1
Credit Default Swaps and Debt
Renegotiation
Ana Quintero Guerra
University of Amsterdam
June, 2015
Abstract: The renegotiation of financial contracts is an important consideration in the area of contract theory. When the parties are unable to meet their debt obligations, there is scope for contract amendment. This paper is related to empirical work examining the renegotiation of credit agreements. Specifically, this paper looks at the impact of credit default swaps on debt renegotiation and the outcomes of renegotiations. This is done by means of a multiple regression analysis, using syndicated loan renegotiation data and CDS net notional amounts for US firms over the period 2008-2013. Although there are signs of a negative relation between CDS and debt renegotiation, the results from this sample show that loan amendment is not significantly influenced by credit default swaps.
Keywords: credit default swaps, debt renegotiation, contract theory
BSc. Economics and Business, specialization Finance Bachelor Thesis
2
Table of contents
1 Introduction
2 Literature review
2.1 Renegotiation of financial contracts 2.2 Credit Default Swaps
3 Methodology and hypothesis
4 Data and sample statistics
5 Results
6 Conclusions and recommendations
3
1. Introduction
The renegotiation of financial contracts is an important consideration in the area of contract theory. When agents design a contract, there are three important assumptions made: full commitment to their obligations, symmetric information and equilibrium efficiency (Hart and Moore, 1999). Assuming these, there is no scope for renegotiation. However, this may be unrealistic in many economic environments. The existence of asymmetric information and the inability of parties to fulfill the requirements agreed at the time of origination, may lead to inefficient outcomes (Beaudry and Poitevin, 1995).
When the parties to a contract are unable to meet their obligations, they can usually rewrite the contract’s terms out of court. The effect of renegotiation is to modify the original agreement. These negotiations are not costless, as amendments are accompanied with a fee that changes depending on the deal size and complexity of the amendment (Roberts and Sufi, 2009). Moreover, amending contracts out of court may not always be successful and may lead to coordination problems (Bolton and Scharfstein, 1996).
This paper is related to empirical work by Roberts and Sufi (2009) examining the renegotiation of credit agreements, its determinants and the determinants of renegotiation outcomes. It contributes to existing literature by looking at the behavior of a financial derivative: credit default swaps (CDS). In a standard CDS contract, a protection buyer pays a periodic premium called credit default swap spread to the protection seller in exchange for compensation in the event that a referenced asset defaults during a specified period of time (Augustin et. al, 2014), thus similar to an insurance contract. Consequently, CDS are an effective proxy to measure bankruptcy risk because they reflect the amount protection buyers are willing to pay to insure against default.
This research attempts to answer the research question: are credit default swaps a determinant in the renegotiation of financial contracts? In order to answer this question, an unbalanced panel dataset is used. This dataset contains 294 firms resulting from a merging process of renegotiation data available in Thomson One and CDS net notional amounts from The Depository Trust & Clearing Corporation (DTCC), from 2008-2013. For each firm there is data for the average yearly CDS net notional amounts traded and the number of renegotiations per year, as well as firm characteristics from financial statements.
The results of this study are part of the growing literature on CDS and the amendment of debt agreements. It highlights another implication of CDS as credit derivatives and the dynamics of contract renegotiation, being thereby important for designing efficient contracts. The remainder of this paper is as follows. Section 2 provides a literature review on renegotiation and credit default swaps. Section 3 explains the methodology and hypothesis. Section 4 describes data
4
collected and sample statistics. Section 5 reports and discusses the results obtained. Finally, in section 6, conclusions, shortcoming and further recommendations are presented.
2. Literature review
Since this paper combines two different topics, the review of the literature will be divided in two subsections. First, renegotiation theory is reviewed. In the second subsection some basics on CDS contracts and markets are presented, as well as the relation of CDS to corporate finance discussed in recent literature.
2.1 Renegotiation theory
Renegotiation is regularly linked to incomplete contracts theory. Incompletes contracts do not optimally use all observable information and will be modified at some point. Accordingly, Maskin and Moore (1999) describe renegotiation as an out of equilibrium problem. At the point where a contract is formulated, agents are interested in ensuring optimal outcomes for both parties. If no contingencies take place, no further reassignment is needed. However, when there is realization of uncertainty, renegotiation occurs. Hence, amending a contract can be seen as a game that takes place when an ex post beneficial outcome exists under the initial terms of the contract (Hart and Moore, 1998). A beneficial outcome can be translated in a decrease in the interest rate, a maturity extension or an increase in the available credit, which may shift the bargaining power in favor of the borrower. Garleanu and Zwiebeln (2009) suggest that, a change in factors improving credit quality will increase the borrower’s power to bargain and obtain more favorable terms when renegotiating.
According to Dewatripoint and Maskin (1990), in agreements that are subject to informational constraints, the possibility of renegotiation may firmly cut welfare because a contract that is ex ante optimal may no longer stay optimal at a later date. Thereby, the agreement would lose its initial optimality and be renegotiated. In contrast, in their theory of incomplete contracts, Hart and Moore (1999) suggest that renegotiation is unavoidable and is mainly driven by the realization of exogenous uncertainty.
