Bank Entry and Bankruptcy. Bank Entry and Bankruptcy. Todd Gormley, Nandini Gupta, and Anand Jha * January Abstract.

Transcription

1 Bank Entry and Bankruptcy Bank Entry and Bankruptcy Todd Gormley, Nandini Gupta, and Anand Jha * January 2011 Abstract The bankruptcy process often involves long delays that erode firm value and raise the cost of capital. One potential contributing factor may be an uncompetitive banking system, where creditors may lack incentives to monitor and exert pressure on defaulting borrowers. We analyze this possibility using a unique dataset on corporate bankruptcy filings and geographic variation in bank entry following deregulation of the banking sector in India. We find that private bank entry in a region is associated with an increase in bankruptcy filings by nearby firms, but a decrease in bankruptcy delays. Foreign bank entry is associated with more bankruptcies, but only when local creditor rights are strong. The findings suggest that ownership and competitiveness of the banking sector may help explain differences in bankruptcy outcomes. Abstract The bankruptcy process often involves long delays that erode firm value and raise the cost of capital. One potential contributing factor may be an uncompetitive banking system, where creditors likely lack incentives to monitor and exert pressure on defaulting borrowers. We analyze this possibility using a unique dataset on corporate bankruptcy filings and geographic variation in bank entry following deregulation of the banking sector in India. We find that private bank entry in a region is associated with an increase in bankruptcy filings by nearby firms, but a decrease in bankruptcy delays. Foreign bank entry is associated with more bankruptcies, but only when local creditor rights are strong. The findings suggest that ownership and competitiveness of the banking sector may help explain differences in bankruptcy outcomes. Keywords: Bankruptcy, bank competition, bank ownership, creditor rights, India Keywords: Bankruptcy, bank competition, bank ownership, creditor rights, India JEL Classification Codes: G21, G23, G28, G38 * Todd Gormley is at The Wharton School, University of Pennsylvania, Phone: (215) , tgormley@wharton.upenn.edu. Nandini Gupta is at the Kelley School of Business, Indiana University, Phone: (812) , nagupta@indiana.edu; Anand Jha is at Texas A&M International University, Phone: (956) , ajha@tamiu.edu. We are grateful to Franklin Allen, Rajesh Chakrabarti (discussant), Richard Rosen (discussant), Irina Stefanescu, Greg Udell, Vijay Yerramilli, and Zhipeng Zhang (discussant) for valuable comments. We also thank participants of the 2010 Financial Intermediation Research Society Annual Conference (Florence), 2010 CAF-ISB Summer Research Conference in Finance (Hyderabad), and the finance department workshops at the Kelley School of Business, The Wharton School, and the Hong Kong University of Science and Technology Corporate Finance Symposium, December 2009.

2 1. Introduction Bankruptcy procedures are a way of enforcing debt contracts, and the efficiency of this process can affect the cost and allocation of capital (Stiglitz, 2001; Hart, 2000). The bankruptcy process, however, can be highly inefficient, especially in developing countries. The average duration of bankruptcies ranges from 3 to 7 years in emerging markets such as Brazil, India, Indonesia, and Russia (World Bank, 2010), and insolvency practitioners estimate that, on average, nearly 50% of firm value is lost because of inefficiencies and delays (Djankov, Hart, McLiesh, and Shleifer, 2008). Understandably, the existing literature focuses on how differences in bankruptcy laws and legal systems may explain many of these inefficiencies, 1 but legal differences are unlikely to be the only contributing factor. In this paper, we analyze whether banking sector entry is related to bankruptcy outcomes. Entry and ownership of the banking sector are likely to affect the bankruptcy process in a number of ways. By increasing incentives to maximize efficiency, an increase in banking sector competitiveness following bank entry may lead creditors to monitor borrowers more aggressively and attempt to recover assets more quickly (Hicks, 1935; Berger and Hannan, 1998). This shift among creditors may increase a firm s likelihood of seeking bankruptcy protection, and may reduce delays in bankruptcy. Moreover, the ownership of creditors may affect their incentive to monitor loans. For example, privately-owned or foreign banks may have a stronger incentive to pursue delinquent firms and recover assets quickly relative to government-owned banks. The impact of bank ownership, however, may vary based on the firm s ownership. Private banks may be less willing (or less able) to purse delinquent government-owned firms. To investigate the potential impact of creditors incentives on bankruptcy outcomes, we examine how banking sector entry is related to corporate bankruptcy outcomes in India. While the bankruptcy process is implemented at the federal level in India, there is significant variation in regional banking markets following banking deregulation in the early 1990s. Deregulation facilitated the entry of private domestic and foreign banks into geographical regions previously dominated by monopolistic government banks. Using India as the empirical context has four advantages. First, the geographical variation of bank entry into India over time facilitates the use of various identification techniques, including the use of regional and year fixed effects. This helps us better determine the potential sources for any observed connection between bank entry and bankruptcy outcomes. Second, exogenous variation in local creditor rights also improves our ability to assess the exact channels through which lenders may influence bankruptcy outcomes. Third, and probably most importantly, we are able to observe detailed data on over 4,000 corporate bankruptcy filings in India between 1991 and We observe the population of bankruptcy filings since 1991 that are filed with the federal bankruptcy court, the Board for Industrial and Financial Restructuring (BIFR). Finally, India s banking system and bankruptcy process are similar to that of other economies, suggesting that an analysis of India may shed light on bankruptcy and bank entry more broadly. For example, the high share of assets held by government banks in India is similar to that of other developing economies (La Porta, Lopez-de-Silanes, and Shleifer, 2002). Making use of variation in both the timing and the extent of bank entry across India s 500 districts, we find that entry of privately-owned banks into the banking sector is associated with an increase in the number of corporate bankruptcy filings. A one standard deviation increase in the share of deposits controlled by private banks in a district is associated with a one standard deviation increase in the number of bankruptcy filings in that district. There is little relationship, on average, between foreign bank entry and local bankruptcy outcomes. The relationship between bank entry and bankruptcy filings also depends on whether the filing firm is government or 1 E.g. Using cross-country evidence, Djankov et al. (2008) argue that extensive appeals of insolvency decisions reduce efficiency, and Claessens and Klapper (2005) show that bankruptcies are used more often in countries with more efficient judiciaries. Hotchkiss, John, Mooradian and Thorburn (2008) provides an extensive survey of this literature. privately-owned. While the entry of private banks in a market is accompanied by an increase in filings by privately-owned firms, the opposite occurs with government-owned firms.

3 Despite the increase in bankruptcy filings, which may increase the burden on the bankruptcy court, private bank entry is also associated with an economically significant decrease in delays in the bankruptcy process. A one standard deviation increase in the share of deposits held by private banks in a district is associated with an average of 422 fewer days to receive a restructuring or liquidation order from the bankruptcy court. Private bank entry is also associated with a shift away from liquidation orders in favor of more restructuring decisions, and a larger decrease in the duration of restructurings. We also examine the association between bank entry, bankruptcy, and creditor rights, which are likely to facilitate lenders ability to influence the bankruptcy process. Starting in 1993, the Indian government introduced specialized courts to speed up the resolution of debt recovery claims. Using the staggered introduction of these courts to capture changes in creditor rights across regions and time, we find that the positive association between private bank entry and bankruptcy filings is larger when creditor rights are stronger. Moreover, while foreign bank entry is largely unrelated to bankruptcy outcomes, on average, it is associated with an increase in filings in regions with stronger creditor rights. Our findings do not appear to be driven by time-varying, district-level characteristics that may be related to both bank entry choices and bankruptcy outcomes. 2 The findings are robust to including controls for various measures of district-level growth, which may be associated with both higher bank entry and firm bankruptcies. The findings are also robust to adding controls for the amount of political loans in a district. This suggests that the increase in bankruptcy filings not driven by private banks entering regions where incumbent government banks are burdened by underperforming political loans. Lastly, bank entry into districts with better enforcement of bankruptcy laws is unlikely to explain our results. Federal bankruptcy regulation applies equally to all firms in India and is enforced at the federal level by one bankruptcy court, the BIFR. The findings also do not appear driven by a simultaneity bias where a shift in bankruptcies instead drives subsequent banking sector entry. We find that the increase in bankruptcy filings occurs between one to two years after bank entry, suggesting that bank entry precedes the observed increase in bankruptcy filings. It is also not clear how simultaneity bias would explain the other results on bankruptcy restructurings, liquidations, and duration. Our findings also do not appear to be driven by bank entry changing the supply or allocation of credit, which may affect product market competition and the type of firms receiving credit. The findings are robust to controlling for the district level supply of credit, and we do not find evidence that bank entry is related to the types of firms filing for bankruptcy. Our main results are also robust to controlling for the financial characteristics of firms filing for bankruptcy. Overall, the evidence instead suggests that bank entry affects the incentive of creditors to aggressively pursue repayment and speedily recover loans. The increase in filings following private bank entry is consistent with firms seeking the protection of bankruptcy to avoid increased creditor scrutiny. 3 The shift to restructurings away from liquidations, which are typically subject to lengthy legal challenges, and the significant decrease in delays also appear to reflect private lenders greater incentives to recover assets more quickly by making concessions to resolve a restructuring decision. The lack of any shift in bankruptcy outcomes among government firms, however, suggests that private banks may be limited in their ability to pursue delinquent government-owned firms. Likewise, the evidence suggests that foreign banks may face more difficulty navigating local bankruptcy procedures in the absence of strong creditor rights. A potential implication of these findings is that less competitive banking sectors and government ownership may contribute to the extreme inefficiency of some bankruptcy systems by reducing managerial incentives to monitor borrowers. This provides an interesting 2 The underlying identification concern is an omitted variable bias arising from banks entry decisions possibly being related to district characteristics that are also related to bankruptcy outcomes. While all our specifications include district-level fixed effects to capture time-invariant differences across districts, and year dummies to capture country-level trends, time-varying omitted variables could still be problematic. 3 It is widely acknowledged that the stay on all creditor claims once a firm files for bankruptcy in India creates an incentive for firms to file in order to avoid creditors, and the automatic stay has been grossly misused by unscrupulous firms (Government of India, 2002).

