Deleveraging of Firms in India: Causes and Consequences. Abstract

Transcription

1 Deleveraging of Firms in India: Causes and Consequences Abstract The paper highlights systematic deleveraging of firms in India and explores the factors contributing to such deleveraging. The paper further highlights that the deleveraging is pervasive in manufacturing and non-manufacturing firms almost equally. The empirical findings of the paper suggest that most consistent theoretical determinants of corporate leverage fails to explain the decline in debt ratios of the firms. However, institutional deficiencies in the form of underdeveloped bond markets and decline in corporate investments were found quite significant in explaining the decline. The results suggest that firms could be credit rationed and hence losing value on account of such deficiencies. Keywords: Deleveraging; corporate leverage; external financing; credit market development. JEL Classification: G32, G20, H25, H62, E62. 1

2 Deleveraging of Firms in India: Causes and Consequences 1. Introduction Despite of seminal concerns against value creation through debt financing, it has been an integral part of a typical firm s capital structure across time and space. Managers do spend a considerable amount of time in designing the capital structure of their firms. Further, empirical evidences suggest that debt financing has contributed significantly to the development and growth of an economy. 1 This is perhaps the reason why trend in resource allocation for the majority of developed and developing countries shows up a consistent increase in debt ratios over time. However, India remains an exception where debt ratios of the firms are consistently declining since economic liberalization of economy in early 1990s. Figure 1 shows the trend in debt ratios, estimated as the ratio of total debt to capitalization (total debt and equity issued along with retained earnings) and total debt to assets, for non-financial firms in India. In contrast, debt ratios in many other emerging markets are showing up an increasing trend (Mitton, 2007). Interesting would be to delve deeper into the findings and know what factors contributes significantly for such an anomaly. Debt ratios 2 could fall as a result of increasing participation of non-manufacturing sector which has lower debt capacity and lesser requirement of capital to grow. Further, the debt ratios can fall when firms choose to offload debt, raise more equity and/or use increasing proportion of internal financing through retained earnings instead of external finance. Figure 2 reflects on all 1 See King and Levine (1993), for seminal empirical work establishing relationship between credit market development and growth. Recent work in this area includes Rousseau and Wachtel (2002) and Loayza and Ranciere (2006). 2 Henceforth, we have used the definition of debt ratio as total debt to capitalization throughout the paper. However, our results are robust to alternate definition of total debt to assets also. Total capital includes short term debt including current portion of long term debt, long term debt, equity capital raised and retained earnings. Total capital in our definition include all possible sources of equity and debt as per the financial statements of the firms. Since the objective of our paper is to analyse the deleveraging effect of firms and its determinants, we do not discriminate between formal and non-formal sources of capital as in Allen et al. (2006). 2

3 such possibilities by plotting the ratios of debt to total capital, equity issued to total capital, and retained earnings to total capital for manufacturing and non-manufacturing non-financial firms in India. 3 Figure 2 shows that debt ratios are consistently declining not only for manufacturing sector but also for non-manufacturing sector. Further, the decline is not attributed to raising of more equity capital as the ratio of equity to total capital is also declining. The decline is attributed to shedding of debt and use of more internal funds by firms to bridge their financing deficits. Interestingly, the pattern in Figure 2 would mean that either firms do not need external finance and choose to finance themselves through internal funds to a large extent or, they find themselves constrained to raise external funds. Although, it looks unlikely that firms in an emerging markets are saturated with capital requirements and therefore do not require external finance for growth, we check for this possibility too. An implicit benefit for firms of taking debt comes in the form of tax deductibility of interest payments. Such tax savings amounts to additional cash flows which in turn increases the value of the firm. However, excess debt and ensuing non-payment of scheduled interest can lead firms to bankruptcy or financial distress. The firms, therefore, strive to balance the benefits and cost of debt financing. Following this, we posit that if firms in India are voluntarily shedding their debt or are saturated with capital to finance their growth mostly with retained earnings, they would not incur a value loss on account of adverse taxations overtime. In other words, after avoiding any potential distress due to debt, firms would maintain their debt ratios so that they may not face an adverse tax drain. An effective measure to evaluate the tax burden of the firms is to estimate the effective taxes paid by the firms as a percentage of their operating income (earnings before depreciation, 3 Manufacturing firms are classified based on their National Industrial Codes (NIC) between to The remaining firms are classified as non-manufacturing firms. The data is gathered from Centre for Monitoring of Indian Economy (CMIE). 3

