Great Expectations or Side Effects? Bankruptcy Law Reforms and Bank Credit for SMEs. Marco Ghitti London, 24th March 2015

Transcription

1 1

2 Great Expectations or Side Effects? Bankruptcy Law Reforms and Bank Credit for SMEs Marco Ghitti London, 24th March

3 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 3

4 Motivation and Opportunities Main Questions How does a lender react to an exogenous change of Creditor Rights following a Bankruptcy Law Reform? Do credit contracts conditions for SMEs depend on Bankruptcy Proceedings, even if a company is not in financial distress? Opportunities for Investors As a new Bankruptcy Law kicks in, there is room for Investors opportunities: i. providing finance to SMEs as a substitute for bank credit; ii. court shopping across countries to restructure SMEs debt; iii. investing in distressed bank credit. 4

5 Motivation and Opportunities 5

6 Executive Summary 1. This work exploits an almost unique case in OECD countries: Italy, with 6 relevant Bankruptcy Law Reforms in the last 10 years. 2. The analysis is conducted at micro-level thanks to a comprehensive and new dataset on Bank Credit to SMEs: proprietary data at single credit (facility) level from one of the biggest banks; more than 6.4 millions Facility x Quarter observations from 2009-Q4 to Q2 allowing to investigate effects of the post-crisis Bankruptcy Law Reforms; data both on cash and non-cash exposure of the Bank, including pricing and non price characteristics (e.g. size, interest rate, maturity, guarantees, ). 3. The research constructs a new Creditor Rights Index (CRI): a. accounts for 17 rights of Creditor studied in the Literature and built across all Bankruptcy Proceedings available for SMEs; b. finds a consistent and statistically significant relationships between CRI, volume of (Bank) Credit and price of (Bank) Credit. Per each unit of decrease in CRI: - average Exposure at Default lower by 2-3%: at an aggregate national level, it corresponds to a magnitude of tens of billions Euros; - average amount of Granted Credit lower by 3-4%: at an aggregate national level, the reduction impacts for an order of tens of billions Euros; - average Interest Rate higher by 6-8 basis points (bps): this corresponds to an aggregate increase of an order of hundreds of millions Euros, every year. 6

7 Executive Summary 4. The analysis shows that a Bankruptcy Law Reform reducing CRI causes a significant drop in volumes of credits provided by a lender. Variable 2010 Reform 2012 Reform 2013 Reform Cumulative Effects Exposure at Default -8.9% -22.6% +8.3% -23.2% Granted Amount -5.2% -5.0% -3.5% -13.7% Utilized Amount -5.5% -17.1% % DID coefficients under Rating Identification (Below vs. Above Median) See section 6 for details. Depending on volumes definition, the negative differential impact of the reforms can reach up to 23% (an order of billions Euros at aggregate level). Guarantees matter following a reduction in CRI: - during credit contraction Unsecured credits may reduce up to 2 times more than Secured. - when volumes revert up, Secured credits increase up to twice as much than Unsecured. Bank credit crunch following a reduction in CRI may increase probability that a firm enters financial distress. Theoretically, this in turn might exacerbate: - Credit Rationing: entrepreneurs are not able to (re)finance positive NPV projects; - Overinvestment: entrepreneurs may invest in risky projects to gamble for resurrection ; - Underinvestment: shareholders do not invest, even for positive NPV projects. 7

8 Executive Summary 5. The analysis suggests that a Bankruptcy Law Reform reducing CRI causes a substantial differential increase in Interest Rate (IR) for firms more exposed to the Reform: 2010 Reform: differential about +7.3 bps for a firm riskier than a median one. This corresponds to an increase of about 22% in the average Interest Rate crosssectional baseline difference between (i) firms whose rating is below the median and (ii) those with a rating above the median Reform: differential increase about +6.4 bps for a firm riskier than a median one. This means an increment of about 20% in the average Interest Rate crosssectional baseline difference. Unsecured Credits tend to be relatively more affected by an IR increase: 2012 Reform: IR on Unsecured Credits increases 1.8 times more than Secured; 2013 Reform: IR on Unsecured Credits up (+10.6bps), while IR on Secured down (-2.7bps). Newer Credits tend to be relatively more sensitive to an augment of Interest Rate caused by a Bankruptcy Law Reform (not always clear-cut, though). Interest Rates increase is combined with the volumes contraction: this strengthens problems like credit rationing, over- and under-investment. 8

9 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 9

10 Increasing level of distress Bankruptcy Proceedings in Italy Proceedings for SME s Four Bankruptcy proceedings applicable to SMEs disciplined by the Bankruptcy Law (L.F. R.D. 267/1942). Each proceeding may have different outcomes. Proceedings Outcomes Foreclosure - PF (art. 67 L.F.) Foreclosure Endorsed by the Court - FC (art. 182-bis L.F.) Reorganization - R (Concordato Preventivo) Liquidation - L (Fallimento) Continuation Liquidation as Going Concern Liquidation as Piecemeal Sale New Proceedings Foreclosure (PF) and Foreclosure Endorsed by the Court (FC) have been introduced only in

11 Bankruptcy Proceedings in Italy Type of Proceedings Different Bankruptcy Proceedings involve different outcomes, actors, timings, steps and Creditor Rights. 11

12 Bankruptcy Law Reforms in Italy: Season of Reforms 2005 Reform 2006 Reform 2008 Reform 2010 Reform 2012 Reform 2013 Reform Post-crisis Reforms During the last 10 years, Italy has intervened sequentially on its Bankruptcy Law, thus changing repeatedly Creditor Rights (either strengthening or weakening them). Law Type / Number Issued Published Applicable by D.L. 35/ L. 80/ D. Lgs. 5/ D. Lgs. 169/ D.L. 185/ (*) D.L. 69/ (*) D.L. 78/ L. 122/ D.L. 83/ L. 134/ D.L. 179/ (*) L. 221/ (*) L. 228/ (*) D.L. 69/ L. 98/ (*) = not considered in the analysis because not affecting rights included in the Creditor Rights Index Three of the six main reforms are implemented after the Lehman Crisis (2008). 12

13 Bankruptcy Law Reforms in Italy: New Philosophy Italian legislation has moved from a pro-creditor to a prodebtor approach, reducing legal protection of creditors. On the contrary, emerging economies (Brazil, China, and Russia) have recently introduced new Bankruptcy Laws increasing the legal protection of creditors, in an attempt to improve firms access to external finance. In 2010, the legislator introduced some additional changes to the Italian Bankruptcy Law, mainly to encourage the use of Pre-Bankruptcy Procedures. (Loan Market Association News July 2012) Italian Bankruptcy Law has been extensively reformed ( ) in order to focus on the reorganization of distressed and failing businesses rather than on their liquidation. (Shaerman and Sterling, 2012) The purpose of the new rules is to facilitate access to procedures for restructuring debt for firms in financial difficulty. (Freshfield Bruckhaus Deringer, 2012) It is now acknowledged as a statement of fact that the reformed Bankruptcy Law shows special preference for and, somehow, even fosters all those solutions which are aimed at the continuation of the business activity ( ). (Barachini, 2014) Effects on SMEs Great Expectations : have they been met for SMEs? Side Effects: what are the effective consequences for SMEs, taking into account Creditors reactions? 13