4 complement to the large literature that analyzes how differences in bankruptcy laws may affect the bankruptcy process (see Hotchkiss, John, Mooradian, and Thorburn (2008) for an extensive survey of this literature). To the best of our knowledge, ours is the first paper to look at the relationship between banking sector characteristics and bankruptcy outcomes. Strahan, 2006), product market competition (Bertrand, Schoar, and Thesmar, 2007), and the proportion of bad loans (Guiso, Sapienza, and Zingales, 2007). There is also evidence to suggest that government ownership of banks is associated with less developed financial markets and slower economic growth (La Porta, Lopez-de-Silanes, and Shleifer, 2002), less efficient banks Our findings also contribute to the law and finance literature by suggesting that an (Barth, Caprio, and Levine, 2004), financing constraints (Beck, Demirguc-Kunt, and increase in creditor incentives to monitor borrowers combined with stronger creditor rights may affect the use of bankruptcy. Claessens and Klapper (2005) find that controlling for judicial efficiency, bankruptcy tends to be used more frequently in countries with stronger creditor rights. Their findings suggest that institutional characteristics, such as the creditor rights, can affect the resolution of financial distress, whereas our findings suggest that the effect of creditor rights may also depend on the competitive characteristics of the financial sector. The evidence presented here also offers insight into why outcomes may vary within a given bankruptcy process. Existing empirical studies have focused on the complexity of debt arrangements (Gilson, John and Lang, 1990), managerial incentives (Eckbo and Thorburn, 2003), and differences across bankruptcy judges (Chang and Schoar, 2007). In contrast, we provide evidence that bank entry and ownership may also affect bankruptcy outcomes and the ex-post monitoring of borrowers. This focus on the role of creditors is related to recent evidence that creditors adjust their lending practices ex-ante at the time of loan origination in response to differences in creditor rights (Qian and Strahan, 2007 and Davydenko and Franks, 2008). Finally, our findings are related to the large literature on the effects of banking sector competition. 4 Recent evidence suggests that banking sector competition may affect firms access to credit (Petersen and Rajan, 1995, Beck, Demirguc-Kunt, and Maksimovic, 2004, and Zarutskie, 2006), entrepreneurship (Black and Strahan, 2002), economic growth (Jayaratne and Strahan, 1996, Cetorelli and Gambera, 2001), firm size and market structure (Cetorelli and 4 There is also a large literature on the effects of financial market liberalization (see Bekaert, Harvey, and Lundblad, 2005). Maksimovic, 2004), politically motivated lending (Sapienza, 2004), and inefficient capital allocation (Morck, Yavuz, and Yeung, 2008). Our study suggests that banking sector entry and ownership may also affect the ex-post monitoring of borrowers and bankruptcy outcomes, which may have important implications for firms financing and investment choices. Our paper is organized as follows: Section 2 describes the Indian bankruptcy process and banking sector, Section 3 describes the data, Sections 4 describes the main results, Section 5 analyzes creditor rights, Section 6 provides robustness checks, and Section 7 concludes. 2 The Bankruptcy Process and Banking Sector in India 2.1 Bankruptcy in India Our analysis of bankruptcy outcomes and bank entry makes use of two regulatory acts in India: the Sick Industrial Company Act (SICA) and the Recovery of Debts due to Banks and Financial Institutions Act (RDDBFIA). The SICA governs bankruptcy procedures in India, while RDDBFIA provides regional variation in creditor rights. We now describe these two acts Sick Industrial Company Act of SICA governs the vast majority of cases and is the most commonly used process for bankruptcy filings in India (Panagariya, 2008). 5 SICA applies to all industrial firms that employ more than 50 workers and have been in operation for over 5 years. While firms filing for 5 Other regulatory acts that govern bankruptcies in India include the Companies Act (2002) and Securitization and Reconstruction of Financial Assets and Enforcement of Security Interests Act (2002) (SARFAESI). The Companies Act was to address weaknesses of SICA but is yet to be implemented because of legal challenges. SARFAESI allowed secured creditors to recover assets without court intervention. We discuss the potential effect of this law later.

5 liquidation and reorganization in the U.S. do so separately under Chapters 7 and 11, SICA provides a platform for both the renegotiation of loans and liquidation of the firm. When a company files under the jurisdiction of SICA, it files at the Board for Industrial and Financial Reconstruction (BIFR), which is the federal bankruptcy court. The BIFR, which is located in New Delhi, oversees all SICA bankruptcy cases in India. 6 It is mandatory for the board of directors to file at the BIFR once the firm is sick, where a firm is determined to be sick when its book value of equity is equal to zero. However, SICA also allows a firm to file at the BIFR even if the balance sheet does not show that the firm is sick. As long as the firm s directors are of the opinion that the firm will meet the criteria of sickness in the near future, the firm may file (SICA, Section 15(1)). Once the BIFR validates that a filing firm is sick, an attempt to reorganize the firm is initiated. 7 Similar to Chapter 11, the board of directors remains in control during this process. The management may propose the initial plan to reorganize the firm. If this plan is acceptable to all creditors, the plan is sanctioned. If creditors do not agree with the management s plan, and if the BIFR believes that it is in the public interest to reorganize the firm, the BIFR appoints an Operating Agency (OA) to examine the turnaround possibility. The OA consults with other creditors while preparing a reorganization proposal to be submitted to the BIFR. The proposal must be accepted by all parties making concessions in order to be approved. If this proposal is not accepted, the BIFR may recommend liquidation of the firm. To do this, the BIFR may either prevents creditors from taking any legal action against the firm until the filing is resolved, but it takes a year, on average, for the BIFR to determine whether a firm is actually in financial distress and will be admitted in the restructuring/liquidation process. 8 This long delay creates incentives for firms to file for bankruptcy to avoid paying creditors (Government of India, 2002), and as evidence of this, about 30% of the filings made to the BIFR are eventually dismissed because the firms do not meet the criteria for financial distress. It is widely acknowledged in India that the bankruptcy system has been abused by firms seeking to avoid their creditors. 9 Creditors may also force a firm to file for bankruptcy, but generally, there is little incentive for creditors to do so because of the stay on assets and long delays. 10 The average time to a court decision about either restructuring or liquidating a firm is 4 years. Restructuring decisions are typically delayed because of a lack of cooperation from creditors when asked to make concessions (Kang and Nayar, , Panagariya, 2008). There are no provisions under SICA to divide creditors into classes or to force them to accept a plan (even if the majority of creditors agree upon a plan). Liquidations, however, involve particularly long delays. Even after the BIFR issues a liquidation order, firms can appeal the decision in the civil courts adding even more time to the process. Panagariya (2008) notes that 48% of these cases take more than 10 years to complete and 10 percent take about 25 years (page 293) Recovery of Debts due to Banks and Financial Institutions Act, The RDDBFIA was enacted in 1993 by the Indian government to strengthen creditor forward its opinion to the civil courts, or it may proceed with the sale of assets and remit the proceeds to the High Court for distribution. A key incentive for firms to file with BIFR is a combination of long delays and an automatic stay on assets that takes place immediately upon filing. The automatic stay of assets 6 For comparison, there are 94 bankruptcy courts in the United States (Hotchkiss, John, Mooradian, and Thorburn, 2008). 7 The initial enquiry to decide whether the firm is sick is made in a meeting that includes representatives from the company, labor unions, financial institutions, and the state and federal governments (Goswami, 1996, page 51). If the BIFR finds that the firm has not met the criteria, the filing is dismissed. 8 The commonly cited reasons for the delays include the BIFR s high workload (and small staff) and delaying tactics used by the firm such as not providing financial records. 9 A news article noted, The best way for Indian corporates to avoid repayment of loans to financial institutions seems to be the BIFR route, (Indian Express, June 27, 1999). And in the case of Richmen Silks, the BIFR dismissed the filing noting in its ruling, The sole motive of filing a reference was to deny the secured creditors the opportunity to recover their dues (The Hindu, March 18, 2001). 10 The same is true in the United States, where bankruptcy filings are mainly initiated by managers (Hotchkiss, John, Mooradian, and Thorburn, 2008). 11 Because of these delays, a majority of banks and financial institutions in India remain highly dissatisfied with the BIFR: In a meeting on August 30 in the banking division, Finance Ministry, under whose supervision BIFR functions, banks and financial institutions almost unanimously asked the Government to down shutters on the board, official sources said, (The Daily Excelsior, September 14, 1999).