Authors such as Aghion and Bolton (1992) suggest that new information is neither necessary nor sufficient for renegotiation to take place, but that, in fact, it requires misalignment of incentives from the agents. However, Roberts and Sufi (2009) found that the accrual of new information concerning credit quality and alternative sources of financing are sensitivity factors of renegotiation. Their results show that changes in credit quality and outside options can generate a beneficial outcome ex post under the origination terms and lead to renegotiation.
5
The authors highlight that in order for these changes to affect renegotiation borrowers need a credible threat to abandon their current lender, otherwise they will have little bargaining power. This is consistent with Benmelech and Bergmann (2008), whom developed a model showing that firms will be able to credibly negotiate for more advantageous terms only when they are in a weak financial position and when the liquidation value of their assets is low. This bargaining position is determined by the existence of outside options or alternative sources of financing, but also by increased competition in product markets (Morellec et al., 2014).
Despite the critical implication of loan amendments for financial contracting theories, there is less empirical evidence on the dynamics of renegotiation. The most relevant references regarding empirical research include Roberts and Sufi (2009), M. Roberts (2012), Nikolaev (2013) and Godlewski (2014).
When examining 1,000 private credit agreements from financial institutions and publicly listed US firms from 1996 to 2005, the study by Roberts and Sufi reveals that the majority of all agreements is renegotiated before they reach maturity, especially for loans of stated maturity of at least one year. Moreover, they highlight that renegotiations generate significant changes to the original terms of the contract. In addition to the accrual of new information and outside options, these changes are motivated by the fluctuations in the credit market and the financial health of commercial banks, which turns renegotiation into a highly pro-cyclical process.
A recently published paper by Nikolaev (2013) investigates a sample of 16,500 debt contract amendments from Dealscan and studies the scope for renegotiation, focusing on time before renegotiation takes place. His results reveal that the likelihood of renegotiation is higher for companies with higher uncertainty, greater agency conflicts and more informationally transparent. Finally, Godlewski (2014) examines bank loan renegotiation in European firms. He finds that deal sizes, maturity and financial covenants are the most renegotiated terms in the contract. This empirical analysis shows that amendments indicate large changes in the borrower´s financial structure, which supports the idea that contract amending can been see as a proxy for borrower´s credit quality.
2.2 Credit Default Swaps
A CDS is a contract in which a debt protection buyer pays a premium (CDS spread) to a protection seller to insure against losses in case of a credit event. When they were introduced for the first time in 1994 by JP Morgan, CDS insurance companies and commercial banks were the main users (Augustin et al, 2014). However, in the past decade, the CDS market has expanded enormously and has participant from almost every sector in the financial world. In recent years, hedge funds
6
and bond mutual funds have increased their participation in the CDS market. According to the International Swaps and Derivatives Association (ISDA), the CDS market has grown from $180 billion in 1997, to reach $60 trillion in 2007 just prior to the financial crisis. After 2008 CDS gross notional amounts decreased, given that derivate was central to the financial crisis (Augustin et al., 2014). Nevertheless, they are still the most frequently traded credit derivative.
Credit default swaps have played an important role in credit risk markets and have positively contributed to its development. Angelini (2012) suggest that CDS enables users to hedge or diversify credit exposure, either default risk arising from holding debt or against concentration risk. They also provide new avenues to speculators by allowing them to speculate on credit exposure. Such speculative activity enhances market liquidity and improves the quality of price discovery. The liquidity of CDS markets also improves the chances of protection buyers and protections sellers finding a contract partner, as well as an increase in price efficiency. Terzi and Uluçay (2011) refer to CDS as “the cleanest available measure of corporate default risk” and as an efficient measure of corporate financial health. These characteristics help investors to enhance market discipline over financial institutions.
However, the financial crisis brought attention to the deficiencies present in CDS markets. Peristiani and Savino (2011) highlight that CDS reduces incentives for lenders to analyze and monitor credit quality, which results in its overall decrease. Angelini (2012) points out that CDS trading raises counterparty risk, which may in turn have an effect on financial system vulnerability. Subrahmanyan, Tang and Wang (2014) state that the likelihood of bankruptcy on reference firms increases when CDS is introduced on written debt. Subrahmanyan et al. (2014) refer to a lack of transparency in CDS markets since they are traded over-the-counter (OTC) and are bilateral trades. Moreover, studies by Peltonen et al (2014) and Atkeson et al. (2014) found that CDS markets are highly concentrated, meaning that the default of any of these dealers may provoke contagion effects and generate systemic risk.
A widely discussed issue regarding credit default swaps is their relation to cost of debt and credit supply. Ashcraft and Santos (2009) found that CDS has a negative effect on the cost of debt for both riskier and less transparent firms, while safe and less opaque firms experience a small reduction in spreads. Saretto and Tookes (2013), however, argue that the introduction of CDS markets allows firms to maintain higher levels of corporate leverage and extend debt maturities even when credit supply constraints are most binding. Hirtle (2009) shows that a greater use of CDS leads to increased bank credit supply, but only for large term loans, for which average maturity increases and spreads decrease. In the case of small term loans, the evidence is mixed and even when there is an impact, the effect is economically small.
