Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 What Does Capital Structure Decision Depend On? Evidence from S&P Cnx Nifty and Nifty Junior Firms N R Parasuraman and P Janaki Ramudu In this paper we attempted to analyze as to how Indian firms went about in designing their capital structure during the period 2002 through 2011. The findings of the study revealed that designing the capital structure depended on profitability in general and on ROCE and RONW in particular. The hypotheses testing revealed that, on an aggregate basis, capital structure decisions of Indian firms were influenced by profitability, solvency, size, tax factor and growth opportunities. The analysis pertaining to the impact of individual factor on leverage decision revealed that there was no link between growth opportunities and the way the firms designed their capital structure. Though other factors impacted the leverage decision, ‘t’ values of beta co-efficients indicated that the influence was not so high. Though the extent of reliance varied among the years, the findings reveal that capital structure decision of Indian firms depended mostly on profitability followed by short-term solvency, size and tax laws in India. The study goes into in-depth analysis of capital structure decisions by Indian firms and made its own contribution to the body of knowledge on the topic. Field of Research: Finance I. Introduction, literature survey and gap in research: Companies take into account several factors before deciding on an ideal capital structure for themselves. Theoretically, a lot of postulates have been put forth to focus on key areas which companies consider in this regard. The theories are based on specific assumptions that they make. While none of the theories can be dismissed in their entirety, it is also a fact that empirical evidence does not suggest a complete validation of any theory. The famous Modigliani-Miller (1958) propositions and other postulates including the signaling theory and the pecking order theory have their supporters, but hardly can these be found suitable for explaining specific practices in companies. Therefore in order to explain the capital structure practices in Indian context we have undertaken this study on the constituent companies of NSE NIFTY and NIFTY JUNIOR. The study is organized in a logical flow of (I) introduction, literature survey and gap in research, (II) objectives and methodology, (III) hypotheses formulation, (IV) model development, (V) results and discussions and (VI) conclusions. For a long time, the theory of target capital structure also found favor with academics and practitioners alike. The theory is elegant in that it suggests that companies have in mind what they want in the long run and work towards it. However, experience tells us that barring a few exceptions, ________________________________________________________________ N R Parasuraman, Sdm Institute for Management Development, India. E-mail: nrparasuraman@sdmimd.ac.in P P Janaki Ramudu, Alliance University School of Business, India. E-mail: pjanakiramudu@yahoo.co.in 1 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 companies have been unable to consistently follow this principle. Also, it is not clear as to how the target can be correctly determined and also, the time frame within which the target needs to be accomplished. The theory of capital structure used to give a lot of importance to targets so much so that the weighted average cost of capital was said to be best taken with target weights in mind. Further, the decision regarding capital structure is also influenced by the extent of present borrowings. A company that is already under heavy borrowing may be reluctant to go in for new capital intensive projects entailing debt, for the simple reason that it fears inability to service interest and repayment schedules on time. Further, banks and lending institutions have their own norms regarding this. Debt capacity depends not merely on the viability of new projects, but also on the ability of the company as a whole to service existing debt. Other considerations like financial distress and agency costs, signaling effects and the pecking order theory also influence the determination of the target capital structure. The pecking order theory looks at the tendency of companies to first utilize short term debt facilities (even for longer term use), and other forms of debt in preference to equity The principle is that equity raising would involve more questions which could be avoided in a debt issue. While there is no unanimity, empirical studies have come to varied results as regards the applicability of these theories in practical decision-making. The results have been different across countries and over time. Raghuram Rajan and Zingales (1995) in their research analyzed financing decisions of the firm in the major industrialized countries. They found that the leverage decision of the firms across G-7 countries was fairly similar and the factors identified by earlier studies in the U S were similarly correlated with other countries. They however found that the theoretical propositions of capital structure decisions were still unresolved. This study has made a significant contribution to the body of knowledge in capital structure theories. John R. Graham (2000) estimated as to how big would be tax benefits of the debt employment and the behavior of COMPUSTAT sample firms. The paper reveals that a firm could continue to be issuing the debt till the point where marginal tax benefits start declining. His research found that large, liquid and profitable firms with low expected distress costs were conservative in using the debt component. The study brought out the existence of debt conservative policy. This paper and its findings are of good relevance to the present study in terms of providing enough support to the validity of research. Laurence and Varouj et al (2001) assessed if the capital structure theory is still applicable across ten developing countries with variations in institutional structures. They evidenced that capital structure decision of developing countries depended on similar factors as in case of developed countries. They however, concluded that the specific country factors would also influence and play a key role in capital structure decision of the country concerned. Though the findings of their study suggested the portability of modern finance theory, they felt that there is much to be done to understand the capital structure decision. 