4 interest and taxes). We would call the ratio as effective tax henceforth. Figure 3 shows the plot of effective taxes measured as described above for all non-financial firms as well as their breakup for manufacturing as well as non-manufacturing firms overtime. As can be seen firms indeed face an increasing tax drain and an ensuing value loss because of it. The increasing effective tax burden on the firms could also be reinforced if we follow the pattern of corporate income tax receipt by the central government and the industrial GDP (contribution from industry and services excluding agriculture in the GDP). Between 1992 and 2012, while the industrial output grew at a rate of 14.85% (CAGR), the corporate tax receipts grew at a rate of 20.13%. This shows that higher taxes paid by the firms do not match with the increased participation of corporate sector in the economy. This can further be ascertained by looking into the relative differences in the growth rate of corporate taxes and industrial output during the market boom experienced between During this period, it can be expected that firms are in a relatively better position to readjust their debt ratios so as to minimize the tax drain. However, the CAGR of industrial output was only 15.88% as compared to the CAGR of 32.25% for corporate tax receipts. These figures along with falling debt ratios during this time is inconsistent with the views put forward in Nagaraj (2013), which argues for a debt led growth of the economy during the boom period of Thus, although the growth opportunities might have improved during the booms, the ability of the firms to raise adequate capital in response of the increased growth opportunities might not have increased. Although, most consistent determinants of corporate leverage across a range of empirical and theoretical studies are found to be firm-specific attributes 4, the increased tax burden of the firms points toward potential systematic constraints faced by the firms to raise capital in India. This 4 See Zingales (2000) for a survey of literature on capital structure theories. See Frank and Goyal (2009), Rajan and Zingales (1995), Harris and Raviv (1991) and Titman and Wessels (1988) for empirical studies on capital structure determinants. 4

5 would mean that firms may be sacrificing their growth on account of systematic underinvestment reflected through the deleveraging of firms in India. Such systematic constraints may show up in the form of underdeveloped capital markets, ineffective law and order, sub-optimal corporate governance structure and weak investors protection. These factors could plausibly affect the capital raising abilities of the firm as discussed in Demirguc- Kunt and Maksimovic (1996), Booth et al. (2001), Agarwal and Mohtadi (2004) and Allen et al. (2006). This paper takes up an extensive inquiry into these observations and investigate the following: (i) to what extent debt ratios and their decline are explained by firm-specific factors, (ii) to what extent debt ratios and their decline are explained by institutional and macroeconomic factors responsible for the development of capital markets, and (iii) what are the possible implications of declining debt ratios to the firms and to the macro-economy. The remainder of the paper is organized as follows. Section 2 describes the cross-sectional determinants of leverage discussed in the past literature. This section also highlight the potential role of institutional and macro-economic factors in influencing the cross-section of debt ratios. Section 3 undertakes the empirical analysis by discussing the data, methodology and results. Section 4 discusses the implications of the findings and section 5 concludes the paper Determinants of Corporate Leverage 2.1 Firm-specific attributes of corporate leverage 5

6 Various theoretical frameworks 5 attempt to describe attributes affecting debt ratios of a firm. These are mainly firm-specific attributes related to tax related benefits, agency costs and bankruptcy costs associated with debt financing. We list the key firm-specific attributes associated with such potential benefits and costs of debt for firms as discussed in the past literature such as Titman and Wessels (1988), Rajan and Zingales (1995) and Frank and Goyal (2009). Size of a firm is thought to be positively correlated with debt ratios in many empirical studies (Zingales, 2000). However, exceptions to these studies are found in Titman and Wessels (1988) where size correlates negatively. More importantly size correlates positively in studies focusing on emerging market firms. 6 Such a relation can be attributed to lowering of bankruptcy costs for larger firms. Indirectly, size somehow measures the growth options for firms as large mature firms tend be bigger in size. Further, size would also correlates positively with accessibility of firms to capital markets. Thus, it would indicate availability of credit to some extent, which might be a decisive factor for corporate leverage in emerging markets. Profitability of a firm, on one hand, indicates the basic earning potential of firm s asset, on the other, the change in profitability over time signifies the volatility of earnings for the firm and thus its business risk. Theoretical predictions with regard to possible correlation of profitability with debt ratios are mixed. Pecking order hypothesis of Myers (1984) would suggest a negative correlation due to information asymmetry between the firm and its investors. Highly profitable firms would resort to more of internal financing and less profitable firms would opt for debt rather than equity for external financing. However, static trade-off model would suggest a positive relation owing to reduced cost of bankruptcy for more profitable firms. Further, highly 5 Important among these frameworks are static trade-off model (Kraus and Litzenberger, 1973), pecking order hypothesis (Myers, 1984) and agency theoretic frameworks (such as Jensen, 1986). 6 See Mitton (2007) and Booth et al. (2001) for empirical work on emerging market firms. 6