6 rights by introducing new rules for the recovery of large debts. 12 Under this act, banks can begin attempting to recover outstanding debts greater than Rs. 1 million (approximately $22,000) by filing a petition before a specialized court known as a Debt Recovery Tribunal (DRT) rather than with a Civil Court, as was previously done. Using a streamlined procedure laid out by the Indian government, the DRTs were intended to speed the process by which banks seek an initial court government spending and serving priority sectors such as agriculture, India s public banks lacked proper lending incentives and exhibited a high number of non-performing loans. Following a balance of payments crisis in 1991, a number of structural reforms were implemented that greatly deregulated many economic activities in India. In 1991, a broad financial reform agenda was established in India by the Committee on the Financial System ruling against a delinquent firm. 13 It is important to emphasize, however, that a lender s attempt (CFS). One of the recommendations was to introduce competition into the banking sector. The to recover outstanding debts through a DRT does not prevent a firm from seeking bankruptcy protection from the BIFR. Similar to when a firm receives an unfavorable ruling in the Civil Courts, they can still seek bankruptcy protection at the BIFR if the DRT rules against the firm. Since the DRTs were intended to speed up the debt recovery process, we use their introduction across the different Indian states in different years to capture an improvement in creditor rights. Such an improvement likely increases the ability of lenders to influence bankruptcy outcomes. Because of legal challenges, however, the DRTs were introduced in a staggered way across the different Indian states after 1993, allowing us to make use of this exogenous regional and time variation in our subsequent analysis. 2.2 Banking Sector Reforms Our analysis also makes use of banking deregulation that facilitated the entry of private and foreign banks into India. Prior to 1991, India s economy and financial markets were heavily regulated. A complicated regulatory regime required firms to obtain licenses for most economic activities, and many industries were reserved for the public sector, including much of the financial system. Bank nationalizations in 1969 and 1980 increased the public sector share of deposits to over 80%, and branch licensing was rigidly controlled. Primarily focused on financing reforms also deregulated deposit rates, reduced requirements that banks invest in government securities, and eliminated regulatory barriers protecting government banks from competition in the market for long-term loans. Another key recommendation of the CFS was to allow entry by private domestic and foreign banks. It was argued that private bank entry would improve the competitive efficiency and productivity of the Indian banking system. New private bank entry guidelines were established in 1993, and in 1994, the Indian government allowed foreign bank entry under the WTO General Agreement on Trade in Services. While there were no restrictions on where foreign banks could establish new branches, their expansion was by de novo branches only as foreign banks were not allowed to own controlling stakes in domestic banks. The reforms led to an increase in domestic private and foreign bank entry, which appears to have led to significant improvements in the efficiency of government banks (Bhaumik and Dimova, 2006). 14 On March 31, 1994 there were 24 foreign banks with 156 branches in India. In the eight years following the acceptance of GATS, the number of foreign banks increased to 41 with 212 branches as of March Private domestic banking exhibited an even larger increase. Twelve new private banks and 1,700 new branches were added between 1993 and The official policy noted that: Recovery of dues due to banks and financial institutions is not given any priority by the civil courts. The banks and financial institutions like any other litigants have to go through a process of pursuing the cases for recovery through civil courts for unduly long periods (Government of India, 2000). 13 See Visaria (2008) for a description of the debt recovery tribunals. 14 An editorial in the Economic and Political Weekly noted: Financial sector reforms forced banks to confront the quality of their loans and wake up to the reality of huge and rapidly growing NPAs [nonperforming assets], (EPW Editorial, July 13, 2002).

7 3. Data To analyze the relationship between banking sector characteristics and bankruptcy in India, we make use of two datasets. The first is a dataset on on banks locations, deposits, and loans in India, and the second is a detailed dataset on the population of bankruptcy filings in India between 1991 and We now describe these two datasets. 3.1 Bank Entry and Ownership Data The data on bank entry and ownership is published by India s central bank, the Reserve Bank of India (RBI). These quarterly data provide the outstanding deposits and loans and the number of bank branches in each district by bank ownership group between the years 1991 and Bank ownership categories include state and nationalized (government) banks, private domestic, foreign, and regional rural banks. Using data from the fourth quarter of each year, we construct several measures of banking sector characteristics at the district and year level. First, Total Banks/Population is the ratio of the number of bank branches in that district and year to district-level population (in millions), which is obtained from India s 2001 census. To distinguish the impact of bank entry by ownership, we use Private Bank % of Deposits, which measures the share of total deposits in that district held by domestic private banks, or the ratio of total deposits held by all private banks in a district to aggregate bank deposits in that district; State Bank % of Deposits, which is the share of deposits held by government banks; and Foreign Bank % of Deposits, which is the share of deposits held by foreign banks in each district and year. From the summary statistics reported in Table 1 we note that on average there are 67 bank branches per million people across all Indian districts during our sample period. Government banks dominate the banking sector, accounting for an average of 82% of all deposits, In Table 2 we compare average banking sector characteristics at the district level in 1991, when the reforms were initiated, and at the end of the sample in Bank entry deregulation led to a substantial increase in the deposit share at the district level of private domestic banks from an average of 4% in 1991 to 18% in 2004, while the deposit share of government banks decreased from 86% to 74% over the same period. While the overall geographical presence of foreign banks increased following deregulation, their relative market share did not increase because their entry was limited to fewer branches Bankruptcy Data The bankruptcy data for our study is collected from the Board for Industrial and Financial Reconstruction (BIFR) in New Delhi, and includes the population of bankruptcy filings at the BIFR since Of the 4,212 firms that filed for bankruptcy during this period 1,327 firms were found to not meet the criteria for financial distress and were dismissed; 1,707 firms were determined to be sick and were admitted into the bankruptcy process; and 842 firms were still pending an initial decision about whether they met the criteria of financial distress. Of the 1,707 firms admitted, 992 firms were approved for liquidation, and the remaining 715 were either approved for restructuring or were still undergoing negotiations with lenders for restructuring. We use the year that a firm files for bankruptcy and the district-level location of its head office to construct a number of bankruptcy outcome variables for each district and year. Our primary measure is total corporate filings for bankruptcy in a given district and year per million persons (Total Filings/Population). We also disaggregate the filings into a number of categories: filings that are dismissed for not meeting the criteria of financial distress (Dismissed Filings/Population); filings where the firm has been validated as financially distressed (Sick Filings/Population); and filings pending determination of financial distress (Pending domestic privately-owned banks account for about 6%, and foreign banks account for 0.15% of total deposits over this period. The remaining share is held by rural banks, which we ignore as they are government sponsored banks that finance small, agricultural loans. 15 Because of limits on the number of new foreign bank branches allowed under India s agreement with the WTO, foreign banks entered just 8 new districts between 1991 and 2002, and 9 additional districts between 2002 and Another 33 foreign bank branches closed during this time period, largely because of closures initiated by ANZ Grindlays Bank Ltd. and Standard Chartered Bank.

