness, and industry character. Our results in this paper provide two major insights into the validity of many mainstream capital structure theories. First, while our results relate directly to India, they are of more general interest. Most of the agency theoretic and information-based models of financial structure have been developed with the aim of explaining the data relating to advanced countries like the U.S. These models usually assume the existence of well functioning liquid financial markets in which investors can diversify risks and the existence of an efficient legal system in which a broad range of property rights can be enforced. These assumptions are often violated in developing countries and, therefore, testing these models with data from economies with less developed financial markets and very different institutions provides a case for testing robustness of these models. In this sense our findings can be viewed as a complement to the work of Titman and Wessels (1988) for the U.S. and Rajan and Zingales (1995) for a sample of several developed countries.
Second, despite the difference in institutional framework in India, our findings point out that both the agency theoretic and information-based models of capital structure can successfully predict the capital structure choice in our sample. Moreover, the explanatory power of the theory is quite strong when compared to many studies using U.S. data (e.g., Titman and Wessels 1988). Our results also indicate a wider set of determinants of the optimal financing choice, which can be attributed to differences in market efficiency and the legal system. Since alternative contractual means of resolving agency conflicts are missing in many developing Countries, the relationships derived from simple agency theoretic models are more likely to have empirical validity here. Therefore, contrary to general intuition, many of the mainstream capital structure theories do conform to the realities of developing countries, despite their differences in institutional characteristics.
1 Firms that underwent large merger and acquisitions during the sample period were dropped from the sample. We also dropped firms that have diversified into various industries from the sample.
2 However, there could be cost associated with long-term borrowing as firms cannot take advantage of revelation of good news. Therefore, as pointed out by Diamond, debt maturity choice involves a trade-off between borrowers’ preference for short-term borrowing due to private information and liquidity risk arising from the opportunistic lenders.
3 We also use proxies like PBDIT/TA and PBDIT/SA to capture the profitability factor in an unreported exercise. The result is, however, invariant to such choice.
4 Another reason for such negative association could be a high correlation between past profitability and dividend payment.
5 CIMM (Corporate Information in Magnetic Medium) database is developed by CMIE (Centre for Monitoring Indian Economy).
6 Growth rate is calculated over 1989 to 1995.
8 However, there could be cost associated with long-term borrowing as a firm cannot take advantage of revelation of good news. Therefore, as pointed out by Diamond, debt maturity choice involves a trade-off between borrowers’ preference for short-term borrowing due to private information and liquidity risk arising from the opportunistic lenders.
9 This result is also consistent with previous findings by Demirguc-Kunt and Maksimovic (1994). However, it is important to note that, this can also be due to the fact that collateral value factor has high loading for land and building, plant and equipment and inventory, but they might not be equally accepted collateral.
10 The proxies related to tax rate, dividend payment behaviour and long term business risk are not captured in Booth et al. (2001) study due to lack of data.
References
Ang, J., Chua, J., and J. McConnell. 1982. “The Administrative Costs of Corporate Bankruptcy: A Note.” Journal of Finance 37: 219-226.
Booth, L., V. Aivazian, A. Kunt-Demirguc, and V. Maksimovic. 2001. “Capital Structure in Developing Countries.” Journal of Finance 56: 87-130.
Cherian, S. 1996. “The Stock Market as a Source of Finance: A Comparison of U.S. and Indian Firms.” The World Bank Policy Research Working Paper, No. 1592.
Chevalier, J. 1993. “Capital Structure and Product Market Competition: An Empirical Analysis of the Supermarket Industry.” Doctoral dissertation, Massachusetts Institute of Technology. Cobham, D., and R. Subramaniam. 1995. “Corporate Finance in Developing Countries: New
Evidence for India.” Discussion Paper No. 9512. U.K.: University of St. Andrews.
DeAngelo, H., and R. Masulis. 1980. “Optimal Capital Structure under Corporate and Personal Taxation.” Journal of Financial Economics 8: 3-29.
Diamond, D. W. 1991a. “Monitoring and Reputation: The Choice between Bank Loans and Directly Placed Debt.” Journal of Political Economy 99: 689-721.
Diamond, D. W. 1991b. “Debt Maturity Structure and Liquidity Risk.” Quarterly Journal of Economics 106: 709-737.
Grossman, S., and 0. Hart. 1982. “Corporate Financial Structure and Managerial Incentives. In The Economics of Information and Uncertainty, edited by J. McCall. Chicago: University of Chicago Press.
Galai, D., and R. Masulis. 1976. “The Option Pricing Model and the Risk Factor of Stock.” Journal of Financial Economics 3: 53-81.
Huisman, J., and N. Hermes. 1997. “Financial Liberalization in India and the Impact on Business Investment.” International Journal of Development Banking 15: 3-14.
Jensen, M., and W. Meckling. 1976. “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure.” Journal of Financial Economics 3: 305-360.
Jensen, G. R., P. Solberg, and T. S. Zorn. 1992. “Simultaneous Determination of Insider Ownership, Debt and Dividend Policies.” Journal of Financial and Quantitative Analysis 27: 247-263.
John, K., and J. Williams. 1985. “Dividends, Dilution and Taxes: A Signaling Equilibrium.” Journal of Finance 40: 1053-1070.
Kim, J., and W. Mueller. 1978. “Factor Analysis: Statistical Methods and Practical Issues.” Sage Publications.
Kunt, A. D., and V. Maskimovic. 1995. “Stock Market Development and Firm Financing Choice.” World Bank Policy Research Working Paper, No.1461.
Maksimovic, V. 1988. “Capital Structure in Repeated Oligopolies.” Rand Journal of Economics 19:389-407.
Mayer, C. 1990. “Financial System, Corporate Finance and Economic Development.” In Asymmetric Information, Capital Market and Investment, edited by G. Hubbard. Chicago University Press for NBER.
Modigliani, F., and M. H. Miller. 1963. “Corporate Income Tax and the Cost of Capital: A Correction.” American Economic Review 53: 433-443.
Miller, M. H., and K. Rock. 1985. “Dividend Policy Under Asymmetric Information.” Journal of Finance 40: 1031-1051.
Myers, S. C. 1977. “Determinants of Corporate Borrowing.” Journal of Financial Economics 9: 147-176.
Myers, S. C., and N. Majluf. 1984. “Corporate Financing and Investment Decisions When Firms Have Information Investors Do Not Have.” Journal of Financial Economics 13: 187-221. Rajan, R. G., and L. Zingales. 1995. “What Do We Know about Capital Structure? Some
Evidence from International Data.” Journal of Finance 50: 1421-1460.
Singh, A. 1995. “Corporate Financing Patterns in Industrializing Economics: A Comparative International Study.” International Finance Corporation Technical Paper, no. 2. Washington, DC: World Bank.
Titman, S. 1984. “The Effect of Capital Structure on a Firm’s Liquidation Decision.” Journal of Financial Economics 13: 137-151.
Titman, S., and R. Wessels. 1988. “The Determinants of Capital Structure.” Journal of Finance 1: 1-19.
Warner, J. 1977. “Bankruptcy Costs: Some Evidence.” Journal of Finance 32: 337-347.
Saumitra N. Bhaduri *
* Saumitra N. Bhaduri, Madras School of Economics, Gandhi Mandapam Road, Chennai-25, sbhaduri hhotnail.com.
Copyright Journal of Economics and Finance Summer 2002
Provided by ProQuest Information and Learning Company. All rights Reserved