7 profitable firms with cash surplus would like to reduce free cash flows at manager s discretion by resorting to higher debt (Jensen, 1986). Notwithstanding the ambiguity in theoretical directions for correlations a good number of studies indicate a negative relationship between profitability and debt ratios. 7 Moreover, empirical work on emerging market firms also approves of such negative correlations. The empirical capital structure literature would further suggest a positive correlation between asset tangibility and debt ratios. This is primarily owing to reduction in bankruptcy and agency costs between debtholders and shareholders associated with hard assets being good collateral for lending. Number of empirical studies, including those on emerging markets, show up this relationship quite consistently. Growth opportunities available with the firm can significantly affect the choice of debt and equity and hence the capital structure. Theoretical underpinning would largely suggest a negative relationship between growth opportunities and debt ratios. Negative correlation is evident considering (i) increase in probability of bankruptcy with uncertain growth in static trade-off model; (ii) reduction in conflict between managers and shareholders by pursuing growth objectives and, (iii) retention of flexibility for growth owing to restrictive covenants associated with debt. Further, the problem of underinvestment would deter firms with growth opportunities to take up large amount of debt (Jensen and Meckling, 1976). However, pecking order hypothesis suggest that such firms in dire need of funds should issue less information sensitive securities, such as debt, to overcome the information asymmetry and hence a positive association of growth opportunities with debt ratios is implied. Further, incremental debt raised 7 See, for example, Frank and Goyal (2009) and Fama and French (2002) for such a relationship. 7

8 by underlevered firms may not sacrifice financial flexibility to a greater extent and hence debt ratios may show a positive correlation with growth options for such firms. Measurement of growth option with a firm in the literature is taken up with different variables. Titman and Wessels (1988) defines growth in terms of growth in total assets and the ratio of capital expenditure to total assets. Many others would define growth in terms of capital expenditure, research and development expenditure and market to book ratio of its equity 8. While growth in assets and capital expenditure to total asset are backward looking and may not capture the prospective growth options, market to book ratio while forward looking could end up measuring the profitability rather than growth opportunities. Further, when we consider year end market values for firms in empirical studies, we are exposing our results to inter-temporal market volatility. Event studies for market returns at financial year closing would indicate that these measures could be biased. 9 This is true for any ratios calculated using market values of equity and further troublesome in volatile equity markets of emerging markets. Further, many of the firms, as in our dataset, do not have a reliable market value owing to thin trading of their stocks or due to the inclusion of private firms deprived of market values. Since interest expenses are tax deductible, presence of taxes would indicate immediate value creation through these tax shields. Higher tax rates, ceteris paribus, would prompt firms to take on higher leverage. Static trade-off model suggests an optimal debt capacity for firms where they trade-off the tax benefits with ensuing distress costs. The associated cost of debt can be minimized by substituting debt oriented tax shields by alternative non-debt tax shields. Non-debt tax shields are expected to be negatively correlated with debt ratios. However, empirical evidences on such correlations are not consistent and 8 See, for example, Myers (1977) which shows that market-to-book equity ratios corresponds to growth opportunities envisaged in Miller and Modigliani (1961) 9 See Keim (1983) for event studies identifying anomalies in stock prices at year end. 8

9 reported being ambiguous. Firms may not increase the proportion of debt in the presence of higher tax rates if they are protected alternatively by non-debt tax shields. However, if firms are credit restrained or are underlevered, we may expect an increase in debt ratios even in the presence of non-debt tax shields. Notwithstanding the tax rates applicable for the firms, the effective taxes paid by the firms as a fraction of their total operating income may lead firms to change their debt ratios. Increasing tax burden may induce firms to take on additional debt so as to cut down on any tax drain. However, institutional constraints may prevent firms to reduce the ensuing tax burden and firms may choose sub-optimal capital structure for themselves Institutional determinants of corporate leverage At the core of most influential theoretical frameworks discussing corporate leverage, institutional factors play hardly any role. In fact, such consideration do not influence any propositions put forth by Modigliani and Miller (1958), where most important theories currently draws their strengths. However, apart from firm specific attributes, empirical studies would indicate an important role of institutional and macro-economic factors for determining corporate leverage. This is especially critical for firms in emerging markets with evolving capital markets and institutions. 10 Booth et al. (2001) shows that just by knowing the nationality of a firm, a high degree of variation in debt ratios could be explained for firms in the emerging markets. Quite pertinently limited to their scope, empirical work testing capital structure theories tend to incorporate institutional factors as mere controls for firm-specific attributes but do not necessarily provide any direction for their expected correlation with debt ratios. 10 See, for example, Demirguc-Kunt and Maksimovic (1996) and Agarwal and Mohtadi (2004) for relative importance of institutional factors in explaining the capital structure of firms in the emerging markets. 9