8 Filings/Population). The sick filings are further disaggregated into those where the firm and lenders are negotiating or have negotiated a restructuring agreement (Workouts/Population), and firms that are ordered to be liquidated by the bankruptcy court (Liquidations/Population). To examine the relationship between bank entry and bankruptcy delays, we construct the average number of days taken for a successful restructuring decision (Average Duration of Workouts); average number of days taken for a liquidation order (Average Duration of Liquidations); and, the average number of days taken for either a workout or liquidation decision to be rendered by the BIFR (Average Duration of Workouts and Liquidations). Table 3 reports the descriptive statistics for these bankruptcy measures. On average, a district has 0.16 bankruptcy filings per million people in a year. Of the firms that file for bankruptcy, about 50% are determined to be financially distressed and are admitted to the BIFR. Another 31% of filings are dismissed for not meeting the criteria for financial distress, and 19% await a decision from BIFR about whether they meet the criteria for financial distress. Among firms that meet the criteria of financial distress and are admitted into the BIFR for bankruptcy proceedings, about one-half are ordered by the court to be restructured and the rest to be liquidated. On average, it takes more than 4 years to obtain a restructuring or liquidation decision once a firm files with the BIFR. But as noted earlier, subsequent to the BIFR s decision, liquidations are carried out in the civil court system, which can take up to 10 years or longer (Panagariya, 2008). Such delays in the civil courts would not be captured by our data. 4. Empirical Specification and Results To examine the relationship between banking sector competition and bankruptcy Bank Environment captures local banking sector entry by ownership in that district and year. We include district fixed effects, d, to control for time-invariant district characteristics that may explain the incidence and outcome of bankruptcies in that district, and we include year fixed t effects,, to control for any country-level changes in the bankruptcy process. 16 Other timevarying controls are included in X dt, and the standard errors are clustered at the district-level. We use a variety of measures to control for other factors that may also affect the incidence and outcomes of bankruptcy, To control for economic growth, we use the state-level GDP in each year and state, Log(GDP of State). As district-level GDP figures are not available in India, we include the number of bank branches per million people, Total Banks/Population as a another proxy. All subsequent results are also robust to using Log(Total Bank Credit) as an alternative proxy for district-level growth. 4.1 Bank entry and bankruptcy outcomes Using this framework, we investigate the relationship between bank entry in a region and corporate bankruptcy outcomes in that region. In Table 4, we find that an increase in the share of deposits held by domestic private banks in a district is associated with an increase in bankruptcy filings (column 2), but an increase by foreign banks is not (column 3). The positive association between the number of filings and private bank entry is robust to controlling for state-level growth, total banks per person, and district and year fixed effects Mirroring the private bank estimates, an increase in government-owned banks share of deposits is associated with a reduction in bankruptcy filings (column 1). 17 The increase in filings appears driven by an increase in dismissed filings rather than an outcomes, we estimate the following OLS district-level panel regression with district and year fixed effects for the period : Bankruptcy Outcome Bank Environment X (1) dt 0 1 dt dt t d dt where Bankruptcy Outcome is the bankruptcy measure of interest for district d, in year t, and 16 Two related reforms undertaken during this period are the RDDBFI Act introducing the debt recovery tribunals beginning in 1993 (which we explore below) and the SARFAESI Act of Since SARFAESI was implemented at the national level, any effect it might have on overall bankruptcies at the BIFR will be captured by the time dummies. Our findings below are also robust to dropping post-2001 bankruptcies. 17 Only 28 of the 565 districts experience an increase in the percent of deposits held by government banks. Since changes in the deposit share of government banks are largely driven by changes in the share of private banks, this coefficient will generally mirror the private bank results.

9 increase in financially distressed firms. An increase in private bank entry is associated with an increase in the number of firms dismissed by the BIFR because they are not financially distressed (Table 4, column 2), but there is no change in the number of financially distressed firms (column 4). There is also an increase in the number of filings still pending determination of financial distress (column 3). The increase in dismissed filings without an increase in financially distressed firms suggests that more firms are seeking protection from creditors by filing for bankruptcy in response to greater creditor monitoring, rather than there actually being an increase in the number of financially distressed firms. As noted earlier, the automatic stay on assets along with the long delays before the BIFR reaches a decision creates an incentive for firms to strategically file for bankruptcy to avoid paying creditors (Government of India 2002; Panagariya, 2008). While there is no increase in the number of sick firms, there is a shift in their outcomes. Of the firms that are declared sick, private bank entry is associated with a decrease in the number of liquidations (Table 4, column 5), and correspondingly an increase in the number of restructuring decisions or workouts (column 6). 18 As shown in Table 6, column 2, a one percentage point increase in the share of deposits held by private banks is associated with an additional workouts per million persons (average workouts for the sample are per million persons). Since the number of sick firms does not increase with private bank entry, the increase in workouts must come from a decrease in liquidations. This is in fact what we observe in column 5 of Table 6. There is no discernible effect of foreign bank entry on the relative use of workouts and liquidations (columns 3 and 6), and as expected, the estimates for government banks largely mirror those of the private banks (columns 1 and 4). The observed increase in bankruptcy filings following private bank entry may increase the burden on the bankruptcy process, leading to more delays. To investigate this possibility, we analyze the impact of bank entry on the average number of days taken for the bankruptcy court to 18 By construction, the number of sick firms matches the sum of workouts and liquidations. After a firm is declared sick, firms and lenders must either agree to enter into negotiations regarding a workout, or the lenders can make a request to have the firm liquidated. sanction a restructuring scheme or a liquidation proposal. The results, reported in Table 7, show that despite the increase in bankruptcy filings, private bank entry is significantly and negatively related to average duration of bankruptcy (column 2). For example, a one standard deviation increase in the share of deposits held by private banks is associated with an average duration that is 422 days shorter. Foreign bank entry is also associated with a drop in average duration, but the coefficient is not statistically significant at conventional levels (Table 7, column 3). The drop in the average duration of a workout or liquidation decision is not just driven by a shift away from liquidations. This can be seen in columns 5 and 8 of Table 7 where we consider average duration for workouts and liquidations separately. Private bank entry is negatively related to duration for both types of outcomes, and the magnitudes are also economically significant. A one standard deviation increase in the deposit shares of private banks is associated with an average decrease of 609 days (20 months) in the duration of workouts, and a decrease of 414 days (13 months) for liquidations. Lastly, the average duration of workouts also appear to be much shorter in growing states, as indicated by the statistically significant negative coefficient of Log(State GDP). This may reflect an improvement in firms ability to repay or a greater willingness of lenders to renegotiate in a growing economy. The relationship between Log(State GDP) and the average duration of liquidations is also negative but not statistically significant. 4.2 Firm ownership and bankruptcy outcomes Bankruptcy outcomes also appear to vary based on the ownership of filing firms. These estimates are reported in Table 8. The share of deposits held by private banks is positively related to the number of filings by private firms (column 2), but negatively related to the number of filings by government firms (column 5). Government bank estimates largely mirror those of the private banks, while the share of deposits held by foreign banks is not significantly related to bankruptcy filings. In unreported results, we find that the shift away from liquidation towards more workouts is driven by the filings of private firms and not government firms. Restricting the