2 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Jianjun Miao (2005) provided an equilibrium model to test the relationship between optimal capital structure and industry dynamics. The study revealed that while the firms make financing and investment decisions subject to idiosyncratic technology shocks, it proved that the capital structure decisions reflect the tradeoff between tax benefits of the debt component and associated bankruptcy and agency costs. The study also found out that the industries with growth opportunities were low in leverage. Ilya A. Strebulaev (2007) tried to find out if the tests of capital structures theory meant what they say. Using a calibrated dynamic trade-off model to stimulate firms’ capital structure paths, he found that the results of standard cross-sectional tests were consistent with those reported in empirical literature. Soku Byoun (2008) raised and tried to answer a research question “How and when do firms adjust their capital structures toward target?” He developed and suggested a financing needs-induced adjustment framework to examine the dynamic process by which firms adjust their capital structures. The results of his research indicated that the firms move towards target capital structure when they face a financial deficit/surplus, but not in the manner hypothesized by the traditional pecking order theory. This paper clearly evidences that the pecking order theory of financing need not be held good all times. Xin Chang and Sudipto Dasgupta (2009) in their paper attempted to find out if the firms had a debt ratio target which is a primary determinant of financing behavior. Up on addressing the issue, the authors found the evidence in support of the target debt ratio based leverage decisions of the firms. However, on an overall basis they evidenced that the existing tests of target behavior did not have any power to reject alternatives. Murray Carlson and Ali Lazrak (2010) developed a model to predict if the leverage of the U S based firms trades off the tax benefit of the debt against the utility cost of expost asset substitution. The authors found a positive relationship between cash to stock and leverage ratios of the firms. Erol (2011) develops what he calls the triangular relationship between firm size, capital structure choice and performance. The paper looks at trade-off theory, pecking order theory and irrelevance theory on a relative basis. The sample relates to the period 1994 to 2003 of 114 firms listed in the Istanbul Stock Exchange. Depending on how the size expansion gets financed, the impact on performance will change. Asset expansions with debt are seen to increase risk, which is what the trade-off theory says in any case. Berumeu, Beyer and Dye (2011) look at a model for determining capital structure, along with the extent of voluntary disclosure and cost of capital. They say “We establish a hierarchy of optimal securities and disclosure policies that varies with the volatility of the firm's cash flows. Debt securities are often optimal, with the form of debt—risk-free, investment grade, or "junk"—varying with the firm's cash flow volatility. Though the model predicts a negative association between firms' cost of capital and the extent of information firms disclose, more expansive voluntary disclosure does not cause firms' 3 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 cost of capital to decline. Mandatory disclosures alter firms' voluntary disclosures, their capital structure choices, and their cost of capital”. Brav (2009) analyses the data from public and private firms in the U.K, and finds that private firms have a heavy dependence on debt financing and are more sensitive to performance changes when it comes to deciding on the capital structure. The desire to maintain control and possible information asymmetry have resulted in private equity being costlier than public equity. Jagadish (2011) examines the impact of financial risk in capital structuring decisions. A sample of 59 companies listed in Indian stock exchanges has been examined for the period 1997 to 2007. Different statistical tools of analysis have been used based on the total asset value of the company. Financial risk in general and the risk of fluctuations in ROE in particular are seen to have an impact on the extent of debt in the capital structure. However, the study cannot be termed as comprehensive enough in the sense that other factors would also influence and the testing may not have fully isolated the financial risk angle completely. A strategic side to capital structure decisions is brought out by Mats (2010). In his paper, the author looks at ways of improving the bargaining position with organized labor using leverage. The author’s argument is that since high cash flow liquidity would result in higher demands from labor, firms would be better off having a higher level of debt and consequently higher demands on debt service. This collective bargaining power is a great influencer for the capital structure decision, according to the findings of the author. However, this would largely depend on the country of study and also the legislation prevalent. Chang and Yu (2010) look at how capital structure choice would have an influence on secondary market stock prices. The authors say “We show that the capital structure decision affects traders' incentives to acquire information and subsequently, the distribution of informed traders across debt and equity claims. When information is less imperative for improving its operating decisions, a firm issues zero or negative debt (i.e., holding excess cash reserves) in order to reduce socially wasteful information acquisition and the liquidity premium associated with it. When information is crucial for a firm's operating decisions, the optimal debt level is one that achieves maximum information revelation at the lowest possible liquidity cost. They say that their model can explain why many firms consistently hold no debt. It also provides new implications for financial system design and for the relationship among leverage, liquidity premium, profitability, and the cost of information acquisition Gill and Nahum (2011) argue that profitability is a crucial factor in capital structure decisions. The sample is from American manufacturing and service companies. A correlation and regression analysis is used to find the relationship of profitability with capital structure parameters. The authors state that the findings of the study show a positive relationship between i) short-term debt to total assets and profitability, ii) longterm debt to total assets and profitability, and iii) total debt to total assets and 4 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 