14 Creditor Rights Index (CRI) Creditor Rights Index Assesses the level of Creditor Rights in Italy - but can be easily replicated for any other country. Score & Scale - 17 Creditor Rights studied in the literature: for each, a score of 0 (pro-debtor) or 1 (pro-creditor) is assessed (Appendix 1). - The sum of the scores across all the rights represents the CRI. - CRI is measured separately for each Bankruptcy Proceeding. - Total CRI is the sum of CRI across each Proceeding. Crosssection Timeseries - CRI is measured from to Bankruptcy Proceeding Creditor Rights Index (year end) Foreclosure (F) Foreclosure Endorsed by the Court (FC) Reorganization (R) Liquidation (L) Total Creditor Rights Index

15 Creditor Rights Index (CRI) CRI Time-Series Total CRI lessens by 24% between 2004 and Reduction is mainly driven by reforms from 2010 on Reform causes the biggest variation. CRI Cross-Section Different impacts across Bankruptcy Proceedings: Liquidation s CRI goes up, while Reorganization s drops. Total CRI variation is driven by changes in Reorganization. 15

16 Number of Bankruptcy Proceedings in Italy Reforms consequences There is a significant increase in the number of new proceedings following recent Bankruptcy Law Reforms. Is it only caused by economic conditions or credit cycle? Increase of Reorganization New Reorganization proceedings have been increasing from 2008, on average, with a sharp rise after the Q3: from 421 (2012-Q2) to 1,600 (2012-Q4). Such trend may be linked to the reduction in CRI: drops to a score of 2 from 7 after 2012 Reform (2012-Q3). 16

17 Number of Bankruptcy Proceedings in Italy Share of Reorganizations After the Reforms, Reorganization is used as a substitute for Liquidation: share of R in Proceedings raises to 29.21% (2012-Q4) from 3.76% (2005-Q2). Effects on Bank Credit Increasing usage of Reorganizations meet Reforms objectives, but what are the effective consequences on Bank Credit s conditions for SMEs? 17

18 Bank Credit Market and Bankruptcy Law Reforms Overall conditions Total Bank Loans grow from 1,670 to 1,818 /Bn (+ 8.9%) between 2009-Q2 and 2011-Q3; subsequently, the market shrinks (-7.0% from the maximum) to a minimum of 1,691 /Bn in 2013-Q4. Performing vs. NPL Contraction affects only performing loans: -13.6% from the maximum (equal to -1.46% average quarterly decay). There is a sharp increase in non-performing loans (NPL): +224%, from 87 to 283 /Bn between 2008-Q4 and Q4, whose market share triples from 5.14% to 16.70%. 18

19 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 19

20 Research Question: Motivation Main Questions How does a lender react to an exogenous change of CRI? Do credit contracts conditions for SMEs depend on Bankruptcy Proceedings, even for non-distressed firms? Interest Rates Example IR on Bank Credit mirrors the trend of Government Bond yield until the 2012 Reform, while diverges thereafter. 20

21 Research Question: Contribution Within Country Perspective The research adopts a within country perspective to study separately the effects of each reform over time, almost unique case amongst OECD countries. We hold constant other institutional settings that might impact the design of financial contracts, while previous main research studies effects of Bankruptcy Proceedings at a cross-country level (see Appendix 2). SMEs The focus is on Small and Medium Enterprises (SMEs). Bank Credit accounts for around 80% of financing sources for SMEs: pointing at the heart of Credit Market Unique Database The research benefits from a unique dataset, which collects credit information at single facility level. The micro-level analysis represents a key contribution to the Literature, allowing to present novel results both on pricing and non-price terms of financial contracts. 21

22 Research Question: Variables of Interest Empirical Analysis on single credit i to firm j in a given quarter t: Exposure at Default (EAD) Granted Utilized (Utl) Interest Rates (IR) Total Exposure at Default (both on and off balance): effective exposure of the Bank according to Basel rules. - Economic concept: reflects changes in Granted, Utilized and Portfolio. Amount of credit granted: maximum nominal amount that the Bank is willing to lend, given a contract. - Nominal contractual amount: needs time to renegotiate/adjust. Amount of credit utilized: effective amount of money that a firm has withdrawn/borrowed from the Bank. - Operational amount: volatile and subject to competing forces. Annual interest rate: nominal annual interest rate. Guarantee New/Old Credit - Splitting the sample between Secured and Unsecured credits, because the Bank could even modify guarantees. - Splitting the sample between Older and Newer facilities, because the Bank could reduce the New facilities. 22

23 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 23

24 Dataset: Source and Description Source Time Frame Coverage One of the biggest Italian Banks, with over 1,500 branches and a stronger presence in the richest regions. Anonymous Dataset based on supervisory requirements. Frequency: Quarterly data. Time Horizon: from 2009-Q4 to 2014-Q2. All firms clients of the Bank, excluding financial firms. Credit level data: data are collected at single facility level (line of credit, loan, bank guarantee, ). Data on performing and non-performing credits. Data on cash and non-cash Bank s exposure. More than 6.4 millions Facility x Quarter observations: average 340,281 facilities and 147,409 firms per Quarter, corresponding to 2.3 facilities per each firm. Interest Rates Data on Interest Rates available for a subsample of 1.4 millions Facility x Quarter observations. 24

25 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 25

26 Volumes of Credit: TS Statistics Exposure at Default - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 151, , , , , ,638 Median 16,376 15,900 15,504 13,953 13,286 13,276 1st Quartile 1,455 1,291 1,125 1, rd Quartile 69,062 65,863 62,085 57,648 56,275 56,692 Standard Deviation 1,269,422 1,208,906 1,202,714 1,086,256 1,103,669 1,068,872 # of Observations 290, , , , , ,779 Granted Amount - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 258, , , , , ,184 Median 27,200 26,436 25,163 20,000 18,455 21,012 1st Quartile 5,000 5,000 5,000 2,067 1,170 3,600 3rd Quartile 100, ,000 95,000 76,972 70,835 80,000 Standard Deviation 2,628,506 2,469,284 2,618,711 2,633,446 1,680,817 1,694,622 # of Observations 290, , , , , ,991 Utilized Amount - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 188, , , , , ,281 Median 17,968 17,274 16,977 14,804 13,769 17,466 1st Quartile 1, ,344 1, ,284 3rd Quartile 73,930 70,053 66,295 59,434 56,301 65,162 Standard Deviation 1,742,701 1,648,105 1,691,536 1,668,457 1,033,344 1,082,999 # of Observations 290, , , , , ,991 More detailed time series statistics are provided in Appendix 3. 26

27 Interest Rate: TS Statistics Interest Rate - Overall Credit Portfolio (data in %) 2009-Q Q Q Q Q Q2 Mean Median st Quartile rd Quartile Standard Deviation # of Observations 43,232 77,241 82,027 86,946 83,424 81,562 Interest Rate - Performing Credits (data in %) 2009-Q Q Q Q Q Q2 Mean Median # of Observations 41,195 73,007 77,323 80,401 75,801 73,951 % of All Observations 95.29% 94.52% 94.27% 92.47% 90.86% 90.67% Interest Rate - Unsecured Credits (data in %) 2009-Q Q Q Q Q Q2 Mean Median # of Observations 3,993 11,739 13,295 15,748 12,325 11,451 % of All Observations 9.24% 15.20% 16.21% 18.11% 14.77% 14.04% 27