Credit default swaps are of great importance in the debtor-creditor relationship. The credit insurance offered by a CDS can partially or fully separate the control rights of the creditor from his right to cash flows. In 2008, Hu and Black brought attention to the consequences of this
7
separation. They argue that CDS may create empty creditors to have incentives to push the referenced firm to default, resulting in inefficient outcomes of bankruptcy. This is called the empty creditor problem. Bolton and Oehmke (2011) suggest that when creditors increase their level of credit protection they will tend to overinsure, making the empty creditor problem more severe. Over-insurance takes places when CDS markets are able to anticipate the behavior of the empty creditors, and results in a greater number of defaults. With regard to the incentives to overinsure, Campello and Matta (2012) argue that the empty creditor problem depends on market conditions and firm characteristics, meaning that it is stronger during economic booms and weaker during economic busts.
3. Methodology and Hypothesis
This study builds on earlier work by Roberts and Sufi (2009), a research pioneering the area of empirical evidence regarding the renegotiation of debt contracts. By examining a sample of private credit agreements, the authors investigate the effects of a selection of firm characteristics on the likelihood of renegotiation and renegotiation outcomes. This paper extends the research of Roberts and Sufi by including a novel variable along with the other firm-specific variables previously used: credit default swaps (CDS). The research question is thus: are CDS a determinant in the renegotiation of financial contracts and/or in the different types of renegotiation outcomes? In order to answer the research question, a multivariate regression model using an Ordinary Least Squares (OLS) estimator is used. Although it may seem as too simplistic, a multiple regression analysis allows the researcher to characterize and identify relationships among the dependent variables and the independent variables described by means of regression analysis (Schneider, Hommel and Blettner, 2010). Moreover, the OLS estimator is the dominant method used throughout economics and the social sciences in general (Stock and Watson, 2012). It has desirable theoretical properties under the correct assumptions, such as unbiasedness and consistency. Finally, a multivariate regression model allows to obtain parameters estimates, which can be used for prognostication purposes.
The regression equation constructed in this study is as following:
𝑅𝑒𝑛𝑒𝑔𝑜𝑡𝑖𝑎𝑡𝑖𝑜𝑛𝑠𝑖,𝑡 = ∝𝑖+ 𝛽1 ( 𝐷𝑒𝑏𝑡 𝐸𝐵𝐼𝑇𝐷𝐴) + 𝛽2 ( 𝐸𝐵𝐼𝑇𝐷𝐴 𝐴𝑠𝑠𝑒𝑡𝑠) + 𝛽3 ( 𝐿𝑜𝑔𝐴𝑠𝑠𝑒𝑡𝑠) + 𝛽4 ( 𝐶𝑎𝑠ℎ 𝐴𝑠𝑠𝑒𝑡𝑠) + 𝛽5 ( 𝑀𝑎𝑟𝑘𝑒𝑡 𝐵𝑜𝑜𝑘 ) + 𝛽6 (𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑜) + 𝛽7( 𝐶𝐷𝑆 𝑡𝑜 𝐴𝑠𝑠𝑒𝑡𝑠) + 𝑒𝑖
The dependent variable Renegotiations stands for the number of renegotiations in firm i in period t, measured in years. The control variables relate to firm-specific characteristics during the period. Firstly, the ratio of debt to EBITDA is used as a measure of financial health. Debt/EBITDA is a
8
commonly used by credit rating agencies to assess the probability of defaulting on issued debt, and is therefore an important variable that should be taken into account in the analysis. The second independent variable, EBITDA to assets, is an indication of profitability. Moreover, total assets is included in the set of control variables. This regressor is important as it is a measure of credit worthiness, meaning that it shows whether you have enough financial stability to take on new debt obligations. Cash to assets, the fourth variable, is included to account for the ability of the firm to pay its short term obligations. In order to measure future investment opportunities, the market-to-book ratio is included (Roberts and Sufi, 2009). The sixth variable is the financial leverage ratio, which measures total debt relative to assets and is therefore a good proxy for credit quality to include in the regression.
