Firm Financing in India: Recent Trends and Patterns Inessa Love and Maria Soledad Martinez Peria* The World Bank Abstract: Using balance sheet information for nearly 6,000 firms between 1994-2003, this study investigates recent firm financing patterns in India. The paper documents the overall use of debt and, in particular, the role of bank financing (short-term and long-term), trade credit, intra- business group borrowing and foreign financing. The study examines financing patterns over time and explores differences across firms by sector, age, ownership type, export orientation, and, in particular, size. In terms of trends, we find that while debt to asset ratios have been relatively stable, nominal debt growth has slowed down in recent years. At the same time, firms' repayment capacity, as measured by the interest coverage ratio has exhibited a U-shaped pattern falling during 1997-99 and recovering in recent years. Throughout the period of study, bank financing as a share of total debt has increased, while borrowing from non-bank financial institutions fell sharply. In terms of differences across firms, the most robust finding is that debt levels increase with firm size. Smaller firms have especially less debt relative to larger firms if they are young (below 10 years since incorporation), if they are in the manufacturing sector, and if they are located in Southern India. Furthermore, while the ratio of debt to assets has been relatively stable for large firms, we observe a significant decline for smaller firms. Overall, the findings presented in the paper provide suggestive (but not definite) evidence of stronger credit constraints for smaller firms. * We would like to thank Priya Basu and Leora Klapper for guidance and suggestions for this study. Niraj Verma and Mahesh Vyas were instrumental in helping us obtain and understand the data we use. Xuxin Alex Yu provided excellent research assistance. 1 I- Introduction Adding to the well-documented link between finance and growth at the macro level (see Beck, Levine and Loayza 2000a, 2000b), recent research has provided evidence of this positive association at the firm level as well. In particular, it has been shown that firms that are financially constrained tend to experience slower growth (see Beck, Demirguc-Kunt, and Maksimovic, forthcoming). During the early 1990s, India embarked on a series of structural reforms including important changes in its industrial, trade, and financial sector policies that stimulated growth and investment (see World Bank, 2003). Between 1992/93 and 1996/7, Indian GDP grew by almost 7%. However, since then growth has slowed down to a rate of 5% and some have argued that remaining bottlenecks and deficiencies in the country's investment climate are partly to blame. In particular, expert committees appointed by the Indian government (see India 1997, 1999, 2000), as well as recent surveys of Indian firms (see Confederation of Indian Industry and World Bank, 2002), have emphasized the lack of adequate and timely financing on competitive terms as a key constraint on firm growth, especially among small and medium firms. Using balance sheet information for nearly 6,000 firms between 1994-2003, this study investigates the recent financing patterns of firms operating in India. In particular, the paper documents the use of debt, the role of bank financing (short-term and long-term), trade credit, intra-business group borrowing, and foreign financing. This study tracks financing patterns over time and explores potential differences across firms by sector, age, ownership, export orientation, and, in particular, size. Univariate t-tests are used to investigate differences in the mean and median financing ratios across firm types. Furthermore, following the empirical literature on firm capital structure (see, for example, Booth et al. 2001, Klapper et al. 2002, and Rajan and 3 Zingales 1995), regression analysis is also used to study the determinants of debt ratios over time and across firms. A number of papers have conducted this type of analysis for Indian firms (see Bhaduri 2002, Bhaduri 2000), however, the existing studies have primarily focused on a relatively small sample of firms for the early to mid-1990s. The analysis in this paper shows that while as a proportion of assets debt levels have been relatively stable, nominal debt growth has slowed down in recent years. At the same time, firms' repayment capacity, as measured by the interest coverage ratio, has exhibited a U-shaped pattern falling during 1997-99 and recovering subsequently. Throughout the period of study, bank financing as a share of total debt has increased, while borrowing from non-bank financial institutions fell sharply. In terms of differences across firms, the most robust finding is that debt levels increase with firm size. Smaller firms have especially less debt relative to larger firms if they are young (below 10 years since incorporation), in the manufacturing sector and located in Southern India. Furthermore, while the ratio of debt to assets has been relatively stable for large firms, we observe a significant decline for smaller firms. Overall, the findings presented in the paper provide suggestive (but not definite) evidence of stronger credit constraints for smaller firms. The remainder of the paper is organized as follows. Section II describes the data analyzed in this study. Section III presents indicators of firm financing patterns and discusses their recent trends. Section IV explores differences in financing patterns across firms by size, sector, age, ownership, and export orientation. This section presents tests of differences in mean and median financing ratios by firm characteristics. Section V discusses the regressions results for the determinants of firms' capital structure. In particular, we study the impact of size on the level of 4 firm debt and on the responsiveness of debt ratios to other factors known to affect debt financing. Finally, section VI concludes. II- Data Description The firm level data used in this study comes from Prowess, an electronic database produced by the Center for Monitoring the Indian Economy. Our sample of study includes 5,781 firms over the period 1994-2003. Thus, in total, the database we analyze contains about 40,000 observations, roughly equally distributed over the years. Table 1 shows how firms in the data are distributed by size, region, age, sector, ownership type, and export orientation. Also, this table presents the distribution of observations by year. Approximately, 27% of the firms in our sample are small (i.e., have gross fixed assets below 50 million Rs.), 15% are classified as medium (with fixed assets between 50 and 100 million Rs.), and the remaining 58% are large firms (with gross fixed assets above 100 million Rs.).1 Most firms in our sample (43%) are based in Western India, the region including the states of Gujarat, Maharashtra, Goa, Madhya Pradesh, and Chhattisgarh. More than half of the firms in this region (24% of the total number of firms) are large, and approximately one third (13% out of the total number of firms) are small. Southern India, encompassing Andhra Pradesh, Karnataka, Kerala, and Tamil Nadu, is the second most populous region in terms of the number of firms. Approximately 1,300 firms or 24% of the total number of firms in the sample are located in this region. As in the West, most of the firms -15% out of the total number of firms in the country- are large, almost 5% are small and 4% are medium-sized. Firms from the Northern and Eastern regions account for almost 20% and 13% of all Indian firms, respectively. The Northern region 1We divide our firms in size categories using the total amount of fixed assets following the methodology adopted by the Indian government in classifying small and medium enterprises. 5 includes the states of Jammu and Kashmir, Himachal Pradesh, Punjab, Haryana, Uttar Pradesh, Rajasthan, Delhi, and Uttaranchal, while Eastern India is made up by the states of Bihar, Orissa, West Bengal, Assam, Meghalaya, Tripura, Mizoram, Manipur, Nagaland, Arunachal Pradesh, Sikkim, and Jharkhand. In both Northern and Eastern India, most firms are classified as large (i.e., more than 50%) and approximately one third are small. In terms of age, 80% of all firms are more than 10 years old. Approximately 49% of all firms in the sample are large and more than 10 years old, almost 20% are small and "old". The second largest age category includes firms between 5 to 10 years old. Approximately 17% of all firms fall in this category, while only 2.9% of firms are less than five years old. The sectoral composition of firms is equally skewed. Approximately 81% of firms are in the manufacturing sector, while the remaining 19% are service firms. Among manufacturing firms, the most important sector corresponds to the production of chemicals. This is the most populated sector among large firms (large chemical firms account for 10.8% of all Indian firms) and the second most important sector for small firms (small chemical firms account for 3.1% of all Indian firms). Approximately a quarter of all firms (23% to be precise) are classified as exporters, using the rule that requires that they have sold more than 10% of their products overseas during at least half of the sample period considered. Out of the 23% of firms that are classified as exporters, almost 70% (or 16.2% of all firms) are large firms. In terms of ownership, the vast majority of firms in the sample are private Indian firms (89.7% of the total). Foreign private firms make up for almost 7% of the firms in the sample, while the remaining firms have ownership ties to the national, state or municipal governments. 6 III- Recent trends in firm financing The focus of this study is to understand the financing patterns of firms in India. We begin with an analysis of some key financing ratios (the main one being the debt to assets ratio) for the period 1994-2003. First, we present some descriptive statistics for these ratios (see Table 2) and we compare them with those observed in other countries over this period. Second, we discuss the behavior of these ratios over time for all firms combined (see Table 3) and, separately, by sector and size (Table 4 and 5). Finally, we also explore changes in debt levels (instead of ratios) during recent years (Table 6, 7 and 8). Our main variable of interest is the ratio of debt (i.e., total borrowings) to assets. Debt includes bank financing, foreign borrowing, borrowing from non-bank financial institutions, corporate financing, public borrowing (debentures and commercial borrowing), government borrowing, and hire purchase. Over the entire period considered, this ratio averaged 0.36, with a median of 0.32. The minimum debt to asset ratio was 0 and the maximum was 2.46. Ninety five percent of firms had debt to asset ratios below 0.87. We also consider ratios of total liabilities to assets, payables to assets, and long-term debt to assets. Total liabilities include total borrowings plus trade credit (i.e. payables), accrued interest, deferred taxes, and provisions (usually for taxes, dividends, retirement benefits, etc.). Over the full sample period, the mean ratio of total liabilities to assets was 0.69 and the median was 0.64. At the same time, the mean ratio of payables to assets was 0.14 and the median was 0.10. The ratio of payables to assets is an indicator of trade credit provided to firms by their suppliers, typically on a short-run basis. Finally, the average ratio of total long-term debt to assets was 0.24 and the median for this ratio was 0.18. 7 Comparing these aggregate ratios with those for other countries, debt levels in India are relatively high. On average, firms operating in India have slightly more debt to assets than firms in the U.S. (with a mean ratio of 0.27), U.K. (0.18), Italy (0.27), France (0.25) and Germany (0.16), but similar debt to assets ratios as firms in Japan (0.35) and Canada (0.32) (see Rajan and Zingales, 1995). In terms of total liabilities, firms in India have ratios that are similar to those observed in 4 out of 7 countries (namely, France, Germany, Italy, and Japan) and are higher than those for firms in 3 out of 7 countries (US, UK, and Canada). Comparing firms in India to firms in developing countries as described in Booth et al. (2001), we find that firms in India fall in the group of high-debt countries, along with Pakistan and South Korea. In addition to looking at aggregate debt ratios, we also examine the interest coverage ratios of firms in India, defined as the share of earnings before depreciation, interest and taxes to interest payments. The interest coverage ratio measures the number of times that company's earnings exceed its interest payments. This ratio shows the ease with which a firm can meet its debt payments (i.e., it measures the repayment capacity) and provides an indication of how far a company's cash flows could fall before the company will have difficulties meeting its interest payments. We find that on average interest coverage ratios during 1994-2003 were relatively high ­ the mean was about 6 for the whole sample. However, the median was approximately 2, a value more in line with what is found for other countries.2 This suggests that the median firm could cover twice its interest payments with its cash flows. In other words, even if the earnings of such a firm were to have fallen by 50%, it would still have been possible for such a firm to meet its interest payments. These are considered healthy interest coverage ratios by most financial industry standards. 2 The significant discrepancy between mean and median interest coverage ratios suggests that only some firms in our sample have very high debt repayment capacities (less than 5% as indicated by the value of the interest coverage ratio for the 95th percentile). These are firms with really little interest payments, relative to their earnings. 8 To capture potential difficulties with debt repayment in the Indian corporate sector, we also consider two additional proxies for possible financial distress. The first indicator measures the percentage of firms with negative net worth. Negative net worth implies that the firm has more liabilities than assets. This measure is often used as a balance sheet measure of insolvency.3 The second indicator measures the percentage of firms with an interest coverage ratio below one. Firms meeting this criterion are likely to encounter difficulties repaying their interest payments with their earnings.4 We find that on average about 12% of all firms have negative net worth, while for about 22% of firms the interest coverage ratio is below one. Finally, approximately 9% of firms have both negative net worth and coverage ratios below unity. Next, we consider a number of ratios that highlight the sources of debt financing for firms in India. Specifically, we separately consider the percentage of total debt that comes from bank borrowing, foreign financing, public borrowing (commercial paper plus debentures), borrowing from non-bank financial institutions, borrowing from corporations and the catch-all category "other borrowing" that includes government loans, promoters, hire purchase and any residual borrowing. These categories cover all sources of borrowing and, therefore, they sum to 100%. We find that bank borrowing represents by far the largest category among different financing sources. On average, over the period 1994-2003 bank borrowing represented 47% of total debt (the median was 44%). The next largest category is borrowing from other financial institutions5 3Negative net worth means that even if the firm sold all its assets (at their book value), it will not have enough funds to cover all its debt obligations. However, the firm might still not be considered immediately insolvent if it is able to meet its interest payment with its earnings. 4In the case of earnings shortfalls, the company could use other sources of cash flows to repay the interest, such as reduce inventories, collect on accounts receivables, or increase accounts payables. In addition, the firm could sell some of the assets or raise additional external finance ­ but these means would only work in the short-term. Thus, an interest coverage ratio below one does not automatically suggest a default, but it is indicative of possible difficulties in debt repayment. 5 This category includes borrowing from all non-bank Indian financial institutions, including development banks (like IDBI, ICICI, IFCI and their subsidiaries), as well insurance companies, such as Life Insurance Corp., General Insurance Corp., and Unit Trust of India. 