Does a Credit Default Swap have an effect on the involved firms
return on stock?
Abstract:
A Credit Default Swap (CDS) is an agreement between two entities to exchange cash-flows and credit risk for a pre-determined period of time. Given this
information, acquiring firms utilize the opportunity to minimize their credit risk exposure by purchasing an insurance on the probability of default of the
reference entity. By paying a quarterly fee to the protection seller, normally a financial institution, the protection insures this in case of an event where the issuer of bonds is unable to pay its interest.
The purpose of this thesis is to determine if the finance construction, deal amount or the announcement of a Credit Default Swap itself, has any significant influence on the return on stock of those companies who were involved in a merge or acquisition event.
To investigate this, two methods have been used. The results of the OLS-regression found that the finance constructing of the CDS and the deal amount did have a statistically significant effect on a target firm’s return on stock, but did not on the acquiring firm’s return on stock. The results of the Event-Study
concluded that the announcements did not carry enough information to force a statistically significant effect on a corporation’s return on stock, within the data used for this thesis.
Author: Sven Calis
Studentnumber:10351272
Supervisor: Dr. R. Almeida da Matta Date: 28/06/2015
Table of Contents
Introduction ... 3
Credit Default Swaps ... 3
Introduction of CDS ... 3
Estimating probability to Default ... 5
Management behavior within the event of a Credit Default Swap ... 6
Hypothesis, Data and Methodology ... 7
Efficient Market Hypothesis... 7
Hypothesis ... 8
Data ... 9
OLS-regression analysis data set ... 9
Event-Study data set ... 10
Methodology ... 10
Ordinary Least Square Regression analysis-method ... 10
Event Study-method ... 11
Results ... 13
OLS-regression output ... 13
Stata output return on stock of acquiring firm ... 13
Stata output return on stock of target firm ... 15
Stata Event-Study output ... 16
Conclusion ... 17
References ... 18
Introduction
In the mid-90s, JP Morgan introduced the first Credit Default Swap. A Credit Default Swap is an agreement between two entities to exchange cash flows and credit risk for a pre-determined period of time (Hull 2004).A protection buyer can transfer the credit risk of a bond it owns to a protection seller by paying the seller a quarterly fee, known as the Credit Default Swap Spread. In case of a default or another negative financial effect by the reference entity, the protection seller will cover the potential credit loss the protection buyer may suffer. The use of the Credit Default Swap has become the most common over-the-counter (OTC) issued credit derivative (O’kane 2003). Due to the swaps dependence on credit events, they have also become a measure of the probability of default of the reference entity and thus gives investors and banks a qualitative,
instantaneous and efficient measure of credit risk (Jacobs, 2010). Because the Credit Default Swaps have become an important measure, it is interesting to study if this credit event has any influence on the value of the firm who is
involved in the specific event. For measuring the value of the firm, the stock price has been taken.
The purpose of this thesis is to determine if the finance construction, deal amount or the announcement of a Credit Default Swap itself, has any significant influence on the return on stock of those companies who were involved in a merge or acquisition event.
Credit Default Swaps
In this particular section we will give a brief introduction of the Credit Default Swaps, how the derivatives were developed and constructed. Besides that, there will be determined what the Credit Default Swap Spread actually is, to give a fair compensation for the probability to default. At the end of this section there will be a view on the management behavior under the event of a Credit Default Swap.
Introduction of CDS
In 1997, JP Morgan introduced the first Credit Default Swap. According to Hull (2004), a credit default swap is an agreement between two entities to exchange Cash-flows and credit risk for a specific period of time. A protection buyer can transfer the credit risk of a bond it owns to a protection seller by paying the seller a (mostly) quarterly fee, known as the CDS spread. When the reference entity (the issuer of bonds) goes into default, the protection seller will cover the credit loss the protection buyer may suffer.” Since their development, CDSs have become the most common over-the-counter issued credit derivatives. According to the British Bankers’ Association Credit Derivatives Survey, it dominates the credit derivatives market with over 72,5% of the outstanding notional in 2003(O’kane 2003). Because of the swaps dependence on credit events, they have also become a measure of the probability of default of the reference entity and thus give investors and banks a qualitative, direct and efficient measure of credit risk (Norden 2004).
As the figure below describes, the Credit Default Swaps have grown significantly since their introduction.