profitability in the manufacturing industry. This paper offers useful insights for the owners/operators, managers, and lending institutions based on empirical evidence” Mukherjee and Mahakud (2012) raise the question of whether the pecking order theory and the tradeoff theories have to necessarily give diverse conclusions. The sample is from Indian manufacturing companies for the period from 1993-94 to 2007-08. The findings suggest that they are actually complimentary to each other. The authors also state that the firms have not been showing a marked tendency to go towards the target and the adjustment speed is around 40%. The testing method of finding target structure adjustment is open to debate. Specific aspects of manager contracts and their influence on capital structure are examined by Bhagat, Bolton and Subramanian (2011). They find that long term debt has a tendency to decline with long term risk; and short term risk shows a declining trend with short term risk. The analysis is comprehensive enough, but the testing methodology is open to question. Fan, Titman and Twite (2012) give an international comparison of capital structure, with emphasis on the time of maturity of debt. The sample is taken from 39 developed and developing countries. The authors find that the legal and tax system in individual companies dictate the preferences of capital structure in many countries. So much so, they find that companies in countries with greater levels of corruption use more of debt and particularly short-term debt. In countries having greater tax deductions from leverage, more debt is seen to be used. Specifically in India, several factors have been influential in capital structure decisions. These factors include volatility in interest rates, changes in legislation on taxation of dividends and differences in planned retention ratios. Following the liberalization policy of the Government in early 1990s companies have become more capital intensive. More emphasis is being laid by companies on shareholder value creation and market value added. The stock markets have also demonstrated great sensitivity to changes in debt patterns and borrowing. Empirical studies have shown that growth companies and companies with high market capitalization have not shown any specific theoretical framework in their capital structure decisions. Academic studies on Capital Structure the world over have looked into many diverse angles. Studies have reached conflicting conclusions on what drives capital structure policy. Most papers have sought to test basic postulates, the Pecking Order theory, the trade-off theory and the theory of Asymmetric Information. Most of these studies pertain to companies in the United States and Europe. The conclusions have been sharply different and have depended on the type of sample and the country from where these are taken. What we have endeavored to do in this paper is to show the broad relationship between actual capital structure and other key parameters like profitability and asset growth etc. Although the model is straightforward, we feel that this will go a long way in showing a basic trend in this policy. 5 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 II. Objectives and methodology: Keeping in view the significance of leverage decision, as it has been highlighted in literature review, the study aims at: 1. Investigating into as to how Indian firms went about in raising the funds to meet with the growing financial needs of the business over the last five years (2007 through 2011). 2. Identifying the factors that influenced Indian firms’ financing decisions and 3. Testing statistically as to which specific factors exercised influence, and to what extent, on Indian firms’ capital structure decisions. While there are many companies constituting corporate India, the researchers have taken the companies constituting S&P CNX NIFTY and CNX NIFTY JUNIOR numbering a total of hundred companies. In order to make the study and its findings more meaningful, banking and financing companies have been eliminated from the sample size as capital structure does not mean much in these firms. Also the companies for which there was no sufficiency of the data were eliminated. Thus after such elimination process, data of seventy three companies have used for analysis purpose. The rationale behind in choosing these companies is that as it has been historically observed, the market capitalization of these indices (popularly known as NIFTY and NIFTY JUNIOR) happens to be high in Indian stock market. Also the companies constituting these two indices have been found to be reasonably large in size. And it is a well known fact that the size of the firm matters a lot while deciding on the capital structure pattern. While it is argued that only seventy three companies may not be sufficient to arrive at the conclusions, the researchers feel that it is a good representative figure to provide a meaningful lead to understand as to how Indian companies went about capital structure decisions. The study depends fully on primary data of the companies for the years 2007 through 2011. While most of the companies reported the results on fiscal basis, the researchers found some companies reporting on calendar basis in some years. In order to ensure the inclusion of the data for all the companies and all the years chosen, the researchers have considered both for the purpose of analysis. It may however be noted that, if a company reported the results both in March and December in a year, the latest data, i.e that of December has been taken into account. This procedure has been followed for all such companies having two reporting figures in a year to eliminate any bias in data consideration. Though we find many corporate databases, we depended on Capitaline database (www.capitaline.com), a leading corporate database in India and therefore we sincerely acknowledge the support of this database in terms of providing required data for our study. III. Hypotheses formulation: Designing of capital structure is a function of many factors like the profitability, liquidity, size, growth and tax proportion of the firm. However there may be differences in the ways that these factors would impact the decision. In order to provide meaningful 6 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 guidance and to answer the research question, the following hypotheses have been formulated and tested. H1: Capital structure design of the Indian firms does not depend on profitability, solvency, size, tax proportion and growth factors. H2: Increasing profitability of the Indian