9 (which averaged 22% of total debt), and other borrowing (which averaged 15% of total debt). Borrowing from corporations is also a relatively large proportion of total borrowing ­ the average over the period was 9%. On the other hand, public borrowing and foreign borrowing are relatively small fractions of total borrowing, averaging less than 1 and 5 percent of total borrowing, respectively. Approximately 70% of all debt held by firms in India is secured or collateralized. The median value is in fact higher, exceeding 80%. Table 3 illustrates the time series behavior of all the financing ratios discussed above over the period 1994-2003. In general, we observe that most financing ratios were fairly stable over time (both in terms of the means and medians). In particular, the ratio of total debt to assets, the share of total liabilities to assets, and the ratio of payables to assets were relatively constant over the period of study. On the other hand, the interest coverage ratio, the percentage of insolvent firms, and the share of firms with interest coverage ratios below 1 changed considerably between 1994-2003. In particular, the mean and median interest coverage ratio exhibited a U-shaped pattern over the period 1994-2003, falling at the beginning of the period and recovering after 1997-99.6 At the same time, the percentage of insolvent firms and the share of enterprises with interest coverage ratios below one increased steadily over this period. There are also some apparent trends in the behavior of other variables. The share of bank debt (in particular long-term debt) and the percentage of foreign and corporate financing increased since 1994.7 On the other hand, borrowing from non-bank financial institutions 6The mean interest coverage ratio fell from 5.92 in 1994 to 4.70 in 1997, only to more than double its value between 1997-2003, reaching 10.62 in this last year. The median ratio declined somewhat between 1994-1999 (going from 2.93 to 1.99 in 1999) and recovered slightly thereafter reaching 2.85 by 2003, a value still slightly below the initial level. Differences in mean and median interest coverage ratio patterns indicate that while some firms improved their financing capacity over this period, the majority of firms saw a slight drop followed by a mild recovery. 7 Total bank debt has increased from an average of 43% of total debt in 1994 to a high of 52% by 2003. Foreign borrowing went from an average value of 0.7% of total debt at the beginning of the sample to 1.2% by the end of the sample. Corporate financing rose from 6% in 1994 to close to 10% in 2002-03. 10 declined from 27% in 1994 to 13% in 2003. The significant decline in financing from non-bank financial institutions (NBFIs) might be connected to the collapse of this sector around 1997. Looking at debt to asset ratios by firm size and sector (see Table 4), we observe that mean and, especially, median debt to asset ratios dropped for small firms. The mean debt to assets ratio declined from 0.28 in 1994 to 0.24 in 2003 for this group of firms, while the median debt ratio fell significantly from 0.23 in 1994 to 0.09 in 2003. In other words, while in 1994 50% of all firms had debt to asset ratios below 0.23, by 2003, this number dropped to 0.09. However, large firms have not experienced any significant decline in debt to assets ratios over this period. In terms of ratios by sector, debt to asset ratios increased for firms in the foods, textiles, and metals sectors, while they have tended to decline in the computers and other services sectors.8 Across most firm sizes and sectors, we observe the aforementioned U-shaped pattern for the behavior of the average interest coverage ratio (see Table 5). After dropping considerably between 1994-1998, the average interest coverage ratio increased significantly for most firms during the period 1999-2003. We observe no differences in the behavior of the interest coverage ratio among small, medium, and large firms. However, there are some exceptions to this pattern across sectors. In particular, firms in the foods and garments and leather sectors saw their average interest coverage decline between 1994 and 1998, peak during 1999 and fall thereafter (with exception of the year 2003 for the latter sector). The median interest coverage ratio also experienced a U-shaped pattern during the period 1994-2003. However, as opposed to what we found for the mean, where the average ratio was 8 Debt to asset ratios in the food sector went from average (and median) ratio of 0.34 in 1994 to a high mean(median) ratio of 0.47(0.4) by 2001, settling at 0.43(0.35) by 2003. In the textile's sector, the average debt ratio went from 0.4 in 1994 to a peak of 0.55 by 2001-02. Firms in the metals sector saw their average debt ratio rise from 0.39 at the beginning of sample to 0.48 by 2001. The average debt to asset ratio for firms in the computer and services sector fell from 0.21 and 0.28 in 1994 to a low of 0.09 and 0.17 towards the end of the sample, respectively. 11 higher in recent years than at the beginning of the sample, for most sectors, and across all firm sizes, the median interest coverage in the most recent years was below its value in 1994. Recent trends in financing ratios could be driven by either changes in the numerator ­ for example, the level of debt- or by fluctuations in the denominator ­ total assets. Given our interest in understanding whether the availability and use of debt financing by firms in India changed recently, it is important to separately examine the growth rate of each of these variables (i.e., the growth in the numerator of the financing ratios). Table 6 shows mean and median nominal growth rates of total borrowing, total liabilities, total payables, total bank borrowing and total assets. Also, in order to trace firm performance and repayment capacity over this paper, we also show the behavior of interest coverage ratios, return on assets, and the growth of profits before interests and taxes. The picture that we get from looking at growth rates is quite different from that described for the ratios. In particular, while debt ratios appeared to be fairly constant in most cases, Table 6 shows that the mean and median growth of total debt, total liabilities, and bank debt declined significantly since the mid-1990s. In particular, the mean growth of firm debt exceeded 20% in nominal terms during 1995-1996, averaged more than 10% during 1997-98, but dropped below 1% during 2002-2003. At the same time, the mean growth in assets also declined from ranges above 30% in 1995 to an average of 3.6% in 2002-3, which explains why the debt to asset ratios seem fairly constant during this period. Average firm profitability (measured by return on assets) also appears to have declined (at least until 2003) during this period, going from levels above 8% over the period 1995-96 to 4% by 2001-02. At the same time as discussed earlier, mean and median interest coverage ratios exhibited a U-shaped pattern. It is difficult to disentangle whether the steady decline in debt financing is the result of demand or supply factors. On the one hand, it is possible that lower profitability and fewer 12 growth opportunities (as reflected in the growth of assets) led firms to demand less debt, and having less debt allowed them to increase their repayment capacity. On the other hand, it is feasible that a negative shock in the supply of finance (driven either by macro events or by the perception of higher risks involved in lending to firms) might have impeded firms from financing the kind of asset growth they had experienced in the past. Unfortunately, without more information (such as survey data or individual loan data), we are not able to decisively pinpoint whether the decline is due to supply or demand factors. Table 7 reports the mean and median growth rate of borrowing by firm size and, separately, firm sector, since the mid-1990s. Both mean and median growth rates are larger for large firms. Nevertheless, whether we look at means or medians, we observe that the growth rate of debt declined across firms of all sizes. However, while the means suggests that this decline was larger for small firms, the medians indicate that it was very similar among firms of different sizes. The growth of firm debt declined steadily across all sectors, both in terms of means and medians. Among manufacturing firms, those in the auto components, chemicals, and electrical goods and components sectors were the hardest hit. For these firms, debt growth became negative in recent years. IV- Financing patterns across firm types While the section above focused on the time series behavior of firm financing ratios, this section explores differences across firm characteristics. In particular, we examine differences in financing patterns by firm size, sector, age, ownership type, and export orientation. First, we investigate these differences by analyzing descriptive statistics (means and medians) across firm 13 groups and by conducting univariate tests to determine whether these differences are statistically significant. Second, following the literature on capital structure, we perform regression analysis to understand the determinants of firm financing ratios in India and to investigate whether there are differences in the factors driving firm financing ratios across firm size and other firm characteristics. Firm financing ratios by size Whether we look at means or median ratios, we see that small and medium firms have lower debt to total assets ratios and lower total liabilities to total assets ratios than large firms (see Table 8). The debt and liabilities ratios are monotonously increasing with firm size (i.e., medium size firms have more debt than small firms, large firms have more debt than medium enterprises). Small firms have a mean debt to asset ratio of 0.25, this ratio is 0.39 for medium firms, and 0.43 for the large ones. The differences are large economically (large firms have about 70% more debt relative to assets than small firms) and are statistically significant. They are even more pronounced when we only focus on long-term debt to assets ­ large firms have about 80% more long term debt to assets relative to small firms (i.e., the mean ratio of long-term debt to total debt is 0.16 for the small firms and it is 0.29 for the large firms). Differences in debt ratios are even larger in the medians: the median debt to assets for large firms is 3 times that of the median for small firms and the median long-term debt to assets ratio is 8 times bigger for large firms than it is for small firms. All these results point in one direction ­ small and medium firms have less debt than large firms. When we compare interest coverage ratios across firms of different sizes we get somewhat mixed results: small firms have slightly higher mean coverage ratios than large firms 14 (the difference is significant at 10%), but they have slightly lower median ratios (the difference is significant at 5%). What these results suggest is that while most small firms have similar or lower interest coverage ratios than large firms (as shown by medians), some small firms have quite large interest coverage ratios, which explains the higher mean for the group. We also observe that medium size firms have lower interest coverage ratios than both small and large firms (according to both ­ means and medians). Thus, there is not a very clear pattern in the relationship between the interest coverage ratio and firm size. (We revisit this issue in the regression analysis below). We also consider two indicators of excess indebtedness and difficulties in debt repayment. We find that the proportion of firms with negative net worth and interest coverage ratios below one is higher among medium-sized firms. In this category we observe that 17% of all firms have negative net worth and 32% have interest coverage ratios below one. These indicators suggest that these firms are likely to have excess leverage and have difficulty repaying their debt obligations. Comparing small and large firms we observe that a slightly smaller percentage of small firms have negative net worth (14% of small firms relative to 15% of large firms), however the opposite is true for the percentage of firms with interest coverage below one (28% of small firms and 25% of large firms have interest coverage ratios below one). Although these differences are statistically significant, the magnitude of the differences is rather small, so they do not appear to be economically significant. Similarly, we do not find any large differences in the average use of trade credit across firm sizes (although the medians show a very slight increase of trade credit usage with size). Next, we investigate the composition of debt by firm size by looking at the percentage of total debt (i.e., borrowing) coming from different sources. We find that small and medium firms 15 have a higher proportion of bank debt relative to large firms. This difference is driven by differences in short-term bank debt, since long-term bank debt is very similar across firm sizes. On the other hand, large firms are able to borrow proportionately more from non-bank financial institutions relative to small firms (3 times as much as small firms). Adding up the share of debt from bank and non-bank financial institutions, we see that small firms are able to rely less on formal financial intermediaries (the sum is 69% for large firms and 58% for small firms). At the same time, large firms have more public borrowing (commercial paper and debentures) and more foreign borrowing, however both of these comprise relatively small portions of total borrowing. Perhaps not surprisingly, we find that small firms borrow significantly more from other corporations ­ 3 times as much as large firms do. We expect that much of this financing comes from companies within the same business groups. Thus, it appears that for small firms intra- group corporate financing serves as substitute for financing from formal financial institutions and public debt markets. Finally, we find that small firms have less secured borrowing than medium and large firms. Smaller firms are less likely to have the appropriate assets that could be used as collateral or are less able/willing to pay the fixed costs of registering such collateral (i.e., in relative terms the costs of collateralizing assets tend to be higher for small firms). Firm financing ratios by age We split firms based on their age -calculated as the number of years since the date of incorporation- into three categories: firms that are less than 5 years old, firms between 5 and 10 years old, and firms over 10 years old. The last category of firms could be considered mature firms, while firms in the other two categories are young firms. The results from comparing mean and median financing ratios by age categories are presented in Table 9. 16 We find that mature firms (those over 10 years old) have more debt relative to total assets than younger firms. Debt levels increase gradually with firm age. The difference is economically significant - while firms that are less than 5 years old finance about 30% of their assets with debt, those older than 10 years finance 38% of their assets with debt, which is about a 25% increase in the ratio. The difference is even larger in the medians. A similar pattern is observed in the ratio of long-term debt to assets, but the differences there are smaller in magnitude. Total liabilities are also smaller for younger firms relative to mature firms, but the difference is quite small and only visible in means (not in the medians). There is no clear pattern in trade credit finance across the different age categories. Also, there are no obvious differences in means and medians for interest coverage ratios (all the differences are not significant). This suggests that firms of different age categories have similar repayment capacity, relative to their debt levels. There is also no difference in the proportion of firms with interest coverage ratios below one. While there is slightly higher percentage of firms with negative net worth among older firms, there are no statistically significant differences between very young (less than 5 years) and mature firms. In terms of the composition of borrowing we find, somewhat surprisingly, that older firms borrow more short-term from banks, while younger firms borrow relatively more long- term from banks. We find that younger firms borrow more from other corporations (more likely from group companies) than older firms do, while older firms borrow more from non-bank financial institutions. This pattern, combined with the above results on debt to assets ratios, is indicative of the possibility that younger firms have more difficulty accessing credit from formal financial intermediaries and, therefore, resort to credit from other corporations, which take on a role of informal credit providers. 