10 Key institutional factors considered in the literature can be classified into factors associated with availability and substitutability of debt, macroeconomic factors and industry specific factors. Development of credit markets could ensure higher accessibility of credit to the firms and hence would lead to higher debt ratios for the firms. On the other hand, development of stock markets could lead to substitutability of debt and hence signifies lower debt ratios for the firms. In general, development of capital markets improves the distribution of risk among several investors and hence would generally lead to improved financing for firms either through debt or equity. Demirguc-Kunt and Maksimovic (1996) shows that development of stock markets is complementary for the development of credits markets for developing countries. This is particularly found to be the case for large firms in emerging markets. Thus, accessibility of credit for firms can be improved by liquidity and depth in stock markets. We capture the effect of bank credit as the ratio of aggregate bank credit for commercial sector to GDP. We have also included the ratio of corporate financing through non-banking sources as a proportion of total corporate assets as an additional measure of credit market development. We measure stock market development as the ratio of total market capitalization to total assets of all nonfinancial firms in India. Further, we also include total trading volumes of listed firms as a proportion of total corporate assets as an additional measure of stock market development. Among macro-economic factors influencing the financing choices of the firms, Agarwal and Mohtadi (2004) highlights the role of GDP growth, corporate investments and foreign cash inflows. Along with these macro-economic factors they considered credit and stock market development related variable to adjudge their influence on cross-sectional debt ratios. We consider these control variables in our study in the form of GDP growth, total corporate investment in a year as a proportion of total corporate assets and total net foreign investment inflows (including direct and portfolio investments). 10

11 Frank and Goyal (2009) found that inflation and industry leverage are also the significant determinants of corporate leverage. We have used the change in consumer price index as a proxy for inflation controlling for general economic conditions. Industry leverage is defined as average debt to capital ratio for each industry for each year in our analysis. Industries are classified in seven classes using National Industrial Classification (NIC) codes in India. These are manufacturing, mining, construction and real estate, electricity and power, services, transportation and diversified. Unlike firm-specific variables, institutional variables are constant for all the firms in a particular year. This restricts the variation of these variables between firms in a given year and across industry. Summary of all the relevant variables is presented in Table Empirical Analysis 3.1 Data We have considered data from all available non-financial firms in CMIE (Centre for monitoring of Indian economy) database from 1992 to CMIE has composite data for a population of more than 20,000 firms. This includes public as well as private firms and firms in diverse sectors such as manufacturing, mining, power and service industries. Number of firms for which data is available in a given year varies across the time period chosen. Data from annual financial statements of the firms is considered for the empirical analysis. We have excluded firms having (i) annual sales less that Rs. 10 Mn, (ii) negative or no debt, (iii) negative total shareholder s equity, and (iv) missing values for any primary variable of interest. After filtering the data on these parameters, we have also removed firm-year observations on tails to the extent of 0.5 percentile on either side for important primary variables of interest. Final sample consists of 8081 firms, each having different total number 11

12 of observation during 1992 to This constitutes the unbalanced panel data for further analysis. Appendix A and B shows the descriptive statistics and pair-wise correlation between all the variables of interest. 3.2 Methodology The focus in this paper is to see the relative impact of most consistent firm-specific and several possible institutional determinants of corporate leverage on the evolution of debt ratios for firms in India. Aggregate or pooled data on explanatory variables may not provide us with insights into firm level decision to increase or decrease corporate leverage. Further, several unknown variables may influence the financing decisions of the firm. Moreover, these unknown variables may be correlated with explanatory variables and hence can bias the estimators 11. In this study, since we are guided by most consistent theoretical variables explaining variation in debt ratios in the literature, an implicit assumption is that other time varying variables are not significant. However, we explicitly control for time-invariant unknown variables for firms which may be correlated with explanatory variables. We have made use of fixed effect panel data model to control for such time-invariant variables. Fixed effect models 12 enable us to look for within-firm effect of subject explanatory variables on debt ratios for firms. This way we can focus on within-firm change in leverage and also control for changing composition of the sample firms over time. Fixed-effect models, therefore, could have greater utility for unbalanced panel data. We employ the following generic panel data fixed-effect model to estimate the influence of subject variables on debt ratios. Debt Ratios Firm. Factors (1) it i it 11 Hausman (1978) specification test for panel data models confirms such correlations to be significant. 12 See Wooldridge (2002) for fixed-effect models in Panel data analysis. 12

13 Where, debt ratio is the ratio of total borrowing to total capital. 13 Total capital is defined as the sum of total borrowings and total equity consisting of total reserves and owner s capital ploughed into the firm. Firm denotes the effect of unknown within-firm time invariant effects and Factors denotes the set of firm-specific and/or institutional explanatory variables under consideration. Error terms have been adjusted by using robust standard errors to control for heteroscedasticity Influence of firm-specific variables on corporate leverage As discussed in the introduction, we suspect a potential value loss for the firms such that they are not being able to rebalance their debt ratios to avoid any potential tax drain. If such is the case we posit that most consistent firm-specific variables would not be able to explain the decline in debt ratios for firms in India. We begin our empirical analysis by observing the within-firm trend of all firm-specific variables using the following within-firm panel model: Factor it Firm. Year i t t it (2) Where Yeart represents the complete set of time period dummies from 1992 to 2012 with dummy omitted for the year The coefficients on Year would then indicate the increase or decrease of subject variable over its standing in the year Thus, the within-firm trend, if identified, is the movement in coefficients of year dummies for the subject period. 13 Market value debt ratios are not considered since market values of equity is not known for many of the firms in the sample. Further, where market values are available for firms, growth variables such as market to book ratios and stock market development variable, market capitalization to total assets are naturally negatively correlated with such debt ratios. In unreported findings we found that such correlations are significant and hence could bias the estimation. Also, support for not using market measures of leverage can be found in Myers (1977) and Graham and Harvey (2001). 13