companies does not influence the capital structure pattern. H3: The Indian firms that are solvent in the short-term do not depend on debt financing. H4: The capital structure pattern of Indian firms does not depend on size H5: The capital structure position of Indian firms does not depend on tax proportion. H6: The Indian firms growing significantly in short-term do not depend on debt financing. H7: Time factor does not have any impact on the capital structure decision of Indian firms. IV. Model development: For the purpose of testing hypotheses and as a part of analysis, multiple regression model has been used to test the results statistically. The following is the description of the model fit used in the study. Model 1: Relationship between leverage as criterion variable and profitability, solvency, size, tax proportion and growth as predictor variables. Leverage = α + β (ROCE) + β (RONW) + β (NPM) + β (CR) + β (Size) + β (TP) + β (Growth) + µ (1) Model 2: Relationship between leverage and profitability. Leverage = α + β (ROCE) + β (RONW) + β (NPM) + µ (2) Model 3: Relationship between leverage and solvency. Leverage = α + β (CR) + µ (3) Model 4: Relationship between leverage and size. Leverage = α + β (Size) + µ (4) Model 5: Relationship between leverage and tax proportion. Leverage = α + β (TP) + µ (5) Model 6: Relationship between leverage and growth. Leverage = α + β (Growth) + µ (6) In the above models Degree of Leverage (here after Leverage or capital structure) is the dependent variable or also known as criterion variable measured as debt up on equity. The following are the Independent variables or also known as predictor variables. ROCE = RONW = NPM CR = = Return On Capital Employed measured as Profit Before Interest and Tax (PBIT) up on Capital Employed (CE). Return On Net Worth measured as Profit After Tax (PAT) up on Net Worth (NW). Net Profit Margin measured as Profit After Tax (PAT) upon Sales. Current Ratio measured as Current Assets (CAs) up on Current Liabilities (CLs). 7 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Size TP = = Growth = Size measured as Logarithm (base 10) of Total Assets. Tax Proportion measured as Tax Expenditure up on Profit Before Interest and Tax (PBIT). Growth Rate in assets measured as Total Assets in the year ‘t’ up on the Total Assets in the year ‘t-1‘ Apart from the above connotations, ‘α’ is the intercept and ‘β’s are the coefficients of the predictor variable concerned which indicate the variance in the criterion variable (i.e leverage) caused by predictor variables and ‘µ’ is the error term of the model concerned. These abbreviations and connotations have been used in the study frequently to denote the variables and their influence. The following table indicates the proxies for various parameters in the above models. Parameter Profitability Solvency Size Tax Factor Growth Proxies ROCE i.e Return On Capital Employed RONW i.e Return On Net Worth NPM i.e Net Profit Margin CR i.e Current Ratio Log 10 of Assets TP i.e tax as the percentage of PBIT Year on Year Growth Rate in total assets The above models have been run in Statistical Package for the Social Sciences (SPSS) for windows version 15.0. The regression has been run using ENTER and STEP methods. While ENTER method considers all the predictor variables in the equation, STEP method considers predictor variables best correlated with the criterion variable in the equation as the first stage. Followed by this, the remaining predictor variables having the highest partial correlation with the criterion variable, controlling for predictor variables are entered and those variables having lowest partial correlation are eliminated. This process in STEP method is repeated at each stage “partialising” for previously entered predictor variables until the addition of a remaining predictor variable does not increase R2 by a significant amount. The linear regression model has been run separately for each year using criterion and predictor variables of all seventy three sample companies. While there may be some arguments that either ENTER method or STEP method would suffice, the idea behind running the models through both the methods is to extract the results and interpret the same in all possible ways. This would help to analyze if the pattern of capital structure depended on any one or more predictor variables significantly throughout the study period. Followed by this ANOVA also has been run across all the years to test H7 and see if the leverage of sample companies varied over a period of time. The strength of the linear relationship between criterion and predictor variables has been analyzed using Co-efficient of Multiple Determination, widely known as R2 the value of which must technically lie between ‘0’ and ‘1’. The value of R2 indicates as to what extent variance 8 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 in criterion variable is explained by predictor variables. Therefore, higher the value of R2, higher is the total variation in criterion variable explained by predictor variables. This in turn indicates that the higher level of significance is greater is the strength of linear regression model. However in order to test the statistical significance of the model and the hypothesis, Sig. F (also known as ‘p’ value) has been used. This statistic would help us know if the equation on an overall basis is significantly fit enough. The ‘p’ value generated in SPSS analysis helps us to accept or reject the null hypotheses. In the present study null hypotheses have been tested at 5% significance level and therefore, if the ‘p’ value is less than 0.05, then the null hypothesis is rejected and concluded that the model holds significant. In other words, this in the present study means that capital structure pattern would depend on profitability, solvency, growth, size and tax factors wherever relevant. As it has been mentioned earlier, the hypotheses are tested across all the companies for every year. Apart from testing the significance of overall model, relative influence of each predictor variable on criterion variable has been tested through ‘t-test’ the values of which are generated in SPSS analysis. This would help us to know as to which predictor variables contribute and do not contribute significantly to the regression model. Testing this at 5% significance level, if the ‘p’ value for the ‘t-test’ is less than 0.05, then the null hypothesis is rejected and concluded that the predictor variable concerned has significant influence on the overall model. In order to study the special contribution of each predictor variable to the