10 sample to bankruptcy filings by government firms, we find that both workouts and liquidations decline with an increase in the share of deposits held in private banks. 5. Interpretation of Evidence In this section, we provide an initial interpretation of the findings. We then make use of the variation in local creditor rights to further test the potential role of creditors. 5.1 The Potential Role of Creditors The observed correlations between changes in banking sector characteristics and changes in bankruptcy outcomes are consistent with the possibility that creditors incentives, which may be influenced by local competition or ownership, affect the bankruptcy process. Lenders in a more competitive environment may have greater incentives to aggressively monitor their loans and pursue delinquent firms. It is also likely that private banks have greater incentives to pursue delinquent loans relative to that of state banks. The evidence appears consistent with both possibilities. The increase in filings, particularly dismissed filings, is suggestive that lenders are more aggressively pursuing delinquent firms when there are more private banks present or when banking sector competition is greater (as measured by the HHI). There are two possible channels by which private bank entry may affect bankruptcy outcomes. One possibility is that private banks more aggressively pursue delinquent firms than government banks. But since private banks account for a small fraction of loans and deposits relative to government banks, our results might reflect more than this. A second possibility is that private bank entry, and the associated increase in competitive pressures, may increase the incentives of incumbent banks to monitor their loans more intensively. The shift away from liquidations and towards restructurings and reduced durations are also consistent with the possibility that lenders incentives affect bankruptcy outcomes. Since liquidation decisions are highly contested and appealed over several years in the courts (Panagariya, 2008), more competition among banks may create an incentive for creditors to make additional concessions in restructuring negotiations so as to avoid lengthy liquidations. The drop in average duration of bankruptcy is also consistent with lenders trying to recover assets more quickly. One of the commonly cited reasons for the long delays at the BIFR is that creditors are slow to respond to requests for making concessions in workouts (Kang and Nayar, ; Panagariya, 2009). The large decrease in duration, particularly among workouts, suggests that making concessions may be one way in which banks attempt to recover assets more quickly. The findings with respect to foreign banks and government-owned firms suggest that there may be limitations to lenders ability to affect bankruptcy outcomes. The lack of an association between foreign bank entry and bankruptcy outcomes is suggestive that foreign banks may be relatively disadvantaged in navigating the bankruptcy process. The differential impact of private bank entry on government firms also suggests that private banks may be limited in their ability to pursue politically connected government-owned firms Additional Tests using Changes in Creditor Rights To further test the possibility that lenders incentives may affect bankruptcy outcomes, we make use of local variation in creditor rights caused by the RDDBFIA. As noted earlier, the Indian government introduced specialized courts, the Debt Recovery Tribunals (DRTs), after 1993 to speed up the initial debt recovery process for lenders. Since the establishment of DRTs was staggered over time across states, we define a variable DRT in State to be equal to one if the government has set up a debt recovery tribunal in that state by that year, and equal to zero otherwise. 20 We then interact this DRT variable with our banking sector measures and cluster the error term at the state level instead of at the district level, so as not to overestimate the impact of the presence of a DRT. If our findings are being driven by lenders affecting bankruptcy 19 This finding may also reflect a lower willingness among private banks to finance government firms, thus reducing the number of future delinquencies by these firms. However, this cannot explain the entire drop since we would then observe an initial increase in government bankruptcy filings as private banks restrict lending to these firms, which we do not. 20 While the courts may have jurisdiction over neighboring states, we consider whether a particular state has a DRT in order to capture the fact that distance from the court may affect the rights of creditors.

11 outcomes, we would expect the correlation between banking entry and the increase in filings to be stronger when creditors have a greater ability to pursue delinquent firms, as captured by the presence of a DRT in the state of the firm. These estimates are reported in Table 9. Consistent with the possibility that lenders influence bankruptcy outcomes, we find that strong creditor rights, as captured by having a DRT present, is positively related to an increase in bankruptcy filings. The positive association between private bank entry and bankruptcy filings becomes even larger when a DRT is present (Table 9, column 3). Moreover, foreign bank entry is associated with an increase in filings when creditor rights are strong but negatively associated when a DRT is not present (column 4). This suggests that creditor rights may be particularly important for foreign banks, and further corroborates our interpretation that creditor incentives to monitor borrowers can affect bankruptcy outcomes The DRTs, however, appear to have little direct effect on bankruptcy filings. The main coefficient for DRT in State is not statistically different from zero in any of the estimates. In unreported results, we find that the increase in filings in districts with both stronger creditor rights and more private bank entry is driven by an increase in either dismissed or still pending filings and not by an increase in distressed firms. This again suggests that these firms file to escape increased creditor scrutiny. We also find that foreign bank entry in districts with stronger creditor rights is associated with an increase in the number of workout decisions but not the number of liquidation decisions made by the bankruptcy court. are also related to bankruptcy outcomes, the specification may suffer from an omitted variable bias. However, the potential source of omitted variables in our estimation is very specific. Timeinvariant differences across districts that are correlated with both bank entry and bankruptcy outcomes will be captured by the district-level fixed effects, and country-level trends in both bank entry and bankruptcy outcomes will be absorbed by the inclusion of year fixed effects. The source of omitted variables is thus limited to district-level characteristics that are related to both banking entry and differential trends in bankruptcy outcomes within districts over time. One such omitted variable may be that private banks are more likely to enter districts with greater future growth potential. If higher growth is also associated with greater firm exit rates, then our result that bankruptcies increase following private bank entry may reflect these banks choice to enter high-growth potential districts. The observed increase in bankruptcy filings, however, does not appear driven by differential growth prospects across districts. We find that using the log of total loans at the district level to capture district-level growth and financial development does not affect our results (Table 10, column 1). The increase in bankruptcy filings is also robust to allowing districts to trend differently based on their level of urbanization (column 2). 21 This suggests that the relation between bank entry and bankruptcy outcomes is not driven by a concentration of bank entry in India s growing, urban districts. 22 In unreported results, we also find that our results are robust to adding literacy-year interactions, which is another measure of areas likely to experience greater future growth. The findings are also robust to using state-year fixed effects to non-parametrically 6. Robustness Tests and Alternative Specifications In this section we discuss the robustness of our results to potential identification issues, and the alternative channels by which bank entry may affect corporate bankruptcy outcomes. 6.1 Identifying the impact of bank entry A potential identification concern is that banks decisions to enter a district are likely to be related to various unobservable characteristics of that district. To the extent that these factors capture differential trends in bankruptcy filings across India s states (column 3). Our earlier findings also suggest that banks decision to enter high-growth areas is 21 Our measure of urbanization is an indicator that is equal to one if a district s share of citizens located in urban areas is in the top quartile according to the 2001 Census. Since the census data is only available for 2001 we interact this indicator with year dummies to control for differential time trends. Our results are similar if we define urbanization more broadly (above median) or more narrowly (top decile). 22 This result also suggests that changes in bankruptcy outcomes are unlikely to be driven by new bank entry being concentrated in urban districts with rising real estate prices, which might affect the incentives of lenders to liquidate firms and recover underlying assets.

12 unlikely to explain the overall increase in filings. If the increase in bankruptcy filings reflected new firm entry driving out unviable firms, we would observe an increase in the number of financially distressed firms, which we do not. It is also unclear why greater growth at the district level leading to more firm exits would be correlated with more workouts, fewer liquidations, and a decrease in the average duration for workouts and liquidations. Lastly, our finding that the increase in bankruptcy filings following bank entry is even greater in states with stronger creditor rights cannot be explained by higher growth districts. Another potential omitted variable bias might arise if bank entry is more likely to occur in districts where incumbent banks are captured by politically connected firms and are burdened by underperforming political loans. These districts may be a target for new banks if there is a large unsatisfied demand for credit in the region. But, these districts may also experience more future bankruptcies, irrespective of bank entry, if politically-connected, underperforming borrowers of incumbent banks are more likely to end up in bankruptcy. Our findings are robust to adding proxies for the presence of political loans in a district, which we measure using the share of borrowings and the share of sales accounted for by government-owned firms in each district-year. We calculate these measures using information on firms covered in Prowess, a firm-level financial dataset covering nearly 80% of industrial output in India. 23 In Table 11 we show that private bank entry remains positively associated with increased bankruptcy filings after controlling for the presence of political loans (column 1). The results are also robust to using the market share of government-owned firms in a district to control as a proxy for the presence of politically-connected firms (column 2). Lastly, as another check on more bank entry occurring in districts with higher growth rates as discussed above, we find that controlling for overall output growth at the district level with sale revenues of firms in that district, Ln(Sales), does not affect our results (column 3). In unreported results we also investigate whether private bank entry is more likely to occur in districts with more politically motivated loans. Specifically, we look at the relationship between average private bank deposit shares at the district level in the post-reform years, and the proportion of bank branches in a district that are government-owned in the pre-reform years. We find that private bank deposit shares in the post-reform era are not significantly related to the share of government bank branches in that district in the pre-reform era. This suggests that private banks do not selectively choose to enter districts dominated by government banks, which typically make more politically motivated loans. Our earlier findings also suggest that selective entry into districts with a disproportionate amount of political loans is unlikely to explain our results. The observed decrease in government firm bankruptcies would appear to contradict this potential explanation. Moreover, private banks entering districts with more political loans cannot easily explain the increase in workouts, the drop in liquidations, or the reduced average duration. Political loans also cannot explain why bankruptcy filings increase more following bank entry in districts with stronger creditor rights. Another issue is whether bank entry is more likely to occur in districts where the enforcement or stringency of bankruptcy requirements are increasing at a greater rate. Such improvements in the law or its enforcement would be directly related to changes in bankruptcy outcomes and could also influence bank entry decisions. Differential changes in bankruptcy requirements or enforcement, however, are unlikely to explain our results. As described earlier, the BIFR was set up under the Sick Industrial Companies Act of 1985, which applies to all firms in the economy. All bankruptcy rules and procedures at the BIFR, including the financial criteria for determining sickness and the procedures for obtaining a restructuring or liquidation order, are done at the national level. Moreover, the members of the BIFR court, which is located in the Indian capital of New Delhi, 23 Compiled by the Centre for Monitoring Indian Economy (CMIE), Prowess is a panel of both listed and unlisted public limited companies with assets plus sales greater than 40 million Rupees (approx. $900,000), and covers between 2,000 to 6,000 firms each year from