Source: http://www.isda.org/statistics/pdf/ISDA-Market-Survey-historical-data.pdf
The figure suggests that the Credit Default Swaps played a significant role in the financial crisis of 2007 until 2009, with a peak in the second half of 2007. The financial crisis caused the notional amount of Credit Default Swap outstanding to decrease drastically after this peak and forced central banks and policy makers to reevaluate the CDS and standardize it in order to create a easy way in
managing and stabilizing the effect of widespread defaults (Qui and Yu 2012). There are a couple of reasons why the use of the Credit Default Swap has increased so drastically. The derivative was originally created in order to
minimize credit and to make their capital in the bank available to use. The banks could use the CDS to reduce their risk by buying protections on the event of a default by a corporation, which had borrowed money from them (Hull 2004). To illustrate this suppose a bank could issue bonds worth of $100 million to a corporation for five years, which is the usual term for CDS (O’kane 2003), at an interest rate of 10%. The corporation would make an interest payment of $10 million to the bank on yearly basis and when the bond comes to its maturity, it would return the principal amount. The bank is receiving its interest on the money of the corporation has borrowed from them, but if the corporation defaults, the bank would stand to lose all, or a big part, of their lending. Which means, that the bank has a potential loss of $100 million on the bonds if the corporation were to default. Because of this big potential loss, the bank is willing to reduce is exposure to this risk by acquiring Credit Default Swaps. The bank could contact an protection seller, often a bank, such as Goldman Sachs, or other financial institution, and buy a Credit Default Swap. As compensation of the risk transferring asset, the bank (the buyer of protection) will make regular
payments to its counterparty (the seller of protection), and when the event of bankruptcy occurs, the protection seller would pay the protection buyer whatever it lost due to this event (O’kane 2003).
The second main reason why banks and other financial institutions would buy Credit Default Swaps is for regulatory issues. Most of the countries have
rules that contain that banks must have a sustainable amount of cash set aside to protect itself in case of bad loan repayments. According to Weistroffer (2003), by buying Credit Default Swaps, banks can abridge the amount of cash they must have set aside, and because of that, they can lend more money to other
enterprises.
At first, before the Credit Default Swap by the ISDA, the International Swaps and Derivatives Association, parties were capable to set any terms they agreed upon in their agreements of the CDS. Later on, the Credit Default Swaps were standardized for faster and clearer use of it (Qui and Yu, 2012). The
constructing of a Credit Default Swap is actually quite simple. The construction is as follows. Company A (protection buyer) buys bonds issued by company C (reference entity). Company A, to protect itself from the credit risk they get exposed to by buying bond from company C, will buy a Credit Default Swap from company B (protection seller). Company C will make regular interest payments to company A and company A will make regular, quarterly payments to company B. When company C is unable to pay the interest, or another credit event occurs, such as bankruptcy, failure-to-pay occurs and restructuring is necessary (O’kane 2003). Company B will compensate company A on any loss they have had. This will continue until the Credit Default Swap matures (as said above, usually five years).
Estimating probability to Default
Norden and Weber(2009) are referring that the market of Credit Default Swaps should reflect pure issuer default risk and no facility or issue specific risk, making them a potentially benchmark for measuring and pricing credit risk. These derivatives have turned out to clearly dominate other types such as credit-linked notes and total return swaps in terms of market volume and (2009). This pure issuer default risk has been reflected into the terms of the CDS as the ‘spread’ of the Credit Default Swap.
According to O’kane and Tumbull (2003), the Credit Default swap spread has been defined as follows: ‘The Credit Default Swap Spread is the amount the protection buyer is required to pay for the protection, given by the protection seller. These spreads are usually quoted in basis point, which are 100th of 1% on
the protection seller bonds’ face value.’
Standardized contracts use mostly semiannually or quarterly payments. lIf the reference entity is facing a negative credit event, there is a accrual
payment covering the period from the previous payment to the date of this negative credit event and after that the payments stop (Hull & White 2002).
The width of the spread of the Credit Default Swap is based on the discounted premium payments combined with the risk-neutral probability that the reference entity defaults between the date of issue of the CDS and the
maturity of it. Although the spread is agreed upon by the two parties and is fixed for the duration of the Credit Default swap, the spread is quoted daily and firms usually mark-to-market their CDSs on a daily basis (O’kane & Tumbull 2003).
There are different ways to determine the spread. Although, to illustrate the way of calculation of the spread, we will discuss the formula of Hull & White made in 2000, to establish a fitting spot price of a spread. But before we show the exact formula, there has to be an estimation of the probability of the case that the reference entity survives, denoted as 𝜋.
Let q(t)δt be the risk-neutral probability of default between time t and t + δt as seen at time zero. The probability that the company will survive to time t is then 𝜋(𝑡) = 1 − ∫ 𝑞(𝜏)𝑑𝜏𝑡=0𝑡
So, with this said, the exact formula according to Hull & White (2000) is as follows: 𝑆 = ∫ [1 − 𝑅̂ 𝑇 𝑡=0 − 𝐴(𝑡) ∗ 𝑅̂] ∗ 𝑞(𝑡) ∗ 𝑣(𝑡) ∗ 𝑑𝑡) ∫𝑇 𝑞(𝑡)[𝑢(𝑡) + 𝑒(𝑡)]𝑑𝑡 + 𝜋(𝑇∗) ∗ 𝑢(𝑇∗) 𝑡=0
S: Spread spot price
u(t): The present value of payments at the rate of $1 per year on the payment dates of the underlying CDS between times T and t (T < t ≤ T*)
e(t): The present value of the final accrual payment that would be required on the CDS at time t in the event of a default at time t if payments were made at the rate of $1 per year2
v(t): The present value of $1 received at time t
A(t): The accrual interest on the reference bond at time t as a percent of the face value
𝑅̂: The expected risk-neutral recovery rate on the reference bond defined as a percentage of the claim amount. The claim amount is assumed to be the face value of the bond plus accrued interest.