equation, beta coefficients (β) of the un-standardized variables are taken into account. Apart from these statistics, Standard Error term (or simply S.E) and Durbin Watson’s coefficient (or simply D.W) have been used to analyze the results. Standard error indicates the variance in the mean values and therefore higher the standard error, more erroneous and less reliable is the model fit and vice versa. Durbin Watson’s coefficient indicates the extent of auto correlation among the error terms of the models of various samples. As a general rule, Durbin Watson’s coefficient closer to 2 indicates that there is no correlation among the error terms which is good sign to justify the model. Although adjusted R2 is captured in the tables, it is not interpreted as the sample size of the firms is not very high. V. Results and discussions: The analysis and interpretation has been carried out in the order of testing the impact of all predictor variables, followed by profitability predictors, solvency predictor, size predictor, tax proportion predictor and growth predictor on leverage position of the sample firms. It may also be noted that STEP method of multiple regression analysis has been carried out in case of overall model (Model 1) and profitability model (Model 2) as there are more than one predictor variable in those cases. As there is only one predictor variable in case of solvency, size, tax proportion and growth analyses only ENTER method has been used. Since it is space consuming activity, the researchers preferred to capture SPSS output and results in form of tables which are mostly self explanatory in nature. However, a few comments and observations are made out of the results for the purpose of better understanding and implications of regression results. 9 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 1: Summary statistics of the model 1 pertaining to overall analysis between leverage as the criterion variable and profitability, solvency, size, tax and growth as predictor variables (ENTER method) Year α R R2 Adj. R2 Std. Error Sig. F D.W 2007 0.048 0.872 0.76 0.734 0.522 0 2.04 2008 0.915 0.683 0.466 0.409 0.417 0 1.91 2009 0.76 0.665 0.442 0.382 0.407 0 0.19 2010 1.075 0.715 0.511 0.458 0.401 0 2.17 2011 0.916 0.649 0.421 0.359 0.413 0 2 Un-standardized beta Coefficients and ‘t’ statistic Predictor ROCE RONW CR NPM Size TP Growth ‘β’ and ‘β’ and ‘β’ and ‘β’ and ‘β’ and ‘β’ and ‘β’ and Year (‘t’ value) (‘t’ value) (‘t’ value) (‘t’ value) (‘t’ value) (‘t’ value) (‘t’ value) -9.15* 0.10* -0.12 0 -0.06 (1.24 2007 0.48 (1.87) (-11.92) (11.87) (-1.35) (-0.19) 0.53) (1.80) -2.87* 0.027* -0.16* 0.00 0.04 -1.25* 2008 0.053 (0.35) (-3.16) (2.56) (2.92) (-1.18) (0.36) (-2.03) -2.27 0.02 -0.09* 0.00 0.09 -1.19* -0.05 2009 (-1.96) (1.25) (-2.99) (-1.32) (0.87) (-2.05) (-0.20) -4.77* 0.04 -0.14* 0.00 0.05 -0.47 -0.21 2010 (-3.48) (2.89) (-2.81) (-1.48) (0.45) (-0.64) (-1.06) -0.87 0.00 -0.11* 0.00 0.15 -1.9* -0.29 2011 (-1.72) (0.78) (-2.45) (-1.56) (1.33) (3.09) (-1.17) * Significant at 5% level 10 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 2: Summary statistics of the model 1 pertaining to overall analysis between leverage as the criterion variable and profitability, solvency, size, tax and growth as predictor variables (STEP method) Adj. Std. R2 Sig. Predi ‘β’ and Year Model R R2 D.W 2 R Error Change F ctors (‘t’ value) 1 0.397 0.158 0.146 0.936 0.158 0.001 2 0.851 0.723 0.715 0.54 0.566 0 2007 2008 3 0.862 0.743 0.732 0.524 0.02 0.025 1 0.417 0.174 0.162 0.496 0.174 0 2 0.533 0.284 0.263 0.465 0.11 0.002 3 0.62 0.385 0.358 0.434 0.101 0.001 4 0.671 0.45 0.417 0.413 0.065 0.006 1 0.472 0.223 0.212 0.458 0.223 0 2 0.561 0.315 0.296 0.433 0.092 0.003 2009 3 0.629 0.396 0.37 0.409 0.081 0.003 1 0.449 0.202 0.191 0.489 0.202 0 2 0.621 0.386 0.369 0.432 0.184 0 2010 3 0.698 0.487 0.465 0.398 0.101 0 1 0.466 0.217 0.207 0.458 0.217 0 2 0.552 0.305 0.285 0.434 0.087 0.004 2011 3 0.61 0.372 0.345 0.416 0.068 0.008 ROCE ROCE RONW ROCE RONW TP TP TP CR TP CR ROCE TP CR ROCE RONW ROCE ROCE RONW ROCE RONW CR ROCE ROCE RONW ROCE RONW CR ROCE ROCE CR ROCE CR TP 1.35 (6.03) -8.72 (-13.27) 0.10 (11.96) -9.47 (-13.21) 0.10 (12.49) 1.49 (2.29) -2.18 (-3.86) -2.58 (-4.75) -0.17 (-3.28) -1.96 (-3.63) -0.19 (-3.83) -0.79 (-3.36) -1.11 (-1.86) -0.16 (-3.23) -3.11 (-3.66) 0.03 (2.83) -1.17 (-4.54) -3.99 (-4.36) 0.036 (3.09) -3.91 (-4.36) 0.03 (3.06) -0.08 (-3.05) -1.17 (-4.27) -5.48 (-5.67) 0.05 (4.61) -5 (-5.56) 0.04 (4.19) -0.17 (-3.71) -1.12 (-4.47) -1.24 (-4.87) -0.14 (-2.98) -0.91 (-3.35) -0.13 (-2.83) -1.61 (-2.74) 1.97 1.91 1.98 2.22 1.90 11 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table one above contains the details pertaining to regression model, through ENTER method that reveal if the leverage decision of the firms depended significantly on all predictor variable together. Significance ‘F’ in all the years was below 0.05 and hence we understand that leverage decision of Indian firms depended significantly on profitability, solvency, size, tax proportion and growth. At the outset, we therefore reject null hypothesis, H1 and prove that capital structure design of Indian firms depended on the profitability, solvency, size, tax proportion and growth factors together in every year. However, while ‘R’ is found to be reasonably high in all the years, ‘R 2’ is found to be low in most of the years. Second part of the table 1 also contains un-standardized beta coefficients and ‘t’ values of predictor variables. The values indicate that the capital structure design of the firms did not depend significantly on size and growth of the firms while it significantly depended on profitability, solvency and tax proportion in most of the years. From the same table we also observe that of all the predictor variables, ROCE had very high influence on capital structure as revealed by its beta co-efficient (B) and ‘t’ statistic while the other predictors’ influence was relatively lower in almost all the years. Table 2 contains the details pertaining to step-wise multiple regression of model 1. As we see, size and growth variables never appeared as the determinants of the capital structure throughout the study period which was a similar observation even in ENTER method. This is therefore an important finding of the study that Indian firms did not consider the size and growth opportunities while designing their capital structure position. This finding is not in line with the theoretical proposition that size and growth factors play a vital role in capital structure decision of the firms. While tax proportion was the most influencing factor in the year 2008, two