13 are typically career bureaucrats appointed by the federal government. 24 This centralized decision process minimizes the chance of enforcement variation at the district level. As we describe in Table 9, our findings are also robust to controlling for a potential source of variation in creditor rights at the district level, the presence of a debt recovery tribunal. 6.2 Identifying the direction of the relationship An alternative interpretation of our results is that the shift in bankruptcies drives banking sector entry, rather than vice versa. This issue may arise if increased bankruptcies adversely affect the health of incumbent banks by reducing their capital reserves, deposit base, and ability to meet local firms demand for loans. New bank entrants, which are not burdened with these underperforming legacy loans, may choose to enter these districts. Bankruptcies of older firms may also clear the way for entry by young, upstart firms, which may then attract the entry of new banks hoping to finance the new firms. While it is difficult to fully exclude this possibility, it is unclear why some of our other estimates with respect to bankruptcy outcomes and duration would also be subject to this type of simultaneity bias. For example, it is unclear how a shift away from liquidations towards workouts for incumbent banks might increase the entry of new banks. It is also unclear why a drop in the average duration of bankruptcy would facilitate new entry since this change in duration is unlikely to reflect any local shift in bankruptcy regulations. As a robustness check, we investigate whether increased bankruptcies preceded future entry by private banks in a district. In Table 12 we add lags and leads of the private bank entry measure to the base specification. If the prior estimates are driven by a change in bankruptcies affecting bank entry choices, then we would expect to find a positive correlation between current bankruptcy levels and future bank entry after controlling for contemporaneous entry (Wooldridge, 2001, page 285). In column 1 of Table 12, we add a measure for private bank entry in t+1, but do not find any evidence that future bank entry is correlated to current bankruptcies. Adding a measure for bank entry in t-1 (column 2), we find that past bank entry does predict current bankruptcies. In fact, we find that our measure for contemporaneous bank entry is primarily capturing the impact of previous bank entry going back two years (column 3). 25 The lagged impact of bank entry is consistent with causality running from bank entry to bankruptcies. 6.3 Channels by which bank entry may affect bankruptcy Our results are consistent with private bank entry shifting the incentive of creditors to aggressively monitor borrowers. The increase in dismissed filings following private bank entry, suggests that firms are seeking the protection of the automatic stay on assets in bankruptcy to avoid increased creditor scrutiny. The shift to restructurings from liquidations, which can take even longer to resolve, and the decrease in delays in the bankruptcy process also indicate that a change in creditor incentives to monitor borrowers may affect the efficiency of the bankruptcy process. The increase in bankruptcies associated with bank entry when creditor rights are stronger is also consistent with bank entry being related to a shift in creditors incentives. We also consider two other potential channels through which bank entry may affect bankruptcy outcomes: an increase in the supply of credit and a shift in the allocation of credit. An increase in the supply of credit associated with bank entry might reduce bankruptcies if it improves credit access, but may also increase bankruptcies if it fosters greater product market competition. A change in the types of firms receiving credit, possibly because foreign and private banks lend to different types of firms, may have similar implications. For example, bank entry might reduce credit access for opaque firms (Petersen and Rajan, 1995) and increase credit access for riskier firms (Dell Arricia and Marquez, 2009). Or, private and foreign banks may be more likely to extend loans to better firms with a higher chance of repayment. Our results, however, do not appear to be driven by changes in the supply of credit. If bank entry is associated with easier access to credit, we should observe an immediate decrease in 24 For more details on the BIFR s organizational structure and procedures see 25 In unreported estimates we find no evidence of an effect for bank entry in year t-3.

14 bankruptcy filings following entry, while we find the opposite. If improved access to credit instead fostered more product market competition, then we should observe an increase in filings by financially distressed firms as weaker firms are pushed out by the entry of new firms. However as noted in Table 4, the number of financially distressed firms does not increase following bank entry. The results are also robust to controlling for the total availability of credit measured by total bank loans at the district level. Lastly, a shift in the supply of credit or product market competition cannot explain the changes in bankruptcy outcomes and duration or their relationship with local creditor rights. A shift in the allocation of credit also does not explain a number of our results. If better firms are funded by entering private banks and everyone else continues to receive credit from incumbent banks then we should observe a decrease in filings, contrary to our results. Alternatively, if a shift in credit to better firms causes worse firms to get less credit, then we should observe an increase in sick firms, which we do not. A shift in credit to better firms also cannot explain the increase in dismissed filings or the relationship with creditor rights. Lastly, we find that the type of firms filing for bankruptcy does not change following bank entry, and controlling for firm-level characteristics does not affect our findings. In Table 13A we examine the effect of bank entry on the characteristics of bankrupt firms and do not find a significant correlation between bank entry and the financial characteristics and workforce size of firms filings for bankruptcy. In Table 13B, we examine the relationship between private bank entry and bankruptcy outcomes, controlling for the characteristics of bankrupt firms in that district. 26 The results are similar, although some effects are less significant because of the smaller sample size. 7. Concluding remarks We investigate whether entry in the banking sector is related to changes in the bankruptcy outcomes. Increased competition among banks may change creditor incentives towards monitoring or exerting pressure on defaulting borrowers, which may in turn affect bankruptcy outcomes. Our results are consistent with this possibility. We find that, on average, more firms file for bankruptcy following bank entry. The evidence suggests this increase is driven by firms trying to avoid heightened creditor scrutiny. The results also show that bank entry is associated with a shift away from liquidations towards more restructurings and a significant decrease in overall delays in the bankruptcy process. This is consistent with banks making more concessions in a more competitive lending environment in order to more quickly recover assets. Our results also show that the ownership of creditors and firms matter. For the most part, the increase in bankruptcy filings and decrease in delays is driven by the entry of private domestic banks rather than foreign banks. Private bank entry is also associated with a shift from liquidations to restructurings among private firms. However, private bank entry is associated with fewer bankruptcy filings among government-owned firms. This suggests that private banks are less willing or able to pursue delinquent government firms. An improvement in creditor rights appears to amplify these findings. We find that stronger creditor rights, in combination with private bank entry, leads to an increase in filings. Creditor rights are particularly important for foreign banks. While foreign bank entry has no association with bankruptcy outcomes, on average, it is associated with increase in bankruptcy filings when creditor rights are strong and a decrease when they are weak. This finding suggests that foreign lenders may face relatively greater hurdles in navigating the local bankruptcy system in the absence of strong creditor rights. Overall, our evidence suggests that banking sector competition and lenders incentives play an important role in addition to bankruptcy regulations and creditor rights in bankruptcy 26 The data on firm characteristics is restricted only to firms that file for bankruptcy, which will reduce the sample size as bankrupt firms are not located in all Indian districts. outcomes. Lenders without a strong incentive to aggressively monitor loans, such as government