17 We find that older firms are somewhat more likely to have secured credit, although the differences are not very large in magnitude. We find no significant differences in foreign, public and other borrowing between young and mature firms. Firm financing ratios by ownership type We separate firms into 3 groups depending on their ownership type: privately owned Indian firms, government-owned enterprises, and foreign-owned firms.9 We compare means and medians in financing ratios for foreign firms relative to private Indian firms, and also government firms relative to private domestic firms (see Table 10). Somewhat surprisingly, we find that foreign-owned firms have less debt and total liabilities than both private Indian and government-owned firms. While the mean debt to asset ratio is 0.25 for foreign-owned firms, the mean ratio is 0.38 for private Indian and 0.43 for government firms. This could be because foreign ownership implies foreign equity, possibly as FDI, and therefore less debt. However, foreign firms use more trade credit (0.19 vis-ŕ-vis 0.14 for private Indian firms, and 0.17 for government owned firms). It is possible that the trade credit available to foreign firms comes from their parent firm. Foreign firms have much larger interest coverage ratios (16.69) relative to domestic firms (6.24), so they could possibly support more debt. However, the larger interest coverage ratios could simply be a result of lower debt levels. We find that foreign firms are less likely to have negative net worth (9% of foreign vs. 14% of private domestic firms have negative net worth) and similarly they are less likely to have interest coverage ratios below one. This suggests that foreign firms have healthier debt levels. Not surprisingly, we find that foreign firms have more foreign borrowing and more public 9The results of this section should be cautiously interpreted since our sample of firms in foreign and government ownership categories are very small (only about 6% of firms in our sample are classified as foreign and 4% are considered to be owned by the government). 18 borrowing, relative to private domestic firms. Foreign firms also borrow more from other corporations and less from financial institutions. The mean debt to asset ratio for government-owned firms (which equals 0.43) is higher than that for the private Indian firms (0.38), but the median is lower (the median for private Indian firms is 0.34 while that for government owned firms is 0.29). This suggests that a few government-owned firms have relatively high levels of debt. This is confirmed by the high proportion of government-owned firms with negative net worth (36% relative to 14% in private firms) and also by the proportion of firms with interest coverage ratios below one (41% relative to 27%). Thus, we find that a large proportion of government-owned firms are possibly overleveraged ­ they have more liabilities than assets and many firms have difficulty meeting their interest payments with their earnings. One possible reason for this pattern is that government-owned firms get a significantly larger proportion of loans from the government (28% relative to 3% for private firms). Thus, at least some of the government directed credit goes to firms that have excess leverage and cannot support it with their assets or earnings. This is indicative of inefficient and unsustainable use of government directed credit in a relatively large proportion of government-owned firms. Finally, we also observe that government firms have significantly less secured borrowing than private firms (40% relative to 70% of total borrowing), suggesting that even when these firms get non-government finance, their loans are likely to be based on "government reputation collateral" rather than hard-assets collateral. 19 Firm financing ratios by export orientation We do not observe any large differences among exporting and non-exporting firms (see Table 11).10 If anything, exporters have slightly lower average debt ratios than non-exporters, however, the medians are the same for both groups.11 Similarly, exporters have lower levels of liabilities and long-term debt to assets and they also have lower levels of trade credit (as measured by payables to total assets). The lower debt levels of exporting firms could possibly be explained by their higher earnings, which allow them to use relatively more internal funds and less external financing. This argument is supported by their higher interest coverage ratios, which are much larger than those for non-exporters (even though their debt levels are only slightly lower). Exporters are also less likely to have negative net worth than non-exporters and are less likely to have interest coverage ratios less than one. Together with the above result this suggests that their financial status is healthier. Interestingly, exporters have the same proportion of foreign or public borrowing as non- exporters; they borrow less from corporations and slightly more from banks and other financial institutions. These results may indicate that while exporters have slightly lower debt levels, they have more access to intermediated sources of finance (banks and other financial institutions) and their lower debt levels could be a result of their lower demand for debt. Firm financing ratios by sector First, we split the firms in the sample into two broad industrial groups ­ manufacturing and services and later we break these groups down into finer sector categories. We find that on 10We define as exporters those firms that have sold more than 10% of their products overseas during at least half of the sample period considered 11The average debt ratio for exporters is 0.35 and the mean for non-exporters is 0.38. The median for both is 0.33. 20 average manufacturing firms have more debt (and total liabilities) than service firms (see Table 12).12 We find a similar, although less pronounced pattern for total liabilities and long­term debt to assets ratios. There is no large difference in access to trade credit. We find that service firms have slightly higher interest coverage ratios than manufacturing firms (compare 9.05 to 6.77), which suggest that they could be able to sustain more debt than they currently have. In line with this, we find that there are fewer service firms that have interest coverage ratios below one (22% relative to 28% in manufacturing) and, also, a smaller proportion of service firms that have negative net worth (12% relative to 15%). In terms of the composition of borrowing sources, we find that manufacturing firms use slightly more bank debt (and, in particular, more short-term debt). Service firms borrow more from other corporations and less from other financial institutions, which could be indicative of their weaker access to intermediated finance. Service firms use less secured debt, probably because they have less fixed assets that could be used as collateral. This could also explain their weaker access to intermediated finance and their overall levels of debt. There is no large difference in foreign or public borrowing between manufacturing and services firms. Breaking down industrial groups into finer sector groupings, we find that among manufacturing firms, those in the textiles sector have the highest levels of debt and liabilities to assets (see Table 13). The average debt ratio for firms in the textiles sector is 0.53 and the ratio of total liabilities is 0.87. In turn, the textile industry has the lowest average interest coverage ratio (2.88), most likely as a result of higher debt levels and relatively lower earnings. This industry has a high proportion of firms with interest coverage below one (33%) and the highest 12The differences are quite large ­ the mean ratio of debt to assets is 0.40 for manufacturing firms and only 0.27 for firms in the service sector. Differences are even larger for medians, in which case manufacturing firms have almost twice as much debt as services firms. 21 proportion of firms with negative net worth (23%). Thus, the proportion of possibly overleveraged firms is the highest for the textile industry. On the other side, the computer industry13 has by far the lowest leverage levels. While the sample average debt to assets is about 0.37, it is only 0.11 for firms in the computer industry. So computer firms finance a very small portion of their assets with debt or other liabilities. Their total liabilities on average are only 0.32, while for the rest of the sample it is 0.72. Thus, computer firms are financed mainly with equity finance, primarily coming from their retained earnings. Given their very low debt levels, the interest coverage ratio of computer firms is very high ­ the median is about 8 (relative to the whole sample median of about 2). Clearly, computer firms could support a lot more debt than they currently have. A few reasons could explain these really low levels of debt for firms in the computer industry. On one hand, demand factors could explain it ­ these firms generate sufficient cash flows to finance their operations and expansion and do not need to resort to debt finance. In addition, these firms are likely to have very high growth opportunities (and high upside potential) and theory predicts that firms with high growth opportunities will tend to avoid high levels of debt (see Myers 1977). On the other hand, supply factors could be at play. To the extent that the computer industry is riskier and more volatile than others, financial institutions might not be willing to give them credit. This might explain why computer firms have relatively lower levels of borrowing from financial institutions and relatively higher level of borrowing from other corporations relative to other manufacturing firms. Because of their very low debt levels, there are very few computer firms that have negative net worth (only 2% of all firms) and only a small fraction that has interest coverage 13 Computer industry includes firms that provide consultancy services (for example hardware installation and configuration), software development, data analysis, database management, multimedia services (websites, internet access, etc), and hardware repair. This industry does not include manufacturing of computers. 22 ratios of less than one (13%). Therefore, relative to other firms, very few computer firms have difficulty sustaining their current debt levels. Firm financing ratios by size and other firm characteristics combined Because one of the main goals of this study is to examine differences in debt structures for small, medium, and large firms, in Tables 14 and 15 we report mean and median debt and interest coverage ratios for each size category by age, region, sector, export orientation, and ownership. Table 14 reports total debt to asset ratios, while Table 15 shows the corresponding interest coverage ratios. We focus on these two ratios since they capture firms' debt obligations and ability to meet them, respectively. Aside from reporting the means and medians, these tables also show t-tests for the differences in the means and medians across firm sizes. From Table 14 it is clear that across all regions, mean and median debt to asset ratios are higher for large firms.14 These differences seem to be more pronounced among firms operating in Southern India. Across all age groups, we also observe a positive relationship between debt ratios and firm size. For a given size category, debt ratios increase with age, but differences between mean and median debt ratios by size are largest for firms in the 5 to 10 years age group.15 Differences in these ratios are statistically significant, as shown by the t-tests reported in Table 14. Among manufacturing and service sector firms, mean and median debt to asset ratios increase with size. However, differences in debt to asset ratios are more pronounced for firms in 14While both the mean and median debt to asset ratios for large firms across all regions hover around 0.35-0.45, the mean debt to asset ratio for small firms is around 0.25 across all regions and the corresponding median is between 0.10 and 0.15. At the same time, mean debt ratios for medium firms range from 0.37 to 0.4 and median debt ratios vary from 0.28 to 0.31. The t-tests confirm that, for all regions, differences in debt to asset ratios by firm size are statistically significant. 15For the small firms in this category the mean debt to asset ratio is 0.19, while it is 0.47 for the large firms. Similarly, the median debt to asset ratio is 0.08 among the small firms and 0.46 for the large firms. 23 the manufacturing sector.16 With some few exceptions, the patterns described distinguishing between manufacturing and service firms, remain true if we consider a finer sectoral classification. In other words, in general, we continue to find that firm debt to asset ratios increase with size across most sectors. The pharmaceutical, auto components, and garment and leather sectors are an exception where medium firms exhibit larger and, at times more statistically significant, debt to asset ratios than the large firms in these sectors. For the trade, metal and computer sectors, differences between medium and large firm debt to asset ratios are negatively signed, but are not statistically significant. Both for exporting and non-exporting firms, mean and median debt to asset ratios increase with firm size. Among private Indian firms, mean and median debt to asset ratios increase with firm size and these differences are consistently statically significant.17 Among government owned firms in India, medium firms have the largest debt to asset ratios.18 Among foreign firms, the pattern of debt to asset ratios by firm size varies depending on whether we look at means or median ratios.19 While mean and median debt to asset ratios across all firm characteristics tend to increase with firm size, the pattern is not as clear when we examine interest coverage ratios. For example, there are no consistently significant differences in interest coverage ratios by firm size across 16Thus, in the manufacturing sector the mean (median) debt to asset ratio is 0.44 (0.41) for the large firms and 0.27(0.16) among the small firms. For firms in the service sector, the mean (median) debt to asset ratio is 0.34(0.29) for large firms and 0.20(0.05) for the small firms. 17The mean (median) debt to asset ratio for small private Indian firms is 0.25(0.13), while this figure is 0.39(0.31) for medium firms and 0.44(0.41) for large firms. 18The mean (median) debt to asset ratio for medium firms is 0.74(0.47), while this figure is 0.42(0.3) for large firms and 0.37(0.16) for small firms. Mean t-tests confirm that these differences between medium and large firms are statistically significant. 19The mean debt to asset ratio is 0.25 for small and large firms, while the median ratio is in the order of 0.08 for small and 0.20 for large firms. 24 regions20 This is also the case among young firms (those under 5 years old). For firms in the 5- 10 years old category we find that both mean and median interest coverage ratios for small firms exceed those of large firms. On the other hand, among mature firms (those that are more than 10 years old), large firms have higher interest coverage ratios than medium and small firms. Both for manufacturing and service firms, mean interest coverage ratios for small firms are larger than those for medium size firms, but not different from those of large firms. On the other hand, median interest coverage ratios for small firms are lower than those of large firms, suggesting that most small firms' repayment capacity is below that of large firms. Looking at a finer sectoral classification, we find that mean interest coverage ratios for small firms exceed those of large firms in the garments and leather, chemicals and food sectors. However, the median ratios for small firms in those industries are below those of large firms. Thus, only some small firms in these industries have a repayment capacity that exceeds that observed for large firms. In the pharmaceuticals, electronic goods and equipment, and computer sectors large firms have statistically and economically higher mean and median interest coverage ratios than those observed for small firms. There is no consistent pattern in the differences in interest coverage ratios for firms of different size groups among exporting and non-exporting firms and also among firms with different ownership structure.21 20 For example, we find that mean interest coverage ratios for small firms in Northern and Eastern India are both economically and statistically higher than those of large firms in these regions. Median ratios are also higher for small firms in Eastern India, but lower in Northern India. On the other side, large firms in Southern and Western India exhibit higher interest coverage ratios than those of other firms (but the differences between small and large firms are neither statistically nor economically significant). 