Besides the exact formula of the credit default swap, there are other important factors that influence the CDS spread. Because the relationship between the Credit Default Swap spread and stock prices is examined, and not how the Credit Default Swap spread is being established, the focus of the determination of the CDS spread lies on the relationship with the stock prices. According to the Merton model, Merton (1974) suggests a negative connection between a firm’s market value of equity and its probability of default. Higher stock returns increase a firm’s value, which theoretically should decrease CDS spreads. So when a protection seeker has a stable high stock return, the spread will be lower than the spread of a protection seeker with a lower stock return. A negative relation is thus expected between stock returns and CDS spreads (Galil 2013).
When there is an event of a Credit Default Swap, there is a default
risk-movement from one company to another. After the establishment of the CDS, the protection seller is obliged to pay the protection seeker when the reference entity goes into default, or another negative credit event appears (Hull 2004). Because of the large CDS transactions on a reference entity, there is a creation of a large inventory positions. The large inventory positions may shock dealers’ exposures to the credit risk of reference assets. They could respond in many different ways. They may widen their quotes afterwards, so they pass their costs of holding inventory risk through to their customers. Additionally, dealers may actively manage inventory risks that they do not want to bear by entering into offsetting contracts that diversify or hedge new risk exposures (SEC 2014).
Doing so requires finding market participants who are willing to act as counterparties to these offsetting contracts. The problem here is that there are only a small number of participants active in the interdealer market, therefore, public dissemination of Credit Default Swap transaction information can be costly for a dealer who actively hedges in his CDS inventory. These
counterparties could respond by strategically widening spreads in an effort to extract rents above those associated with competitive liquidity supply (SEC 2014).
Hypothesis, Data and Methodology
This section will discuss the theory on which this thesis is based, followed by an explanation of the hypothesis. Besides that, this section will examine the data and will explain the methodology that had been used during this thesis.
Efficient Market Hypothesis
According to the founder of the Efficient Market Hypothesis, Eugene Fama, an efficient market can be described as a market where the price of the securities that represent ownership of firm’ will always reflect all available information at any moment in time, given certain assumptions (Fama, 1970). The market has to be an active market with many profit-maximizing firms who all have access to the latest information to make this hypothesis valid (Fama, 1970).
The reason why markets are efficient is based on the assumption that all investors seek to maximize profit. According to Fama, profit-maximizing
investors will always search for all available information before making any investment decisions, to make sure that they get the maximum profit out of it. For stock prices, it is exactly the same. An investor will search for information about the firm and will judge if the stock price is valued right. If this is not the case, the investor will buy or sell his stock according to his believes about the price level of the stock compared to the value he estimated with all available information (Fama, 1970).
The Efficient Market Hypothesis can be divided into three levels depending on the nature of the information.
The highest level of information availability is the so called ‘strong-form markets’. Here the asset prices fully reflect all available information, public and private. This implies that not even by trading insider information will results in an profit-creating opportunity for an investor. There is no expectation that such
an extreme market will be an exact description of the world, but it can be used as a benchmark, which ensures that market efficiency can be judged by
acknowledging the importance of deviations (Fama, 1970).
The mid-level market is called the semi-strong-form market. According to Fama (1970), these markets do exist. In these markets, the security prices reflect all publicly available information. Because this market is based on publicly available information, there has to be a price movement when there is an
announcement of an event (Fama, 1970). This announcement of an event can be an announcement of a merger or acquisition. The only two ways to outperform the market are to get access to insider-information or to be lucky.
The lowest level is called the weak-form market. Here the prices fully reflect all information from the past, such as historical prices and returns. Weak-form markets of the efficient market model are the easiest and occurring
markets, according to Fama (1970).
The relation from these different levels and the efficient market hypothesis to the thesis subject is that it assumes that there has to be a stock price movement when an event of an announcement occurs.
Assuming that the stock price market is a semi-strong efficient market, this assumption is important when analyzing the deviation of stock returns compared to their expected return throughout certain time periods. This level takes the assumption that when an announcement or credit event occurs, there will be a price movement of the stock.
According to Norden(1997), within efficient markets, default risk of firms should be reflected by market prices of financial claims on these firms. Theory suggests that stock value depends on the distribution of the market value of the firm’s assets (Norden 1997). So it is also very interesting to look at how the Credit Default Swap is being created, because it can be based on the firms’ debt or it can be created on the firms’ equity. A second parameter to take into consideration is the deal amount of the Credit Default Swap.
The focus of this thesis will be on the relationship between M&A’s, which are financed with a Credit Default Swap, and stock prices. In this thesis there will also be an examination of the announcement of the M&A’s are carrying a high information content to establish an significant price movement of the return of the stocks.
Hypothesis
In this section the hypothesis is being discussed. First, the hypothesis for the Event-Study method, followed by the hypothesis for the OLS-regression analyses. Event-Study hypothesis
H1: The announcement of the M&A will cause a significant reaction on the return on stock.
In the light of the literature, the first hypothesis is in agreement with the efficient market hypothesis of Fama. The new information that is coming available at the time of the announcement has not been included in the stock prices. So, the information content of the announcement is high enough to see a significant change in the return of stock.