15 banks and banks that face relatively little competition in their local market, may contribute to the inefficiency of some bankruptcy systems. An important implication of these findings is that bankruptcy reforms that focus on changes in the law should also take into account local financial market factors such as the competitiveness of the banking sector. References Barth, James R, Gerard Caprio, and Ross Levine, 2004, Bank regulation and supervision: what works best? Journal of Financial Intermediation, 13 (2), Beck, Thorsten, Asli Demirguc-Kunt, and Vojislav Maksimovic, 2004, Bank Competition and Access to Finance: International Evidence, Journal of Money, Credit, and Banking, 36(3), Beck, Thorsten, Asli Demirguc-Kunt, and Ross Levine, 2006, Bank Concentration, Competition and crises: First Results, Journal of Banking and Finance, 30(5), Bekaert, G., Harvey, C., Lundblad, C., Does financial liberalization spur growth? Journal of Financial Economics 77, Berger, Allen N., Asli Demirguc-Kunt, Joseph. G. Haubrich, and Ross Levine, 2004, Bank Concentration and Competition: An Evolution in the Making, Journal of Money, Credit, and Banking, 36(3), Berger, Allen N. and Timothy H. Hannan, 1998, The Efficiency Cost of Market Power in the Banking Industry: A Test of the Quiet Life and Related Hypotheses, The Review of Economics and Statistics, 80 (3), Bertrand, Marianne, Antoinette Schoar, and David Thesmar, 2007, Banking Deregulation and Industry Structure: Evidence from the French Banking Reforms of 1985, Journal of Finance, 62 (2), Bhaumik, Sumon K. and Ralitza Dimova, 2006, How Important is Ownership in a Market with Level Playing Field? The Indian Banking Sector Revisited, Journal of Comparative Economics, 32(1), Black, Sandra E., and Philip E. Strahan, 2002, Entrepreneurship and Bank Credit Availability, Journal of Finance, 57(6), Cetorelli, Nicola and Michele Gambera, 2001, Banking Market Structure, Financial Dependence and Growth: International Evidence from Industry Data, Journal of Finance, 56(2), Cetorelli, Nicola and Philip E. Strahan, 2006, Finance as a Barrier to Entry: Bank Competition and Industry Structure in Local U.S. Markets, Journal of Finance, 61(1), Chang, Tom, and Antoinette Schoar, 2007, Judge Specific Differences in Chapter 11 and Firm Outcomes, mimeo. Claessens, Stijn, and Leora F. Klapper, 2005, Bankruptcy around the World: Explanation of its relative use, American Law and Economics Review, 7(1), Davydenko, Sergei A. and Julian R. Franks, 2008, Do Bankruptcy Codes Matter? A Study of

16 Defaults in France, Germany, and the U.K., Journal of Finance, 63(2), Dell Ariccia, Giovanni and Robert Marquez, 2009, Risk and the Corporate Structure of Banks, forthcoming, Journal of Finance. Djankov, Simeon, Oliver Hart, Caralee McLiesh, and Andrei Shleifer, 2008, Debt Enforcement Around the World, Journal of Political Economy, 116 (6), Eckbo, B. Espen, and Karin S. Thorburn, 2003, Control Benefits and CEO Discipline in Automatic Bankruptcy Auctions, Journal of Financial Economics, 69 (2), Editorial, July 13-19, 2002, NPAs and Beyond, Economic and Political Weekly, 37(28), Gilson, Stuart C., Kose John, and Larry H. P. Lang, 1990, The Empirical Study of Private Reorganization of Firms in Default, Journal of Financial Economics 27 (2), Goswami, Omkar, 1996, Corporate Bankruptcy in India: A Comparative Perspective (OECD Press, Paris, France). Gormley, Todd, 2008, The Impact of Foreign Bank Entry in Emerging Markets: Evidence from India, forthcoming, Journal of Financial Intermediation. Government of India, Companies Act, 1956, Government of India Publications: New Delhi, India. Government of India, Sick Industrial Company Act, 1985, Government of India Publications: New Delhi, India. Government of India, 2000, Justice Eradi Committee Report, Government of India Publications: New Delhi, India. Government of India, 2002, Industrial Development, Planning Commission of India, Government of India Publications: New Delhi, India. Government of India, 2002, Securitisation and Reconstruction of Financial Assets and Enforcement of Security Interest (SARFAESI) Act 2002, Government of India Publications: New Delhi, India. Guiso, Luigi, Paola Sapienza, and Luigi Zingales, 2007, The Cost of Banking Regulation, mimeo. Hart, Oliver, 2000, Different Approaches to Bankruptcy, National Bureau of Economic Research, Working paper Hicks, John, 1935, Annual Survey of Economic Theory: The Theory of Monopoly, Econometrica 3(1), Hoshi, Takeo, Anil Kashyap, and David Scharfstein, 1990, The Role of Banks in Reducing the Costs of Financial Distress in Japan, Journal of Financial Economics 27, Hotchkiss, Edith S., Kose John, Robert M. Mooradian, and Karin S. Thornburn, 2008, Bankruptcy and the Resolution of Financial Distress, in Handbook of Corporate Finance: Empirical Corporate Finance, Volume 2, Ed: B. Espen Eckbo, The Netherlands, Elsevier. Jayaratne, Jith and Philip E. Strahan, 1996, The Finance-Growth Nexus: Evidence from Bank Branch Deregulation, The Quarterly Journal of Economics, 111(3), Kang, Nimrit and Nitin Nayar, 2004, The Evolution of Corporate Bankruptcy Law in India, Money and Finance, Oct 03-March 04, La Porta, Rafael, Florencio Lopez-de-Silanes, and Andrei Shleifer, 2002, Government Ownership of Banks, Journal of Finance, 57(1), Megginson, William, L. 2005, The Economics of Bank Privatization, Journal of Banking and Finance, 29(8-9), Morck, Randall, Deniz Yavuz, and Bernard Yeung, 2009, Bank Control, Capital Allocation, and Economic Performance, mimeo. Panagariya, Arvind, 2008, India: The Emerging Giant, Oxford University Press. Petersen, Mitchell, and Raghuraman Rajan, 1995, The Effect of Credit Market Competition on Lending Relationships, Quarterly Journal of Economics, 110(2), Qian, Jun, and Philip E. Strahan, 2007, How Laws and Institutions Shape Financial Contracts: The Case of Bank Loans, Journal of Finance, 62(6), Sapienza, Paola, 2004, The Effects of Government Ownership on Bank Lending, Journal of Financial Economics, 72(2), Stiglitz, Joseph, 2001, Bankruptcy Laws: Some Basic Economic Principles, in Stijn Claessens, Simeon Djankov, and Ashoka Mody, eds, In Resolution of Financial Distress, World Bank Institute, Washington, D.C., pp Tarapore, Savak, 1999, Indian Banking: Preparing for the Next Round of Reform, in James A. Hanson and Sanjay Kathuria eds., India: A Financial Sector for the Twenty First Century, Oxford University Press, New Delhi, Thorburn, Karin S., 2000, Bankruptcy Auction: Costs, Debt Recovery and Firm Survival, Journal of Financial Economics, 58(3), Visaria, Sujata, 2008, Legal Reform and Loan Repayment: The Microeconomic Impact of Debt Recovery Tribunals in India, AEJ: Applied Economics, 1(3), Villas-Boas, J. Miquel, and Udo Schmidt-Mohr, 1999, Oligopoly with Asymmetric Information: Differentiation in Credit Markets, RAND Journal of Economics, 30(3), The World Bank, 2010, Doing Business Database, The World Bank: Washington D.C. Wooldridge, Jeffrey, M., 2001, Econometric Analysis of Cross-Section and Panel Data,

17 Cambridge, MA: MIT Press. Zarutskie, Rebecca, 2006, Evidence on the Effects of Bank Competition on Firm Borrowing and Investment, Journal of Financial Economics, 81(3), Table 1: Describing Banking Sector Entry This table provides summary statistics for the main banking variables used in later analyses. Total Banks is the total number of banks in that district and year, and Population is the district-level population (in millions) in State Bank % of Deposits is the percent of deposits held by state-owned and nationalized banks in that district and year; Private Bank % of Deposits is the percent of deposits held by domestic private banks; Foreign Bank % of Deposits is the percent of deposits held by foreign banks. Total Banks / Population 1-Herfindahl- Hirschman Index of Deposits State Bank % of Deposits Private Bank % of Deposits Foreign Bank % of Deposits Mean Standard Deviation Minimum Maximum Number of Observations 7,187 7,187 7,187 7,187 7,187