The seventh and key variable of this research is CDS to assets. It is important to highlight that the variable CDS to assets is a relative measure insurance. Because firms with higher total assets tend to have more CDS, using total assets as a scaling factor for CDS amounts seems appropriate. Furthermore, it is critical to note that this variable is measured in net notional amounts of credit default swaps traded. According to the International Swaps and Derivatives Association (ISDA), for CDS, the term notional refers to “the amount of credit protection bought or sold, equivalent to debt or bond amounts, and is used to derive the premium payment calculations for each payment period and the recovery amounts in the event of default”. The use of notional amounts is driven by the simplicity to identify and gather, as well as it consistency over time. Moreover, notional amounts overstate the level of new activity because it represents cumulative total of past transactions and therefore one should be careful when using them as a measurement tool for risk (ISDA, 2015). The Depository Trust & Clearing Corporation (DTCC) separates notional amounts into two categories in order to account for this imperfection. The first category is gross notional values, which stand for the sum of CDS contracts bought (or sold) for contracts in the database in aggregate, by sector or for existent single reference firms. Category two refers to net notional values, stating the sum of the net protection bought by net buyers (or net protection sold by net sellers (DTCC, 2011). The aggregate net notional amounts represent “the maximum possible net funds transfers between net sellers of protection and net buyers of protection that could be required upon the occurrence of a credit event relating to particular reference entities”.
As a method for controlling for omitted variable bias in panel data, industry and time fixed effects are incorporated to the regression. Industry fixed effects control for variables that are constant over time but differ across entities. Time fixed effects control for variables that change over time but remain constant across firms. Combining these industry and time fixed effects eliminates omitted variable bias arising from both unobservable variables that are constant over time and variables that are constant across entities, leading to much more statistically reliable results (Stock and Watson, 2012).
9
A second part of this research is to examine the role of CDS as a determinant of outcome renegotiations. In their paper, Roberts and Sufi split renegotiation outcomes into three different types: changes in maturity, changes in loan amounts and changes in loan interests spreads. More specifically, a change in these characteristics is defined as the difference between the terms of the contract at origination and the terms after a loan amendment occurs. Mathematically, this can be represented as follows:
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒𝑖 = 𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑡𝑖𝑜𝑛 𝑡𝑒𝑟𝑚𝑠 − 𝐴𝑓𝑡𝑒𝑟 𝑙𝑜𝑎𝑛 𝑎𝑚𝑒𝑛𝑑𝑚𝑒𝑛𝑡 𝑡𝑒𝑟𝑚𝑠
, where i is the type of renegotiation outcome. Using these differences three new multiple regression equations are derived in order to test whether CDS amounts, along with the other firm characteristics as control variables, has an impact on the type of renegotiation outcome. The regression models are represented by the equations presented below.
𝐷𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒𝑖 = ∝𝑖+ 𝛽1 ( 𝐷𝑒𝑏𝑡 𝐸𝐵𝐼𝑇𝐷𝐴) + 𝛽2 ( 𝐸𝐵𝐼𝑇𝐷𝐴 𝐴𝑠𝑠𝑒𝑡𝑠 ) + 𝛽3 ( 𝐿𝑜𝑔𝐴𝑠𝑠𝑒𝑡𝑠) + 𝛽4 ( 𝐶𝑎𝑠ℎ 𝐴𝑠𝑠𝑒𝑡𝑠) + 𝛽5 ( 𝑀𝑎𝑟𝑘𝑒𝑡 𝐵𝑜𝑜𝑘 ) + 𝛽6 (𝐿𝑒𝑣𝑒𝑟𝑎𝑔𝑒 𝑟𝑎𝑡𝑖𝑜) + 𝛽7( 𝐶𝐷𝑆 𝑡𝑜 𝐴𝑠𝑠𝑒𝑡𝑠) + 𝑒𝑖 This model is also controlled for biased results using industry and time fixed effects.
In general, it is expected that the amount of credit default swaps have an impact on debt renegotiation and its outcomes. According to the empty creditor problem described by Hu and Black (2008), after CDS is introduced, holders of debt may no longer be willing to renegotiate debt obligations out-of-court, even when liquidation of the firm is economically more costly. In this case, CDS-insured lenders may push firms in financial distress into bankruptcy, even though continuation would have resulted in an efficient outcome. Bolton and Oehmke (2011) highlight that when investors are free to choose their own levels of credit protection they have an incentive to over-insure. Consequently, expectations are that there is negative relationship between CDS amounts and the renegotiation of financial contracts.
4.
Data and sample statistics
For the aims of this study, a novel data set is built based on the merging of two different sets of information. The first dataset is compilation by Matta, R. on syndicated loan renegotiation data from the former SDC Platinum, now Thomson One. This file contains 4450 firms from the US over the period 1989-2013 (fyear is the time variable), and shows all loans outstanding at each firm (uniquely identified by gvkey codes) in each year. Also, each debt agreement is identified
10
with a facility id and a tranche id, since a loan may have multiple tranches (for example, a loan may have a one credit line and one term loan). Also, amended loans have an amended facility id, which identifies them after a renegotiation takes place.
The key variable in the dataset is amendment year. This binary variable equals 1 if the loan is a renegotiation and was issued in that year. For example, when a loan is issued in year 2005 and matures in 2010, then the dataset has one observation for that loan in 2005, 2006, 2007, 2008, 2009 and 2010. However, if the loan is renegotiated in 2009, then observations from the original loan in 2009 and 2010 are not included in the dataset. Instead, there is an observation for the amendment loan in 2009 and subsequent years, and amendment year is set equal to 1 for that loan in 2009. In order to find out how many renegotiations occur in a year in each firm, sum amendment year by gvkey and fyear and check whether the firm renegotiated one or more loans.