14 We then proceed to estimate the extent and direction of influence traditional firm-specific variables would have on debt ratios. Specifically we use the following fixed effect panel estimation model: Debt Ratios Firm. Firm Factors it i a it (3) Where Firm Factors denotes the firm specific factors measured as follows: Size - log of annual sales; Profitability - ratio of operating income before interest, taxes and depreciation to total assets; Asset tangibility - ratio of net fixed assets to total assets; Growth - ratio of total capital expenditure to total assets as well as a ratio of market to book value of equity for firms with available market values of equity; Tax rate - taxes paid as a fraction of total earnings before taxes, where, tax paid for a firm is estimated by subtracting after tax earnings from earnings before tax 14 ; Non-debt tax shields - ratio of depreciation and amortization expenses to operating income before depreciation and tax; Effective tax burden - ratio of total tax paid in a year to the total operating income of the firm. In the recent debate on dynamic nature of firms existence, firms choose their capital structure dynamically to maintain a balance between their investment opportunities on the one hand, and the financing possibilities in the form of internally generated funds and external financing on the other. During this process, firms incur adjustment costs that affect their financing choices. In order to accommodate such adjustment costs, past literature has used dynamic panel models for the testing of target debt ratio following behavior of the firms. These models use past debt ratios at time t-1 (ADRt-1) to predict the future debt ratios at time t (ADRt). 15 In the spirit of 14 A more pertinent measure of tax rates would be the marginal tax rate than average tax rate as suggested by Graham (1996). This is because the most relevant tax rate for the profitable firms would be marginal tax rate to decide on incremental leverage for the firms. Further, average tax rates for the firms may vary with their pre-tax earnings and hence would rather be a measure of profitability. Lack of availability of data prevents us to use marginal tax rates as the proxy of tax variable. 15 See, for example, Faulkender et al. (2012), Huang and Ritter (2009), Lemmon et al. (2008), Kayhan and Titman (2007), Flannery and Rangan (2006), and Hovakimian et al. (2001), among others, for the use of dynamic panel 14

15 these dynamic panel models, we also estimate the significance of firm-specific variables by including lagged debt ratios as an explanatory variable as follows: Debt Ratios it Firm i Debt Ratios i, t 1 a. Firm Factors it (4) Using the within-firm trend and the estimated direction of influence for the firm-specific variables, we then proceed to seek an explanation for the contribution of these variables in the decline in debt ratios. For example, a downward within-firm trend in a particular variable along with a positive influence on debt ratios would contribute toward the decline in debt ratios and vice versa Influence of institutional factors on corporate leverage Since we observe an increasing tax burden for firms in India through Figure 3, we posit that the decline in debt ratios could be a consequence of adverse effect of systematic factors influencing these firms. We estimate the relative influence of institutional and macro-economic factors on corporate leverage through the following conventional and dynamic panel models: Debt Ratios it Firm i a Firm Factors b.. Inst Factors (5) it Debt Ratios it Firm i Debt Ratios. Firm Factors. i, t 1 a b Inst Factors (6) it Where Firm Factors are as explained above and Inst Factors denote institutional factors measured as follows: Industry leverage (IND D/(D+E)) average debt to capitalization ratio of firms in each of the seven industries classified as per the NIC codes for each year; GDP models. Agarwal and Mohtadi (2004) has also used dynamic panel model to estimate the effect of financial market development on firms financing decisions. 15

16 growth (GDP) annual growth rate in GDP of India; Investments (INV) total annual corporate capital expenditure as a proportion of total assets of non-financial firms in India; Bank Credit (BC/GDP) ratio of bank credit to commercial sector as a proportion of GDP; Other Borrowings (OTHB/TA) ratio of total annual corporate borrowings from sources other than banks (sum of corporate debentures and bonds, inter corporate loans, foreign commercial borrowing, borrowing from other financial institution and commercial papers) as a proportion of total assets of non-financial firms in India; Market capitalization (MC/TA) ratio of total market capitalization of listed firms as a proportion of total corporate assets of non-financial firms in India; and, Trading volume (TRDV/TA) ratio of total trade volumes of the listed stocks in a given year as a proportion of total corporate assets of non-financial firms in India; Foreign inflows (FDI) ratio of total net foreign inflows (including direct and portfolio investments) as a proportion of GDP. In addition to identifying the statistical significance of these variables, similar to our analysis for firm-specific variables, we take guidance from the trend of institutional factors to evaluate their relative economic contribution on declining debt ratios. However, since the values of these institutional variables remain constant for all the firms in a given year, we could not observe the within-firm trend but observe their trend from the change in their numerical values overtime. 3.3 Results Influence of firm-specific variables on corporate leverage We estimate the within-firm trend of the firm-specific variables using equation (2). The results are reported in Table 2. It can be seen from the table that a within-firm trend is quite visible in the some of the variables of interest. Specifically, we see a downward trend in debt ratios, 16