28 Exposure at Default: CS Statistics Sample Observations (#) Mean Exposure at Default (Euro) 1st 3rd Median Quartile Quartile Min Max Standard Deviation All 6,465, ,106 14,960 1,177 61, ,000,000 1,188,821 By Status: Performing 5,511, ,862 14, , ,000,000 1,235,101 Sofferenze 604,749 80,686 18,259 4,365 54, ,344, ,529 Incaglio 287, ,588 14,884 1,527 57, ,661, ,247 Restructured 22, ,436 77,083 8, , ,627,664 4,122,570 Past Due 39,209 72,988 7, , ,531, ,018 By Guarantee: Unsecured 3,167, ,854 5, , ,000,000 1,438,709 Mortgage 509, , ,545 63, , ,157,144 1,719,163 Pledge 160, ,866 38,962 12, , ,661,808 1,719,805 Confidi 460,172 71,818 31,054 14,142 69, ,286, ,486 Personal 2,704,617 99,830 20,991 4,740 67, ,320, ,539 Other 50, ,767 24,928 7, , ,022,352 1,905,454 By New Facilities: New 625,767 92,358 8, , ,000,000 1,019,234 Old 5,548, ,710 15,586 1,385 64, ,000,000 1,202,029 n/a 290, ,865 16,376 1,455 69, ,627,664 1,269,422 Similar statistics for Granted and Utilized are provided in Appendix 3. 28

29 Interest Rate: CS Statistics Sample Observations (#) Mean Interest Rate (%) 1st Median Quartile 3rd Quartile Min Max Standard Deviation All 1,460, By Status: Performing 1,359, Sofferenze 2, Incaglio 83, Restructured 3, Past Due 11, By Guarantee: Unsecured 221, Mortgage 433, Pledge 41, Confidi 348, Personal 845, Other 6, By New Facilities: New 74, Old 1,342, n/a 43,

30 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 30

31 Econometric Setting: Identification Exposure to Reforms Literature A plain comparison of firms financing conditions over time may be deceptive (for unobserved economics). We need to distinguish firms according to their exposure to the design of Bankruptcy Proceedings. Firms heterogeneity with respect to the: i. Risk of Default (Panetta et. al., 2009; Rodano et al., 2013). ii. Court Efficiency (Jappelli et al., 2005; Gennaioli and Rossi, 2010) Identification Focus on the Risk of Default (later for Court Efficiency). Capturing the time-varying exposure using an average measure for each firm over the sample period. Rating - A first measure of Risk of Default: Rating, as estimated for risk management purposes according to Basel Rules. Probability of Default - A second measure Risk of Default: Probability of Default (PD) in 1 year time, as estimated according to Basel rules. 31

32 Econometric Setting: Identification Rating and Interest Rates There is a positive relationship between Rating and IR: - Rating 1: average IR equal to 3.25%; - Rating 9: average IR equal to 3.88%. Rating and IR over time When a Bankruptcy Law Reform reduces CRI, we observe an upward shift of the curve plotting the relationship between average IR and Rating. The shift is more evident for more risky firms: - Rating 1: 47 bps shift b/w before 2010 and after 2013 Reform; - Rating 9: 79bps shift b/w before 2010 and after 2013 Reform. 32

33 Econometric Setting: Specification DID Framework Specification Yijt Expi CRI The econometric analysis is structured according to a Difference-in-Difference (DID) framework. The analysis is run both (i) excluding CRI and (ii) including CRI as explanatory variable. Estimation is done under pooling cross-section OLS. Y ijt = α + βexp i + κ Exp i Ref10 t + γ Exp i Ref12 t + δ Exp i Ref13 t + Χ ijt Ω + F j t 1 Λ + D jt Φ + η Exp i Cycle t + μcri t + Q Y + ε ijt - Output variable of interest (EAD, Granted, Utl, or IR) for the facility i (e.g. loan) to firm j at time t. - Time invariant indicator capturing a firm exposure to each reform (either Rating or PD according to Identification): i. Case 1: Exp i = PD i average or Exp i = Rating i (average) 1, PD > median PD or Rating > median Rating ii. Case2: Exp i = 0, PD median PD or Rating median Rating - Creditor Rights Index (CRI) in quarter t. Others - Please see Appendix 6 for other variables description. 33

34 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 34

35 Empirical Results: EAD Dependent Variable: Log Total Exposure at Default Exposure (Exp) (0.006) (0.006) ** (0.033) ** (0.033) Exp x Ref *** (0.003) *** (0.003) *** (0.011) *** (0.011) Exp x Ref *** (0.004) *** (0.004) *** (0.012) *** (0.012) Exp x Ref *** (0.002) *** (0.002) *** (0.009) *** (0.009) Mortgage Guarantee *** (0.034) *** (0.034) *** (0.034) *** (0.034) Pledge Guarantee *** (0.022) *** (0.022) *** (0.023) *** (0.023) Confidi Guarantee *** (0.011) *** (0.011) *** (0.011) *** (0.011) Personal Guarantee *** (0.012) *** (0.012) *** (0.013) *** (0.013) Other Guarantee *** (0.212) *** (0.212) *** (0.210) *** (0.210) Non Performing (Sofferenza) *** (0.096) *** (0.096) *** (0.091) *** (0.091) Non Performing (Incaglio) *** (0.052) *** (0.052) *** (0.046) *** (0.046) Non Performing (Restructured) * (0.232) * (0.232) (0.223) (0.223) Non Performing (Past Due) *** (0.031) *** (0.031) *** (0.029) *** (0.029) Non Cash Facility *** (0.062) *** (0.062) *** (0.061) *** (0.061) New Facility *** (0.015) *** (0.015) *** (0.015) *** (0.015) Medium-Term Maturity *** (0.061) *** (0.061) *** (0.061) *** (0.061) Long-Term Maturity *** (0.062) *** (0.062) *** (0.061) *** (0.061) Log Revenues (0.009) (0.009) (0.009) (0.009) Log Assets *** (0.018) *** (0.018) *** (0.018) *** (0.018) Leverage (0.000) (0.000) (0.000) (0.000) EBITDA % (0.000) (0.000) (0.000) (0.000) Bank Debt / Total Liabilities (0.000) (0.000) (0.000) (0.000) Bank Debt / Net Debt *** (0.043) *** (0.043) *** (0.043) *** (0.043) Exp x Credit Cycle *** (0.006) *** (0.006) *** (0.020) *** (0.020) Creditor Rights Index (CRI) ** (0.005) *** (0.002) Quarter x Year FE Industry FE Facility Nature FE Segment Size FE Province FE Independent Variable Coefficient Rating Identification Average Rating (1) (2) Cluster SE Coefficient Cluster SE Coefficient Rating Identification Below Median vs. Above Median Rating (3) (4) Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 3,780,287 3,780, ,780,287 3,780, Firm ID Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 35

36 Empirical Results: EAD Dependent Variable: Log Total Exposure at Default Independent Variable Probability of Default Identification Average PD Average PD (5) (6) Coefficient Cluster SE Coefficient Cluster SE Coefficient Probability of Default Identification Below Median vs. Above Median PD (7) (8) Cluster SE Coefficient Exposure (Exp) *** (0.052) *** (0.052) *** (0.036) *** (0.036) Exp x Ref ** (0.033) ** (0.033) *** (0.011) *** (0.011) Exp x Ref *** (0.063) *** (0.063) *** (0.012) *** (0.012) Exp x Ref * (0.035) * (0.035) *** (0.010) *** (0.010) Non Cash Facility *** (0.061) *** (0.061) *** (0.061) *** (0.061) New Facility *** (0.015) *** (0.015) *** (0.015) *** (0.015) Medium-Term Maturity *** (0.061) *** (0.061) *** (0.061) *** (0.061) Long-Term Maturity *** (0.061) *** (0.061) *** (0.061) *** (0.061) Log Revenues (0.010) (0.010) (0.009) (0.009) Log Assets *** (0.019) *** (0.019) *** (0.018) *** (0.018) Leverage (0.000) (0.000) (0.000) (0.000) EBITDA % (0.000) (0.000) (0.000) (0.000) Bank Debt / Total Liabilities (0.000) (0.000) (0.000) (0.000) Bank Debt / Net Debt *** (0.043) *** (0.043) *** (0.043) *** (0.043) Exp x Credit Cycle *** (0.087) *** (0.087) *** (0.022) *** (0.022) Creditor Rights Index (CRI) *** (0.002) *** (0.003) Cluster SE Guarantee Controls Non-Performing Controls Quarter x Year FE Industry FE Facility Nature FE Segment Size FE Province FE # of Observations Adjusted R-squared Cluster Variable 3,780,287 3,780, Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 3,780, Firm ID 3,780, Firm ID 36