proxies of profitability (i.e. ROCE and RONW) were the most dominant determinants of capital structure in the rest of the individual years. On the other hand solvency (CR) and another proxy of profitability (i.e. NPM) had very least influence on capital structure decision of the firms. In fact NPM never appeared as the determinant variable throughout out the study period. The Durbin Watson statistic in both tables 1 and 2 in every year indicates that auto correlation among the error terms was within the limits. This in turn indicates that the models are justified to explain as to what factors have influenced the leverage of Indian firms and to what extent also. 12 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 3: Summary statistics of the regression model 2 between leverage and profitability (ENTER Method). Adj. Std. Sig. Predictor ‘β’ and D.W Year Α R R2 (‘t’ value) R2 Error F ROCE -8.74* (-13.31) RONW 0.10* (12) 2007 0.43 0.85 0.72 0.71 0.53 0 2.05 NPM -0.001 (-1.1) ROCE -4.15* (-5.62) RONW 0.04* (4.5) 2008 0.62 0.61 0.37 0.34 0.43 0 1.74 NPM -0.001 (-1.16) ROCE -3.96* (-4.13) RONW 0.03* (3.01) 2009 0.70 0.57 0.32 0.29 0.43 0 1.83 NPM 0 (-1.11) ROCE -5.58* (-5.78) RONW 0.05* (4.71) 2010 0.66 0.63 0.40 0.38 0.42 0 2.04 NPM -0.001 (-1.63) ROCE -1.47* (-3.21) RONW 0.003 (0.75) 2011 0.73 0.48 0.23 0.20 0.46 0 1.78 NPM 0 (-0.90) * Significant at 5% level Table 4: Summary statistics of the regression model 2 between leverage and profitability (STEP Method). 2 Year Model R R Adj. Std. R2 Sig. Predictors ‘β’ and (‘t’ value) R2 Error Change F -2.61 (-3.64) 1 0.39 0.15 0.14 0.93 0.15 0 ROCE ROCE -8.72 (-13.27) 2007 2 0.85 0.72 0.71 0.54 0.56 0 RONW 0.10 (11.96) -0.97 (-3.88) 1 0.41 0.17 0.16 0.49 0.17 0 ROCE ROCE -4.14 (-5.59) 2008 2 0.6 0.36 0.34 0.44 0.18 0 RONW 0.04 (4.49) -1.17 (-4.55) 1 0.47 0.22 0.21 0.45 0.22 0 ROCE ROCE -3.93 (-4.10) 2009 2 0.56 0.31 0.29 0.43 0.08 0 RONW 0.03 (2.97) -1.18 (-4.30) 1 0.45 0.20 0.19 0.48 0.20 0 ROCE ROCE -5.43 (-5.58) 2010 2 0.62 0.38 0.36 0.43 0.17 0 RONW 0.05 (4.51) -1.21 (-4.48) 2011 1 0.47 0.22 0.21 0.45 0.22 0 ROCE D.W 2.04 1.76 1.86 2.04 1.85 Tables 3 and 4 contain model summary statistics and beta co-efficients along with ‘t’ values of the regression model between leverage and profitability. While table 3 contains the details of model through ENTER method, table 4 contains that of STEP 13 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 method. A few important observations about details through ENTER method (in table 3) are as follows. The strength of the model, as revealed by ‘R2’ was reasonably high in the years 2007, 2008 and 2010 while in the rest of the years it was very weak. This indicates, that the relationship explained between the leverage (criterion variable) and the profitability (predictor variables being ROCE, RONW and NPM) was very low in these years. However, the ‘Sig. ‘F’ in all the years was less than 0.05 and hence we reject null hypothesis, H2 and prove that in case of Indian firms, the increasing profitability influenced leverage decision significantly as a whole during the period under review. But beta coefficients and their ’t’ values indicate that ROCE and RONW were the two predictor variables of profitability that influenced leverage significantly throughout the study period except RONW being insignificant only in the year 2011. This is in fact in line with the observation made in the overall analysis as we can see in tables 1 and 2. It is therefore understood that Indian firms did take NPM into account as a factor to be concerned about, while designing their capital structure position. The ‘t’ values of beta co-efficients indicate that the impact of ROCE on leverage was the highest when compared to that of RONW. A few important observations from STEP method as captured in table 4 are as follows. Like in the case of ENTER method, in STEP method also, the strength of the model, as aptly revealed by R2, was found to be very weak during the entire period of study barring the second model in the year 2007. The highest portion of the relationship explained by the independent variables was as low as 0.15. The standard error was also very high which indicates that the model was highly erroneous throughout the study period. When it comes to the question of predictor variables, like in the case of ENTER method, only two proxies of profitability (i.e ROCE and RONW) had significant impact on leverage throughout the study period. Based up on these observations, we understand that the leverage decision of Indian firms depended on ROCE and RONW in the years 2007 through 2011. However, the beta co-efficients and their ‘t’ values indicate that the impact of profitability variables was relatively low. The Durbin Watson’s co-efficient in both the methods (i.e ENTER and STEP methods in tables 3 and 4) was close to two and hence we conclude that there was no auto co-relation among the error terms of the models during the study period and hence the models being justified. This in turn means that the models were found to be fit enough to explain and prove that the leverage depended on profitability variables. Thus from all angles point of view, as revealed by the statistical analysis we understand that the capital structure design of Indian firms very much depended on ROCE and RONW significantly. 14 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 5: Summary details of regression model 3 between leverage and solvency Std. ‘β’ and Year α R R2 Adj. R2 Sig. F Predictor (‘t’ value) Error -0.05 2007 0.71 0.04 0.00 -0.02 1.01 0.73 CR (-0.33) -0.12 2008 0.68 0.23 0.05 0.04 0.53 0.05 CR (-1.99) -0.07* 2009 0.59 0.24 0.06 0.04 0.50 0.03 CR (-2.14) -0.16* 2010 0.75 0.30 0.09 0.08 0.52 0.00 CR (-2.71) -0.12* 2011 0.64 0.26 0.07 0.05 0.50 0.02 CR (-2.30) *Significant at 5% level. Table 5 contains the details pertaining to the relationship between leverage, the criterion variable and the solvency (i.e. CR) the predictor variable. Values of R and R2 reveal that the strength of the model in every year has been very weak. However as Sig. ‘F’ in the years 2009, 2010, 2011 was less than 0.05, we reject null hypothesis, H3 and prove that Indian firms that are solvent in the short-run depended on debt financing in these years. However, Sig. ‘F’ indicate that the leverage did not depend on solvency in the years 2007 and 2008. Though the solvency influenced the leverage significantly in a few years, the coefficient and their beta values of CR (proxy for solvency) indicate that the impact was very low. This in turn reveals that the solvency indeed did not have any major impact on capital structure decision