The same dataset also contains information on loan sizes (deal amounts), interest spreads and years to maturity for each individual credit agreement. This information is important to determine renegotiation outcomes. In order to see the results of a renegotiation, match the facility id of the loan when it was first issued to the amended facility id after the initial terms of the contract were modified. Moreover, this set of information contains other variables related to debt covenants and pricing grids but they will not be used in this study, although it will be certainly interesting to take them into account for future research.
The second dataset is related to CDS trading and is obtained from the Depository Trust & Clearing Corporation (DTCC). It contains 891 US firms also identified by gvkey codes, from 2008 to 2013. The main variables in this file are gross and net notional amounts of CDS measured in dollars, and the number of contracts these amounts represent. DTCC provides weekly transaction activity on the quantities of CDS (DTCC, 2011). In order to simplify the analysis and have a better match with the syndicated loan renegotiation data, weekly data is transformed into yearly data. By merging the gathered data, a new dataset is composed. The resulting merged dataset contains 294 firms over the period 2008-2013. For each firm there is data on the number of renegotiations per year and the total net notional amounts of CDS traded. Moreover, this dataset also includes firm-specific and financial data collected from the Compustat database through Wharton Research Data Services.
Table I shows summary statistics for firm characteristics and total number of renegotiations for the period 2008- 2013, for all firms in the final sample. There sample accounts for a total of 139 renegotiations for the time space analyzed. As interesting remark note that in 2008 the scope for renegotiation was larger, meaning that more debt contracts were amended during that year. This high frequency of renegotiation is combined with the lowest measure of relative insurance, i.e. the largest net notional amounts of traded CDS. Cash-to-assets, leverage ratios and EBITDA-to-assets remain relatively constant during the whole period. Debt-to-EBITDA and
market-to-11
book ratio fluctuate, which can be explained by fluctuations in market conditions affecting the credit quality of the firm.
Table I
Summary statistics for firm-specific financial data and renegotiations
This table presents a means summary of firm characteristics and number of renegotiations for the period 2008-2013 for all firms in the merged databases. Financial data is collected from Compustat. CDS data is obtained from DTCC. Syndicated loan renegotiation data is compiled by Matta, R. and is obtained from Thomson One. 2008 2009 2010 2011 2012 2013 Debt-to-EBITDA 2.353 3.353 2.532 2.264 -5.798 6.377 EBITDA-to-Assets 0.126 0.120 0.124 0.127 0.127 0.129 Leverage ratio 0.349 0.346 0.347 0.344 0.355 0.348 Cash-to-assets 0.081 0.072 0.076 0.078 0.078 0.073
Market to book ratio 2.591 1.791 3.604 4.933 1.998 1.135
CDS scaled by assets 0.250 0.435 0.455 0.429 0.352 0.370
Renegotiations 30 18 17 19 27 28
Firms 294 294 294 294 294 294
With the purpose of getting a first approximation to the relation between CDS amounts and the renegotiation of financial contracts and using the total number of contracts amended for all firms over the complete period, a graph is plotted. In Figure 1, each point represents a firm with its corresponding number of loans renegotiated and the CDS amounts scaled by total assets of the firm. Colors identify the number of renegotiations firms have, i.e. firms with no amended contracts are represented in red and blue identifies firms with one renegotiation. The highest number of contracts renegotiated by a firm in this sample is 5, with only one observation and shown by the color black.
Although there are many firms that do not renegotiate their contracts in any year during the period, the graph suggests an opposite relation between CDS amounts and debt obligations restructured out-of-court. Together with the analysis of the summary statistics presented above, this is a reference to the empty-creditor problem presented by Hu and Black (2008). The empty creditor hypothesis states that after CDS is introduced, lenders may lose their incentives to restructure debt obligations of distressed firms out-of-court, thereby pushing the firm into bankruptcy even when firm continuation would be optimal. This because when a firm goes formally bankrupt, immediate compensation is triggered, resulting in large profits for the lenders.
12 Figure 1
Renegotiations vs. CDS scaled for the period 2008-2013
This graph plots the total number of CDS net notional amounts and total number of contracts renegotiated for the period 2008-2013. CDS data is obtained from DTCC and renegotiation data is from a compilation by Matta, R. collected from Thomson One.
13 Table II
Summary loan contract terms
This table summarizes average interest spreads, deal sizes and stated maturity of in the sample of syndicated loans renegotiation during the period 2008-2013. Data comes from a compilation by Matta, R. and is obtained from Thomson One.
An overview of the average renegotiation terms is presented in Table II. A loan contract has an average interest spread across the period around 146 basis points and a stated maturity of 4.78 years. Borrowing amounts have a mean of $ 1336 million. It can be seen that these terms have not been subject of substantial changes during the period examined.