37 Empirical Results: EAD Secured vs. Unsecured Dependent Variable: Log Total Exposure at Default Exposure (Exp) (0.006) * (0.005) (0.011) Exp x Ref *** (0.003) *** (0.005) (0.004) Exp x Ref *** (0.004) (0.004) *** (0.007) Exp x Ref *** (0.002) *** (0.003) *** (0.004) Exp x Credit Cycle *** (0.006) *** (0.007) *** (0.009) Creditor Rights Index (CRI) ** (0.005) *** (0.006) (0.007) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls # of Observations Adjusted R-squared Cluster Variable Independent Variable Overall Sample Rating Identification Rating Identification Rating Identification Average Rating Average Rating Average Rating (2) (9) (10) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Secured Credit Unsecured Credit 3,780,287 2,301,161 1,479, Firm ID Firm ID Firm ID Dependent Variable: Log Total Exposure at Default Independent Variable Exposure (Exp) *** (0.052) *** (0.068) *** (0.078) Exp x Ref ** (0.033) (0.063) (0.048) Exp x Ref *** (0.063) *** (0.061) *** (0.094) Exp x Ref * (0.035) ** (0.036) ** (0.072) Exp x Credit Cycle *** (0.087) ** (0.114) *** (0.116) Creditor Rights Index (CRI) *** (0.002) *** (0.002) *** (0.003) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls Overall Sample Secured Credit Unsecured Credit PD Identification PD Identification PD Identification Average PD Average PD Average PD (6) (11) (12) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 3,780,287 2,301,161 1,479, Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 37

38 Empirical Results: EAD New vs. Old Dependent Variable: Log Total Exposure at Default Independent Variable Exposure (Exp) (0.006) (0.023) *** (0.006) Exp x Ref *** (0.003) *** (0.025) (0.003) Exp x Ref *** (0.004) *** (0.007) *** (0.004) Exp x Ref *** (0.002) (0.007) *** (0.003) Exp x Credit Cycle *** (0.006) *** (0.020) *** (0.006) Creditor Rights Index (CRI) ** (0.005) *** (0.017) *** (0.005) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls Overall Sampe New Credit Old Credit Rating Identification Rating Identification Rating Identification Average Rating Average Rating Average Rating (2) (13) (14) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 3,780, ,776 3,371, Firm ID Firm ID Firm ID Dependent Variable: Log Total Exposure at Default Independent Variable Exposure (Exp) *** (0.052) (0.304) *** (0.052) Exp x Ref ** (0.033) ** (0.338) *** (0.034) Exp x Ref *** (0.063) *** (0.104) *** (0.065) Exp x Ref * (0.035) ** (0.157) (0.033) Exp x Credit Cycle *** (0.087) * (0.260) *** (0.086) Creditor Rights Index (CRI) *** (0.002) *** (0.006) *** (0.002) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls Overall Sample New Credits Old Credits PD Identification PD Identification PD Identification Average PD Average PD Average PD (6) (15) (16) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 3,780, ,776 3,371, Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 38

39 Empirical Results Summary: EAD 2010 Reform Secured New/Old 2012 Reform Unsecured New/Old 2013 Reform Guarantee New/Old CRI Bank reduces its average exposure (CRI s reduction): - Rating: differential about -8.9% for firms riskier than median one; - PD: differential about -11.4% for firms riskier than median one. Appear slightly more impacted (not always statistically significant though). Newer Credits see bigger drop: -12% for riskier firms (Average Rating). Highest impact - Substantial contraction of CRI and EAD: - Rating: average differential -22.6% for above median risky firms; - PD: average differential -19.8% for above median risky firms. Greater reduction: about average -9.5% (average Rating identification). Older Credits apparently more affected (option not to give New Credit). Partial reversal of previous reforms - Increase of EAD: - Rating: differential +8.3% for firms riskier than the median; - PD: average differential +8.9% for firms riskier than the median. Relative increase mainly for Secured Credits (double than Unsecured). No clear-cut and statistically significant effects. Economically/Statistically significant: increase of 1 unit in CRI corresponds to a 2-3% growth of average EAD. 39

40 Empirical Results Summary: Granted 2010 Reform Guarantees New/Old 2012 Reform Unsecured New/Old 2013 Reform Secured New/Old CRI Decrement of average Granted credit, but lower than EAD: - Rating: differential about -5.2% for above median risky firms; - PD: differential about -4.9% for above median risky firms. Reduction appears to be slightly driven by Unsecured Credits. More impact on Newer Credits: -6% for more risky firms (average Rating). Greatest impact (CRI plummets) - Contraction of Granted: - Rating: differential about -5.0% for firms riskier than median one; - PD: differential about -3.8% for firms riskier than median one. Drive the contraction: about -4.4% (average Rating identification). Older Credits relatively more affected (option not to give New Credit). Confirms Negative impact of previous reforms on Granted: - Rating: average differential -3.5% for above median riskier firms; - PD: average differential -3.2% for above median riskier firms. More affected (twice as much, but lower cumulative effects of reforms). Older Credits appear slightly more impacted (Bank s option not to issue). Economically/Statistically significant: an increase of 1 unit in CRI implies a 3-4% growth of the average Granted amount. 40

41 Empirical Results Summary: Utilized 2010 Reform Guarantees New/Old 2012 Reform Unsecured New/Old 2013 Reform Secured New/Old Decrement of average Utilized credit, for riskier firms: - Rating: average differential -5.5% for above median risky firms; - PD: average differential -7.0% for above median risky firms. No clear-cut and statistically significant effect. Higher impact for Newer Credits: -12% for riskier firms (average Rating). Strongest impacts (CRI drops) - Relevant decrease of Utl: - Rating: average differential -17.1% for firms riskier than median; - PD: average differential % for firms riskier than median. Drive the reduction: average difference 2.5 times bigger than Secured Older Credits relatively more affected (option not to give New Credit). Negative impact (not always statistically significant though): - Average Rating: differential of -1.1% for each additional Rating unit; - Average PD: difference of -2.6% for each incremental 10 p.p. of PD. Relatively more affected: -2.2% average difference (Average Rating). Newer Credits appear more impacted (Older Credits previously affected?). CRI No univocal and always statistically significant impacts. 41

42 Empirical Results Summary: Volumes Volumes EAD Granted Utilized Guarantees The analysis suggests that a Bankruptcy Law Reform reducing Creditor Rights also causes a contraction of credit volumes provided by a lender. The size of the contraction depends on volumes definition, but cumulatively the studied reforms have negative impacts. - more timely sensitive: down when CRI reduces and up otherwise. The analysed reforms have a cumulative negative impact: -23.2%. - tends to react slower: it needs time to renegotiate a contract. The analysed reforms have a cumulative negative impact: -13.7%. - tends to react faster, yet is subject to competing effects. The analysed reforms have a cumulative negative impact: -22.6%. Guarantees matter in how a lender reacts to a reform: - Unsecured Credits tend to be affected earlier and more than Secured, up to 2 times bigger differential contraction; - If volumes go up, it is mainly for Secured Credit (twice as much). New/Old Newer Credits tend to be impacted relatively earlier and more; yet, it is not easy to disentangle effects linked to the Bank s option not to issue new credit at all. 42