21 Thus, among exporting firms, mean and median interest coverage ration are higher for small firms relative to the rest. On the other hand, among non-exporting firms small firms ratios surpass those of medium and large firms, but large firms have better coverage ratios on average than those of medium firms. The t-tests for median interest coverage ratios for non-exporters suggest that large firms surpass both medium and small firms in this dimension. Both mean and median interest coverage ratios for large government owned firms exceeds those of small and 25 In summary, the analysis in this section suggests that while debt to asset ratios tend to increase with firm size across all other firm characteristics, there is no clear pattern when it comes to interest coverage ratios. V. Regression analysis An alternative way to study the recent financing patterns of firms in India is to use regression analysis. In particular, we conduct estimations of the form: Ratioit = 1Zit + 2Xi +3Regioni+4Sectori+5Yeart+it Ratio is one of the two variables ­ the proportion of debt to assets or the interest coverage ratio. These two ratios complement each other in the sense that the first one captures the stock of debt (relative to the stock of firm's assets) and the second one captures the flow of interest payments, relative to the firm's cash-generating capacity. Xi is a vector of firm-specific characteristics such as age, size, exporter status and ownership type; Zit is a vector of time- varying firm characteristics that are linked to existing theories of the determinants of a firm's capital structure (as discussed below). In addition, we include dummy variables to capture firm's region (Northern India is the omitted category), firm's sector (16 sectoral dummies) and year for each observation (year 1997 is the omitted category22). We estimate the model using ordinary least squares under the assumption that the error terms, it, could be correlated across years for each firm (i.e., we use clustering of errors on the firm level). medium firms. Among the private Indian firms, the mean interest coverage ratio of small firms exceeded that of medium and large firms. Among, foreign firms, large firms seem to have higher interest coverage ratios. 22We omit year 1997 as it was the year with major changes in the financial sector. In this way prior years and later years are compared to year 1997. 26 An extensive literature exists that analyzes the determinants of a firm's capital structure ­ i.e., the decision to finance firm operations with debt as opposed to equity (see Harris and Raviv, 1991 and Frank and Goyal, 2003 for a review of this literature). Two very important theories of capital structure are the tradeoff theory and the pecking order theory. According to the tradeoff theory, firms weigh the benefits of debt increases (such as tax deductions and reductions in agency costs between managers and equity holders) against the costs of increased leverage (the deadweight bankruptcy costs and agency problems between firm owners and debt holders) to determine an optimal level of debt. An important benefit of debt financing is the tax savings that it entails, since debt interest payments are usually tax deductible. Thus, other things equal, the tradeoff theory predicts a negative association between taxes and debt to asset ratios (DeAngelo and Masulis, 1980). Agency problems between managers and owners, on the one hand, and debt holders and equity holders, on the other, also affect the costs and benefits of debt financing under the tradeoff theory (see Jensen and Meckling, 1976). Conflicts between managers and equity holders arise because managers have a tendency to deviate firm resources for their own consumption (i.e., they will want to consume more than the optimal level of perquisites). Higher debt levels diminish this tendency because of the increased threat of bankruptcy and because managers of highly levered firms will be less able to consume excessive perquisites, since debt holders are inclined to closely monitor such firms. Firms where the potential for conflict between managers and equity holders is larger will have higher levels of debt financing. Managers of firms with few tangible­ easy to monitor- assets will be more likely to get away with misusing funds and, therefore, these firms will resort to debt financing as a vehicle to mitigate this agency problem. 27 While debt financing can alleviate conflicts between equity holders and managers, it can at the same time create agency problems between equity holders and debt holders. Debt contracts are such that if an investment yields returns above the nominal value of the debt, equity holders capture the surplus, but if the investment fails, because of limited liability, debt holders bear the consequences. Thus, equity holders of levered firms have an incentive to make suboptimal investment decisions (Myers 1977). In this case, theory predicts that firms with fewer growth opportunities will be more highly levered, since in this case there will be less opportunities for asset substitution and the cost from passing up certain projects will be lower. At the same time, firms with a larger proportion of tangible assets that can be used as collateral will also have higher debt levels, since the potential loss of these assets in the event of default helps to reduce the incentives equity holders might have to misuse firm funds and serves to protect debt holders in case of bankruptcy. In the pecking order theory of capital structure, asymmetric information between firm managers and outsiders imply that external financing is relatively more expensive than the use of internal funds (Myers and Majluf, 1984). Thus, firms prefer to finance their operations first with internal funds, then with debt, and lastly with equity. This theory predicts a negative relationship between firm profitability (as a measure of internal funds) and debt financing. Like other empirical studies on the determinants of the capital structure (see Rajan and Zingales 1995, Booth et al. 2001, Klapper et al. 2002), we control for a number of variables that the theories described above predict should affect firm financing ratios. Asset tangibility measures the fraction of total assets that are in the form of plant, machinery and equipment. To the extent that tangible assets help to mitigate conflicts between debt holders and equity holders, we expect this variable to have a positive impact on firm financing ratios. Return on assets is 28 included as a measure of profitability. If the pecking order hypothesis holds, we expect this variable to be negative. We use two proxies for growth opportunities: sales growth and market to book value (the later is only available for publicly listed firms and therefore the sample is reduced by about 50% when this variable is included in the regression). Because agency conflicts between equity holders and debt holders are larger for firms with higher growth opportunities, we expect these variables to have a negative impact on firm financing ratios. We also include measures of the tax rate and of business risk (defined as the standard deviation of return on assets). In theory we expect the tax rate to have a positive impact on firm debt ratios, since firms facing higher tax rates have an incentive to take on more debt, given that debt payments are tax deductible. To the extent that firms with higher volatility of profits are more difficult to monitor by debt holders and have a higher probability of financial distress, we expect business risk to be negative to enter the regressions with a negative sign. Firm-specific characteristics can also help to capture differences across firms in their relative use of external finance and their ability to service their debt. The main focus of this analysis is on firm size. We define two dummy variables: Small and Medium, to differentiate small and medium firms from large firms (the large firms in the omitted category in this case). The dummies are defined the same way as size categories throughout the paper ­ firms with less than 50 million Rs in gross fixed assets are defined as small, while firms with 50-100 million Rs in gross fixed assets are defined as medium. All other firms are considered to be large.23 According to the tradeoff theory of debt, size is expected to be positively correlated with debt ratios, since smaller firms tend to have a higher likelihood of bankruptcy. At the same time, since small and medium firms tend to be more opaque than large firms, the cost of external 23We have also used a continuous variable to represent firm size such as log of gross fixed assets. The continuous size variable avoids arbitrary cutoffs in defining small and medium firms. All the results are similar when we use this continuous variable instead of the size dummies, therefore we only present our result using size dummies. 29 financing will likely be higher for these firms (as a result of larger asymmetric information problems) and therefore, according to the pecking order hypothesis small firms should have lower levels of debt (they will prefer internal financing relatively more than large firms). Finally, our estimations also allow other firm characteristics to influence firms' use of debt and repayment capacity. The age of the firm (measured as the number of years since incorporation) enters the model non-linearly, to allow for possible changes in the age effects over the life of the firm. Young firms are likely to have different behavior from more mature firms, while really old firms might also differ from the firms in their "prime" years in a sense that their size has reached optimal levels and their growth slowed down to a maintenance level. Firm ownership dummies allow us to examine whether there are differences in financing of foreign and government owned firms relative to the private Indian firms, the omitted category in the regressions. We also include a dummy to allow for differences between firms that export a significant portion of their total output and those that do not. Exporters might have more access to debt finance for several reasons: their revenues might be less affected by negative domestic demand shocks, since they sell part of their goods abroad; they might use their contracts with foreign customers as a form of collateral (or guarantee), and they might be able to obtain foreign financing. Our baseline regression results are presented in Table 16 for the debt to assets ratio and Table 20 for the interest coverage ratio. Size The regression results on firm size confirm the patterns observed in the univariate tests discussed earlier. In particular, we find that debt to asset ratios are smaller for small and medium 30 firms relative to those for the large firms. Small firms finance about 12% less of their assets with debt relative to the large firms. Medium firms finance about 6% less of their assets with debt, relative to the large firms. These effects are both statistically and economically significant. Given that the average ratio of debt to assets in the sample is 0.37, small firms have almost 30% less debt than the large firms.24 There are several reasons why small firms will have less debt. As discussed above, small firms may not be able to support more debt because they are more likely to go bankrupt. Second, it could be that small firms demand less debt, because the cost of external financing might be relatively higher for them as a result of the fact that they tend to be more opaque. Unfortunately, our data and methodology does not allow us to firmly distinguish demand from supply factors. However, we are able to shed some light on this issue by contrasting the results on the debt to assets ratio with those on the interest coverage ratio. The latter captures firms' repayment capacity ­ in other words it shows how much interest payments the firm can support from its operating cash flows. Our regression results indicate that small and medium firms have significantly higher interest coverage ratios relative to large firms. This suggests, that on average, small and medium firms have better repayment capacity than large firms, relative to their respective debt levels. Clearly, one reason that interest coverage ratios are higher for smaller firms is their lower debt levels. However, these interest coverage ratios suggest that small firms are able to support larger interest payments with their cash flows. While this is suggestive of the possibility that small and medium firms suffer from credit constraints (i.e. inability to raise more debt) the results are not definite in the sense that we do not fully observe the demand for debt. 24An alternative definition of size using the log of gross fixed assets implies that a one standard deviation change in the log of fixed assets (which is equal to 2.1, and corresponds to about 80 million Rs. ) would imply an increase of 0.37 in the debt to assets ratio, which is equal to the average level of debt to assets. 31 Other firm characteristics We find that firm age is non-linearly related to debt and that the relationship represents an inverted U-shape. Very young firms have lower debt levels and the debt levels steadily increase with firm age until firms reach the age of about 20 years since incorporation, at which point debt levels start to decline gradually. The lower levels of debt for older firms could be explained by the fact that older firms are more likely to have reached their optimal scale and to exhibit a lack of growth opportunities. The results on the interest coverage ratio do not show any significant differences in this variable with regards to age. This suggests that firms of different ages have similar interest payments relative to their cash generating capacity. Contrary to expectations, we do not find any significant differences in debt ratios for exporters and non-exporters. However, the interest coverage ratio of exporters is higher than that of non-exporters, which suggest that they could support higher levels of debt with their cash flow generating capacity. It is plausible that their demand for debt is lower because their exports bring them additional revenues to cover their expenses and investment or they might be better able to access equity markets as a result of their exporting status, possibly because they are more visible and known to investors (especially to foreign investors). We find that foreign owned firms have lower debt levels than private Indian firms, and the difference is about 8% of total assets. However, foreign firms have significantly higher interest coverage ratios, which suggests that they could sustain higher interest payments with their cash flows. One plausible reason for this relationship is the easier access to foreign equity finance among foreign owned firms. Foreign ownership implies that foreigners own some equity 32 in the firm, possibly as a result of FDI. Such inflows of foreign equity could provide sufficient external finance to reduce the level of debt that these firms demand.25 Contrary to the univariate tests where government ownership seemed to be associated with higher levels of debt, the regression results do not point to any consistent significant relationship between government ownership and debt financing (in model 1 we observe a positive relationship while in model 3 we observe a negative sign). Differences between the univariate and regression results might be driven by other firm characteristics that we do not control for in the former, but that are captured by other regressors in the latter.26 Two firm characteristics that are known to be associated with debt levels- asset tangibility and profitability- are strong predictors of debt ratios in our data. Firms with more tangible assets have more debt, as expected. More profitable firms have less debt, which is likely because their need for external debt is lower or, as stressed by the pecking order hypothesis, because debt financing is more expensive relative to financing with internal resources. Interest coverage results parallel those on debt ratios for these variables. The measures of growth opportunities - sales growth and market to book ratio- are not significant predictors of debt levels. However, the market to book value has a positive and significant impact on the interest coverage ratio. Contrary to the predictions of the tradeoff theory, the tax rate is negatively related to debt levels. This could be due to a mechanical relationship between debt levels and taxes ­ firms with more debt get more interest payment deduction and, therefore, pay fewer taxes. Alternatively, as suggested by Booth et al. (2001), the 25However, the same caveat about our methodology would apply here ­ the results are not definite and we cannot exclude the possibility that foreign firms are more credit constrained and have more difficult time in raising debt finance. 26Note that our sample of government and foreign owned firms is very small and therefore any results related to ownership type should be treated with caution. 33 tax rate could simply be another proxy for profitability, which enters negatively in the debt regression. Firms with higher business risk have more debt. This is possibly an indication of their higher demand for debt. However, these firms also have higher interest coverage ratios. This suggests that firms with more risk need to have lower interest payments relative to their average cash flows to ensure that they will be able to meet their interest payments in years with low cash flows. In terms of regional differences, we find that firms in Southern India are more likely to have larger levels of debt to assets, relative to those in Northern India. However the magnitude of the difference is rather small (only 2% of total assets). There are no differences in debt levels between Eastern, Western and Northern India, once we control for other firm and time factors. In terms of interest coverage ratios, the only consistent finding is that firms in Western India seem to have a better repayment capacity. In terms of time patterns, we observe that debt levels were larger in all years relative to 1997, except for 2003. Thus, 1997 corresponds to the lowest debt levels in our sample, which is expected as this year was marked by the collapse of the NBFI sector. The interest coverage ratios show a steady and significant increase in all years since 1997. This suggests that firms have consistently increased their debt-repayment capacity, likely as a result of better cash flows in more recent years. While the cash flows (relative to interest payments) are increasing, the debt levels do not increase as fast, which leaves the firms with extra cash flows with which they could potentially sustain more debt. This pattern could suggest that in recent years the credit expansion was not fast enough to match the increased cash flow generation. (The same caveats about the demand apply to this argument). 