18 Table 2: Bank Entry in India from This table provides the total deposits and share of deposits by year and by type of bank for 1991 and Percent changes in shares from 1991 to 2004 are also calculated. Private Banks Deposits in 1991 (Rs. Crores) Deposits in 2004 (Rs. Crores) Share of Total Deposits in % Share of Total Deposits in % Change in Deposit Share ( ) 13.65% State and Nationalized Banks Deposits in 1991 (Rs. Crores) Deposits in 2004 (Rs. Crores) Share of Total Deposits in % Share of Total Deposits in % Change in Deposit Share ( ) % Foreign Banks Deposits in 1991 (Rs. Crores) Deposits in 2004 (Rs. Crores) Share of Total Deposits in % Share of Total Deposits in % Change in Deposit Share ( ) -2.87% Table 3: Describing Bankruptcy Outcomes This table provides summary statistics for the bankruptcy outcomes used in later analyses. Total Filings is the number of bankruptcy filings by all firms in that year and district and Population is the district-level population (in millions) in 2001; Dismissed Filings is the number of firms filing in that district and year that are dropped, dismissed, or declared non-maintainable by the BIFR as they do not meet the criteria for financial distress; Pending refers to filings in that district and year that are still pending determination of financial distress by the BIFR; Sick is the number of firms filing in that district and year that meet the definition of financial distress by the BIFR; Workouts are the number of firms filing in that district and year that the BIFR ruled should be restructured; Liquidations refers to the number of firms filing in that district and year that the BIFR ruled should be liquidated. Average Duration of Workouts is the average number of days it takes for a firm filing in that district and year to receive a restructuring decision from the BIFR. Average Duration of Liquidations is the average number of days it takes for a firm filing in that district and year to receive a liquidation decision from the BIFR. Average Duration of Workouts and Liquidations is the average number of days it takes for a firm filing in that district and year to receive a restructuring or liquidation decision from the BIFR. Total Filings /Population Dismissed Filings /Population Pending Filings /Population Sick Filings /Population Workouts Liquidations /Population /Population All Firms Average Duration of Workouts Average Duration of Liquidations Avg. Duration of Workouts and Liquidations Mean Standard Deviation (0.526) (0.184) (0.099) (0.250) (0.128) (0.130) (728.21) (688.04) (674.54) Minimum Maximum Number of Observations Private Firms Mean Standard Deviation (0.507) (0.179) (0.098) (0.237) (0.124) (0.120) (616.14) (579.20) (580.24) Minimum Maximum Number of Observations Government Firms Mean Standard Deviation (0.040) (0.019) (0.004) (0.026) (0.012) (0.019) ( ) ( ) ( ) Minimum Maximum Number of Observations

19 Table 5: Type of Bank Entry & Bankruptcy Filings Table 4: The Relationship between Competition in the Banking Sector and Bankruptcy Outcomes This table describes the results from an OLS district-level panel regression with district and year fixed-effects. Total Filings is the number of bankruptcy filings by all firms in that year and district and Population is the district-level population (in millions) in 2001; Dismissed Filings is the number of firms filing in that district and year that are dropped, dismissed, or declared non-maintainable by the BIFR as they do not meet the criteria for financial distress; Pending refers to filings in that district and year that are still pending determination of financial distress by the BIFR; Sick is the number of firms filing in that district and year that meet the definition of financial distress by the BIFR; Workouts are the number of firms filing in that district and year that the BIFR ruled should be restructured; Liquidations refers to the number of firms filing in that district and year that the BIFR ruled should be liquidated. Average Duration of Workouts and Liquidations is the average number of days it takes for a firm filing in that district and year to receive a restructuring or liquidation decision from the BIFR. Total Banks is the total number of banks in that district and year. HHI of Deposits is the Herfindahl-Hirschman Index of Deposits of all banks in that district and year. Log (State GDP) is the log of the gross domestic product at the state level. The standard errors clustered at the district-level are reported in parentheses. ***,**,* denotes significance at 1, 5 and 10 percent respectively. Dependent Variable = Total Filings /Population Dismissed Filings /Population Pending Filings /Population Sick Filings /Population Workouts /Population Liquidations /Population Average Duration of Workouts and Liquidations (1) (2) (3) (4) (5) (6) (7) Total Banks / Population ** *** ** (0.0015) (0.0006) (0.0008) (0.0007) (0.0003) (0.0006) (3.109) 1 - HH1 of Deposits 1.237*** 0.521*** 0.812*** *** *** -2,983.9** (0.414) (0.179) (0.203) (0.136) (0.086) (0.106) (1174.5) Log (State GDP) * (0.130) (0.049) (0.094) (0.063) (0.031) (0.042) (646.2) Year fixed effects X X X X X X X District fixed effects X X X X X X X Number of District-Years 7,187 7,187 7,187 7,187 7,187 7, Number of Districts R-squared This table describes the results from an OLS district-level panel regression with district and year fixed-effects. The dependent variable is Total Filings/Population, where Total Filings is the number of bankruptcy filings by all firms in that year and district and Population is the district-level population (in millions) in Total Banks is the total number of banks in that district and year. State Bank % of Deposits is the percent of deposits held by state-owned and nationalized banks in that district and year ; Private Bank % of Deposits is the percent of deposits held by domestic private banks; Foreign Bank % of Deposits is the percent of deposits held by foreign banks. Log (State GDP) is the log of the gross domestic product at the state level. The standard errors clustered at the district-level are reported in parentheses. ***,**,* denotes significance at 1, 5 and 10 percent respectively. Dependent Variable = Total Filings / Population (1) (2) (3) Total Banks / Population 0.004** 0.002** 0.005*** (0.002) (0.001) (0.002) State Bank % of Deposits *** (0.005) Private Bank % of Deposits 0.041*** (0.011) Foreign Bank % of Deposits (0.120) Log (State GDP) (0.130) (0.134) (0.132) Year fixed effects X X X District fixed effects X X X Number of District-Years 7,187 7,187 7,187 Number of Districts R-squared

20 Table 6: Workouts & Liquidations by Type of Bank Entry This table describes the results from an OLS district-level panel regression with district and year fixed-effects. In Columns (1)-(3), the dependent variable is Workouts/Population, where Workouts are the number of firms filing in that district and year that the BIFR ruled should be restructured. In Columns (4)-(6), the dependent variable is Liquidations/Population where Liquidations refers to the number of firms filing in that district and year that the BIFR ruled should be liquidated. Total Banks is the total number of banks in that district and year. State Bank % of Deposits is the percent of deposits held by state-owned and nationalized banks in that district and year; Private Bank % of Deposits is the percent of deposits held by domestic private banks; Foreign Bank % of Deposits is the percent of deposits held by foreign banks. Log (State GDP) is the log of the gross domestic product at the state level. The standard errors clustered at the district-level are reported in parentheses. ***,**,* denotes significance at 1, 5 and 10 percent respectively. Dependent Variable = Workouts / Population Liquidations / Population (1) (2) (3) (4) (5) (6) Total Banks / Population (0.000) (0.000) (0.000) (0.001) (0.001) (0.001) State Bank % of Deposits *** 0.003** (0.001) (0.001) Private Bank % of Deposits 0.007*** *** (0.002) (0.002) Foreign Bank % of Deposits (0.020) (0.024) Log (State GDP) * * (0.031) (0.033) (0.030) (0.042) (0.041) (0.043) Year fixed effects X X X X X X District fixed effects X X X X X X Number of District-Years 7,187 7,187 7,187 7,187 7,187 7,187 Number of Districts R-squared Dependent Variable = Table 7: Duration of Workouts and Liquidations This table describes the results from an OLS district-level panel regression with district and year fixed-effects. The dependent variables Average Duration of Workouts & Liquidations is the average number of days taken by the BIFR to render a restructuring or liquidation decision for all firms filing for bankruptcy in that district and year; Average Duration of Workouts is the average number of days taken by the BIFR to render a restructuring decision for all firms filing for bankruptcy in that district and year; Average Duration of Liquidations is the average number of days taken by the BIFR to render a liquidation decision for firms filing for bankruptcy in that district and year.total Banks is the total number of banks in that district and year and Population is the district-level population (in millions) in 2001; State Bank % of Deposits is the percent of deposits held by state-owned and nationalized banks in that district and year; Private Bank % of Deposits is the percent of deposits held by domestic private banks; Foreign Bank % of Deposits is the percent of deposits held by foreign banks. Log (State GDP) is the log of the gross domestic product at the state level. The standard errors clustered at the district-level are reported in parentheses. ***,**,* denotes significance at 1, 5 and 10 percent respectively. Average Duration of Workouts & Liquidations Average Duration of Workouts Average Duration of Liquidations (1) (2) (3) (4) (5) (6) (7) (8) (9) Total Banks / Population * * 9.58** 6.92 (4.44) (4.53) (4.47) (10.51) (15.57) (9.69) (4.38) (4.68) (4.52) State Bank % of Deposits (20.50) (31.78) (22.35) Private Bank % of Deposits ** * * (16.23) (26.82) (18.00) Foreign Bank % of Deposits *** (16.23) (21.45) (19.13) Log (State GDP) ,387*** -3,484*** -3,779*** (633.5) (637.3) (634.8) (1238) (1239) (1164) (687.6) (691.7) (687.7) Year fixed effects X X X X X X X X X District fixed effects X X X X X X X X X Number of District-Years Number of Districts R-squared