43 Empirical Results: Interest Rate Dependent Variable: Interest Rate Independent Variable Exposure (Exp) *** (0.003) *** (0.003) *** (0.012) *** (0.012) Exp x Ref *** (0.003) *** (0.003) *** (0.012) *** (0.012) Exp x Ref *** (0.004) *** (0.004) *** (0.014) *** (0.014) Exp x Ref (0.005) (0.005) (0.012) (0.012) Mortgage Guarantee *** (0.014) *** (0.014) *** (0.013) *** (0.013) Pledge Guarantee (0.020) (0.020) (0.020) (0.020) Confidi Guarantee *** (0.009) *** (0.009) *** (0.009) *** (0.009) Personal Guarantee * (0.010) * (0.010) *** (0.010) *** (0.010) Other Guarantee * (0.040) * (0.040) *** (0.038) *** (0.038) Non Performing (Sofferenza) *** (0.078) *** (0.078) *** (0.074) *** (0.074) Non Performing (Incaglio) *** (0.029) *** (0.029) *** (0.030) *** (0.030) Non Performing (Restructured) *** (0.230) *** (0.230) ** (0.246) ** (0.246) Non Performing (Past Due) (0.030) (0.030) *** (0.031) *** (0.031) Non Cash Facility New Facility *** (0.012) *** (0.012) *** (0.012) *** (0.012) Medium-Term Maturity *** (0.024) *** (0.024) *** (0.023) *** (0.023) Long-Term Maturity *** (0.025) *** (0.025) *** (0.025) *** (0.025) Log Granted Amount *** (0.004) *** (0.004) *** (0.004) *** (0.004) Log Revenues *** (0.004) *** (0.004) *** (0.004) *** (0.004) Log Assets *** (0.006) *** (0.006) *** (0.006) *** (0.006) Leverage (0.000) (0.000) (0.000) (0.000) EBITDA % * (0.000) * (0.000) ** (0.000) ** (0.000) Bank Debt / Total Liabilities *** (0.000) *** (0.000) *** (0.000) *** (0.000) Bank Debt / Net Debt *** (0.016) *** (0.016) (0.016) (0.016) Exp x Credit Cycle *** (0.004) *** (0.004) *** (0.014) *** (0.014) Creditor Rights Index (CRI) *** (0.004) *** (0.002) Quarter x Year FE Industry FE Facility Nature FE Interest Rate Kind FE Segment Size FE Province FE Rating Identification Rating Identification Average Rating Below Median vs. Above Median Rating (1) (2) (3) (4) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 1,063,856 1,063,856 1,063,856 1,063, Firm ID Firm ID Firm ID Firm ID 43

44 Empirical Results: Interest Rate Dependent Variable: Interest Rate Independent Variable Probability of Default Identification Probability of Default Identification Average PD Average PD Below Median vs. Above Median PD (5) (6) (7) (8) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Coefficient Exposure (Exp) *** (0.041) *** (0.041) *** (0.011) *** (0.011) Exp x Ref *** (0.042) *** (0.042) *** (0.010) *** (0.010) Exp x Ref *** (0.082) *** (0.082) *** (0.013) *** (0.013) Exp x Ref *** (0.074) *** (0.074) (0.010) (0.010) Non Cash Facility New Facility *** (0.012) *** (0.012) *** (0.012) *** (0.012) Medium-Term Maturity *** (0.024) *** (0.024) *** (0.024) *** (0.024) Long-Term Maturity *** (0.026) *** (0.026) *** (0.026) *** (0.026) Log Granted Amount (0.005) (0.005) ** (0.004) ** (0.004) Log Revenues *** (0.005) *** (0.005) *** (0.005) *** (0.005) Log Assets *** (0.007) *** (0.007) *** (0.006) *** (0.006) Leverage (0.000) (0.000) (0.000) (0.000) EBITDA % ** (0.000) ** (0.000) ** (0.000) ** (0.000) Bank Debt / Total Liabilities *** (0.000) *** (0.000) *** (0.000) *** (0.000) Bank Debt / Net Debt (0.017) (0.017) (0.016) (0.016) Exp x Credit Cycle (0.069) *** (0.013) *** (0.013) Creditor Rights Index (CRI) *** (0.002) *** (0.002) Cluster SE Guarantee Controls Non-Performing Controls Quarter x Year FE Industry FE Facility Nature FE Interest Rate Kind FE Segment Size FE Province FE # of Observations Adjusted R-squared Cluster Variable 1,063,856 1,063,856 1,063,856 1,063, Firm ID Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 44

45 Empirical Results: Interest Rate Secured vs. Unsecured Dependent Variable: Interest Rate Overall Sample Secured Credit Unsecured Credit Rating Identification Rating Identification Rating Identification Average Rating Average Rating Average Rating (2) (9) (10) Independent Variable Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Exposure (Exp) *** (0.003) *** (0.003) *** (0.008) Exp x Ref *** (0.003) *** (0.003) *** (0.007) Exp x Ref *** (0.004) *** (0.004) *** (0.008) Exp x Ref (0.005) (0.005) *** (0.009) Exp x Credit Cycle *** (0.004) *** (0.006) *** (0.008) Creditor Rights Index (CRI) *** (0.004) *** (0.004) *** (0.008) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls # of Observations Adjusted R-squared Cluster Variable 1,063, , , Firm ID Firm ID Firm ID Overall Sample Secured Credit Unsecured Credit Dependent Variable: PD Identification PD Identification PD Identification Interest Rate Average PD Average PD Average PD (6) (11) (12) Independent Variable Coefficient Cluster SE Coefficient Cluster SE Robust Exposure (Exp) *** (0.041) *** (0.049) *** (0.100) Exp x Ref *** (0.042) *** (0.041) (0.081) Exp x Ref *** (0.082) *** (0.074) *** (0.091) Exp x Ref *** (0.074) *** (0.071) *** (0.156) Exp x Credit Cycle (0.069) (0.129) *** (0.094) Creditor Rights Index (CRI) *** (0.002) *** (0.002) *** (0.004) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls # of Observations Adjusted R-squared Cluster Variable 1,063, , , Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 45

46 Empirical Results: Interest Rate New vs. Old Dependent Variable: Interest Rate Overall Sampe New Credit Old Credit Rating Identification Rating Identification Average Rating Average Rating (2) (13) Rating Identification Average Rating (14) Independent Variable Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Exposure (Exp) *** (0.003) *** (0.020) *** (0.003) Exp x Ref *** (0.003) *** (0.030) *** (0.003) Exp x Ref *** (0.004) ** (0.019) *** (0.005) Exp x Ref (0.005) *** (0.017) (0.005) Exp x Credit Cycle *** (0.004) ** (0.069) *** (0.004) Creditor Rights Index (CRI) *** (0.004) *** (0.010) *** (0.004) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls # of Observations Adjusted R-squared Cluster Variable 1,063,856 52,551 1,011, Firm ID Firm ID Firm ID Dependent Variable: Interest Rate Independent Variable Overall Sample New Credit Old Credit PD Identification PD Identification Average PD Average PD (6) (15) PD Identification Average PD (16) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Exposure (Exp) *** (0.041) (0.303) *** (0.042) Exp x Ref *** (0.042) *** (0.459) *** (0.038) Exp x Ref *** (0.082) (0.280) ** (0.086) Exp x Ref *** (0.074) *** (0.405) *** (0.071) Exp x Credit Cycle (0.069) ** (1.060) ** (0.068) Creditor Rights Index (CRI) *** (0.002) *** (0.007) *** (0.002) Facility Charachteristics Controls Firm Charachteristics Controls Fixed Effects Controls # of Observations Adjusted R-squared Cluster Variable 1,063,856 52, Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 1,011, Firm ID 46