OLS-Regression analyses hypothesis
H1: There is a significant change in the return on stock prices on both the target and acquiring firm.
The first hypothesis for the OLS-regression analyses is in agreement with the efficient market hypothesis. The ratio of the deal amount or the constructing of the Credit Default Swap has a significant influence on return of the stock of both the acquiring and target firms.
H2: There is a significant change in the return on stock prices of the target firm.
This second hypothesis for the OLS-regression analyses is not fully in agreement with Fama’s efficient market hypothesis. The ratio of the deal amount or the constructing of the Credit Default Swap has a significant influence on return of the stock of the target firms only.
H3: There is a significant change in the return on stock prices of the acquiring firm.
Also the third hypothesis for the OLS-regression analyses is not fully in
agreement with the efficient market hypothesis. The ratio of the deal amount or the constructing of the Credit Default Swap has a significant influence on return of the stock of the acquiring firms only.
Data
Before discussing the used dataset matching each method, the list of companies used have a few things in common. The quantity of the selection criteria for the regression analysis was 150 and for the Event-Study 50. The time-interval for both methods were the same, namely from 2010 up until the end of December 2014 for listed firms. The observations are all from origin American or listed on a U.S. stock exchange market. The completion or success of the merger or acquisition was not a criterion for this research since the focus relies on the effect of the announcement of the Credit Default Swap, how it is constructed and the ratio of the net notional amount of the deal to the total assets of the acquiring firm.The announcement dates that were in weekends or holidays were adjusted to the first following working day to match the data from the returns on stock.
OLS-regression analysis data set
To provide the appropriate data for the OLS-regression analyses, data has been obtained from two different sources. The historical stock price data and returns were obtained from Wharton Research Data Services. Whether the CDS was financed on debt or equity, the ratio net notional amount divided by total assets, Tobin’s q and the log of the market capitalization at the time of the
announcement, were provided by the database of Thomson Reuters’s Datastream.
To construct an appropriate regression, 150 biggest M&A’s, financed with CDS with useful data (provided by Thomson Reuters’s Datastream), have been taken to run this regression analyses. In this regression, there are four variables created: whether the CDS is financed on debt or equity, a ratio of the net notional
amount of the deal to total assets of the acquiring firms and stock prices of both acquiring and target firm.
In the same regression, there have been added two controls: Tobin’s q and the logarithm of the market capitalization at the time of the announcement. Each finance regression must have controls, such as the Tobin’s q and the logarithm of the market capitalization (Norden, 2004).
To make the data appropriate for running the OLS-regression, the separate datasets were being merged with the date of the announcement event as comparing variable. Because of incomplete data, the analysis was performed without 37 target firms and 44 of the acquiring firms.
Event-Study data set
To provide the appropriate data for the Event-Study method, data is obtained from the same two sources as the OLS-regression.
Wharton Research Data Services provided the return on stock and the return of the S&P500-index, matched with the same date as the return on stock. The database from Thomson Reuters’s Datastream provided the announcement dates for the 50 target and acquisition firms used in the Event-Study analysis.
To make the data appropriate for running the Event-Study analysis, the separate datasets were being merged with the date of the announcement event as comparing variable. Because of further incomplete data, 22 companies were excluded from the analysis to ensure higher result significance.
Methodology
In this section two methods that were extensively used will be discussed to determine whether there is a relationship between Credit Default Swap spread and stock prices. First presenting the Ordinary Least Square (OLS) regression analyses. Second, examining the method of the Event Study.
Ordinary Least Square Regression analysis-method
Here there will be an analysis of the movement of the return on stock due to an constructing of the Credit Default Swap, at the moment of announcement. The aim lies on the stock return, whether the CDS is created on debt of equity, the ratio of the net notional amount to total assets and the development of the stock price for both acquiring and target firms.
For the regression analyses, the following regression equations are used:
𝑦𝑡= 𝛼 + 𝛽1∗ 𝐶𝐷𝑆𝑎+ 𝛽2∗ 𝐷𝑢𝑚𝑚𝑦𝑁𝑁𝐴𝑇𝐴 + 𝛾 ∗ 𝑋𝑎+ 𝜃 ∗ 𝑋𝑡+ 𝜇 ∗ 𝑇𝑞 + 𝜌 ∗ ln(𝑚𝑎𝑟𝑐𝑎𝑝) + 𝑝𝑟𝑜𝑓𝑖𝑡𝑎𝑏𝑖𝑙𝑖𝑡𝑦 + 𝜀𝑡𝑎
𝑦𝑎= 𝛼 + 𝛽1∗ 𝐶𝐷𝑆𝑎+ 𝛽2∗ 𝐷𝑢𝑚𝑚𝑦
𝑁𝑁𝐴
𝑇𝐴 + 𝛾 ∗ 𝑋𝑎+ 𝜃 ∗ 𝑋𝑡+ 𝜇 ∗ 𝑇𝑞 + 𝜌 ∗ ln(𝑚𝑎𝑟𝑐𝑎𝑝) + 𝑝𝑟𝑜𝑓𝑖𝑡𝑎𝑏𝑖𝑙𝑖𝑡𝑦 + 𝜀𝑡𝑎
Explaining the variables:
𝑦𝑡: Return on stock of the target firm, calculated as follows: ((stock price after
the announcement date / stock price before the announcement date)/stock price before the announcement date)
𝑦𝑎: Return on stock of the acquiring firm, calculated as follows: ((stock price after
the announcement date / stock price before the announcement date)/stock price before the announcement date)
𝐶𝐷𝑆𝑎: Whether the acquiring firm has the Credit Default Swap financed on his debt( if so CDS = 1, if not CDS = 0)
𝑁𝑁𝐴
𝑇𝐴 : Dummy variable, A ratio, Net Notional Amount of the Credit Default Swap to
Total Assets of the acquiring firm
𝑋𝑎: Stock price of acquiring firm on the announcement date
𝑋𝑡: Stock price of the target firm on the announcement date
𝑇𝑞: Tobin’s q, finance regression control variable, is ratio between the asset’s value and its replacement value.