of Indian firms during the study period. Table 6: Summary details of regression model 4 between leverage and size Adj. Std. Sig. ‘β’ and Year α R R2 Predictor (‘t’ value) R2 Error F 0.23 2007 -0.21 0.12 0.01 0.00 1.01 0.29 Size (1.04) 0.27* 2008 -0.52 0.27 0.07 0.06 0.52 0.02 Size (2.37) 0.34* 2009 -0.84 0.35 0.12 0.11 0.48 0.00 Size (3.24) 0.37* 2010 -0.97 0.36 0.13 0.12 0.51 0.00 Size (3.28) 0.36* 2011 -1 0.36 0.13 0.12 0.48 0.00 Size (3.35) *Significant at 5% level. Table 6 contains the details pertaining to the relationship between leverage and size of sample firms. As in the case of relationship between leverage and solvency, the strength of the model between leverage and size of the firms has been found to be very 15 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 weak as revealed by R2. Since Sig. ‘F’ is less than 0.05 in all the years except in the year 2007, we reject the null hypothesis, H4 and prove that leverage depended on size significantly during the study period except in the year 2007. The beta co-efficients and their ‘t’ values of CR also indicate that the impact of size on leverage was significant in all the years except in the year 2007. It is also worth noting that the impact of size on the leverage was very high as revealed by beta co-efficients. Thus we conclude that, though the strength of the model is lesser, the capital structure design of Indian firms depended significantly on size. Table 7: Summary details of regression model 5 between leverage and tax proportion Sig. ‘β’ and Year α R R2 Adj. R2 Std. Predictor F (‘t’ value) Error 2007 1.03 0.22 0.04 0.03 0.99 0.06 TP 2008 0.89 0.42 0.17 0.16 0.49 0.00 TP 2009 0.79 0.39 0.15 0.14 0.48 0.00 TP 2010 0.90 0.37 0.14 0.13 0.50 0.00 TP 2011 0.93 0.45 0.21 0.19 0.46 0.00 TP -2.06 (-1.91) -2.18* (-3.89) -1.84* (-3.56) -2.24* (-3.44) -2.55* (-4.35) * Significant at 5% level Table 7 contains the details pertaining to the relationship between leverage and tax proportion (TP) of Indian companies. The values of R2 reveal that the model in all the years has been very weak as the explained portion of the relation was very low. However as the Sig. ‘F’ is less than 0.05 in all the years except in the year 2007 we reject the null hypothesis, H5 and prove that leverage decision of Indian firms depended on tax proportion in all the years except in the year 2007. This is also confirmed by the beta co-efficients of tax proportion being significant throughout the study period except in the year 2007. This finding is exactly in line with the findings pertaining to the impact of solvency and size. We also notice that the co-efficients of TP throughout the study period have been found to be reasonably higher than those of solvency and size factors indicating greater influence of tax factor on leverage when compared to that of solvency and size. 16 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 8: Summary details of regression model 6 between leverage and growth Adj. Std. Sig. ‘β’ and Year α R R2 Predictor (‘t’ value) R2 Error F 1.02* 2007 -70 0.27 0.07 0.06 0.97 0.01 Growth (2.45) 0.13 2008 0.31 0.09 0.00 -0.00 0.54 0.44 Growth (0.77) 0.34 2009 0.03 0.13 0.01 0.00 0.51 0.26 Growth (1.11) -0.18 2010 0.70 0.08 0.00 -0.00 0.54 0.45 Growth (-0.75) -0.27 2011 0.76 0.11 0.01 -0.00 0.51 0.34 Growth (-0.94) Details pertaining to the relationship between the leverage and the growth have been summarized in table 8. The strength of the model between the leverage and growth is very weak throughout the study period as revealed by R2. The change in the leverage could not be explained properly by the growth factor. The standard error term is also high in all the years indicating that the model was unreliable in terms of studying the extent of change in leverage due to change in growth. However, Sig. ‘F’ indicates that the impact of growth on leverage was significant in the year 2007 only. Based on these observations, we reject null hypothesis, H6 and prove that the leverage depended on growth in the years 2007 only but not in the rest of the years. While this being worth noting observation, the coefficients and their ‘t’ values indicate that the percentage of impact exercised by the growth on leverage has been very low. Thus, we conclude that the capital structure design of Indian firms did not depend much on growth factor. Table 9: ANOVA results pertaining to leverage position of Indian firms during the years 2007 through 2011. Sum of Mean Squares df Square F Sig. Between 1.844 4 .461 1.070 .371 Groups Within 155.125 360 .431 Groups Total 156.969 364 Table 9 captures the details pertaining to ANOVA of leverage position of Indian firms during the years 2007 through 2011. Since Sig. ‘F’ is greater than 0.05, we accept the null hypothesis, H7 and conclude that the leverage position of Indian firms did not vary significantly over the years 2007 through 2011. This in turn implies that the time period did not have any impact on the leverage position of Indian firms. 17 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Table 10: Descriptive statistics (mean and S.D) pertaining to criterion variable (DER) and predictor variables (ROCE, RONW, CR, NPM, SIZE, TP and Growth) Variable 2007 2008 2009 2010 DER Mean 0.64 0.48 0.47 0.48 S.D 1.01 0.54 0.52 0.55 Mean 27.20 27.90 24.90 25.00 ROCE S.D 15.30 23.40 21.00 21.00 Mean 25.86 24.84 21.80 21.97 RONW S.D 12.03 18.61 17.25 17.05 Mean 1.53 1.63 1.76 1.71 CR S.D 0.81 1.04 1.70 1.04 Mean 29.72 29.65 30.33 28.27 NPM S.D 116.17 104.46 120.06 98.46 Mean 3.63 3.75 3.85 3.92 SIZE S.D 0.53 0.55 0.55 0.54 Mean 19.00 19.00 18.00 19.00 TP S.D 11.00 10.00 11.00 09.00 Mean 31.00 33.40 27.00 20.00 Growth S.D 27.60 38.30 19.80 26.10 2011 0.44 0.52 22.70 20.00 18.00 25.54 1.72 1.10 29.20 117.19 3.99 0.53 20.00 09.00 18.30 20.60 Table 10 contains the mean and the standard deviation values of both criterion and predictor variables in every individual year. On an average basis, the leverage position of Indian firms did not vary much rather remained almost same in a range of 0.436 times being the lowest in the year 2011 and 0.639 times being the highest in the year 2007. The mean values of predictor variables like CR, NPM, Size and TP did not vary much but remained almost same in all the years. However, other predictor variables like ROCE, RONW and Growth decreased slightly in the years 2010 and 2011. The standard deviation of DER, ROCE, CR, NPM, Size and TP remained almost same in all the years and that of RONW and Growth factors varied slightly in the later years of the study period. Of the standard deviations of all variables, that of NPM was exceptionally high