17 profitability and growth opportunities when measures as the ratio of capital expenditure to total assets. While size and effective tax burden shows a perceptible upward within-firm trend, asset tangibility, market to book ratio of equity, tax rate and non-debt tax shield do not show any perceptible trend. To statistically confirm a decline (increase) in the trend of debt ratios, profitability, growth opportunities (size and effective tax burden), we have employed two-sample Wilcoxon (1945) rank-sum (or Mann-Whitney, 1947) test for year 1992, 1993 and 1994 versus year 2010, 2011 and 2012 to test for the equality of medians in any combination of a year from the first set (initial years) with any year from the second set (later years). We find that the median debt ratios, profitability and growth opportunities (size and effective tax burden) of the firms in the later years are significantly lower (higher) than that of the medians in the initial years. This confirms a significant trend in these variables over time. The direction of trends in firm-specific variables such as size, profitability and growth opportunities are very similar to the trends in these variables for other emerging markets as reported in Mitton (2007). On such an account, it is difficult to infer that firm-specific constitution for firms in India differs from firms in other emerging markets. However, unlike other emerging markets, the trend in debt ratios for Indian firms is in stark contrast. While firmspecific variables significantly account for the explanation of increasing trend in debt ratios for firms in emerging markets, this suggests that firm-specific attributes with similar trends may not explain the declining trend in debt ratios of firms in India. In other words, applying theoretical frameworks explaining corporate leverage in India could be challenging. However, as suggested by Booth et al. (2001), we may expect an increasing role of institutional factors in explaining debt ratios for Indian firms. We will formally test for such eventualities now. 17

18 We begin by testing for significance of firm-specific explanatory variables using equation (3) and (4). Following Mitton (2007) we have also used an interaction term, a product of tax rate and non-debt tax shield, to see if tax rate interacts with non-debt tax shields. A negative and significant effect of such interaction term would mean that although firms may take more debt owing to higher taxes, they may substitute the tax savings out of interest payments when they are also exposed to higher non-debt tax shields and vice versa. The results are reported in Table 3 in column (1) and (2) respectively. A couple of interesting observations are as follows. First, the signs for all conventional variables showing significance, except growth opportunities, are as expected and as identified in other emerging market studies such as Booth et al. (2001) and Mitton (2007). Tax rates, nondebt tax shields and their interaction effect is found to be non-significant. However, as hypothesized earlier, effective tax burden negatively and significantly influence the debt ratios. Second, similar to previous studies all of the firm-specific variables are found to be significantly influencing debt ratios. Thus, we may infer that theoretical frameworks explaining debt ratios in other countries are portable to an extent in India also. However, economic significance of these variables suggests an increase in debt ratios rather than a decrease. This can be seen from looking onto the within-firm trend in these variables in table 2 and their respective coefficients in the estimation models in table 3 simultaneously. If we multiply the change in these variables from 1992 to 2012 with their respective coefficients in a model, we can identify the change in debt ratios attributed to change in these variables. For instance, Log(sales) increased by 252% from 1992 to A coefficient of indicates an increase in debt to total asset of 6.80% over time. However, effective tax burden, unlike other conventional firm-specific variables, is found to be a significant contributor to the decline in debt ratios. Along with the increase in effective taxes of about 5.78% between 1992 and 2012, 18

19 effective tax burden, with its negative and significant coefficient, contributes more than 2% in decline of debt ratios overtime. Upon aggregating results for all such changes in significant firm-specific variables in column 1, we see that firm-specific variables jointly account for an increase of 5.97% in debt ratios between 1992 and Our results for the significance and trend of firm-specific factors are largely consistent with the findings in Mitton (2007). However, unlike their findings, the significance and trend of these firm-specific factors do not explain the trend in corporate leverage in India. Consistent with the prior literature these factors indicate towards an increase in corporate leverage rather than explaining the observed decline in debt ratios in India. This leads us to explore the effect of institutional and other systematic factors in explaining the decline in debt ratios Influence of institutional factors on corporate leverage We estimate the relative influence of institutional factors in explaining the cross-sectional variation as well as the decline in debt ratios overtime. Specifically, since firm-specific factors do not contribute towards the decline in debt ratios, we expect a larger contribution from institutional and macro-economic factors to explain such decline overtime. Similar to the analysis above we observe the time series trend in the institutional and macroeconomic variables described in section 2. However, instead of reporting within-firm trend we report the change in these variables as a percentage of their numerical values at the start of the observed time in Since the data for the variable TRDV/TA is not available before 1996, we report its variation starting from Also, since FDI is heavily correlated with other variables of interest we drop it from further analysis. The trend in these variables is shown in Table 4. As can be seen variable for corporate investments (INV), bank credit to commercial sector (BC/GDP) and borrowing from sources other than banks (OTHB/TA) shows consistent 19