ln (𝑚𝑎𝑟𝑐𝑎𝑝): Logarithm of market capitalization of the acquiring firm, finance regression control variable, price of stock multiplied by the quantity of
outstanding stock.
𝑝𝑟𝑜𝑓𝑖𝑡𝑎𝑏𝑖𝑙𝑖𝑡𝑦: EBITDA / Revenue of the acquiring company, finance regression control variable, a ratio of profitability.
𝜀𝑡𝑎: Error-term, gives the sum of in the deviations of the target firm stock return
𝜀𝑡𝑎: Error-term, gives the sum in the deviations of the target firm stock return
After the regression has been done, the p-value of the test statistic will determine if our hypothesis is correct or not.
Event Study-method
For the Event Study, the steps of the method of Craig MacKinlay with his work, Event Studies in Economic and Finance (MacKinlay 1997) have been followed. While MacKinlay mentioned that there is no such thing as a unique structure for an Event Study, there is a general flow of analysis. This general flow is being discussed in this section.
According to MacKinlay(1997), the initial task of conducting an Event Study, is to define the event of interest and identify the period over which the security prices of the firms involved in this event will be examined, the so called ‘time-window.’ In this particular case we want to examine if the announcement has had an immediate impact on the stock price. It is usual to define the event window to be larger than the specific period of interests. It allows us to examine periods
surrounding the event. Normally, the event-window is expanded with a couple of days. Mostly the event-window consists of the day before the announcement and the day after the announcement. Those days capture the shock in the stock price due to the announcement (MacKinlay 1997). To get a better view of the actual impact of the announcement, a couple of days before are the event are also interesting to take into account, the same for a couple of days after the
announcement. This is the event-window, which is ten days prior- and after the announcement date.
After the identification of the event-window, there has to be determined what the selection criteria will be. These selection criteria may involve
restrictions. In this case, there has been chosen to select 50 random acquiring and target firms out of the 150 M&A’s that were being used for the OLS-regression analyses.
For a valuation of the impact of the impact of the event, there measure for abnormal returns is used (MacKinlay 1997). The most fundamental part of conducting an event study is the measurement of normal performance, from which abnormal performance then can be derived (MacKinlay, 1997). One of the most established models to measure normal performance is the market model
(MacKinlay, 1997). Following Greatrex (2009), this study uses the market model to obtain the normal performance.
The abnormal return is the actual ex-post return of the stock over the event-window minus the normal return of the firm over the event-window (Mackinlay, 1997). The normal return has been defined as the expected return without conditioning on the event that is taking place (Mackinlay, 1997). There are two common choices for modeling the normal return, which are the
constant-mean return model and the market model. The constant mean return model assumes that the mean return of a given security is constant through time. The market model assumes a stable linear relationship between the return of the market and the return of the stock (Mackinlay, 1997).
Given the selection of a normal performance model, the estimation-window has to be defined. The most common choice is using the period prior to the event window, which has been used in this case (Mackinlay, 1997). In this Study-Event analyses there is an estimation-window of 60 days prior to the announcement date, excluding the event-window of ten days. With the
parameter estimates for the normal performance model, the abnormal return can be calculated.
For this event study, the market model method has been used. The daily abnormal return is computed by subtracting the predicted normal return from the actual return for each day in the event window. The sum of the abnormal returns over the event window is the cumulative abnormal return (Mackinlay, 1970).Then we set the cumulative abnormal return equal to the sum of the abnormal returns for each company.
The next step is to run a test statistic to check whether the average abnormal return for each stock is statistically different from zero.
Test statistic: 𝑇 =
∑ 𝐴𝑅 𝑁 𝐴𝑅𝑆𝐷
√𝑁
AR: Abnormal return
ARSD: Abnormal return standard deviation.
Comparing the absolute value of the test statistic result to the standard normal distribution value with an alpha of five percent, there can be determined whether the average abnormal return for that stock is statistically different from zero. This comparing value is 1.96, which is obtained from this standard normal distribution of five percent. If the absolute value of test is greater than 1.96, then the average abnormal return for that stock is significantly different from zero at the five percent alpha level.