5. Results
The results of the multivariate regression analysis are presented in Table III. In this table, the dependent variable is the total number of renegotiated contracts by each firm in all years. The control variables are Debt-to-EBITDA, EBITDA-to-assets, (log)Assets, market-to-book ratio, cash-to-assets, leverage ratio and CDS notional amounts scaled by total assets. The first column presents the coefficients and t-statistics obtained without controlling for industry and fixed effects. The second column presents the corresponding statistical results, this time taken into account changes across industry and time.
Loan terms 2008 2009 2010 2011 2012 2013
Interest spreads (in
basis points) 106.14 123.83 141.95 165.60 173.96 164.55
Deal amount (in
millions) 1208.99 1323.36 1359.46 1430.83 1440.63 1254.54
Duration (in years) 4.77 4.85 4.65 4.74 4.78 4.91
14 Table III
Determinants of renegotiations
This table presents the estimates of a regression model with multiple regressors. Column 1 reports without controlling for industry fixed effects and industry fixed effects. Column 2 reports the results of the regression including industry and fixed effects. The dependent variable is renegotiations per year in any firm and. The sample period is from 2008 to 2013. T-statistics are reported in parentheses. Statistical significance at the 10%, 5% and 1% levels are denoted by *, ** and *** respectively.
Table III presents the estimated coefficients and t-statistics for control variables. The first variable, the ratio of Debt-to-EBITDA has a particularly low coefficient and it is statistically not significant. This corresponds to Roberts and Sufi (2009), which find that this financial ratio does not have a large impact on renegotiation. EBITDA to assets has a positive coefficient of 0.493, which means that the number of renegotiations increases when this profitability measure raises.
Dependent variable: Renegotiations
(1) (2) Constant 0.606 1.860 (1.59) (2.97) Debt-to-EBITDA 0.0000162 0.0000120 (0.63) (0.08) EBITDA-to-assets 0.493* 0.116 (2.49) (0.60) Log Assets 0.0160 0.00777 (1.04) (0.40) Cash-to-assets -0.317* -0.418 (-2.65) (-1.90) Market-to-book ratio 0.000122 0.000313 (0.55) (0.73) Leverage ratio 0.178 0.0470 (2.49) (0.50) CDS to assets -0.233 -0.0129 (-2.08) (0.69)
Industry fixed effects No Yes
Year fixed effects No Yes
N 1493 1493
15
The same conditions hold for the financial leverage ratio. According to Roberts and Sufi (2009), these variables reflect the importance of new information concerning the ability of the borrower to repay his debt obligations and the creditor´s ability to recover their investment in case the firm is declared bankrupt. However, this study finds that the significance of EBITDA-to assets in the regression is statistically small and that there is no impact coming from the ratio of debt to assets.
The total number of assets shows a positive coefficient of 0.0160. These results suggest that an increase in the firm´s assets raises the number of contracts amended. Although the positive relation of assets to renegotiation coincides with empirical research done by Roberts and Sufi, the impact of this variable in this study turns out to be not statistically significant. The levels of market-to-book ratio have also no effect on the number of renegotiations that take place in a firm. In the case of cash-to-assets, a decrease in cash holdings may raise the debt agreements restructured in the firm. Benmelech and Bergmann (2008) state that renegotiation is more frequent when cash flows are low. However, Acharya, Davydenko and Strebulaev (2012) state that although by intuition firms with higher cash holdings are safer from bankruptcy, higher levels of cash are positively related to an increase in credit risk, which may in turn accelerate renegotiation. However, the results of this paper reveal that the statistical significance is very small.
The key variable of this research is the total net notional amounts of CDS (scaled by total assets). The estimates in Table III for this variable show a negative coefficient of -0.233. This is consistent with expectations. A negative relation between CDS notional amounts and the number of debt contracts amended suggests that when firms are more insured with credit default swaps, there is tendency to renegotiate a higher number of contracts. This interpretation is also consistent with Figure 1. This results are in line with the empty creditor problem introduced by Hu and Black (2008) and Yavorsky (2009), in which lenders that are protected by CDS have less incentives to be involved in out-of-court debt restructuring. In an analysis of the empty creditor problem, Bolton and Oehmke (2011) suggest that CDS have important ex-ante commitment benefits, such increasing investment and making existing projects more efficient. However, they also show that creditor may tend to over-insure, resulting in an inefficient bankruptcy o when continuation will have provided a beneficial outcome for both parties.
Although there is a sign of negative relation in the regression estimates, CDS does not appear to be statistically significant as a determinant of renegotiation. There are three issues that might be affecting the results of the multiple regression analysis. First, the sample is particularly small in comparison to previous empirical research regarding renegotiation (see Nikolaev (2013) and Roberts and Sufi (2009) and Godlewski (2014)). As a second point, a sample selection bias should be considered as a threat to the results. Because the information on the datasets does not make a distinction between distressed and non-distressed firms, results relating to the empty creditor problem may not be consistent. Lastly, estimates may be subject of functional form misspecification. This means that Renegotiations may be a non-linear dependent variable. An
16
indication of the non-linearity of this variable is observable in the low values of R-squared obtained in the regressions. Although controlling for industry and fixed effects improves the fitness of the model, R-squared values are still considerably low.