34 There are some differences in debt ratios across sectors, which mirror those observed in the univariate tests. Textiles have the highest level of debt and the lowest interest coverage ratios. This could possibly indicate that at least some textile firms have too much leverage (relative to their cash flow generating capacity). Other sectors with high debt levels include Electronic goods and equipment, Metals and Glass and Ceramics. The same industries have correspondingly lower interest coverage ratios. On the other hand, firms in the Service sectors, and especially Computer-related services have significantly lower debt levels and higher interest coverage ratios. These firms might have more difficulty accessing credit or, alternatively, they may have the lowest demand for credit. Sample splits by size Next we study debt patterns for different groups of firms. First we split the sample based on firm size and run our main regression separately for small, medium and large firms. The results are presented in Table 17 for debt to asset ratios and Table 21 for interest coverage ratios. We find that the effect of age is less pronounced for small firms, relative to medium and large firms. Differences between foreign and Indian (be they private or government) firms are only observed among medium and large firms. Medium and large foreign firms seem to have lower debt levels and higher interest coverage ratios than their domestic counterpart. Again, this is likely because of the increased access to equity by foreign owned medium and large firms. Among Indian firms, we find that small and medium sized government firms have more debt but not significantly different interest coverage ratios. Asset tangibility and ROA are significant predictors of leverage for all 3 groups of firms, although the economic magnitude of these 35 variables is increasing with size. In other words these variables have a larger effect on larger firms. Across the regions, we note that the Southern region effect is only significant for medium and large firms. Thus, some large firms in the South are receiving more debt finance than large firms in the North. Finally, we observe that a decline in debt in the last year of the sample (year 2003) is only pronounced for the large firms, while increases in debt in 2000 and 2001 are only significant for small and medium firms.27 Sample splits by other firm characteristics In Tables 18, 19, 22 and 23 we present results for the baseline regression run for different groups of firms and over different time periods. Tables 18 and 19 focus on debt to asset ratios, while tables 22 and 23 present results on interest coverage ratios. In particular, we investigate differences by export orientation, age, region, sector and time period. Since one of the main objectives of this paper is to study differences by firm size, we only discuss the impact of the size dummies across different firm characteristics. We find that the size effect is more pronounced for younger firms. In other words, differences in debt to asset ratios between large and small firms are larger among younger firms. These differences are sizeable in magnitude and are statistically significant. We also see that interest coverage ratios are somewhat larger for smaller firms in the young group relative to the old group (but this difference across age groups is not statistically significant). This could suggest that smaller and younger firms have more difficulty accessing credit, which is in line with previous stylized facts from other countries. 27All these differences in coefficients across size groups are statistically significant, except the regional differences. 36 We also find that size effects are more pronounced in the manufacturing sector relative to the services sector. In other words, among manufacturing firms differences in debt and interest coverage ratios between small and large firms are more significant that those found in the service sector. Regarding firms' export orientation, we find that size has about the same effect on exporters and non-exporters (the difference is very small and is not statistically significant). In terms of regional differences, we find that the size effect is strongest in Southern and Western India and lowest in Northern India. This parallels earlier results on size splits that indicated that large firms have more debt in the South, which generates this larger disparity between large and small firms in this region Finally, we find that the size effect is slightly lower in the later period ­ in other words the difference in the debt finance for small and large firms is reduced in the later period (i.e. 2001-2003 relative to two previous periods). VI. Conclusions This study investigated the financing patterns of Indian firms in recent years. In particular, this paper examined financing patterns over time and explored potential differences across firms with different characteristics (such as sector, age, ownership, export orientation). In particular, the main focus of the paper was to study differences in financing patterns by firm size, comparing small and medium to large firms. In terms of time trends, we found that while debt to asset ratios appear to have been fairly stable during the period 1994-2003, debt growth rates fell considerably during this time. At the same time, firms' repayment capacity, as measured by the interest coverage ratio, exhibited a U- 37 shaped pattern falling during 1997-99 and recovered in recent years. As a share of total debt, bank financing rose during this period, while non-bank debt declined, especially after 1997. In terms of differences in debt financing across firm characteristics, we uncovered a number of interesting patterns. For example, we found that young firms tend to have lower debt ratios than older firms. Since there are no clear differences in these firms' repayment capacities, a possible interpretation of this pattern is that young firms tend to be more constrained by virtue of being more opaque. Also, we showed that foreign firms have less debt than both private and government owned Indian firms. This is consistent with foreign firms having greater access to foreign equity and financing from their parent companies. Across sectors, we found that manufacturing firms have higher debt ratios than service firms. Firms in the textile industry appeared to be overleveraged relative to their repayment capacity. On the other hand, computer firms had the lowest levels of debt and the highest interest coverage ratios. With the data available it is hard to disentangle whether this is driven by demand or supply factors. Finally, yet most importantly, we uncovered very significant and robust differences in firm financing patterns across firm size. In particular, we found that small firms have significantly lower debt to asset ratios and lower growth rates of debt relative to large firms. These differences are very large in magnitude (for example, looking at the medians small firms have up to 3 times less debt than large firms) and are statistically significant. Furthermore, these differences persist even after controlling for firm-specific characteristics (such as age, location, ownership, export status) and factors commonly found to be associated with access to debt such as profitability, asset tangibility, growth and risk. Smaller firms have especially less debt 38 relative to larger firms if they are young (below 10 years since incorporation), if they are in the manufacturing industry and located in Southern India. Relative to large firms, small firms seem to rely less on financing from formal financial institutions (banks and non-banks) and markets (since they have lower levels of public debt) and resort more to borrowing from other (most likely intra-group) corporations. These financing patterns suggest that small firms might tend to be more credit constrained. Partially as a result of lower debt levels, small and medium firms have higher interest coverage ratios, which indicates that their earnings are able to sustain higher debt levels than they currently have. This pattern is also consistent with the possibility that small and medium firms experience difficulties in accessing credit (i.e. they are potentially credit constrained). However, it could also be the case that due to a lack of profitable growth opportunities small and medium firms have lower demand for debt and, therefore, lower debt to asset ratios. Without additional information we cannot decisively pinpoint whether this pattern is due to supply or demand factors. In other words, the evidence presented in this paper is suggestive (but not definite) of stronger credit constraints for smaller firms. 39 References Beck, Thorsten, Ross Levine, and Loayza, Norman (2000a), "Financial Intermediation and Growth: Casuality and Causes," Journal of Monetary Economics 46, August 2000, pp 31- 77. Beck, Thorsten, Ross Levine, and Loayza, Norman (2000a) "Finances and the Sources of Growth," Journal of Financial Economics 58, October 2000, pp. 261-300. Beck, Thorsten, Asli Demirguc-Kunt, and Vojislav Maksimovic, "Financial and Legal Constraints to Firm Growth: Does Firm Size Matter?," Journal of Finance, forthcoming. Bhaduri, Saumitra (2000), "Liberalization and Firms' Choice of Financial Structure in an Emerging Market: The Indian Corporate Sector," Development Policy Review, Vol. 18, 413-434. Bhaduri, Saumitra (2002), "Determinants of Capital Structure Choice: A Study of the Indian Corporate Sector," Applied Financial Economics, 12, 655-665. Booth, Laurence, Varouj Aivazian, Asli Demirguc-Kunt, and Vojislav Maksimovic (2001), "Capital Structures in Developing Countries," Journal of Finance, Vol. 56 (1). Confederation of Indian Industry and World Bank (2002), "Competitiveness of Indian Manufacturing: Results from a Firm-Level Survey," New Dehli. World Bank (2003), India: Sustaining Reform and Reducing Poverty. Washington, D.C. India, Planning Commission (1997), Report of the Abid Hussain Committee on Small Scale Enterprises. New Delhi. India (1999), Report of the S.L. Capoor High Level Credit Committee on SMEs. New Dehli. India (2000), The Interim Report of the S.P. Gupta Study Group on Development of Small Enterprises. New Dehli. 40 Jensen, Michael, and William Meckling (1976), "Theory of the FirmL Managerial Behavior, Agency Costs, and Capital Structure," Journal of Financial Economics 3, 305-360. Klapper, Leora, Virginia Sarria-Allende, and Victor Sulla (2002), "Small- and Medium-Size Enterprise Financing in Eastern Europe," World Bank Policy Research Working Paper Series 2933. Myers, Stewart (1977), "Determinants of Corporate Borrowing," Journal of Financial Economics 5, 147-175. Myers, Stewart and Nicholas Majluf (1984), "Corporate Financing and Investment Decisions When Firms Have Information that Investors Do Not Have," Journal of Financial Economics 13, 187-221. Rajan, Raghuram and Luigi Zingales (1995), "What Do We Know about Capital Structure? Some Evidence from International Data," Journal of Finance, vol 50(5), 1421-146. 41 Table 1. Distribution of firms in the study Number of firms Proportion of firms Small Medium Large Total Small Medium Large Total Panel A. Distribution of firms Total for all sample 1,563 873 3,345 5,781 27.0% 15.1% 57.9% 100.0% Eastern India 206 106 386 698 3.7% 1.9% 7.0% 12.7% Northern India 240 152 693 1,085 4.4% 2.8% 12.6% 19.7% Region Southern India 308 218 814 1,340 5.6% 4.0% 14.8% 24.3% Western India 734 334 1,314 2,382 13.3% 6.1% 23.9% 43.3% 0-5 50 20 98 168 0.9% 0.3% 1.7% 2.9% Age (in years) 5-10 367 185 433 985 6.4% 3.2% 7.5% 17.1% 10+ 1,140 668 2,805 4,613 19.8% 11.6% 48.6% 80.0% Industry Manufacturing 1,021 731 2,948 4,700 17.7% 12.6% 51.0% 81.3% Type Services 542 142 397 1,081 9.4% 2.5% 6.9% 18.7% Food 94 89 332 515 1.6% 1.5% 5.7% 8.9% Textiles 60 65 329 454 1.0% 1.1% 5.7% 7.9% Garments & Leather 40 18 47 105 0.7% 0.3% 0.8% 1.8% Chemicals 177 157 622 956 3.1% 2.7% 10.8% 16.5% Pharmaceuticals 82 45 150 277 1.4% 0.8% 2.6% 4.8% Electronic good & equipment 20 14 80 114 0.3% 0.2% 1.4% 2.0% Electrical goods 37 18 112 167 0.6% 0.3% 1.9% 2.9% Auto components 18 29 188 235 0.3% 0.5% 3.3% 4.1% Sector Machine tools 98 60 212 370 1.7% 1.0% 3.7% 6.4% Metals 90 91 321 502 1.6% 1.6% 5.6% 8.7% Trade 334 64 117 515 5.8% 1.1% 2.0% 8.9% Computer 150 53 86 289 2.6% 0.9% 1.5% 5.0% Services 60 15 42 117 1.0% 0.3% 0.7% 2.0% Paper & printing 27 24 118 169 0.5% 0.4% 2.0% 2.9% Glass & Ceramics 21 21 126 168 0.4% 0.4% 2.2% 2.9% Construction 65 20 66 151 1.1% 0.3% 1.1% 2.6% Others* 190 90 397 677 3.3% 1.6% 6.9% 11.7% Export Non-exporter 1,266 688 2,388 4,342 22.4% 12.2% 42.3% 77.0% Orientation Exporter 219 166 914 1,299 3.9% 2.9% 16.2% 23.0% Private Indian 1,466 832 2,865 5,163 25.5% 14.5% 49.8% 89.7% Ownership Foreign Private 63 25 290 378 1.1% 0.4% 5.0% 6.6% Government 32 15 165 212 0.6% 0.3% 2.9% 3.7% Panel B. Distribution of observations by year 1994 1,142 530 1,668 3,340 2.8% 1.3% 4.1% 8.2% 1995 1,429 661 2,004 4,094 3.5% 1.6% 4.9% 10.0% 1996 1,353 694 2,278 4,325 3.3% 1.7% 5.6% 10.6% 1997 1,258 675 2,377 4,310 3.1% 1.6% 5.8% 10.5% 1998 1,181 641 2,448 4,270 2.9% 1.6% 6.0% 10.4% Year 1999 1,295 634 2,625 4,554 3.2% 1.5% 6.4% 11.1% 2000 1,338 707 2,799 4,844 3.3% 1.7% 6.8% 11.8% 2001 1,209 614 2,802 4,625 3.0% 1.5% 6.8% 11.3% 2002 905 485 2,458 3,848 2.2% 1.2% 6.0% 9.4% 2003 584 319 1,836 2,739 1.4% 0.8% 4.5% 6.7% Total 11,694 5,960 23,295 40,949 28.6% 14.6% 56.9% 100.0% Breakdown by firm size is based on fixed assets. Small firms are those with fixed assets less than 5 Cr. Rs., medium firms are those with fixed assets between 5-10 Cr. Rs. and large firms are those with fixed assets above 10 Cr. Rs. * This category includes mining and quarrying; extraction of crude petroleum and natural gas; manufacture of coke, refined petroleum products and nuclear fuel; electricity, gas, stream and hot water supply; manufacture of furniture and manufacturing N.E.C.; Hotels and restaurants; Diversified; etc. e Max 178 2.46 0.88 4.66 2.00 82,536 186.96 100% 100% 100% 100% 100% 100% 100% 100% 100% 0.39 59.0 .752 14.76 0.74 0.83 th is p95 p95 411 69 0.87 0.41 1.46 24.43 0.67 100% 99% 56% 0% 78% 62% 32% 82% 100% 0.22 18.0 .860 4.13 0.44 0.25 tile, percen p75 49 32 0.49 0.19 0.80 .744 0.34 73% 58% 10% 0% 39% 4% 0% 18% 97% 0.12 75.0 .270 75th 1.07 0.19 0.10 e th of e p50 14 19 0.32 0.10 0.64 .422 0.18 44% 32% 0% 0% 7% 0% 0% 3% 83% 0.07 93.0 .090 0.48 0.00 0.06 luav e th is 4 p25 12 0.14 0.04 0.46 .211 0.05 19% 9% 0% 0% 0% 0% 0% 0% 57% 0.01 22.0 .070- 0.24 0.00 0.03 p75,eulav 5p 0 8 0.00 0.00 0.15 .291- 0.00 0% 0% 0% 0% 0% 0% 0% 0% 0% 0.12- 30.0 diane 0.66- 0.09 0.00 0.01 m e th is n 0 1 Mi 0.00 0.00 0.00 14.67- 0.00 0% 0% 0% 0% 0% 0% 0% 0% 0% .510- 00.0 .342- 00 p50 0. 0.00 0.00 tile, .cte, se dra n atioi 20 1,358 0.31 0.13 0.47 15.75 0.25 33% 41% 33% 31% 20% 7% 27% 21% 13% 25% 32% 0.11 32.0 percen 0.50 1.68 0.16 0.09 hacrup Stand Dev 25th er e hi th Mean 162 26 0.36 0.14 0.69 6.12 0.24 12% 22% 47% 37% 10% 1% 22% 9% 5% 15% 71% 0.06 04.0 of e s,reto 0.10 1.05 0.11 0.08 luav morp e fo sn th mo fr erb atiov m 1994-2003 ser 40,949 40,908 40,503 40,503 40,503 35,214 40,503 40,949 35,214 38,122 38,118 38,113 38,122 38,122 38,122 38,122 38,122 34,336 40,094 205,04 is ngi p25 30,313 17,881 37,842 40,474 Nu Ob tile, e. luav esrut worrob, percen nte mu deben statistics, 1 5th e an th mixa )htro d mnr an ve th m w go : of e e th less etn( mo fr descriptive egarev from luav is paperla ngi sn e ityu rcie ngi tio th Max eq le is d e m )s assets co com stitu an p5 com worrob ratios: Variab R.r tal 1 ste in to s ngi snoi e, ter w tile, tivagen se C n (i assets mrift in luav esdu as ith clni udlcni nancingi stes assets en w borro k k 2 corporat 3 mu cenrep os tal e totbed ialcnanif s ngi gni lesb gni ) assets tal lv ngi to ban gni F as so mini rati to rage m ban total morf omrf ngi ria stes ht59 denifed 2. dexif tal to as to book m e earsy to ter-gn ht ist of term worrob g in gni borrow ilitybig e borrow n borrow Val on ot e th en s (i tot osr eg ggregate Table G A A Deb eslbayaP covt lo infot mriffot knab worrobr ilities ort- tan owrg ci red et thfo lv bl abi reset ci worrobr rat is talo L In T cenreP cenreP term-gno talo Sh L gnier worro e bl het ntro het Percent T Fo B Borrow Pu O ecuS rnute sset Co R A selaS xa Mark T Risk Min luav sonI Pu O 1 2 3 2003 0.30 0.11 46.0 0.16 2.85 53% 34% 0% 0% %0 0% 0% 3% 78% 2002 0.32 0.11 56.0 0.17 2.29 50% 33% 0% 0% %0 0% 0% 3% 79% 2001 0.33 0.11 56.0 0.18 2.14 45% 32% 0% 0% %1 0% 0% 3% 80% 2000 0.33 0.10 46.0 0.18 2.14 44% 33% 0% 0% %3 0% 0% 3% 82% 1999 0.34 0.10 56.0 0.19 1.99 44% 32% 0% 0% %8 0% 0% 3% 83% Median 1998 0.34 0.10 46.0 0.19 2.06 43% 33% 0% 0.0% %11 0% 0% 3% 84% 10 1997 0.33 0. 