In addition, there will be a cross-sectional test computation of the
cumulative abnormal return for all companies treated as a group. The P-value on the constant of this regression will give the significance of the cumulative
abnormal return across all companies. If the p-value is below 0.05, it is valid to say that there is a significant influence of the event of the return on the stocks.
Results
In this particular section, there will be an examination of the obtained output results. First, the output of the OLS-regression analyses will be examined. After that, the output of the Event-Study will be discussed.
OLS-regression output
The OLS-regression results in two outputs. One involves the return on stock of the acquiring firm and another involving the return on stock of the target firm.
Stata output return on stock of acquiring firm
The first thing that is being discussed is the R-squared. R-squared is the proportion of variance in the dependent variable (return on stock price of the acquiring firm), which can be explained by the independent variables (Norden, 2014) .
This is an overall measure of the strength of association and does not reflect the extent to which any particular independent variable is associated with the dependent variable. As displayed, the R-squared is very low. This means that the return on stock of the acquiring is not influenced by all the independent variables together (Norden, 2011).
The regression equation is presented in many different ways. Below, every independent variable will be taken into consideration and explained what it means if they are statistically different from zero (tested with alpha of 5%). - CDS: for every unit increase in CDS, we expect a 0.2402283 unit increase in the return on the stock of the acquiring firm, holding all other variables
constant. Since CDS is coded as 1 or 0 (1=Debt financed CDS, 0=Equity financed CDS) the interpretation is more simply: for debt financed CDS, the predicted return on stock of the acquiring firm would be $0.2402283 higher than for the acquiring firms that doesn’t have a CDS on there debt.
- Dummy_NT: The coefficient for the ratio of the net notional amount of the deal to the total assets of the acquiring firm is -0.007528. So for every unit increase in the 1% increase in ratio, there will be a predicted decrease of $0.007528 on the return of the acquiring’s stock.
- Stock_a_on: The coefficient for the stock price of the acquiring firm is -0.0007528. So for every $1,00 increase in the stock price of the acquiring firm,
the predicted return on stock of the acquiring firm during this event will decrease by $0,0007528.
- Stock_t_on: The coefficient for the stock price of the target firm is 0.003417. So for every $1,00 increase in the stock price of the target firm, the predicted return on stock of the acquiring firm during this event will increase by $0,003417
- Tobin_q: The coefficient for Tobin’s q is 0.6279754. Recall that Tobin’s q stands for the ratio between a total asset's market value and its replacement value. When this ratio increases by 1%, the predicted return on stock of the acquiring firm will increase by $0.6279754.
Ln(mar_cap): The coefficient for the profitability ratio of the acquiring firm is -0.149568. So, if the logarithm of the market capitalization increases by 1, the predicted return on stock of the acquiring firm will decrease with $0.149568. - I = Profitability: The coefficient for the profitability index is -0.0004259. So, if the profitability index increases by 1, the predicted return on stock of the acquiring firm will decrease with $0,0004259.
- _cons: -0.338854. These are the standard errors associated with the
coefficients. As can be seen, this variable has the highest prediction coefficient, so there can be concluded that there are a lot of standard errors in this model.
To see if there is a variable that has a significant influence on the return on stock of the acquiring firm, P > |t| has to be taken into consideration. If this value is below 0.05 (tested with an alpha of 5%), the independent variable has a significant influence on the dependent value. In this regression output there is no independent variable with a p-value below 0.05. None of the independent
variables are not statistically significantly different from 0 because all p-values are definitely larger than 0.05. So, there is no independent variable that has a significant influence on the return on stock of the acquiring firm.
Stata output return on stock of target firm
As displayed, the R-squared is a lot higher than the R-squared of the regression with the return on stock of the acquiring firm, but still not quite high. This means that the return on stock of the target firm is more influenced by all the
independent variables together, in comparison with the acquiring firm return on stock.
The values for the regression equation for predicting the dependent variable are displayed beneath y_a_ta from the independent variables. The regression equation is presented in many different ways. Below, every independent variable will be taken into consideration and explained what is means if they are statistically different from zero (tested with alpha of 5%). - CDS: for every unit increase in CDS, there is an expected increase of
$0.0424417 in the return on the stock of the acquiring firm, holding all other variables constant. Since CDS is coded as 1 or 0 (1=Debt financed CDS, 0=Equity financed CDS) it is valid to say that for debt financed CDS, the predicted return on stock of the acquiring firm would be $0.0424417 higher than for the target firms that does not have a CDS on there debt.
- Dummy_NT: The coefficient for the ratio of the net notional amount of the deal to the total assets of the acquiring firm is -0.0013616. So for every unit increase in the 1% increase in ratio, there will be a predicted decrease of $0.0013616 on the return of the targets’ stock.
- Stock_a_on: The coefficient for the stock price of the acquiring firm is -0.0000711. So for every $1,00 increase in the stock price of the acquiring firm, the predicted return on stock of the target firm during this event will decrease by $0,0000711.