20 and significant decline. Importantly, these variables are related to credit availability and its deployment by firms in India and somehow represent the critical development of credit markets in India. The movement in other variables such as GDP growth (GDP), inflation (INFL), market capitalization (MC/TA) and trading volume (TRDV/TA) do not show any perceptible trend. We test for relative influence of these variables in the presence of other firm-specific variables using equation (4) and (5). We drop effective taxes as the variable of interest in the regressions involving institutional factors as it is found to be heavily correlated with the variable Other borrowings (OTHB/TA). If firms are losing value on account of their inability to raise capital, we expect these firms to underinvest. Such inadequacy of funds can be tested by observing the economic effect of variables associated with credit and stock market development. While bank credit to commercial sector and borrowing from sources other than bank represent credit market development, market capitalization and trading volumes represent stock market development. Similarly, underinvestment by firms, if any, can be tested by observing the economic effect of aggregate corporate capital investments on debt ratios of the firms. In addition to their individual effect, we further suspect that inadequacy of credit through credit markets would relate positively with overall decline in corporate investments. In order to check for such possibility we include an interaction term for the product of corporate investments (INV) and borrowing from sources other than bank (OTHB/TA). When both these variables are declining overtime, a positive and significant coefficient of such interaction term would mean a complementary but adverse effect of underinvestment and inadequacy of credit to the firms. The results are reported in column (3) and (4) of Table 3 for model in equation (4) and (5) respectively. 20

21 As expected, coefficients for all the institutional factors are significant in both the models except average industry leverage. The adverse economic effect of underinvestment and inadequacy of credit for growth of the firms could be verified by the positive and highly significant coefficients for investments (INV), Bank Credit (BC/GDP) and Other borrowings (OTHB/TA) and their respective decline in values overtime. This economic effect of these three variables can be estimated by multiplying their respective coefficients with their percentage movement overtime. Using the model (4) and its results in column (3) in Table 3, we find that investments, Bank Credit and Other borrowings contributes 11.33%, 14.99% and 27.36% to the decline in debt ratios respectively overtime. These contributions are quite larger when compared to the magnitude of contributions by firm-specific factors. Moreover, unlike firm-specific factors, the contributions from these three institutional variables of interest actually signifies the decline in debt ratios. Furthermore, we find a positive and significant coefficient for the interaction term between investments and other borrowing. This suggest that while decline in corporate investments overtime causes debt ratios to decline, the decline in borrowings from sources other than banks aggravate the decline in debt ratios further. This confirms our conjecture that firms do face difficulties in availing capital for their growth from credit markets and hence underinvest. This also implies that such adversities may be quite serious in magnitude such that firms are not able to rebalance their debt ratios so as to preclude any adverse tax drain amounting to the value loss. Importantly, such underinvestment due to difficulty in availing credit for growth is quite different from the underinvestment hypothesized in Jensen and Meckling (1976) owing to the debt overhang problem that surface due to excessive debt with the firms. While market capitalization (MC/TA) is found to be negatively and significantly influencing the debt ratios, no perceptible trend could be seen in it overtime in Table 4. Similar effect could 21

22 be seen for Trading volume (TRDV/TA) estimated through a smaller sample of firm-year observations from 1996 to This is shown in column (5) in Table 3. The expected signs on bank related and stock market related variables are consistent with Demirguc-Kunt and Maksimovic (1996) and Agarwal and Mohtadi (2004). Importantly, the significance of these stock market related variables is found smaller and varying when compared to the stronger significance of credit market related variables. This suggest non-substitutability of credit markets with vibrant stock markets in availing of capital in India. Our results are robust for different size of the firms, across different time periods and alternate proxies for growth options. Notwithstanding the results, the stability and consistency of the coefficients for institutional factors needs to be interpreted cautiously. The effects of these variables may be restrictive in fixed-effect models due to the limited variability of these variables across firms in a given year. In the next section we undertake a series of robustness checks to confirm the validity of our findings. 4. Discussion and Implications The results in the previous section reveal significant role of institutional factors in explaining the declining debt ratios of firms in India. Even though, we used different measures of credit and stock market development in our empirical analysis, the role of credit market variables seem to dominate the explanation. Both the variables viz. bank credit to commercial sector as a proportion of GDP and corporate borrowings as a proportion of total assets shows a declining trend overtime. This could suggest a scenario of falling aggregate capital requirements and underdeveloped credit markets consistent with the observations of Mitra (2009) and Raghvan and Sarvano (2012). Further, these variables positively and significantly influence the debt ratios and hence found to be contributing comprehensively towards its decline. The results are 22