26.0 0.18 2.17 43% 31% 0% 0.0% %31 0% 0% 3% 85% 1996 0.31 0.10 16.0 0.18 2.79 43% 32% 0% 0.0% %31 0% 0% 3% 86% 1995 0.31 0.10 46.0 0.17 3.36 41% 31% 0% 0.0% %51 0% 0% 3% 86% 1994 0.35 0.11 96.0 0.20 2.93 38% 31% 0% 0% %81 0% 0% 3% 84% 2003 0.36 0.14 37.0 0.23 10.62 15% 21% 52% 38% 14% 1.2% %31 9% 4% 17% 66% 2002 0.38 0.15 47.0 0.25 8.39 17% 27% 50% 37% 13% 1.1% %41 10% 5% 17% 66% 39 2001 0. 0.15 57.0 0.25 7.27 17% 29% 47% 36% 11% 0.9% %81 9% 5% 17% 68% .cte, 2000 0.38 0.15 37.0 se 9% 0.25 6.96 16% 27% 47% 37% 10% 0. %02 9% 5% 16% 70% 1999 0.37 0.14 17.0 hacrup 25 0. 5.94 14% 29% 47% 37% 10% 1.2% %22 9% 5% er 14% 71% hi Mean 1998 0.37 0.14 86.0 0.24 5.21 11% 26% 46% 37% 9% 1.3% %32 8% 5% 14% 73% s,reto 1997 0.34 0.13 56.0 0.23 4.70 8% 22% 45% 36% 9% 0.9% %42 9% 4% 14% 74% morp 1996 0.33 0.14 46.0 mo 0.22 5.68 7% 14% 46% 36% 9% 0.6% %52 fr 9% 4% 14% 74% ngi 1995 0.33 0.14 76.0 0.22 6.79 8% 11% 45% 36% 8% 0.6% %62 8% 5% 14% 74% 1994 0.36 0.15 27.0 0.24 5.92 9% 12% 43% 36% 7% 0.7% %72 esrut worrob, 6% 6% 13% 74% nte 1 anht )htro deben d mnr an ve w go ssel : etn( mo fr year e from paperla ityu by stes rage ngi sn rcie ngi eq tio e m le ratios aslatot covt com stitu 1 s ngi in snoi tivagen com worrob se Variab stes ot assets mrift resetni w esdu as clni eary os stes hti udlcni en w borro k k 2 debt 3 tal e ialcnanif corporat lv s gni nancingi to ban gni gni gni ngi F by rati n 3. aslatot aslatot so ban ngi denifed ot to ist borrow borrow borrow dow ot es ilities bl eakr ggregate bte bla abi rmet-gnollato rage total morf omrf k g covt infot rmif oft of rmet- in gni borrow en ng lv ci worrobr reset ci worrobr red Table B A D Pay L T In cenreP banlato orthS rmet-gno L Percen Percent T reioF worro bl het het B Borrow Pu O ecuS sonI Pu O 1 2 3 6 e 2003 .240 0.35 .40 22. 33. .320 .320 52. 0.25 0.29 0.39 0.02 .360 .440 31. 22. 0.14 0.10 .090 53. 0.26 rgal d 9 2002 .230 0.37 .40 702. 503. .320 .320 402. 0.31 0.31 0.39 0.03 .370 .450 601. 801. 0.15 0.08 .090 703. 0.27 an.s 0 R.r 2001 .240 0.40 .50 403. 803. .320 .370 702. 0.33 0.35 0.41 0.03 .380 .470 701. 901. 0.16 0.07 .120 903. 0.31 C 9 10 2000 .240 0.37 .40 403. 803. .300 .390 902. 0.32 0.34 0.41 0.03 .380 .460 701. 701. 0.16 0.05 .130 004. 5- 0.32 een 0 w 1999 .240 0.35 .50 902. 903. .340 .360 003. 0.33 0.37 0.41 0.06 .400 .470 501. 901. 0.16 0.16 .160 204. 0.31 bet 7 Median 1998 .230 0.37 .40 303. 803. .330 .370 35 103. 10 0. 0.35 0.39 0. .400 .460 16 701. 801. 14 0. 0. .150 104. stes 0.34 as 5 1997 .230 0.38 .40 003. 703. .290 .380 602. 0.32 0.36 0.38 0.10 .350 .420 601. 801. 0.16 0.16 .150 004. 0.32 dexif 2 50 6 80 hti 1996 .230 0.38 .40 303. .30 .20 .320 602. 0.31 0.33 0.35 0.10 .300 .380 501. 0.16 .10 0.17 .150 803. w 0.32 e 1 osht 1995 .220 0.33 .40 103. .350 .240 .340 502. 0.31 0.34 0.34 0.11 .310 .400 501. 0.16 .190 0.15 .170 803. 0.31 are 0 s 1994 .240 0.34 .40 303. .390 .300 .370 802. 0.35 0.39 0.38 0.16 .370 .460 002. 0.20 .190 0.23 .230 004. 0.35 rmif 2 40 90 mu 2003 .370 0.43 .50 103. .390 .320 .400 103. 0.30 0.32 0.45 0.11 .380 .470 0.32 .20 .300 0.17 .240 0.37 .30 die 5 m 34 17 2002 0.4 0.46 .50 803. .420 .350 .370 103. 0. 0.36 0.46 0.10 .400 .490 0.34 .240 .250 0. .230 0.37 .420 .,s 5 R.r C 2001 0.41 0.47 .50 304. .440 .360 .410 303. 0.35 0.40 0.48 0.09 .430 .480 0.36 .270 .260 0.18 .260 0.41 .430 5 4 1 2000 0.41 0.44 .50 504. .430 .350 .40 303. 0.36 0.40 0.47 0.09 .420 .480 0.35 .250 .250 0.19 .250 0.39 .440 anht ssel 2 1999 .40 0.40 .50 004. .410 .360 .390 403. 0.35 0.40 0.45 0.13 .400 .470 0.37 .240 .240 0.23 .240 0.37 .430 stes as Mean 8 4 40 1998 .390 0.40 .40 703. .400 .30 .420 0.37 0.39 .30 0.42 0.16 .420 .450 0.37 .220 .230 0.27 .230 0.37 .420 dexif 5 50 1997 .370 0.38 .40 .30 .380 .300 .420 0.34 0.38 .300 0.40 0.15 .360 .440 0.36 .220 .240 0.25 .220 0.35 .410 hti w e 2 1996 .350 0.35 .40 .350 .360 .270 .380 0.32 0.38 .290 0.38 0.17 .330 .410 0.34 .220 .230 0.24 .220 0.34 .390 osht size 0 are 1995 .350 0.34 .40 .320 .360 .250 .400 0.31 0.39 .290 0.37 0.18 .340 .420 0.33 .210 .230 0.24 .230 0.33 .390 s and 0 rmifl 1994 .380 0.34 .40 .330 .400 .300 .410 0.33 0.44 .320 0.39 0.21 .390 .490 0.36 .230 .230 0.28 .280 0.37 .410 al sector, t mS.stes .s R.r en as C year, pmi by dexif 10 e equ re & on abov st g ics de ratios leath en tin tses g bas as inr & ticals oodsg oodsg pon cesi ts ci tools prin si ceram asset cal e tionc en icals aceu ron ri com in & & dexif to actufu zeis oodF tilesxe sla ertup s emh T Garm C armhP ect ect otu mo El El A Mach Met C peraP cesi ader trusno rvesreht mu e ith Glas Man rveS all T C O rga rmif w e Sm Medi L Debt by os n th 4. dow are Size s Sector Table Break rmif 9 e 2003 2.77 1.77 .72 56. 60. .143 .372 .552 .484 70. 84. 2.37 .642 .342 07. 15. 2.69 .253 2.50 2.62 2.92 rgal d 4 2002 2.25 1.53 .71 436. 534. .802 .372 .102 .133 437. 186. 1.93 .562 .981 521. 627. 2.33 .763 2.19 2.18 2.31 an.s 4 R.r 2001 2.09 1.55 .71 425. 120. .392 .202 .402 .622 522. 66.2 1.74 .382 .811 020. 220. 2.40 .094 2.02 2.10 2.18 C 7 10 2000 2.08 1.78 .71 324. 529. .672 .152 .132 .832 024. 21.4 1.85 .122 .701 820. 537. 2.45 .163 5- 2.21 2.25 2.09 een 3 w 1999 1.95 1.92 .51 021. 419. .062 .851 .192 .532 322. 815. 1.65 .771 .671 020. 124. 2.20 .302 2.00 1.83 1.99 bet 3 Median 2 1998 1.92 .81 728. 410. .032 .701 .132 .762 322. 064. 1.75 .921 .641 33 522. 328. 2. .961 stes 2.00 1.70 2.13 as 5 1997 .152 1.75 .81 912. 422. .182 .641 .442 .193 225. 854. 2.00 .272 .142 422. 329. 2.26 .921 2.07 1.94 2.25 dexif 9 12 82 hti 1996 .692 1.98 .02 411. .72 .443 .042 .952 .203 522. 040. 2.76 .813 .872 522. 3.27 .43 .053 w 3.00 2.67 2.77 e 3 osht 1995 .243 2.60 .72 426. .273 .354 .692 .503 .063 737. 555. 3.00 .953 .862 231. 4.07 .883 .173 3.71 3.37 3.18 are 7 s 1994 .842 2.93 .72 030. .982 .583 .772 .862 .582 937. 371. 2.59 .083 .212 543. 3.39 .483 .034 3.29 3.10 2.72 rmif 4 43 25 mu 2003 .019 4.37 .85 043. .219 12.1 .375 .188 72.1 13.3 7.77 44.12 .887 .386 8.12 .08 .585 12.7 10. 9.01 8.93 die 6 m 2002 .297 3.27 .23 895. .137 10.8 .497 513. 10.0 9.35 7.22 17.5 .858 .914 7.16 .196 .115 15.1 .587 .,s 8.47 7.32 9 R.r C 2001 .186 3.15 .12 444. .844 .049 .247 099. 6.36 7.62 4.69 27.7 .585 .104 6.63 .395 .037 11.3 .147 5.39 6.45 5 7 4 2000 .825 4.02 .52 449. .414 .567 .27 282. 5.57 5.89 4.23 30.4 .154 .203 8.62 .997 .217 14.1 .878 anht 4.96 5.67 ssel 9 1999 .075 5.90 .22 663. .454 .005 .094 371. 5.71 5.96 3.02 15.3 .745 .452 5.85 .925 .184 8.58 .117 4.48 4.80 stes as Mean 9 6 9 17 1998 .34 5.36 .02 276. .873 .94 .672 4.16 5.89 .25 4.06 12.5 .774 .622 7.19 .746 .188 7.83 .166 3.56 4.91 dexif 9 74 ze 1997 .374 3.36 .52 .23 .743 .975 .862 3.43 6.44 .655 3.57 15.0 .075 .354 5.39 .924 .286 6.24 .675 hti 3.73 4.58 w si e 9 d 1996 .035 4.85 .92 .416 .044 .436 .692 4.49 5.57 .956 5.25 13.1 .946 .114 7.74 .947 .077 7.91 .736 5.06 5.37 osht an 2 are 1995 .096 5.49 .84 .469 .095 .199 .793 5.93 6.46 .098 4.87 20.2 .106 .883 9.19 .659 .857 9.03 .897 s 7.29 6.05 9 sector, rmifl 1994 .245 5.59 .24 .199 .155 .268 .433 6.00 4.62 .765 4.03 11.4 .235 .223 7.64 .577 .327 8.71 .427 al 6.03 5.05 year, yb t mS.stes s R.r en as C os pmi dexif 10 e rati equ re & on abov st g ics de leath en tin tses g bas as inr & ticals oodsg oodsg pon cesi coverage ts ci tools prin si ceram cal e en icals aceu & & actufu ron ri com in tilesxe sla ertup tionc dexif s zeis emh ect ect cesi mu ith terest oodF T Garm C armhP otu mo El El A Mach Met C peraP ader trusno rvesreht Glas Man rveS all T C O egr rmif w e Sm Medi La In by os n th 5. elba dow are Size s Sector T Break rmif 2003 0.0% 3.3% 0.3% 3.9% 1.7% 9.1% 5.8% 2.83 2002 0.0% 5.8% 3.8% 2.7% 1.3% 0.6% 5.2% 2.29 2001 1.9% 4.7% 3.0% 2.1% 2.0% 1.4% 5.1% 2.14 2000 3.7% 6.3% 5.1% 7.0% 3.5% 10.5% 5.8% 2.14 0% 1999 3.5% 5.8% 4.1% 5.8% 2.6% 2. 5.7% 1.99 Median 1998 7.8% 8.1% 8.5% 5.3% 5.1% 2.8% 6.4% 2.06 1997 8.2% 8.7% 8.0% 8.3% 7.3% 4.7% 7.3% 2.17 1996 15.6% 16.0% 19.7% 19.9% 16.8% 21.4% 8.9% 2.79 1995 17.4% 19.3% 21.0% 24.8% 24.8% 30.3% 9.5% 3.36 2003 0.0% 5.6% 3.6% 3.9% 3.7% 7.5% 5.0% 9.13 2003 3.7% 2002 0.6% 7.5% 7.5% 3.9% 3.6% 0.4%- 4.1% 7.50 2002 4.3% 2001 4.7% 6.5% 6.5% 3.2% 5.5% .1%0- 4.0% 6.44 2001 3.6% 2000 6.2% 8.0% 8.9% 10.6% 7.4% 11.9% 4.9% 6.24 eary 2000 3.9% e th Mean 1999 5.8% 7.4% 7.3% 6.5% 5.3% 0.0% 4.6% 5.26 of 1999 4.6% d En 1998 10.7% 10.1% 14.0% 6.5% 7.8% 2.2% 5.5% 4.98 1998 12.4% 1997 12.8% 11.2% 11.3% 11.1% 10.7% 0.2% 6.5% 4.70 1997 6.9% 1995-2003 s.e 1996 24.1% 21.5% 27.8% 27.0% 23.9% 21.4% 8.4% 5.68 1996 8.6% tax d an 1995 26.6% 27.9% 29.7% 36.1% 33.8% 37.1% 9.2% 6.79 1995 9.7% ste nancing,if ter in rmif fo zeis erofeb te.ar of d Rate th an th th wo gni n * owr eary **e itsforp worg gni e to G by n Grlani atiolfnI 6. m dow worrob borrow ilities k assets* ragevo no srefer thton, liab No eakr talo talo banlato se assets Ct blay * T nr TI se talo reset BP rep Table B T T T Pa T PBI Retu In * * s 2003 0.0% 0.0% .1%0- .2%0- 2.0% .1%1- .0%0 1.5%- 1.8% 0.1%- .8%5- 3.1%- 0.6%- .3%0- 3.2% 0.1% .6%0- 0.0% .3%0- 4.7% .6%1- rmif e 2002 0.0% 0.9% 0.4% 0.0% 4.7% 0.2% .6%0- .7%1- 1.9% .6%3- 0.3% .4%0- .0%2- 2.6% 0.0% 0.2% 0.0% 1.4% 0.4% 5.6% 0.0% rgal d an.s 2001 0.0% 0.4% 3.1% 2.4% 4.0% 4.3% 5.5% 0.4% 4.0% 1.4% 3.6% 3.2% 2.5% 1.4% 6.0% 1.5% 1.3% 0.6% 0.7% 5.1% 2.5%- R.r C 2000 2.0% 4.5% 4.0% 3.7% 10.9% 1.9% 3.1% 3.7% 1.9% 1.2% 3.1% 4.0% 1.3% 5.4% 0.0% 5.6% 2.7% 4.1% 4.2% 5.2% .3%0 10 5- een 0% 0% 7% 1999 1.7% 1.2% 4.3% 3.7% 4.7% 4.6% 4.2% 3.3% 9. 4.5% 1.0% 4.8% 3.0% 4.7% 0. 3. 0.2% 1.9% 2.6% 2.0% .4%1 w Median bet 1998 3.6% 6.5% 9.1% 8.4% 4.1% 9.8% 11.7% 7.6% 14.5% 3.0% 9.1% 6.1% 6.6% 10.1% 13.2% 12.8% 8.6% 2.3% 1.0% 9.5% .1%0- stes as 4% dexif 1997 5.2% 6. 9.2% 8.0% 4.7% 4.9% 7.8% 7.6% 13.4% 4.1% 7.9% 9.9% 10.3% 10.3% 21.2% 12.6% 8.4% 7.6% 5.0% 10.0% 5.3% hti w e 1996 10.5% 17.2% 16.2% 16.0% 16.4% 16.2% 34.1% 16.0% 26.2% 7.2% 20.7% 16.3% 16.2% 15.6% 8.0% 18.9% 4.2% 17.4% 17.0% 29.5% 6.8% osht are s 1995 11.7% 18.2% 18.9% 17.5% 27.0% 26.3% 30.9% 17.7% 27.9% 11.0% 14.3% 9.4% 8.9% 15.9% 21.5% 17.4% 6.6% 19.2% 19.2% 24.7% 8.9% rmif 7% 2003 3.4%- 0.0% 0. 0.5%- 6.1% 2.2%- .0%2- .8%4- .5%3 3.7%- .9%0- 5.3%- 1.8%- .5%1- mu 12.6% 11.1% 2.4%- 0.8% .8%3- 9.5% 2.1% 2003 3.7% die m 2002 .9%3- 0.5% 1.7% .5%0- 7.6% .5%1- 1.4% 5.1%- .8%1- .9%3- 2.0% .4%3- .6%2- 1.1% 12.6% .2%1- .1%0 .,s 5.0% 4.6% 4.5% 7.5% 2002 4.3% R.r C 2001 1.9% 1.7% 6.2% 4.6% 7.6% 6.6% 5.0% .3%0- 5 4.1% 4.9% 11.3% 5.4% 0.9% 1.4% 15.6% 11.4% .0%9 0% 6.4% 5.1% 8.4% 9. 2001 3.6% anht 2000 5.0% 8.2% 6.1% 6.2% 18.1% 2.4% 10.7% 5.4% 3.4% 2.3% 3.9% 5.9% 2.6% 6.1% .3%3 7.1% .3%5 7.3% 9.0% 6.5% 0.4% eary 2000 3.9% ssel e stes Mean 1999 7.3% 2.1% 6.1% 5.1% 7.5% 7.5% 1.9% 4.6% 8.9% 3.7% 2.1% 8.3% 3.5% 6.2% .6%9- 6% 7.5% .1%1- th as 7.6% 4.8% 8.8% 19.3% of 1999 4. d dexif En hti 1998 5.1% 12.1% 12.4% 11.4% 5.2% 13.2% 9.0% 12.3% 17.6% 4.8% 8.3% 5.7% 8.5% 14.4% 20.9% 16.9% 12.9% 6.0% 1.1% 17.7% 7.1% 1998 12.4% w e osht 1997 10.5% 11.5% 14.0% 13.0% 7.7% 10.2% 0.1% 10.6% 20.4% 10.1% 14.1% 11.6% 20.2% 15.5% 29.0% 13.7% 14.5% 11.4% 6.1% 24.8% 11.4% 1997 6.9% sector are s and 1996 21.0% 27.7% 24.3% 24.5% 22.9% 24.3% 38.9% 25.3% 31.4% 17.5% 26.4% 26.0% 23.0% 24.3% 28.4% 26.0% 11.1% 26.6% 25.8% 25.8% 32.1% 1996 8.6% rmifl al size 1995 9.7% by 19.8% 29.0% 28.4% 26.4% 36.0% 31.4% 27.3% 27.2% 34.2% 19.0% 23.3% 13.8% 20.7% 25.0% 35.0% 30.6% 14.1% 30.3% 29.9% 36.8% 20.2% 1995 mS.stes ngi t as .s en pmi dexif R.r C equ on 10 borrow re & de e ts g ics total zeis leath en tin bas abov n tear & ices si of d gni ticals oodsg oodsg ts ci pon tools r inrp ceram tses th an th urtc cal e zeis as en icals aceu tiocu ron ri com in utep & s owr eary l mu e al tilesxe worg emh ect ect tou m ices eda trsn rvesre rmif rga dexif e T Garm C El El A Mach Metals Co Glas Tr Co Oth G by n mS nufaa oodF armhP &repaP by Medi L M Serv ith n w 7. n e thton, dow dow os se eakr Size Sector Table B atiolfnI th rep Break are * .sv s egr all *** *** *** *** ** *** S N *** *** *** rmif La e t. t Sm tes- m T egr rgal d La *** *** *** *** *** *** *** *** S ** N an.s Mediu vs. Median .sv R.r canifingiston m C are all *** *** *** *** S N *** *** *** *** 10 5- ces Sm Mediu een renef Median e w arg bet e L 0.39 0.11 86.0 dif 0.24 .212 43% 31% 0% 0% %51 0% 0% %3 83% stes th at as th m ediu 0.31 0.10 06.0 0.13 .002 60% 43% 0% 0% %0 0% 0% %4 83% dexif s ean M hti m w NS. all ely Sm 0.13 0.08 84.0 e 0.04 .072 54% 33% 0% 0% %0 0% 0% %2 72% osht .sv are s egr all *** *** *** *** * ** *** *** *** *** *** *** *** *** *** *** rmif tivcepser La t Sm mu %01 tes- m egr die T La *** dna, ** ** *** *** *** *** *** *** S N *** *** *** *** *** *** m .,s Mediu vs. Mean .sv R.r %5, m C all *** ** *** *** *** *** *** *** *** *** S N *** *** * *** *** 5 Sm Mediu anht Mean e ssel %1tatnacifingis .cte, se hacrup er hi arg L 0.43 0.14 77.0 0.29 7.19 15% %52 45% 34% 11% 1.5% %42 6% 7% %41 73% stes as erataht s,reto m ediu 0.39 0.14 47.0 0.23 5.91 17% %23 55% 44% 12% 0.4% %51 10% 1% %61 dexif 71% hti M w e secnereffid morp mo fr all Sm 0.25 0.15 16.0 0.16 7.75 14% %82 50% 40% 10% 0.2% %8 18% 1% %12 osht 58% are s otrefer ngi rmifl * esrut worrob, al dna nte 1 an size th mS.stes **, )htro deben d mnr an ve : as ***,.s w go mo rmif less etn( fr sn C ityu by le stes egarev from dexif R.r paperla ngi rcie ngi eq tio on 01 m co de e ratios Variab aslatot com bas worrob 1 ste stitu s ngi in snoi evoba ter si tivagen com se stes ot assets mrift w in zeis stessa esdu as clni os stes ith udlcni en w borro k k 2 debt 3 tal e ialcnanif corporat lv s gni nancingi to gni gni rmif dexif gni ngi F rati 8. aslatot aslatot so ngi denifed by ot to n ist borrow borrow w borrow ot es en bla ilities ci dow ci worrobr red lv bl ggregate bte abi rmet-gnollato rage total morf omrf hti k g covt infot mriffot ban ban of rmet- in gni ng borrow reset worrobr Table A D Pay L T In cenreP cenreP banlato orthS rmet-gno esoht L Percent T reioF worro bl het het B Borrow Pu O ecuS Break era sonI Pu O 1 2 3 earsy *** S N NS *** S e N * *** *** *** S N th t at >10 th tes- s T 10 5- earsy *** *** *** *** S S N N *** *** *** *** ean m Median NS. earsy *** *** ** S S N N * *** *** *** ely 5< tiv ecp Median res earsy 0.33 0.11 56.0 0.18 2.17 46% 34% 0% %0 %3 0% 0% 4% 81% 10% >10 d an, 10 5- earsy 0.30 0.07 95.0 0.16 2.09 46% 25% 0% %0 %0 0% 0% 2% 86% 5%, 1% earsy 0.24 0.11 56.0 att 0.14 2.45 40% 17% 0% %0 %0 0% 0% 1% 76% <5 ancifingis earsy *** S N *** *** S S S N N N *** *** *** S N *** *** S S N N *** t >10 erataht tes- T 10 5- earsy *** *** *** S S * N N *** * * *** *** ** *** ** *** ** Mean secnereffid earsy *** *** *** ** S N ** NS *** *** S se * N *** *** *** S N *** <5 otrefer .cte, * hacrup Mean earsy 0.38 0.15 47.0 er 0.25 7.13 16% %72 48% 38% 10% 1.1% %91 dna hi 9% 5% 15% 69% **, >10 s,reto 10 5- earsy 0.35 0.11 16.0 0.25 7.01 11% %82 49% 33% 15% 0.9% %02 10% 3% 16% 70% earsy 0.30 0.15 76.0 0.22 7.78 14% %82 43% 28% 15% 1.6% %41 17% 5% 17% 60% ***,.noitaroprocni morp mo fr ngi <5 ecnis esrut worrob, sra nte 1 an th yeforeb )htro deben d mnr an ve w go age : mo rmif less mun etn( fr sn ityu by le stes egarev from no paperla ngi tio co desab rcie ngi eq e m ratios Variab aslatot com t. 1 ste stitu s ngi in snoi ter w puorg tivagen com worrob se stes ot assets mrift in ega esdu as clni os stes ith udlcni en w borro k k 2 debt 3 tal e ialcnanif corporat lv s gni nancingi to ban gni gni mrif canifingiston gni ngi F rati 9. aslatot aslatot so ngi total yb denifed ot to n ist borrow borrow ces borrow ot es ilities bl ggregate bte bla abi rmet-gnollato rage morf omrf are k g covt infot mriffot ban of rmet- in gni en ng borrow lv ci worrobr reset rmet-gno ci worrobr red Table A D Pay L T In cenreP cenreP banlato orthS L Percent T reioF worro bl het renef het B Borrow Pu O ecuS wodkaer sonI Pu O B dif 1 2 3 tnemnrevoG t .sv naidnI S S tes- etavirP *** N *** N *** *** *** *** *** *** T ngieroF otrefer * .sv naidnI Median etavirP *** *** *** *** *** *** *** *** *** *** dna **, tnemnrevoG 0.29 0.10 47.0 0.21 1.76 18% 12% 0% 0% %0 0% 0% 19% 0% 28% Median ***,.srengierof ngieroF 0.17 0.16 06.0 0.08 3.74 40% 24% 0% 0% %0 0% 0% 0% 0% yb 65% den ow naidnI etavirP 0.34 0.10 46.0 0.18 2.12 47% 34% 0% 0% %3 0% 0% 3% 0% 83% or,tne m tnemnrevoG .sv naidnI t. t *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** rnevog tes- etavirP an T di In ngieroF eht Mean .sv naidnI ** S canifingiston .cte, etavirP *** *** *** *** *** *** *** *** *** *** *** *** *** N *** *** by den se are ces tnemnrevoG er Mean 0.43 0.17 20.1 hacrup 0.34 11.49 36% %14 ow,s 30% 24% 6% an renef 5.1% %8 12% 8% 36% 28% 39% di hi dif In e by th s,reto ngieroF 0.25 0.19 56.0 at th 0.17 9% 16.49 %02 den 45% 36% 9% 3.0% %11 12% 9% 15% 6% 57% s morp ow y ean mo el m fr naidnI etavirP 0.38 0.14 17.0 atv 0.25 6.24 14% %72 49% 38% 11% 0.8% %02 NS. ngi 9% 4% 15% 3% 72% pri si ely type rmif esrut worrob, nership 