47 Empirical Results Summary: Interest Rate 2010 Reform Unsecured New/Old 2012 Reform Unsecured New/Old 2013 Reform Guarantee New/Old Rises average IR (CRI reduces), primarily for riskier firms: - Rating: average differential +7.3 bps for above median risky firms; - PD: average differential +6.1 bps for above median risky firms. Relatively (slightly) more adversely affected (+0.5bps) than Secured Credits Newer Credits see bigger increase: +139bps for riskier firms. Most relevant (CRI plummets) Substantial increase of IR: - Rating: average differential +6.4bps for firms riskier than median; - PD: average differential bps for firms riskier than median. Greater relative increase: about 1.8 times more than Secured. Newer credits are slightly more affected: +0.6 bps (Average Rating). Light reversal of previous reforms, although results not always statistically significant - Slight reduction of IR: - Differential about -2.2 bps per each incremental 10 p.p. of PD. Decrease on Secured (-2.8 bps), increase for Unsecured Credits (+10.6 bps). Reduction for Older (-2.7 bps), while increase for Newer Credits (+14.0 bps). CRI Economically/Statistically significant: increase of 1 unit in CRI corresponds to a reduction of IR around 6-8 bps. 47

48 Empirical Results Summary: Implications Economics Credit Rationing Overinvestment Underinvestment Volumes contraction and increase of Interest Rates, following a reduction in CRI, may increase probability that a firm enters financial distress. As firms approach/enter financial distress, they might face: - exacerbation of credit rationing problems: entrepreneurs are not able to (re)finance positive NPV projects. - increase of overinvestment issues: entrepreneurs may gamble for resurrection, in an attempt to continue the firm; - emergence of underinvestment problems: shareholders do not find incentive to invest new funds, even for positive NPV projects. This might drive entrepreneurs to strategically file for Renegotiation, in an attempt to cut-off debt and continue as a going-concern. If banks anticipate entrepreneurs move, there is a further credit contraction / increase of interest rate: a vicious cycle might start. Last generation A Bankruptcy Law Reform aiming to facilitate debt renegotiation may suffer from a last generation problem: firms financed before the reform might suffer more volumes contraction / increase in interest rates, if they are not able to fully benefit from the reform itself. 48

49 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 49

50 Robustness Analysis: Summary Robustness both under Rating Identification and PD Identification. All regressions defined as in the baseline case (regressions n. 3 and n. 7). Variables defined as in the Econometric Setting Section (see Appendix 6). Tests for Robustness: Substituting Value Added % with EBITDA %; Including ROE as control variable; Including ROA as control variable; Including GDP Growth and Inflation as control variables; Including GDP Growth and Unemployment Growth as control variables; Including Unemployment Growth and Inflation as control variables; Including Bank Tier 1 ratio as control variable; Excluding Segment Size as explanatory variable; Excluding Province as explanatory variable; Excluding Industry as explanatory variable; Clustering Standard Errors at Province Level. Results are reported in Appendix 5. 50

51 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 51

52 Court Efficiency and Bankruptcy Law Reforms Law vs. Law Enforcement Court Efficiency Literature underlines importance to separate the effects of the Law (namely, Bankruptcy Law Reforms) from the quality of Law Enforcement (Gennaioli, 2013; Ponticelli, 2013): - Court Efficiency can be seen as a proxy for quality of Law Enforcement (Jappelli et al., 2005). - Duration of Liquidation is an indicator of Court efficiency in a Bankruptcy Proceeding (Djankov et al., 2003). Duration of Liquidation proceeding in Italy across the 26 Judicial Districts (# of days) Stat Median 4,201 4,139 3,739 3,432 3,463 3,444 4,279 3,749 3,322 3,197 2,845 2,817 Average 4,752 4,499 4,361 3,960 3,784 4,050 4,721 4,277 3,719 3,459 3,465 3,257 IQ range 2,645 2,278 2,040 1,760 1,649 2,080 1,922 2,609 2,011 1,773 1,459 1,385 Min 2,133 2,023 1,947 1,567 1,489 1,601 1,690 1,686 1,542 1,546 1,829 1,849 Max 11,875 9,730 10,766 7,396 9,701 9,089 11,626 9,594 8,121 7,753 9,167 7,665 Identification Strategy Significant dispersion in the average duration of a Liquidation: the maximum can reach up to 6.88 times the minimum duration. Current Work in Progress: exploits cross sectional dispersion of Court Efficiency as a source of different firm s exposure to the change in the Bankruptcy Law. 52

53 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 53

54 Conclusions Within country study of the effects of Bankruptcy Law Reforms on the Bank Credit Market for SMEs, exploiting recent reforms in Italy. New Creditor Rights Index, replicable for any country: - there is a strong positive relation between CRI and volumes of Bank Credit, - there is a significant negative relation between CRI and Interest Rate. Detailed and novel dataset on Bank Credit, collected at single-credit level. The analysis shows that a Bankruptcy Law Reform reducing Creditor Rights (and thus CRI): - causes an economically significant contraction of volumes of Bank Credit; - produces a relevant increase of the average interest rate; - affects earlier and primarily Unsecured Credits, Such impacts may increase probability of a firm entering financial distress, which in turns might exacerbate three well-known Corporate Finance problems: credit rationing, overinvestment, and underinvestment. «Great Expectations» that followed through Bankruptcy Law Reforms aiming to facilitate debt renegotiation seem to be neglected by the Side Effects that SMEs suffered as a consequence of the Reforms. 54

55 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 55

56 Appendix 1: CRI s constituents CRI s variables Right No Automatic Stay 17 relevant Creditor Rights mentioned in the literature. A score of 0 (pro-debtor) or 1 (pro-creditor) for each right. The sum of the score of each right across all the Bankruptcy Proceedings gives the CRI. Description 1 = there is no automatic stay when the restructuring proceeding starts 0 = there is automatic stay when the restructuring proceeding starts 1 = secured creditors are paid first when liquidating the collateral Secured creditors paid first Restrictions for going into procedure Management does not stay No Debtor-in-Possession Financing 0 = secured creditors are not paid first when liquidating the collateral Excluding Court expenses which are always paid first, if any 1 = management needs creditors consent to file for starting the restructuring proceeding 0 = management can unilaterally file for starting the restructuring proceeding 1 = management must leave the company when it enters a restructuring proceeding 0 = management can continue to run the company even after starting a restructuring proceeding 1 = it is not explicitly allowed to issue debt more senior to the existing one after starting the restructuring proceeding 0 = it is explicitly allowed to issue debt more senior to the existing one after starting the restructuring procedure L L S V 56