23 consistent with Agarwal and Mohtadi (2004) where they suggest that credit market development favors debt financing and hence its under-development could diminish the ability of firms to finance themselves with debt. Although, we find the effect of stock markets being influential in explaining the cross-section of debt ratios, the trend in the variables representing stock market development overtime do not provide a possible explanation for the decline in debt ratios. This is consistent with our initial understanding from Figure 2 of simultaneous decline in the ratio of equity paid-in capital along with debt as a fraction of total capital which reflects their non-substitutability. The results are also in line with the findings in Allen et al. (2006) that despite rapid developments in stock markets in India, they have not contributed enough in resource allocation and financing of the firms. According to Allen et al. (2006), the external financing of the firms in India has being dominated by banking and alternative sources. Since the decline in debt ratios could not be attributed to raising of equity capital to substitute debt effectively, we were prompted to check if overall corporate investment correlates with the declining debt ratios. The results suggest that such is indeed the case. We find a declining trend in corporate investments as a proportion of their total assets overtime which is further found positively and significantly influencing the debt ratios of the firms. We further check if the effect of corporate borrowing from sources other than banks substitute or complement the corporate investments by using an interaction effect. Our results suggest that the two effect complement each other and add to adversities of the firm. The results point towards the credit market adversities faced by the firms to avail much needed credit for growth. An important question at this juncture would be why firms would have to face such adversities? Why corporate markets in India could not be developed enough? Why corporate capital investments are declining overtime? We attempt to suggest some possible 23

24 answers to these questions here following the results in the previous section and state of the affairs of Indian economy. The empirical analysis in this paper suggest that firms in India may be credit rationed owing to the possibility that falling debt ratios are not being substituted by raising more equity capital. Such effective non-substitution can be seen from the empirical findings and also from revisiting Figure 2 showing the aggregate trend in funds provided by equity capital in India. Figure 2 shows the break-up of total equity into total paid-up equity capital and reserves and surpluses as a percentage of total capital for non-financial firms in India between Total paidup capital as a fraction of total capital (E/(D+E+RE)) declined consistently from 24% (17.7%) in 1992 to 12.7% (5.8%) in 2012 for manufacturing (non-manufacturing) firms. On the other hand, reserves and surpluses as a fraction of total asset (RE/(D+E+RE)) grows from 21% (22.7%) in 1992 to 42% (49.7%) in 2012 for manufacturing (non-manufacturing) firms. The credit rationing may be reassuring by observing the trends closely between where stock markets grew by almost three folds. Stock market boom in this period would have led firms to issue more equity as a replacement of debt. However, the trend suggests that even in this period firms did not raise more equity capital and grew largely by the increase in reserves and surpluses. This may be because raising equity for firms in India might be prohibitively costly. Owing to the lack of sound debt markets, banks in India dominate the credit markets. Figure 4 shows data pertaining to the composition of borrowings by firms in India through banks and otherwise. As can be seen, bank borrowings as a proportion of total assets increases sharply overtime. 16 Moreover, reliance on bank credit remained strong even during the economic booms such as between On the other hand, the effect of underdeveloped bond markets 16 As noticed before, bank credit as a proportion of GDP has declined overtime. 24

25 can be seen more clearly in the pattern of borrowings from sources other than banks as a proportion of total assets, which shows continuous decline. The story is true even for larger listed firms (not shown here). Even though banks dominates the credit channel for private sector, the availability of credit is severely restricted and rationed due to several factors. First, government borrowings competing for funds from public savings that can serve as credit to the commercial sector. Government securities issued to finance the fiscal deficit consume a large share of public savings. Figure 5 shows that government debt receipts (excluding external assistance) as a fraction of total financial household savings remain high and averages about 47% between 1992 to Higher share of public borrowings limits the supply of overall funds available to be disbursed through banking channel for entrepreneurial production. Second, commercial banks are supposed to put up 23% of their net time and demand liabilities as SLR securities which are largely treasury instruments, public sector bonds and other high credit rating instruments. This shrinks the availability of funds even further to non-government enterprises and provides more funding to the government treasury indirectly. Moreover, 4% of total deposits of a bank are to be retained as non-interest bearing CRR in the form of cash parked with reserve bank of India; 40% of the aggregate advances by any commercial bank need to be allocated as priority sector allocations 17 away from mainstream competitive firms. Finally, quantitative restrictions in the form of priority sector lending, prudential norms and caps for investments in corporate debt instruments, dominance of public sector banks in the banking sector and high interest rates on government securities lead banks to invest more than SLR requirements in credit risk free assets. 17 Categories under priority sector include agriculture, micro and small enterprises, education loans and housing loans. 25