1 ehtreh ectivp nte res deben d mnr 10% an ve ow an eth )htro d w go th : w an rmif less on etn( mo fr by le stes egarev from de 5%, paperla ngi sn ityu bas rcie ngi eq tio 1%, e e m ratios Variab aslatot co com att don 1 ste stitu worrob s ngi in snoi si ter tivagen com se stes ot assets mrift w in gni pih canifingis esdu as clni nancingi os stes ith udlcni en w borro k k 2 rsen debt 3 tal e lv s gni ialcnanif corporat gni gni ow are gni ngi F rati 10. aslatot aslatot to ban rage so ban ngi at denifed by ot to m n ist borrow borrow ces borrow ot es en bla ilities ci dow ci worrobr red lv bl ggregate bte abi rmet-gnollato total morf omrf borrowtne th k g covt infot mriffot of rmet- in gni ng borrow reset rmet-gno worrobr Table A D Pay L T In cenreP cenreP banlato orthS rnevo L Percent T reioF worro bl het G renef het B Borrow Pu O ecuS sonI Pu O Break dif 1 2 3 tse T na * *** *** *** *** *** *** *** *** *** are at edi th M ces re renef portxe- 0.33 0.11 66.0 0.18 2.04 45% 31% 0% 0% %0 dif 0% 0% 4% 78% Median on N torefer * re portx 0.33 0.09 95.0 dna 0.17 2.54 51% 37% 0% 0% %7 0% 0% 2% 89% **, E t est- *** *** *** *** *** *** *** *** *** S N *** *** *** S N *** *** T ***,.sretropxe .cte, se t. hacrup re er hi Mean portxe- 0.38 0.15 57.0 deredisnoc 0.25 6.42 17% %92 46% 36% 11% 1.0% %81 10% 5% 17% 67% era canifingiston on el s,reto N are mpas morp ces re fr port 0.35 0.12 36.0 0.22 9.22 8% %02 53% 42% 11% 1.0% %12 5% 5% 11% 78% Ex ehtfoflahtsaelta mo renef dif ngi e th at th esrut worrob, s nte 1 orientation gnirud ean m an )htro deben d mnr an ve th NS. w go : ely mo export less etn( from paperla fr by ngi sn ityu le stes egarev selasriehtfo ectivps rcie ngi eq tio re e m ratios Variab aslatot co com 10% worrob 1 ste stitu s ngi in snoi %01 d ter tivagen com se stes ot assets mrift w naht in er an, esdu as clni nancingi os stes ith udlcni k k 2 debt borro 3 tal e en w lv s gni ialcnanif corporat gni gni 5%, gni ngi F rati 11. aslatot aslatot to ban so ban ngi 1% denifed ot to att ist borrow borrow borrow ot es ilities bl ggregate bte bla abi rmet-gnollato rage total morf omrf k g covt infot mriffot of rmet- in gni en ng borrow motropxetaht lv ci worrobr reset ci worrobr red s Table A D Pay L T In cenreP cenreP banlato orthS rmet-gno L Percent T reioF worro bl het het B Borrow Pu O ecuS mriF canifingis sonI Pu O 1 2 3 tse T *** *** *** *** *** *** *** *** *** *** Median %01 dna, cei ervS 0.18 0.08 06.0 %5, 0.08 2.50 34% 9% 0% 0% %0 0% 0% 2% 70% Median g %1tatnacifingis inr actufu 0.35 0.11 56.0 0.20 2.12 47% 35% 0% 0% %5 0% 0% 3% 83% erataht Man secnereffid t tes- *** *** *** *** *** *** *** *** *** *** S N *** *** ** *** *** T otrefer * dna **, .cte, se cei ervS 0.27 0.15 46.0 ***,. 0.20 9.05 12% %22 41% 29% 12% 1.2% %31 hacrup 17% 4% 20% 57% er Mean ectorss hi ices rves s,reto g or morp inr g actufu 0.40 0.14 47.0 mo 0.26 6.77 15% %82 49% 38% 11% 1.1% %02 inr fr 8% 5% 15% 72% Man actufuna ngi t. m e th groupings in te 1 canifingiston esrut worrob, nte sector an )htro deben d mnr an ve opera th are w go ina : ey mo m less ces etn( fr by ngi renef ityu le stes egarev from thre paperla sn tio ethh rcie ngi eq w dif e m e ratios Variab aslatot co com to th g worrob 1 ste stitu s ngi in snoi com at ter w th tivagen se stes ot assets mrift in s esdu as accordin ean clni nancingi os stes ith udlcni en w borro k k 2 debt 3 m tal e lv s gni ialcnanif corporat gni gni gni ngi F rati 12. aslatot aslatot to ban so ban ngi dedi total NS. denifed ot to div borrow borrow ely ist borrow ot es ilities bl ggregate bte bla abi rmet-gnollato rage morf omrf k g covt infot mriffot of rmet- in gni en ng borrow are lv ci worrobr reset rmet-gno ci worrobr red s Table A D Pay L T In cenreP cenreP banlato orthS L Percent T reioF worro bl het ectivp het B Borrow Pu O ecuS sonI Pu O Firm res 1 2 3 secivres rehtO 0.20 0.10 95.0 0.17 12.88 12% 26% 37% 20% 17% 0.7% %12 16% 3% 20% 55% cesi noitcurtsnoC ervS 0.25 0.18 57.0 0.19 6.11 13% 15% 45% 30% 15% 1.0% %31 13% 4% 20% 63% edarT 0.25 0.18 06.0 0.15 7.28 11% 24% 49% 41% 8% 0.5% %9 17% 2% 19% 57% scimareC & ssalG 0.48 0.12 58.0 0.35 3.95 21% 34% 39% 29% 10% 0.3% %03 6% 7% 14% 72% gnitnirp & repaP 0.41 0.13 47.0 0.30 7.30 15% 26% 42% 29% 12% 0.6% %72 8% 3% 18% 71% retupmoC 06 0.11 0. 23.0 0.08 2% 24.90 13% 44% 27% 16% 0.7% %21 11% 4% 26% 64% slateM 0.46 0.16 48.0 0.27 5.23 20% 31% 50% 41% 9% 1.0% %12 7% 5% 14% 74% sloot enihcaM 0.33 0.18 57.0 0.20 8.65 14% 26% 53% 45% 9% 1.1% %31 8% 5% 14% 68% Mean stnenopmoc otuA g 0.38 0.18 77.0 0.26 8.65 13% 18% 47% 34% 13% 1.6% %81 6% 6% 18% 70% inr sdoog lacirtcelE actufu 0.35 0.17 96.0 0.21 7.81 12% 23% 52% 43% 9% 0.6% %61 5% 5% 20% 73% tnempiuqe Man & sdoog cinortcelE 41 0. 0.16 08.0 0.26 6.45 19% 28% 46% 39% 7% 1.6% %32 6% 5% 15% 70% .cte, se slacituecamrahP 0.35 0.15 76.0 0.21 9.31 12% 24% 51% 43% 8% 0.6% %81 7% 5% 15% 71% hacrup slacimehC 0.42 0.14 57.0 28.0 5.82 14% 29% 47% 37% 10% 1.1% %22 er 8% 6% 14% 71% hi rehtaeL & stnemraG 0.41 0.15 47.0 32.0 5.72 19% 33% 63% 55% 8% 2.1% %41 s,reto 8% 3% 9% 77% morp selitxeT 0.53 0.14 78.0 53.0 2.88 23% 33% 49% 37% 12% 0.6% %72 7% 3% mo 12% 79% fr dooF 0.43 0.13 87.0 62.0 ngi 4.97 18% 32% 55% 43% 12% 0.3% %71 9% 3% 13% 73% esrut worrob, nte grouping 1 anht )htro deben d mnr an ve w go sector all ssel : etn( e from paperla mo fr by rage ngi sn ityu rcie ngi eq le tio e m assets covt com stitu com worrob ratios Variab tal 1 in to s ngi snoi tivagen w esdu se ectorss stes assets mrift resetni hti udlcni en w borro k k ialcnanif as clni 2 corporat 3 nancingi all os stes totbed tal e lv s gni gni gni gni ngi F by rati n 13. aslatot aslatot to m ban so ban ngi denifed ot to ter-gn rage total morf omrf k g ist borrow borrow borrow dow ot es lo covt infot rmif oft of rmet- in gni borrow en bla ilities ng lv ci worrobr bl eakr ggregate bte abi talo reset ci worrobr red Table B A D Pay L T In cenreP banlato orthS rmet-gno L Percen Percent T reioF worro bl het het B Borrow Pu O ecuS sonI Pu O 1 2 3 secivres rehtO 0.10 0.04 64.0 0.05 2.95 21% 3% 0% 0.0% %0 0% 0% 2% 69% cesi noitcurtsnoC ervS 0.19 0.14 17.0 0.10 2.72 43% 17% 0% 0.0% %0 0% 0% 5% 78% edarT 0.16 0.11 85.0 0.04 2.14 50% 36% 0% 0.0% %0 0% 0% 1% 71% scimareC & ssalG 0.46 0.09 17.0 0.32 1.78 35% 23% 0% 0.0% %72 0% 0% 6% 81% gnitnirp & repaP 0.38 0.10 66.0 0.26 2.17 36% 26% 0% 0.0% %61 0% 0% 8% 80% retupmoC 03 0.04 0. 22.0 0.02 8.29 36% 1% 0% 0.0% %0 0% 0% 2% 87% slateM 0.40 0.13 27.0 0.22 1.84 48% 39% 0% 0.0% %01 0% 0% 5% 83% sloot enihcaM 0.28 0.15 46.0 0.12 2.39 54% 42% 0% 0.0% %0 0% 0% 3% 78% Median stnenopmoc otuA g 0.35 0.16 66.0 0.22 2.88 44% 30% 1% 0.0% %3 0% 0% 6% 79% inr sdoog lacirtcelE actufu 0.32 0.15 36.0 0.16 2.25 53% 40% 0% 0.0% %0 0% 0% 8% 83% tnempiuqe Man & sdoog cinortcelE 36 0. 0.13 96.0 0.21 2.09 43% 36% 0% 0.0% %8 0% 0% 3% 81% .cte, se slacituecamrahP 0.32 0.12 16.0 0.14 2.38 51% 41% 0% 0.0% %1 0% 0% 4% 82% hacrup slacimehC 0.38 0.11 56.0 23.0 2.10 44% 33% 0% 0.0% %21 er 0% 0% 4% 80% hi rehtaeL & stnemraG 0.31 0.10 95.0 01.0 2.50 72% 58% 0% 0.0% %0 s,reto 0% 0% 1% 94% morp selitxeT mo oupings 0.49 0.09 07.0 03.0 1.81 46% 35% 0% 0.0% %52 0% 0% 2% 90% fr gr dooF ortc 0.37 0.10 07.0 91.0 ngi 1.74 57% 42% 0% 0.0% %2 0% 0% 3% 85% se esrut worrob, all nte by ositar )htro deben d mnr an ve w go : etn( mo fr from paperla ing ngi sn ityu rcie ngi eq le tio e m nanci assets com F. stitu Variab tal ngi in snoi d) to tivagen com worrob w esdu se as inue ectorss stes assets clni udlcni k k 2 borro corporat 3 ntoc all os stes totbed tal e gni ialcnanif to m ban ban gni gni gni ngi by rati n 13( aslatot aslatot ngi denifed ot to ter-gn rage total morf omrf k g ist borrow borrow borrow dow ot es lo covt of rmet- in gni borrow en bla ilities ng lv ci worrobr bl eakr ggregate bte abi talo reset banlato orthS rmet-gno ci worrobr red L Table B A D Pay L T In Percent T reioF worro bl het het B Borrow Pu O ecuS sonI Pu O 1 2 3 ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ** ** ** ** ** ** ** ** ** ** L S- 62. 72. 82. 52. 63. 83. 42. 52. 42. 51. 23. 01. 22. 02. 90. 41. 70. 41. 32. 50. 61. 43. 22. 42. 71. 72. 52. 82. 21. 41. -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** L M- 80. 90. 11. 60. 72. 31. 80. 80. 90. 80. 01. 90. 80. 11. 80. 50. 80. 02. 40. 31. 01. 80. 11. 01. 01. -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.07 -0 0.01 -0 -0 0.00 -0 -0 0.00 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.18 ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ** M S- 71. 81. 71. 91. 80. 52. 61. 71. 41. 70. 12. 81. 31. 22. 10. 40. 70. 60. 71. 50. 80. 41. 81. 11. 70. 02. 41. 91. 10. 23. -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 Median L 0.36 0.40 0.42 0.37 0.37 0.46 0.39 0.41 0.29 0.42 0.52 0.33 0.42 0.37 0.40 0.35 0.35 0.32 0.44 0.06 0.41 0.49 0.28 0.28 0.20 0.40 0.39 0.41 0.20 0.30 M 0.28 0.31 0.32 0.31 0.10 0.33 0.31 0.33 0.20 0.34 0.42 0.41 0.33 0.38 0.32 0.25 0.35 0.24 0.38 0.06 0.33 0.29 0.24 0.15 0.10 0.32 0.28 0.31 0.09 0.47 S 0.10 0.13 0.15 0.12 0.01 0.08 0.15 0.16 0.05 0.26 0.21 0.23 0.20 0.17 0.31 0.21 0.28 0.18 0.21 0.02 0.26 0.15 0.06 0.04 0.03 0.12 0.14 0.13 0.08 0.16 ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ** ** ** ** ** ** ** L S- 81. 81. 91. 51. 52. 72. 51. 71. 31. 11. 71. 41. 31. 41. 80. 20. 60. 40. 71. 50. 80. 42. 21. 31. 01. 81. 81. 02. 50. -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.00 -0 ** ** ** * ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** L M- 40. 60. 50. 20. 81. 01. 30. 30. 50. 30. 50. 50. 01. 90. 60. 20. 70. 91. 10. 90. 01. 30. 70. 60. 90. -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.03 -0 0.12 -0 -0 0.14 -0 0.01 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.32 ** ** ** ** ** ** ** ** ** ** ** ** ** ** ** * ** ** ** ** ** ** M S- 41. 21. 41. 31. 80. 71. 31. 31. 80. 80. 21. 71. 80. 72. 02. 81. 30. 10. 50. 11. 40. 51. 11. 41. 73. Mean -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 -0 0.02 0.07 -0 0.02 -0 -0 -0 -0 -0 -0 0.00 -0 -0 -0 0.09 -0 L characteristics 0.41 0.43 0.45 0.40 0.40 0.47 0.42 0.44 0.34 0.46 0.56 0.46 0.45 0.37 0.44 0.36 0.37 0.35 0.49 0.14 0.44 0.52 0.33 0.32 0.27 0.44 0.41 0.44 0.25 0.42 rmif M 0.37 0.38 0.40 0.38 0.22 0.36 0.39 0.40 0.29 0.43 0.50 0.49 0.40 0.50 0.34 0.27 0.51 0.29 0.49 0.12 0.37 0.33 0.32 0.23 0.17 0.41 0.33 0.39 0.16 0.74 other S 0.22 0.25 0.26 0.25 0.14 0.19 0.27 0.27 0.20 0.35 0.39 0.32 0.32 0.23 0.35 0.34 0.31 0.31 0.32 0.09 0.35 0.28 0.21 0.19 0.17 0.25 0.22 0.25 0.25 0.37 vis-a-vis t en size pmi by equ re & ts g ics leath en ratios gni gni & ticals oodsg tin oodsg cesi ts ci pon tools r inrp n ceram re ate icals aceu cal e nte asset rne nr & tiocu en ron ri com s port re ivrP etavirP he ices m ices mnr le le to ernt rth ut nrets urtc urtc in utep eda trsn rvesreh nufaa nufaa oodF tilesxe emh T Garm C armhP ect ect tou El El A Mach Metals Co &repaP Glas Tr Co Ot ex-no port dian gnier lev lev veo Eas No So We 0~5 5~10 10+ M Serv M Serv N Ex In Fo G 5% 10% ) Debt pey p att att n 14. earsy T edt hisr gio (in ryts port ne Re eg Sector Ex Status du Orien w O canifingis canifingis Table A In * ** ** ** * ** ** ** ** ** ** ** ** ** ** ** ** ** ** L S- 23. 81. 22. 32. 61. 33. 02. 15. 83. 77. 88. 23. 17. 66. 94. 72. 91. 94. 42. 90. 40. 61. 98. 0.30 -0 -0 -0 0.33 0.59 -0 -0 -0 0.22 -0 -0 -0 -0 -0 -0 -0 -0 0.16 -5 -0 0.63 -0 -0 -1 -0 0.38 -0 -2 -1 ** ** ** ** ** ** ** ** ** * ** * * * ** * ** L M- 50. 92. 42. 60. 12. 02. 81. 42. 36. 72. 12. 65. 29. 96. 15. 31. 24. 91. 31. 49. 81. 0.18 -0 -0 -0 1.14 -0 -0 -0 -0 0.02 -0 -1 -0 -1 0.44 -0 -0 0.43 0.04 -2 -0 -1 -0 0.08 1.18 -0 0.19 -0 -0 -2 ** * * ** ** ** * ** ** M S- 72. 18. 20. 41. 11. 13. 90. 97. 85. 24. 12. 0.12 -0 0.11 0.02 -0 0.65 -0 0.04 -0 0.20 0.04 1.12 -0 0.44 -1 0.24 0.21 -1 0.13 -2 0.24 1.77 0.23 -0 -2 0.10 0.19 0.09 -1 0.29 Median L 2.01 2.32 2.14 2.30 2.20 2.00 2.23 2.16 2.66 1.70 1.84 2.82 2.18 2.77 2.17 2.37 3.04 2.43 1.82 11.19 2.27 1.79 2.28 2.83 3.32 2.09 2.47 2.14 3.98 1.98 M 02. 2.19 2.27 1.85 2.06 3.33 1.94 2.02 1.96 2.48 1.72 1.60 1.19 1.91 1.56 2.61 1.81 2.12 2.86 1.85 8.49 1.76 0.66 1.85 2.91 4.50 1.90 2.66 2.01 3.04 -0 S 2.31 2.00 1.96 2.08 2.52 2.59 2.00 2.00 2.33 1.92 1.64 2.31 1.80 2.00 1.30 2.05 2.33 1.76 1.98 5.70 2.00 2.43 2.08 2.34 2.08 2.00 2.85 2.11 1.82 0.10 ** ** ** ** * ** ** ** * ** ** ** ** ** ** L S- 85. 38. 12. 58. 77. 08. 75. 39. 73. 51. 61. 26. 01. 60. 08. 71. 10. 4.30 2.25 -0 -0 4.40 3.91 -0 0.33 -0 3.47 -0 3.25 1.96 -6 -5 -3 -2 -1 -0 -9 -5 4.87 -0 2.09 -2 0.11 4.10 1.75 -5 -8 * ** ** ** ** * ** ** ** ** ** ** ** L M- 93. 02. 86. 97. 49. 00. 07. 52. 05. 24. 49. 02. 81. 95. 68. 58. 84. 27. 83. 65. 51. 52. 0.03 -1 -1 -1 5.78 1.02 -1 -0 -4 2.26 -0 0.54 -1 -9 -3 -1 -4 -0 -0 -2 -4 -1 -3 -1 5.32 -1 0.93 -0 -3 10.1- ** ** ** ** ** ** ** ** * ** ** ** ** ** M aracteristics S- 83. 70. 51. 99. 89. 67. 52. 21. 39. 4.27 3.64 0.62 0.85 Mean -1 2.89 1.58 1.27 3.15 1.21 -0 2.70 3.21 2.70 -2 -1 1.83 -0 0.42 -6 -0 6.35 3.66 3.47 -8 1.68 3.16 1.91 -1 2.06 ch rmif L 5.39 6.25 7.72 8.22 6.28 5.70 7.41 6.85 9.91 4.03 3.04 4.51 5.73 12.31 7.68 8.67 9.25 8.92 5.34 29.06 8.59 3.75 7.82 5.58 13.08 6.64 8.57 5.90 17.26 12.93 er M oth 5.43 4.86 6.52 6.54 12.06 6.72 5.62 5.91 5.91 6.29 2.34 5.05 4.48 2.82 4.26 6.73 5.05 8.75 4.76 26.20 3.75 2.27 4.10 4.20 18.40 5.07 9.51 5.75 14.01 2.86 S 9.69 8.50 7.14 7.39 10.68 9.62 7.19 7.18 9.06 7.50 2.27 7.76 7.69 5.51 2.11 4.75 6.88 7.77 5.18 19.44 3.50 8.62 7.76 7.66 10.28 6.75 12.67 7.65 12.09 4.92 vis-a-vis ze si t yb en pmi equ ratios re & ts g ics leath en gni gni & ticals oodsg tin oodsg cesi ts ci pon tools r inrp n ceram re ate aceu cal e nte coverage rne nr & tiocu etavirP le le ernt he rth ut nrets urtc urtc en icals ron ri com in utep s ices ect ect m ices eda trsn rvesreh port re ivrP mnr nufaa nufaa oodF tilesxe emh T Garm C armhP tou El El A Mach Metals Co &repaP Glas Tr Co Ot ex-no port dian gnier lev lev veo terest Eas No So We 0~5 5~10 10+ M Serv M Serv N Ex In Fo G 5% 10% ) In pey p att att n 15. earsy T edt hisr el gio (in ryts port ne Re eg Sector Ex Status du Orien w O canifingis canifingis Tab A In * ** Table 16. Baseline regressions for debt to asset ratios (1) (2) (3) (4) (5) Log of Age 0.117 0.081 0.052 0.101 0.098 [0.000]*** [0.003]*** [0.056]* [0.000]*** [0.000]*** Square of Log of Age -0.019 -0.014 -0.011 -0.015 -0.016 [0.000]*** [0.001]*** [0.015]** [0.000]*** [0.000]*** Small -0.121 -0.123 -0.165 -0.11 -0.131 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Medium -0.058 -0.068 -0.089 -0.055 -0.063 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Exporter -0.001 -0.003 0.013 -0.009 -0.001 [0.923] [0.626] [0.034]** [0.122] [0.872] Foreign Sector -0.079 -0.081 -0.106 -0.064 -0.082 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Government Sector 0.045 0.039 -0.055 0.051 0.036 [0.072]* [0.125] [0.014]** [0.043]** [0.141] Asset Tangibility 0.208 0.21 0.194 0.17 0.218 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Return on Assets -0.797 -0.834 -0.236 -0.68 -0.698 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Sales Growth 0.002 [0.686] Market to Book 0.001 [0.315] Tax Rate -0.322 [0.000]*** Business Risk 0.567 [0.000]*** Eastern India -0.014 -0.012 -0.017 -0.012 -0.008 [0.218] [0.312] [0.130] [0.291] [0.476] Western India -0.006 -0.007 -0.006 -0.005 -0.005 [0.465] [0.351] [0.397] [0.542] [0.475] Southern India 0.022 0.024 0.016 0.026 0.022 [0.011]** [0.006]*** [0.038]** [0.003]*** [0.008]*** Food 0.049 0.043 0.036 0.05 0.044 [0.000]*** [0.002]*** [0.011]** [0.000]*** [0.001]*** Textiles 0.111 0.108 0.101 0.107 0.107 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Garments & Leather 0.063 0.057 0.014 0.059 0.041 [0.005]*** [0.010]** [0.517] [0.010]*** [0.049]** Chemicals 0.057 0.057 0.05 0.056 0.053 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Pharmaceuticals 0.026 0.034 0.017 0.03 0.024 [0.048]** [0.014]** [0.208] [0.023]** [0.063]* Electronic goods & equipment 0.075 0.069 0.039 0.074 0.069 [0.001]*** [0.003]*** [0.045]** [0.001]*** [0.002]*** Electrical goods 0.033 0.028 0.029 0.043 0.033 [0.048]** [0.115] [0.073]* [0.007]*** [0.044]** Table 16 (Continued). Baseline regressions for debt to assets ratios (1) (2) (3) (4) (5) Auto components 0.064 0.064 0.029 0.076 0.061 [0.001]*** [0.001]*** [0.053]* [0.000]*** [0.001]*** Machine tools 0.027 0.025 0.014 0.031 0.02 [0.052]* [0.076]* [0.270] [0.023]** [0.150] Metals 0.08 0.077 0.056 0.079 0.079 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Trade -0.012 -0.008 -0.01 -0.003 -0.012 [0.361] [0.580] [0.446] [0.812] [0.374] Computer -0.119 -0.123 -0.12 -0.133 -0.141 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Services -0.083 -0.08 -0.072 -0.075 -0.086 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Paper & printing 0.026 0.018 0.021 0.022 0.025 [0.173] [0.365] [0.259] [0.242] [0.201] Glass & Ceramics 0.09 0.097 0.067 0.084 0.079 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Construction -0.019 -0.016 -0.004 -0.005 -0.003 [0.350] [0.402] [0.813] [0.793] [0.887] Year 1994 0.051 0 0 0.049 0.044 [0.000]*** [.] [.] [0.000]*** [0.000]*** Year 1995 0.019 0.039 -0.094 0.013 0.014 [0.000]*** [0.000]*** [0.272] [0.001]*** [0.000]*** Year 1996 0.006 0.019 -0.003 -0.005 0.002 [0.060]* [0.000]*** [0.243] [0.112] [0.526] Year 