Table IV presents the results of the regressions to examine the determinants of renegotiation outcomes. In Column 1, the dependent variable is the change in interest spreads at the time the loan was originated and after an amendment took place. In Column 2, the dependent variable is the change in deal size (measured in millions) at initial terms of the contract and after it was renegotiated. The third column refers to the change in stated maturity before and after renegotiation.
Table IV Determinants of renegotiation outcomes
Columns 1, 2 and 3 present estimates for the types of renegotiation outcomes: interest spread, deal amounts and duration. Data is compiled by Matta, R and obtained from Thomson One. The sample period is from 2008 to 2013. T-statistics are reported in parentheses. Statistical significance at the 10%, 5% and 1% levels are denoted by *, ** and *** respectively.
Interest spread Deal amount Duration
Constant 69.523 72.327 -2.721 (0.241) (0.52) (-0.97) Debt-to-EBITDA 4.723 0.0573 -0.003 (1.05) (0.63) (-0.20) EBITDA-to-assets -0.128 0.274 1.223 (-1.05) (0.32) (1.17) Log Assets -0.004 0.0421 0.093 (-1.44) (0.37) (0.80) Cash-to-assets 0.005 0.008 0.056 (0.14) (0.13) (0.21) Market-to-book ratio 2.659 0.040 -0.006 (0.62) (0.43) (-0.28) Leverage ratio -0.123 -0.0299 1.147 (-0.26) (-0.33) (2.72) CDS to assets -0.071 0.045 -0.044 (-0.75) (0.37) (-0.41)
Industry fixed effects Yes Yes Yes
Year fixed effects Yes Yes Yes
N 139 139 139
R-squared 0.0375 0.0701 0.0846
From Table IV it can be seen that debt-to-EBITDA, cash-to-assets and market to book ratios have are positively related to interest spreads. This means that when firms with stronger financial positions, higher cash holdings and larger investment opportunities renegotiate their debt obligations, interest spreads are most likely to increase. In contrast, an increase in total assets results in reductions of interest spreads, which is confirmed by Roberts and Sufi (2009). When changes in deal amounts is the dependent variable, debt-to-EBITDA, EBITDA-to-assets, the cash
17
ratio and the market to book ratio have a positive sign. Moreover, when the total assets of a company grow, renegotiations increase by 4%. Regarding the firm´s assets, Roberts and Sufi (2009) found that when the company experienced a growth, a renegotiation of a debt contract results in an increase of the available credit. In the last column, debt-to-EBITDA has a negative relation with changes in stated maturity, while an increase in the other financial ratios results in an extension of contract maturity. Finally, reductions in CDS amounts result in increases in interest spreads and more time to repay the debt obligations. Also, when firms have more credit protection, credit becomes more available, which can be seen in increases in deal sizes.
Although these estimates are closely related to previous empirical research, none of these variables show statistically significant influence as determinants of renegotiation outcomes. This issue can be assessed by increasing the sample of credit agreements examined. Previous empirical studies use Thomson Reuters-Dealscan as a source for renegotiation data, which also provides information on deal purposes and loan types. Moreover, implementing a non-linear model could improve the low values of R-squared present in the regressions.
6. Conclusions
This paper is related to empirical work examining the renegotiation of credit agreements, its determinants and the determinants of renegotiation. It contributes to existing literature by looking at the behavior of a financial derivative: credit default swaps (CDS).
This research attempts to answer the research question: are credit default swaps a determinant in the renegotiation of financial contracts? In order to answer this question, an unbalanced panel dataset is used. This dataset contains 294 firms resulting from a merging process of renegotiation data available in Thomson One and CDS net notional amounts from The Depository Trust & Clearing Corporation (DTCC), from 2008-2013. For each firm there is data for the average yearly CDS net notional amounts traded and the number of renegotiations per year, as well as firm characteristics from financial statements.
A multiple regression analysis is used to identify a causal relation between the renegotiation of debt agreements and credit default swaps. A second linear regression model examines the impact of credit default swaps on three types of renegotiation outcomes: stated maturity, interest spread and deal amounts. The estimated coefficient for CDS shows that when lenders have higher levels of credit protection, renegotiation is less likely to occur. These results may be indicating a relation to the empty creditor problem introduced by Hu and Black (2008), in which lenders that are protected by CDS have less incentives to be involved in out-of-court debt restructuring. Bolton and Oehmke (2011) show that creditors may tend to over-insure, resulting in an inefficient bankruptcy even when continuation will have provided a beneficial outcome for both parties.
18
Moreover, increases in CDS net notional amounts may reduce time to maturity and markups, while CDS reductions may affect credit availability.