in all the years indicating higher volatility in reported NPM of Indian firms. VI. Conclusions: In this paper we demonstrated as to how Indian firms went about in designing their capital structure positions. While there are a host of studies available on the topic, we came across, through literature survey, that the research on this topic has been focused primarily on the U S and European firms providing a scope for research on Indian firms. Also, the topic of capital structuring has been at a forefront of corporate finance manager in the light of changing economic environment. Using multiple regression models, we tried to demonstrate if there were any specific factors that influenced the capital structure decision of Indian firms during the years 2007 through 2011. The analysis revealed that the capital structure decisions of Indian firms depended largely 18 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 on profitability in general and ROCE and RONW in specific in most of the years. A deeper insight into statistical results indicates that the impact of various factors on leverage position of Indian firms varied across the years though it was not much significant. To be specific with respect to testing of hypotheses, the findings of the study reveal that: H1 can’t be accepted and therefore it is concluded that the leverage position of Indian firms depended on profitability, solvency, tax factor and growth opportunities together. H2 can’t be accepted and therefore it is concluded that increasing profitability influenced the leverage position of Indian firms. H3 can’t be accepted and therefore is concluded that leverage position of Indian firms depended on their short-term solvency position. H4 can’t be accepted and therefore it is concluded that capital structure design of Indian firms depended on their size. H5 can’t be accepted and therefore it is concluded that tax factor had an impact on the leverage position of Indian firms. H6 can’t be rejected (except in the year 2007) and therefore it is concluded that there was no link between the way Indian firms designed their capital structure and their growth opportunities. H7 can’t be rejected and therefore it is understood that the time factor did not have any significant impact on the changing leverage position of the firms during the study period. To sum up, the study revealed that capital structure design in Indian context is still a matter of concern as it depends on a host of factors specific to the firm and specific to the market. Despite certain limitations that we faced, we strongly feel that this study bought out significant observations which provide meaningful implications to the manager, lender and any stakeholder concerned with respect to financing pattern of Indian firms. We also feel that the study adds up significantly to the body of knowledge on capital structure practices as it has provided empirical evidence as to how Indian firms design their capital structure position. While there would have been scope for further improvement of the paper, we strongly feel and recommend that there is a need to extend the research on this topic and therefore we welcome any researcher to furthering the research on this topic in any possible and value adding manner. References: Bertomeu and Anne etal (2011), Capital Structure, Cost of Capital, and Voluntary Disclosures, Accounting Review, Apr2011, Vol. 86 Issue 3, Pp.857-886. Bolton and Brian etal (2011), Manager Characteristics and Capital Structure: Theory and Evidence, Journal of Financial & Quantitative Analysis, Dec2011, Vol. 46 Issue 6, p1581-1627, 47p. Brav (2009), Access to Capital, Capital Structure, and the Funding of the Firm, Journal of Finance, Feb2009, Vol. 64 Issue 1, p263-308, 46p. 19 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Chang and Xiaoyun (2010), Informational Efficiency and Liquidity Premium as the Determinants of Capital Structure, Journal of Financial & Quantitative Analysis, Apr2010, Vol. 45 Issue 2, p401-440, 40p. Erol (2011), Triangle Relationship among Firm Size, Capital Structure Choice and Financial Performance, Journal of Management Research (09725814), Aug2011, Vol. 11 Issue 2, p87-98. Fan, Joseph P. H. etal (2012), Twite, Garry, An International Comparison of Capital Structure and Debt Maturity Choices, Journal of Financial & Quantitative Analysis, Mar2012, Vol. 47 Issue 1, p23-56, 34p. Franco Modigliani and Merton H Miller, The cost of Capital, Corporation Finance and Theory of Investment, American review, Vol. 48, June 1948. Gill and Biger etal, (2011), The Effect of Capital Structure on Profitability: Evidence from the United States, International Journal of Management, Dec2011, Vol. 28 Issue 4, p3-15, 13p. Ilya A. Srebulaev (2007), Do Tests of Capital Structure theory Mean What they say?, Journal of Finance, Vol. LXII, No. 4, August 2008, Pp. 1747-1787. Jagdish (2011), The Impact of Financial Risk on Capital Structure Decisions in Selected Indian Industries: A Descriptive Analysis, Advances in Management, Nov2011, Vol. 4 Issue 11, p24-30, 7p. Jianjun Miao (2005), Optimal capital Structure and Industry Dynamics, Journal of Finance, Vol. LX, No. 6, December 2005, Pp. 2621-2659. John R. Graham (2000), How Big Are the Tax Benefits of Debt, Journal of Finance, Vol. LV, No. 5, October 2000, Pp. 1901-1941. Laurence Booth and Varouj Aivazian et al (2001), Capital structures in Developing Countries, Journal of Finance, Vol. LVI, No. 1, February 2001, Pp. 87-130. Matsa (2010), Capital Structure as a Strategic Variable: Evidence from Collective Bargaining, Journal of Finance, Jun2010, Vol. 65 Issue 3, p1197-1232, 36p. Mukherjee and Mahakud et al, (2012), Are Trade-off and Pecking Order Theories of Capital Structure Mutually Exclusive? Journal of Management Research (09725814), Apr2012, Vol. 12 Issue 1, p41-55, 15p. Murray Carlson and Ali lazrak (2010), Leverage Choice and Credit Spreads when Managers Risk Shift, Journal of Finance, vol. LXV, No. 6, December 2010, Pp.2323-2362. 20 Proceedings of 8th Asian Business Research Conference 1 - 2 April 2013, Bangkok, Thailand, ISBN: 978-1-922069-20-7 Raghuram G. Rajan and Luigi Zingales (1995), What Do We know about Capital Structure? Some evidence from international Data, Journal of Finance, Vol. L, No. 5, December 1995, Pp. 1421-1460. Soku Byoun (2008), How and When Do Firms Adjust Their Capital Structures toward Target, Journal of Finance, Vol. LXIII, No.6, December 2008, Pp. 3069-3096. Xin Chang and Sudipto Dasgupta (2009), Target Behavior and Financing: How Conclusive Is the Evidence, Journal of Finance, Vol. LXIV, No.4, august 2009, Pp.1767-1796. 21