57 Appendix 1: CRI s constituents Right Early Automatic Stay Court Direct Supervision when Automatic Stay starts Creditors vote directly No Cram-down Procedure No Silent Consent Creditors approve administrator/supervisor Description 1 = management is required to file a full proposal to start Automatic Stay 0 = management can start automatic stay first with a light filing and subsequently submit a full /detailed proposal for the company restructuring 1 = Court has always the right to appoint an administrator / supervisor when Automatic Stay starts 0 = Court does not have always the right to appoint an administrator / supervisor when Automatic Stay starts 1 = creditors can vote directly on the liquidation / reorganization plan 0= creditors can vote in committee or not at all on the liquidation / reorganization plan 1 = if voting is required, each individual creditor can make an independent choice about company's restructuring proposal 0 = if voting is required, there is a kind of Cram-down procedure which forces individual creditors to accept what is decided by the majority or by the Court 1 = if voting is required, no vote is considered a contrary vote 0 = if voting is required, no vote is considered a positive vote 1 = creditors has the right to approve the appointment of the administrator/supervisor during the proceeding 0 = only the Court, the debtor and/or other participants appoint the administrator /supervisor during the proceeding 57

58 Appendix 1: CRI s constituents Right Creditors dismiss administrator/supervisor No Minimum Payment No Automatic Loss of Judicial Mortgage Automatic Stay on Lawsuit No Unilateral Termination of Contracts No Restrictions to Bankruptcy Repetition Description 1 = creditors may dismiss or must approve the dismissal of the administrator/supervisor during the proceeding 0 = only the Court, the debtor and/or other participants has the right to dismiss the administrator/supervisor during the proceeding 1 = there is no minimum payment to be guaranteed to unsecured creditors in order to endorse the proceeding 0 = there is a minimum payment to be guaranteed to unsecured creditors in order to endorse the proceeding 1 = when a judicial mortgage is legally endorsed, it remains despite the start of a proceeding 0 = when a judicial mortgage is legally endorsed, it may be automatically cancelled, under certain conditions, upon starting the proceeding 1 = lawsuits against the debtor are automatically stayed upon starting of the proceeding 0 = lawsuit against the debtor continues upon starting the proceeding 1 = the debtor cannot unilaterally terminate a contract when starting the proceeding 0 = the debtor can, under certain conditions, unilaterally terminate a contract when starting the proceeding 1 = acts / payments executed in a given period before and/or during the proceeding may be subject to repetition in case of subsequent Liquidation 0 = acts / payments executed in a given period before and/or during the proceeding are excluded from repetition in case of subsequent Liquidation 58

59 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 59

60 Appendix 2: Literature Review Summary Law & Finance Legal Origins Law Enforcement Creditor Rights Index Seminal work: Cross-country differences in laws pertaining to investor protection have consequences for Corporate Finance (La Porta et al., LLSV). Countries whose legal rules originate in the common law tradition tend to protect investors considerably more than countries whose laws originate in the civil law. Law enforcement differs substantially around the world; German-civil law and Scandinavian countries have the best quality of law enforcement. Common law countries offer creditors stronger legal protection against manager, while French-civil law countries offer the weakest protection. Creditor Rights Evolution over time Creditors rights are influenced by legal origins mainly through Bankruptcy Law and its Enforcement (La Porta et al., 2008). Djankov et al. (2007) use LLSV Creditors Rights index to test its evolution across 129 countries over 25 years. They conclude that: (i) there is very little convergence in creditor rights score across legal origins, and (ii) private credit rises after improvements in Creditors Rights. 60

61 Appendix 2: Literature Review Summary Evolution over time Firms funding contracts depend on Reorganization and Liquidation in Bankruptcy, when looking at 2005 and 2006 reforms of Italian Bankruptcy Law (Rodano et al., 2013). Law Enforcement Evolution over time Unlike legal rules, which do not appear to depend on the level of economic development, the quality of enforcement tends to be higher in wealthier countries (La Porta et al., 2000; La Porta et al., 2006). Djankov et al. (2003) show that procedural formalism by Courts represents a measure of contract enforcement. Improvement in judicial efficiency should reduce credit rationing and increase lending, by opening the credit market to borrowers with less collateral (Jappelli et al., 2005). Debt enforcement s efficiency predicts cross-country debt market development (Djankov et al., 2008). Lengthy of Bankruptcy procedures determines higher interest rates on bank loans and negatively affects credit availability (Giacomelli et al., 2013). 61

62 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 62

63 Exposure at Default: TS Statistics Exposure at Default - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 151, , , , , ,638 Median 16,376 15,900 15,504 13,953 13,286 13,276 1st Quartile 1,455 1,291 1,125 1, rd Quartile 69,062 65,863 62,085 57,648 56,275 56,692 Standard Deviation 1,269,422 1,208,906 1,202,714 1,086,256 1,103,669 1,068,872 # of Observations 290, , , , , ,779 Exposure at Default - Performing Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 161, , , , , ,081 Median 16,654 16,038 15,494 13,072 11,842 11,848 # of Observations 251, , , , , ,472 % of All Observations 86.45% 87.48% 86.71% 84.27% 82.42% 81.53% Exposure at Default - Unsecured Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 142, , , , , ,800 Median 10,041 5,211 3,996 6,974 6,050 5,750 # of Observations 185, , , , , ,301 % of All Observations 63.82% 53.88% 50.52% 39.60% 38.59% 38.60% 63

64 Granted Amount: TS Statistics Granted Amount - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 258, , , , , ,184 Median 27,200 26,436 25,163 20,000 18,455 21,012 1st Quartile 5,000 5,000 5,000 2,067 1,170 3,600 3rd Quartile 100, ,000 95,000 76,972 70,835 80,000 Standard Deviation 2,628,506 2,469,284 2,618,711 2,633,446 1,680,817 1,694,622 # of Observations 290, , , , , ,991 Granted Amount - Performing Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 286, , , , , ,776 Median 36,617 34,701 31,660 26,459 25,000 30,453 # of Observations 251, , , , , ,842 % of All Observations 86.45% 87.48% 86.71% 84.27% 82.42% 80.21% Granted Amount - Unsecured Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 284, , , , , ,843 Median 20,545 16,000 15,000 15,827 15,000 17,474 # of Observations 185, , , , , ,839 % of All Observations 63.82% 53.88% 50.52% 39.60% 38.59% 39.76% 64

65 Utilized Amount: TS Statistics Utilized Amount - Overall Credit Portfolio (data in Euro) 2009-Q Q Q Q Q Q2 Mean 188, , , , , ,281 Median 17,968 17,274 16,977 14,804 13,769 17,466 1st Quartile 1, ,344 1, ,284 3rd Quartile 73,930 70,053 66,295 59,434 56,301 65,162 Standard Deviation 1,742,701 1,648,105 1,691,536 1,668,457 1,033,344 1,082,999 # of Observations 290, , , , , ,991 Utilized Amount - Performing Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 198, , , , , ,991 Median 18,521 17,681 17,136 14,638 13,202 17,945 # of Observations 251, , , , , ,842 % of All Observations 86.45% 87.48% 86.71% 84.27% 82.42% 80.21% Utilized Amount - Unsecured Credits (data in Euro) 2009-Q Q Q Q Q Q2 Mean 187, , , , , ,505 Median 11,118 5,430 4,876 7,920 7,000 8,212 # of Observations 185, , , , , ,839 % of All Observations 63.82% 53.88% 50.52% 39.60% 38.59% 39.76% 65