1998 0.008 0.005 0.007 0.001 0.009 [0.003]*** [0.117] [0.004]*** [0.740] [0.002]*** Year 1999 0.011 0.005 0 0.003 0.01 [0.004]*** [0.253] [0.886] [0.372] [0.004]*** Year 2000 0.023 0.012 -0.006 0.017 0.024 [0.000]*** [0.009]*** [0.078]* [0.000]*** [0.000]*** Year 2001 0.022 0.011 -0.021 0.017 0.024 [0.000]*** [0.026]** [0.000]*** [0.001]*** [0.000]*** Year 2002 0.01 -0.004 -0.035 0.021 0.012 [0.067]* [0.502] [0.000]*** [0.000]*** [0.026]** Year 2003 -0.006 -0.018 -0.05 0.013 -0.004 [0.371] [0.004]*** [0.000]*** [0.057]* [0.558] Observations 37307 28261 16797 34794 37077 R-squared 0.245 0.258 0.357 0.273 0.273 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 17. Regressions for debt to assets ratios: size split (1) (2) (3) Small firms Medium firms Large firms Log of Age 0.148 0.252 0.046 [0.013]** [0.000]*** [0.120] Square of Log of Age -0.016 -0.04 -0.011 [0.113] [0.000]*** [0.026]** Exporter 0.002 -0.01 -0.006 [0.861] [0.507] [0.392] Foreign Sector 0.014 -0.056 -0.097 [0.608] [0.066]* [0.000]*** Government Sector 0.167 0.286 0.002 [0.038]** [0.013]** [0.941] Asset Tangibility 0.136 0.126 0.256 [0.000]*** [0.001]*** [0.000]*** Return on Assets -0.445 -0.859 -1.001 [0.000]*** [0.000]*** [0.000]*** Eastern India -0.027 0.02 -0.021 [0.231] [0.474] [0.128] Western India -0.009 0.012 -0.004 [0.570] [0.520] [0.661] Southern India 0.007 0.041 0.027 [0.701] [0.043]** [0.005]*** Year 1994 0.049 0.075 0.041 [0.000]*** [0.000]*** [0.000]*** Year 1995 0.014 0.022 0.023 [0.093]* [0.023]** [0.000]*** Year 1996 0 0.012 0.009 [0.966] [0.115] [0.013]** Year 1998 0.009 0.007 0.002 [0.158] [0.384] [0.503] Year 1999 0.024 -0.001 -0.002 [0.005]*** [0.960] [0.642] Year 2000 0.034 0.042 0.002 [0.000]*** [0.001]*** [0.646] Year 2001 0.028 0.045 0.005 [0.007]*** [0.004]*** [0.401] Year 2002 0.015 0.023 -0.005 [0.231] [0.187] [0.414] Year 2003 0.019 0.025 -0.028 [0.244] [0.282] [0.000]*** Constant -0.11 -0.109 0.303 [0.223] [0.330] [0.000]*** Observations 10291 5368 21648 R-squared 0.106 0.214 0.297 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 18. Regressions for debt to assets ratios: export, age, and period splits (1) (2) (3) (4) (5) (6) (7) exporter non-exporter age<=10 age>10 1994-1997 1998-2000 2001-2003 Log of Age 0.043 0.148 0.148 0.109 0.125 [0.295] [0.000]*** [0.000]*** [0.001]*** [0.000]*** Square of Log of Age -0.01 -0.023 -0.021 -0.019 -0.022 [0.137] [0.000]*** [0.000]*** [0.001]*** [0.000]*** Small -0.116 -0.121 -0.174 -0.107 -0.12 -0.122 -0.103 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Medium -0.07 -0.054 -0.087 -0.052 -0.051 -0.069 -0.043 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.005]*** Foreign Sector -0.081 -0.077 -0.074 -0.079 -0.059 -0.095 -0.082 [0.000]*** [0.000]*** [0.001]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Government Sector -0.113 0.059 0.095 0.042 0.085 0.02 0.033 [0.022]** [0.029]** [0.235] [0.105] [0.003]*** [0.462] [0.292] Asset Tangibility 0.265 0.198 0.218 0.211 0.13 0.214 0.326 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Return on Assets -0.66 -0.835 -0.451 -0.862 -0.574 -0.812 -0.992 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Eastern India -0.018 -0.018 -0.019 -0.021 -0.014 -0.012 -0.02 [0.365] [0.188] [0.352] [0.097]* [0.317] [0.390] [0.193] Western India -0.001 -0.007 0.023 -0.012 -0.016 -0.001 0.005 [0.905] [0.429] [0.183] [0.158] [0.047]** [0.875] [0.660] Southern India 0.029 0.017 0.043 0.019 0.015 0.029 0.021 [0.022]** [0.103] [0.020]** [0.048]** [0.098]* [0.005]*** [0.091]* Year 1994 0.041 0.054 0.005 0.052 [0.000]*** [0.000]*** [0.828] [0.000]*** Year 1995 0.007 0.022 -0.024 0.024 [0.255] [0.000]*** [0.074]* [0.000]*** Year 1996 0.005 0.006 -0.022 0.01 [0.320] [0.090]* [0.012]** [0.002]*** Year 1998 0.01 0.007 0.031 0.004 [0.030]** [0.033]** [0.000]*** [0.191] Year 1999 0.006 0.012 0.034 0.006 [0.330] [0.009]*** [0.000]*** [0.142] Year 2000 0.005 0.028 0.039 0.017 [0.447] [0.000]*** [0.000]*** [0.000]*** Year 2001 0.006 0.027 0.043 0.015 [0.465] [0.000]*** [0.000]*** [0.004]*** Year 2002 -0.022 0.021 0.036 0.002 [0.011]** [0.002]*** [0.012]** [0.691] Year 2003 -0.047 0.009 0.02 -0.012 [0.000]*** [0.252] [0.252] [0.080]* Exporter 0.018 -0.004 0.01 0.003 -0.019 [0.219] [0.547] [0.119] [0.638] [0.037]** Constant 0.273 0.082 0.297 0.323 0.102 0.168 0.127 [0.000]*** [0.131] [0.000]*** [0.000]*** [0.097]* [0.003]*** [0.017]** Observations 8987 28320 5449 31889 14728 12348 10231 R-squared 0.36 0.227 0.321 0.237 0.199 0.279 0.282 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 19. Regressions for debt to assets ratios: region and sector splits (1) (2) (3) (4) (5) (6) Eastern India Western India Southern India Northern India Manufacturing Services Log of Age 0.083 0.119 0.123 0.097 0.176 0.098 [0.215] [0.002]*** [0.047]** [0.123] [0.000]*** [0.140] Square of Log of Age -0.012 -0.019 -0.021 -0.011 -0.026 -0.014 [0.246] [0.002]*** [0.038]** [0.298] [0.000]*** [0.218] Small -0.095 -0.123 -0.146 -0.078 -0.142 -0.103 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Medium -0.008 -0.058 -0.07 -0.061 -0.061 -0.055 [0.749] [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.003]*** Exporter 0.044 -0.003 0.004 -0.007 -0.002 -0.002 [0.032]** [0.695] [0.737] [0.616] [0.795] [0.915] Foreign Sector -0.081 -0.053 -0.11 -0.114 -0.1 -0.017 [0.002]*** [0.001]*** [0.000]*** [0.000]*** [0.000]*** [0.532] Government Sector 0.176 -0.073 0.059 0.02 0.063 -0.024 [0.007]*** [0.019]** [0.229] [0.634] [0.055]* [0.432] Asset Tangibility 0.122 0.223 0.203 0.251 0.256 0.169 [0.014]** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Return on Assets -1.023 -0.743 -0.759 -0.754 -0.846 -0.599 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 1994 0.071 0.049 0.053 0.039 0.058 0.043 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.001]*** Year 1995 0.044 0.016 0.002 0.025 0.024 0.011 [0.002]*** [0.003]*** [0.834] [0.006]*** [0.000]*** [0.299] Year 1996 0.023 0.006 -0.003 0.004 0.008 0.002 [0.030]** [0.156] [0.642] [0.534] [0.014]** [0.806] Year 1998 0.015 0.012 0.005 0.007 0.008 0.001 [0.081]* [0.005]*** [0.349] [0.299] [0.007]*** [0.881] Year 1999 0.013 0.015 0.009 0.012 0.01 -0.002 [0.217] [0.007]*** [0.200] [0.182] [0.015]** [0.849] Year 2000 0.015 0.032 0.024 0.012 0.023 0 [0.172] [0.000]*** [0.006]*** [0.205] [0.000]*** [0.964] Year 2001 0.011 0.029 0.023 0.022 0.021 0.004 [0.368] [0.000]*** [0.022]** [0.034]** [0.000]*** [0.690] Year 2002 0.005 0.025 -0.007 0.01 0.009 -0.005 [0.700] [0.003]*** [0.538] [0.436] [0.158] [0.697] Year 2003 0.007 0.008 -0.03 0 -0.013 0.001 [0.712] [0.434] [0.026]** [0.992] [0.067]* [0.954] Eastern India -0.021 -0.009 [0.115] [0.710] Western India -0.016 0.016 [0.053]* [0.391] Southern India 0.003 0.059 [0.721] [0.012]** Constant 0.182 0.135 0.165 0.111 0.082 0.117 [0.113] [0.028]** [0.095]* [0.271] [0.101] [0.240] Observations 4533 16621 8870 7283 30847 6460 R-squared 0.265 0.244 0.28 0.261 0.227 0.132 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 20. Baseline regressions for interest coverage ratios (1) (2) (3) (4) (5) Log of Age 1.032 1.53 1.462 1.133 0.968 [0.466] [0.329] [0.552] [0.443] [0.495] Square of Log of Age -0.177 -0.235 -0.246 -0.247 -0.167 [0.445] [0.362] [0.547] [0.305] [0.470] Small 1.724 1.788 3.339 1.693 1.588 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Medium 1.01 1.056 2.117 0.97 0.968 [0.001]*** [0.002]*** [0.000]*** [0.001]*** [0.001]*** Exporter 0.808 0.815 -0.059 1.109 0.772 [0.018]** [0.032]** [0.893] [0.001]*** [0.024]** Foreign Sector 4.42 4.996 4.966 4.181 4.382 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Government Sector 5.862 6.028 6.855 6.037 5.668 [0.000]*** [0.000]*** [0.005]*** [0.000]*** [0.000]*** Asset Tangibility -3.195 -4.072 -2.832 -2.091 -3.096 [0.000]*** [0.000]*** [0.008]*** [0.003]*** [0.000]*** Return on Assets 46.681 49.554 58.646 41.593 48.513 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Sales Growth 0.02 [0.924] Market to Book 0.643 [0.000]*** Tax Rate 10.381 [0.000]*** Business Risk 8.286 [0.000]*** Eastern India 0.221 0.146 -0.167 0.229 0.239 [0.645] [0.790] [0.827] [0.639] [0.618] Western India 0.756 0.645 0.729 0.708 0.728 [0.034]** [0.111] [0.166] [0.053]* [0.041]** Southern India 0.03 0.012 -0.491 -0.02 -0.031 [0.940] [0.978] [0.394] [0.960] [0.939] Food -2.119 -2.268 -2.663 -2.229 -2.16 [0.001]*** [0.001]*** [0.001]*** [0.000]*** [0.001]*** Textiles -2.864 -2.794 -2.541 -2.826 -2.889 [0.000]*** [0.000]*** [0.001]*** [0.000]*** [0.000]*** Garments & Leather -1.336 -1.488 0.181 -1.296 -1.432 [0.233] [0.187] [0.908] [0.259] [0.212] Chemicals -2.513 -2.45 -2.799 -2.622 -2.55 [0.000]*** [0.000]*** [0.001]*** [0.000]*** [0.000]*** Pharmaceuticals -0.778 -0.665 -0.541 -1.021 -0.796 [0.386] [0.515] [0.660] [0.263] [0.375] Electronic goods & equipment -2.982 -2.752 -2.417 -3.097 -3.069 [0.000]*** [0.004]*** [0.042]** [0.000]*** [0.000]*** Electrical goods -2.293 -2.228 -2.038 -2.722 -2.428 [0.008]*** [0.020]** [0.066]* [0.002]*** [0.005]*** Table 20 (Continued). Baseline regressions for interest coverage ratios (1) (2) (3) (4) (5) Auto components -1.606 -1.537 -0.611 -2.019 -1.625 [0.076]* [0.116] [0.655] [0.029]** [0.075]* Machine tools -1.297 -1.04 -1.011 -1.513 -1.325 [0.096]* [0.231] [0.369] [0.058]* [0.091]* Metals -2.052 -1.844 -1.531 -2.052 -2.061 [0.003]*** [0.016]** [0.114] [0.004]*** [0.003]*** Trade -1.636 -1.866 -1.94 -1.959 -1.541 [0.034]** [0.030]** [0.060]* [0.015]** [0.047]** Computer 8.714 9.219 9.558 9.275 8.416 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Services 3.453 3.68 3.746 2.888 3.457 [0.041]** [0.040]** [0.120] [0.099]* [0.041]** Paper & printing -0.86 -0.794 -0.625 -0.893 -0.903 [0.351] [0.440] [0.600] [0.339] [0.330] Glass & Ceramics -2.707 -2.626 -2.928 -2.793 -2.857 [0.000]*** [0.001]*** [0.009]*** [0.000]*** [0.000]*** Construction -1.773 -2.227 -2.375 -2.318 -1.555 [0.047]** [0.018]** [0.023]** [0.013]** [0.082]* Year 1994 -0.75 0 0 -0.501 -0.818 [0.016]** [.] [.] [0.118] [0.010]*** Year 1995 0.504 0.158 -5.495 0.897 0.465 [0.063]* [0.608] [0.000]*** [0.001]*** [0.090]* Year 1996 0.001 -0.174 -0.158 0.406 -0.019 [0.996] [0.474] [0.536] [0.089]* [0.932] Year 1998 0.874 0.799 0.783 1.06 0.926 [0.000]*** [0.001]*** [0.002]*** [0.000]*** [0.000]*** Year 1999 1.446 1.41 1.108 1.678 1.503 [0.000]*** [0.000]*** [0.001]*** [0.000]*** [0.000]*** Year 2000 2.35 2.299 3.194 2.601 2.439 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2001 2.966 3.122 4.192 3.193 3.095 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2002 3.953 4.178 5.531 3.848 4.064 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2003 5.149 5.318 6.238 4.66 5.258 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Observations 32938 25927 15278 31432 32752 R-squared 0.141 0.148 0.168 0.15 0.142 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 21. Regressions for interest coverage ratios: size split (1) (2) (3) Small firms Medium firms Large firms Log of Age -2.315 0.559 1.804 [0.475] [0.896] [0.228] Square of Log of Age 0.359 -0.092 -0.29 [0.498] [0.888] [0.249] Exporter 2.508 0.88 0.301 [0.006]*** [0.220] [0.446] Foreign Sector 1.291 4.944 4.984 [0.435] [0.054]* [0.000]*** Government Sector -0.904 0.261 7.089 [0.459] [0.892] [0.000]*** Asset Tangibility -2.638 -2.739 -3.848 [0.038]** [0.051]* [0.000]*** Return on Assets 52.299 38.327 45.455 [0.000]*** [0.000]*** [0.000]*** Eastern India 1.05 0.779 -0.281 [0.419] [0.428] [0.599] Western India 0.399 0.885 0.761 [0.668] [0.202] [0.065]* Southern India -1.047 -0.595 0.277 [0.261] [0.452] [0.578] Year 1994 -0.983 -0.295 -0.877 [0.179] [0.722] [0.014]** Year 1995 0.436 1.938 0.03 [0.459] [0.013]** [0.928] Year 1996 0.577 0.378 -0.357 [0.281] [0.516] [0.200] Year 1998 0.452 0.347 1.092 [0.438] [0.562] [0.000]*** Year 1999 1.318 1.789 1.31 [0.042]** [0.009]*** [0.000]*** Year 2000 3.317 1.414 2.149 [0.000]*** [0.026]** [0.000]*** Year 2001 2.212 2.462 3.137 [0.006]*** [0.000]*** [0.000]*** Year 2002 2.977 5.004 3.847 [0.003]*** [0.000]*** [0.000]*** Year 2003 5.141 5.48 4.942 [0.000]*** [0.000]*** [0.000]*** Constant 6.976 3.371 0.459 [0.143] [0.624] [0.841] Observations 7356 4901 20681 R-squared 0.139 0.168 0.146 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 22. Regressions for interest coverage ratios: export, age, and period splits (1) (2) (3) (4) (5) (6) (7) exporter non-exporter age<=10 age>10 1994-1997 1998-2000 2001-2003 Log of Age 4.069 -0.176 -0.424 1.939 0.31 [0.122] [0.917] [0.820] [0.312] [0.890] Square of Log of Age -0.653 0.001 -0.062 -0.262 0.046 [0.125] [0.999] [0.825] [0.406] [0.905] Small 3.1 1.294 2.174 1.707 1.798 1.494 1.382 [0.003]*** [0.002]*** [0.016]** [0.000]*** [0.000]*** [0.008]*** [0.092]* Medium 1.52 0.723 1.783 0.949 0.863 0.479 1.734 [0.048]** [0.016]** [0.016]** [0.003]*** [0.027]** [0.206] [0.011]** Foreign Sector 6.481 3.358 4.185 4.445 1.725 5.216 7.467 [0.000]*** [0.000]*** [0.023]** [0.000]*** [0.023]** [0.000]*** [0.000]*** Government Sector 13.034 4.867 7.594 5.724 4.686 5.451 7.54 [0.019]** [0.000]*** [0.043]** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Asset Tangibility -2.974 -3.12 -1.296 -3.233 -1.402 -4.28 -5.203 [0.084]* [0.000]*** [0.453] [0.000]*** [0.110] [0.000]*** [0.000]*** Return on Assets 58.01 42.937 47.913 46.455 41.155 42.95 58.191 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Eastern India -0.362 0.353 1.959 -0.047 0.498 0.979 -0.491 [0.685] [0.530] [0.119] [0.923] [0.323] [0.144] [0.548] Western India 1.617 0.501 1.325 0.654 0.516 0.805 1.189 [0.035]** [0.213] [0.141] [0.089]* [0.171] [0.064]* [0.103] Southern India 0.917 -0.29 -0.38 0.099 -0.246 0.194 0.341 [0.303] [0.522] [0.661] [0.822] [0.579] [0.674] [0.676] Year 1994 -1.117 -0.62 3.475 -0.923 [0.105] [0.074]* [0.060]* [0.004]*** Year 1995 1.419 0.22 2.839 0.343 [0.041]** [0.428] [0.012]** [0.219] Year 1996 0.153 -0.034 1.23 -0.139 [0.776] [0.889] [0.158] [0.552] Year 1998 0.892 0.885 0.886 0.877 [0.098]* [0.001]*** [0.176] [0.001]*** Year 1999 1.715 1.36 0.749 1.586 [0.002]*** [0.000]*** [0.245] [0.000]*** Year 2000 3.524 1.99 3.067 2.253 [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2001 4.739 2.409 3.311 2.967 [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2002 5.496 3.466 3.555 4.093 [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 2003 5.557 4.995 4.95 5.25 [0.000]*** [0.000]*** [0.000]*** [0.000]*** Exporter 1.268 0.75 0.794 0.615 1.037 [0.168] [0.043]** [0.041]** [0.145] [0.117] Constant -5.873 3.816 -1.079 2.931 5.433 0.858 4.118 [0.156] [0.136] [0.464] [0.000]*** [0.077]* [0.767] [0.226] Observations 8192 24746 4155 28803 13246 10955 8737 R-squared 0.191 0.121 0.194 0.135 0.119 0.161 0.152 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1% Table 23. Regressions for interest coverages: region and sector splits (1) (2) (3) (4) (5) (6) Eastern India Western India Southern India Northern India Manufacturing Services Log of Age 0.048 1.105 3.401 -1.114 -0.901 -3.106 [0.986] [0.624] [0.218] [0.793] [0.520] [0.544] Square of Log of Age -0.035 -0.209 -0.53 0.183 0.091 0.26 [0.935] [0.564] [0.258] [0.809] [0.692] [0.750] Small 3.306 1.516 1.143 1.707 2.207 1.168 [0.014]** [0.004]*** [0.116] [0.097]* [0.000]*** [0.234] Medium 1.68 1.202 0.584 0.975 1.277 -0.033 [0.057]* [0.005]*** [0.345] [0.132] [0.000]*** [0.973] Exporter -1.126 1.035 1.724 -0.031 1.071 1.776 [0.177] [0.047]** [0.018]** [0.965] [0.002]*** [0.138] Foreign Sector 4.966 3.555 5.12 4.698 5.283 0.15 [0.002]*** [0.000]*** [0.002]*** [0.003]*** [0.000]*** [0.907] Government Sector 2.416 6.635 6.978 6.293 5.575 7.038 [0.072]* [0.008]*** [0.007]*** [0.004]*** [0.000]*** [0.018]** Asset Tangibility -1.281 -4.201 -4.796 0.292 -4.66 -0.789 [0.425] [0.000]*** [0.001]*** [0.847] [0.000]*** [0.614] Return on Assets 43.133 50.581 43.473 42.94 45.723 58.679 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** Year 1994 0.407 -0.605 -1.339 -0.993 -1.056 0.241 [0.680] [0.219] [0.021]** [0.143] [0.001]*** [0.844] Year 1995 0.014 0.617 0.234 0.81 0.241 1.442 [0.986] [0.120] [0.669] [0.216] [0.381] [0.123] Year 1996 -0.378 0.033 0.164 -0.095 -0.175 0.635 [0.493] [0.918] [0.761] [0.860] [0.451] [0.415] Year 1998 1.426 1.056 0.728 0.228 0.912 0.887 [0.038]** [0.004]*** [0.148] [0.609] [0.000]*** [0.283] Year 1999 1.826 1.47 1.433 0.82 1.77 0.202 [0.009]*** [0.000]*** [0.010]*** [0.136] [0.000]*** [0.791] Year 2000 2.386 2.755 2.122 1.472 2.373 2.705 [0.004]*** [0.000]*** [0.000]*** [0.008]*** [0.000]*** [0.005]*** Year 2001 0.885 3.425 3.71 2.303 3.087 2.863 [0.158] [0.000]*** [0.000]*** [0.003]*** [0.000]*** [0.004]*** Year 2002 2.727 4.375 4.247 3.335 4.141 3.226 [0.002]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.006]*** Year 2003 3.734 5.984 4.747 4.553 5.484 3.619 [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.000]*** [0.002]*** Eastern India 0.295 -0.239 [0.530] [0.885] Western India 1.031 -0.688 [0.005]*** [0.534] Southern India 0.751 -2.161 [0.062]* [0.118] Constant 2.756 0.594 0.502 3.723 2.744 9.252 [0.529] [0.863] [0.905] [0.524] [0.206] [0.243] Observations 4029 14444 7954 6511 28047 4891 R-squared 0.125 0.152 0.152 0.157 0.135 0.103 Robust p values in brackets * significant at 10%; ** significant at 5%; *** significant at 1%