However, the findings of this paper show no significant impact of credit default swaps on the renegotiation of financial contracts nor renegotiation outcomes. This may be caused by the small size of the sample, since there are only 139 renegotiations. This could be an insufficient number of observations to give significant results. Future research may approach to this issue by using a different source for renegotiation data such as Thomson Reuters-Dealscan. As a second point, a sample selection bias should be considered as a threat to the results. Because the information on the datasets does not make a distinction between distressed and non-distressed firms, results relating to the empty creditor problem may not be consistent. Lastly, it could be that there is a non-linear relation between the variables, which would explain the low values of R-squared obtained in the regression analysis. Future research could address this problem by designing a methodology that uses non-linear model.
References
Acharya, V., Davydenko, S. a., & Strebulaev, I. a. (2012). Cash holdings and credit risk. Review of Financial Studies, 25(12), 3572–3609.
Aghion, P., & Bolton, P. (1992). An incomplete contracts approach to financial contracting. The
review of economic Studies, 59(3), 473-494.
Angelini, E. (2012). Credit default swaps (CDS) and their role in the credit risk market. International Journal of Academic Research in Business and Social Sciences, 2(1), 584– 593.
Ashcraft, A. B., & Santos, J. a C. (2009). Has the CDS market lowered the cost of corporate debt? Journal of Monetary Economics, 56(4), 514–523.
Augustin, P., Subrahmanyam, M. G., Tang, D. Y. and Wang, S. Q (2014). Credit Default Swaps: A Survey. Foundations and Trends® in Finance, 9(1-2), 1–196.
Beaudry, Paul, & Poitevin, Michel. (1995). Contract Renegotiation: A Simple Framework and Implications for Organization Theory. The Canadian Journal of Economics / Revue Canadienne d’Economique, 28(2), 302–335.
Benmelech, E., & Bergman, N. K. (2008). Liquidation Values and the Credibility of Financial Contract Renegotiation: Evidence from U.S. Airlines. The Quarterly Journal of Economics, 123(4), 1635–1677.
Bolton, P., & Oehmke, M. (2011). Credit default swaps and the empty creditor problem. Review
19
Bolton, P. and D. Scharfstein (1996). Optimal Debt Structure and the Number of Creditors. Journal of Political Economy 104, 1(25).
Campello, M., and Matta, R. (2012). Credit default swaps and risk-shifting. Economics Letters, 117, 639-641.
Depository Trust & Clearing Corporation (2009). Explanation of Trade Information Warehouse Data. Retrieved from: http://www.dtcc.com/repository-otc-data.aspx
Dewatripont, M., & Maskin, E. (1990). Contract renegotiation in models of asymmetric information∗. European Economic Review, 34(2-3), 311–321.
Gârleanu, N., & Zwiebel, J. (2009). Design and renegotiation of debt covenants. Review of Financial Studies, 22(2), 749–781.
Godlewski, C. J. (2015). The certification value of private debt renegotiation and the design of financial contracts: Empirical evidence from Europe. Journal of Banking & Finance, 53, 1– 17.
Hart, O., & Moore, J. (1998). Default and Renegotiation: A Dynamic Model of Debt. The Quarterly Journal of Economics, 113(1), 1–41.
Hirtle, B. (2009). Credit derivatives and bank credit supply. Journal of Financial
Intermediation, 18(2), 125-150.
Hu, H. T., & Black, B. S. (2008). Equity and debt decoupling and empty voting II: Importance and extensions. As published in University of Pennsylvania Law Review, 156, 625-739.
ISDA. (2009). The Empty Creditor Hypothesis. ISDA Research Notes, (3).
ISDA (2015). Understanding notional amounts. Retrieved from:
http://www.isdacdsmarketplace.com/market_overview/understanding_notional_amount
Nikolaev, V. (2013). Scope for Renegotiation in Private Debt Contracts (SSRN Scholarly Paper No. ID 2012526). Social Science Research Network. Rochester, NY.
Peristiani, S and Savino, V. (2011) Are Credit Default Swaps Associated with Higher Corporate Defaults ? Federal Reserve Bank of New York Staff Reports (494).
Roberts, M. R., & Sufi, A. (2009). Renegotiation of financial contracts: Evidence from private credit agreements. Journal of Financial Economics, 93(2), 159–184.
Saretto, A., & Tookes, H. E. (2013). Corporate leverage, debt maturity, and credit supply: The role of credit default swaps. Review of Financial Studies, 26(5), 1190-1247.
Schneider, A., Hommel, G., & Blettner, M. (2010). Linear regression analysis: part 14 of a series on evaluation of scientific publications. Deutsches Arzteblatt International, 107(44), 776– 782.
Subrahmanyam, M. G., Tang, D. Y., & Wang, S. Q. (2014). Does the Tail Wag the Dog?: The Effect of Credit Default Swaps on Credit Risk. Review of Financial Studies, 27(10), 2927-2960.
20
Terzi, N., & Uluçay, K. (2011). The role of credit default swaps on financial market stability. Procedia - Social and Behavioral Sciences, 24, 983–990.
Yavorsky, A., Bauer, G., Gates, D., & Marshella, T. (2009). Analyzing the potential impact of credit default swaps in workout situations. Moody’s Investors Services, New York, NY.