- Stock_t_on: The coefficient for the stock price of the target firm is 0.0001631. So for every $1,00 increase in the stock price of the target firm, the predicted return on stock of the acquiring firm during this event will increase by $0,0001631
- Tobin_q: The coefficient for Tobin’s q is -0.0019801.When this ratio increases by 1%, the predicted return on stock of the acquiring firm will decrease by $0.0019801
- Ln(mar_cap): The coefficient for the logarithm of the market capitalization of the acquiring firm is 0.0030813. So, if the logarithm of the market capitalization increases by 1, the predicted return on stock of the acquiring firm will increase with $0.0030813
- I = Profitability: The coefficient for the profitability index is -0.0002238. So, if profitability index increases by 1, the predicted return on stock of the acquiring firm will decrease with $0,0002238.
_cons: -0.0143854. These are the standard errors associated with the coefficients.
In this regression output, there two independent variables that have a significant influence on the return on stock of the target firms, the CDS variable and the stock price of the target on the announcement date. Both coefficients are significantly different from zero, because both p-values are smaller than 0.05 (tested with an alpha of 5%).
Stata Event-Study output
As showed in the output, there are two companies whose test statistic is below or above 1.96, thus having a p-value below 0.05. So, it is valid to say that there is a statistically significant influence of the event on the return on stocks if you only look at these two companies.
The cross-sectional test shows us a p-value of 0.21, that is definitely larger than 0.05. Concluding that within the Event-Study test with 28 observations, there is no statistically influence of the event on the return on stock of those companies, who faced a merge or acquisition.
Conclusion
In this analysis, there has been examined how the returns on stock of firms, who were subject to an event of a merge or acquisition financed with a Credit Default Swap, changes due to this particular event. The 150 largest M&A’s during the period of January 2010 up until the end of December 2014 were taken into consideration, questioning whether the construction of a Credit Default Swap (financed on debt or on equity), the ratio of the net notional amount of the deal to the total assets of the acquiring firm and the announcement of a Credit Default Swap carries new information to the market to establish a abnormal return on stock of those companies.
First, by examining the returns on stock of the acquiring firm through an OLS-regression analysis, the results did not give enough evidence statistically to conclude that the construction of a Credit Default Swap or the ratio of the net notional amount of the deal to the total assets of the acquiring firm have a significant influence on the return on stock of the acquiring firm. This result is not consistent with the Efficient Market Hypothesis.
Second, by examining the returns on stock of the target firm throughout an OLS-regression analysis, the results gave statistically enough evidence to conclude that the construction of a Credit Default Swap has a significant influence on the return on stock of the target firm. A part of this result is consistent with the Efficient Market Hypothesis.
Third, using an Event-Study method, there can be said that some firms are significantly influenced by the information content of the announcement. But overall, the result of the cross-sectional test shows us that the information content of the announcement has all but a significant influence on the return of stocks. The result is also not consistent with the Efficient Market Hypothesis.
Comparing this with the hypothesis for the Event-Study, the results of the Event-Study do not match the hypothesis. So, within the available data in this these, it is valid to say that the information content of the constructing, amount and announcement data is (mostly) already included in the stock prices, so there is no abnormal return due to the event of a credit announcement.
Considering the hypothesis for the OLS-regression analysis, the third hypothesis does not match the results of the OLS-regression on the return on stock of the acquiring firm either. So, its appropriate to say that the return on stock of the acquiring firms are not statistically significant influenced by the event of the Credit Default Swap. This also rules out our first hypothesis for the OLS-regression analysis
Taking the second hypothesis for the OLS-regression, this hypothesis matches the result of the OLS-regression on the return on stock of the target firm. This implies that the finance construction of the Credit Default Swap and the stock price of the target at the announcement dates does have a statically significant influence on the return on stock of the target firm.
For further research, a more specific methodology to capture the differences occurring due to the time interval has to be developed. The time at the announcement date of the announcement itself as an event window has to be specified instead of the whole day of the announcement to compare it with the other dates and using a larger database with other listed stocks from other indexes from around the world is desirable for obtaining a model that is closer to the reality.
References
Collin-Dufrense, P. & Goldstein, R.S. & Martin, J. (2001). The Determinants of
Credit Default Spread Changes. Journal of Finance, Vol. 56, No. 6, 2177-2207 Data CDS Notional Amount. Retrieved from
http://www.isda.org/statistics/pdf/ISDA-Market-Survey-historical-data.pdf (11 June 15).