66 Granted Amount: CS Statistics Sample Observations (#) Mean Granted Amount (Euro) 1st 3rd Median Quartile Quartile Min Max Standard Deviation All 6,408, ,621 24,076 4,000 88, ,615,113 2,422,904 By Status: Performing 5,457, ,434 30,000 10, , ,615,113 2,592,081 Sofferenze 604,349 1, ,628,000 51,083 Incaglio 285, ,661 3, , ,171,057 1,001,135 Restructured 22,802 1,148,411 91,190 11, , ,031,057 5,120,643 Past Due 38,937 79,464 2, , ,555, ,157 By Guarantee: Unsecured 3,152, ,177 15, , ,615,113 3,202,802 Mortgage 509, , ,004 58, , ,945,659 2,551,913 Pledge 160, ,838 50,000 20, , ,171,057 2,149,589 Confidi 460,144 72,621 30,541 13,291 71, ,946, ,763 Personal 2,662, ,631 26,788 8,000 81, ,000, ,191 Other 50, ,410 46,667 15, , ,633,214 3,492,424 By New Facilities: New 607, ,022 12, , ,632,213 2,046,736 Old 5,510, ,650 25,000 5,000 94, ,615,113 2,449,576 n/a 290, ,112 27,200 5, , ,146,712 2,628,506 66

67 Utilized Amount: CS Statistics Sample Observations (#) Mean Utilized Amount (Euro) 1st 3rd Median Quartile Quartile Min Max Standard Deviation All 6,408, ,290 16,140 1,121 64, ,000,000 1,565,514 By Status: Performing 5,457, ,153 16, , ,000,000 1,633,090 Sofferenze 604,349 74,924 15,696 3,644 46, ,229, ,025 Incaglio 285, ,124 15,633 1,647 62, ,831,468 1,052,279 Restructured 22,802 1,220,472 84,057 11, , ,627,662 5,405,173 Past Due 38,937 99,268 8, , ,686, ,346 By Guarantee: Unsecured 3,152, ,647 6, , ,000,000 1,888,950 Mortgage 509, , ,025 63, , ,945,659 2,462,753 Pledge 160, ,828 43,662 15, , ,831,468 1,818,531 Confidi 460,144 71,585 31,025 14,099 69, ,946, ,601 Personal 2,662, ,662 22,311 5,423 69, ,640, ,864 Other 50, ,911 40,200 12, , ,909,537 2,765,563 By New Facilities: New 607, ,307 9, , ,000,000 1,279,024 Old 5,510, ,696 16,882 1,426 67, ,000,000 1,584,127 n/a 290, ,342 17,968 1,084 73, ,627,662 1,742,701 67

68 Agenda 1. Executive Summary 2. Bankruptcy Proceedings and Creditor Rights 3. Research Question 4. Dataset 5. Descriptive Statistics 6. Econometric Setting 7. Empirical Results 8. Robustness Analysis 9. Court Efficiency and Bankruptcy Law Reforms 10. Conclusions A. Appendix 1: Creditor Rights Index s constituents B. Appendix 2: Literature Review Summary C. Appendix 3: Detailed Descriptive Statistics D. Appendix 4: Granted and Utilized: Empirical Results Tables E. Appendix 5: Robustness Analysis Results F. Appendix 6: Econometric Specification and Variables Definition 68

69 Empirical Results: Granted Exposure (Exp) *** (0.006) *** (0.006) *** (0.032) *** (0.032) Exp x Ref *** (0.003) *** (0.003) *** (0.010) *** (0.010) Exp x Ref *** (0.004) *** (0.004) *** (0.011) *** (0.011) Exp x Ref *** (0.004) *** (0.004) *** (0.009) *** (0.009) Mortgage Guarantee *** (0.034) *** (0.034) *** (0.035) *** (0.035) Pledge Guarantee *** (0.022) *** (0.022) *** (0.022) *** (0.022) Confidi Guarantee *** (0.010) *** (0.010) *** (0.010) *** (0.010) Personal Guarantee *** (0.011) *** (0.011) *** (0.011) *** (0.011) Other Guarantee * (0.220) * (0.220) ** (0.209) ** (0.209) Non Performing (Sofferenza) (0.133) (0.133) *** (0.135) *** (0.135) Non Performing (Incaglio) (0.073) (0.073) *** (0.071) *** (0.071) Non Performing (Restructured) *** (0.193) *** (0.193) (0.188) (0.188) Non Performing (Past Due) *** (0.032) *** (0.032) *** (0.028) *** (0.028) Non Cash Facility *** (0.077) *** (0.077) *** (0.076) *** (0.076) New Facility *** (0.013) *** (0.013) *** (0.013) *** (0.013) Medium-Term Maturity (0.080) (0.080) (0.080) (0.080) Long-Term Maturity *** (0.082) *** (0.082) *** (0.082) *** (0.082) Log Revenues (0.009) (0.009) * (0.009) * (0.009) Log Assets *** (0.020) *** (0.020) *** (0.019) *** (0.019) Leverage *** (0.000) *** (0.000) *** (0.000) *** (0.000) EBITDA % * (0.000) * (0.000) (0.000) (0.000) Bank Debt / Total Liabilities (0.000) (0.000) (0.000) (0.000) Bank Debt / Net Debt *** (0.047) *** (0.047) *** (0.047) *** (0.047) Exp x Credit Cycle *** (0.005) *** (0.005) ** (0.016) ** (0.016) Creditor Rights Index (CRI) (0.004) *** (0.002) Quarter x Year FE Industry FE Facility Nature FE Segment Size FE Province FE Dependent Variable: Log Granted Credit Independent Variable Rating Identification Rating Identification Average Rating Below Median vs. Above Median Rating (1) (2) (3) (4) Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE Coefficient Cluster SE # of Observations Adjusted R-squared Cluster Variable 4,049,081 4,049,081 4,049,081 4,049, Firm ID Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 69

70 Empirical Results: Granted Dependent Variable: Log Granted Credit Independent Variable Probability of Default Identification Average PD Average PD (5) (6) Probability of Default Identification Below Median vs. Above Median PD (7) (8) Cluster SE Coefficient Cluster SE Coefficient Cluster SE Coefficient Coefficient Exposure (Exp) *** (0.052) *** (0.052) *** (0.032) *** (0.032) Exp x Ref (0.046) (0.046) *** (0.009) *** (0.009) Exp x Ref *** (0.070) *** (0.070) *** (0.011) *** (0.011) Exp x Ref *** (0.085) *** (0.085) *** (0.008) *** (0.008) Non Cash Facility *** (0.075) *** (0.075) *** (0.076) *** (0.076) New Facility *** (0.013) *** (0.013) *** (0.013) *** (0.013) Medium-Term Maturity (0.078) (0.078) (0.079) (0.079) Long-Term Maturity *** (0.080) *** (0.080) *** (0.081) *** (0.081) Log Revenues ** (0.010) ** (0.010) * (0.009) * (0.009) Log Assets *** (0.021) *** (0.021) *** (0.019) *** (0.019) Leverage *** (0.000) *** (0.000) *** (0.000) *** (0.000) EBITDA % * (0.000) * (0.000) (0.000) (0.000) Bank Debt / Total Liabilities ** (0.000) ** (0.000) (0.000) (0.000) Bank Debt / Net Debt *** (0.046) *** (0.046) *** (0.047) *** (0.047) Exp x Credit Cycle (0.100) (0.100) *** (0.015) *** (0.015) Creditor Rights Index (CRI) *** (0.002) *** (0.002) Cluster SE Guarantee Controls Non-Performing Controls Quarter x Year FE Industry FE Facility Nature FE Segment Size FE Province FE # of Observations Adjusted R-squared Cluster Variable 4,049,081 4,049,081 4,049,091 4,049, Firm ID Firm ID Firm ID Firm ID *** significant at 1%; ** significant at 5%; * significant at 10%. Results rounded at the third decimal place. 70