Galil, Koresh & Shapir, Offer Moshe (2014). The determinants of CDS spreads. Journal of Banking and Finance, Vol.41, 271-282
Greatrex, C.A. (2009). Credit default swap market determinants. Journal of Fixed Income, Vol.18, No. 3, 18-34
Griffin, Paul A (2014). The market for credit default swaps: new insights into
investors' use of accounting information? Accounting & Finance, Vol.54, No. 3,
847-883
Hull, J., & Predescu M. (2004). The relationship between credit default swap
spreads, bond yields, and credit rating announcements. Journal of Banking and
Finance, Vol.28, No. 11, 2789-2811
Hull, John & White, Alan (2003). The valuation of credit default swap options Journal of Derivatives, Vol.10, No. 3, 40-51
Jamshidian, Farshid (2004). Valuation of credit default swaps and swaptions. Finance and Stochastics, Vol.8, No. 3, 343-371
Jorion. P. & Zhang. G. (2007). Good and bad credit Contagion: Evidence from Credit
Default Swaps. Journal of Financial Economics, Vol. 84, 860-883
Juurikkala, Oskari , Weistroffer (2012). Credit Default Swaps and the EU Short
Selling Regulation: A Critical Analysis. European Company and Financial Law
Review, 2012, Vol.9, No. 3, 307-341
Leland & Toft. (1996). Calibrating structural models: a new methodology based on
stock and credit default swap data. Journal of Finance, 51, 987–1019
MacKinlay, A.C. (1997). Event Studies in Economics and Finance. Journal of Economic Literature, Vol. 35, 13-19
Markosea, S.& Giansanteb, S. & Shaghaghic, A.R. (2012). ‘Too interconnected to
fail’ financial network of US CDS market: Topological fragility and systemic risk.
Journal of Economic Behavior & Organization, Vol. 83, 627–646
Mengle, David (2007). Credit derivatives: an overview. Economic Review (Atlanta, Ga.), Oct, 2007, Vol.92, No 4, 1-25
Norden, L.,& Weber, M. (2004) . Informational efficiency of credit default swap
and stock markets: The impact of credit rating announcements. Journal of Banking
and Finance, Vol.28, No. 11, 2813-2843
Norden, L. & Weber, M. (2009). Co-movement of Credit Default Swap, Bond and
Stock Markets: an Empirical Analysis. European Financial Management, Vol. 15,
No. 3, 529–562
O’Kane, D. & Turnbull, S. (2003). Valuation of Credit Default Swaps. Fixed Income Quantitative Credit Research. Lehmann Brothers.
Thomson Reuters Datastream. Retrieved from
http://www.http://thomsonreuters.com/en.html (2015, 11 June).
Trutwein, P. & Ramchander, S. (2011). Jumps in Credit Default Swap Spreads and
Stock Returns. Journal of Fixed Income, Vol.20(3), 56-71
Qiu, J. & Yu, F. (2012). Endogenous liquidity in credit derivatives. Journal of Financial Economics, Vol. 103, 611–631
Wharton Data Research center. Retrieved from
Appendix
List of companies
Acquiring Company Name Target Company Name
Potash Corp of Saskatchewan BHP Billiton PLC
Swiss Reinsurance Co Ltd Swiss Reinsurance Co Ltd
Mobile Telecommunications Co Emirates Telecommunications
British Sky Bdcstg Grp PLC News Corp
CareFusion Corp Becton Dickinson & Co
Polyus Zoloto KazakhGold Group Ltd
Berkshire Hathaway Inc Lubrizol Corp
Microsoft Corp Skype Global Sarl
Lihir Gold Ltd Newcrest Mining Ltd
Perrigo Co Elan Corp PLC
Bausch & Lomb Inc Valeant Pharmaceuticals Intl
Zain Africa BV Bharti Airtel Ltd
AMP Ltd AXA Asia Pacific Holdings Ltd
Talisman Energy Inc Repsol SA
International Power PLC GDF Suez Energy Services Inter
Level 3 Communications Inc tw telecom inc
VimpelCom Ltd Weather Investments Srl
Applied Materials Inc Tokyo Electron Ltd
Duke Energy Corp Progress Energy Inc
Prudential PLC AIA Group Ltd
Exelon Corp Constellation Energy Group Inc
ConAgra Foods Inc Ralcorp Holdings Inc
Nalco Holding Co Ecolab Inc
Petroleo Brasileiro SA Brazil-Oil & Gas Blocks
Kinder Morgan Inc El Paso Corp
Bristol-Myers Squibb Co Amylin Pharmaceuticals Inc
Vivendi SA SFR
PPL Corp E.ON US LLC
AT&T Inc T-Mobile USA Inc
Sanofi-Aventis SA Genzyme Corp
Eaton Corp Cooper Industries PLC
Walgreen Co Alliance Boots GmbH
ProLogis AMB Property Corp
ASX Ltd Singapore Exchange Ltd
Nippon Steel Corp Sumitomo Metal Industries Ltd
Hewlett Packard Co Hewlett Packard Co
CenturyLink Inc Qwest Commun Intl Inc
AT&T Inc DirecTV Inc
DISH Network Corp Sprint Nextel Corp
AIG AIA Aurora LLC
Lowe's Cos Inc Lowe's Cos Inc
Comcast Corp NBCUniversal Media LLC
Allegheny Energy Inc FirstEnergy Corp
Tyson Foods Inc Hillshire Brands Co
Roche Holding AG Illumina Inc
Ameriprise Financial Inc Ameriprise Financial Inc
Takeda Pharmaceutical Co Ltd Nycomed Intl Mgmt GmbH
BP PLC NK Rosneft'
Reliance Industries Ltd-21 Oil BP PLC
Community Health Systems Inc Health Management Assoc Inc
America Movil SAB de CV Carso Global Telecom SAB de CV
Express Scripts Inc Medco Health Solutions Inc
Novartis AG Alcon Inc
Ensco PLC Pride International Inc
Time Warner Cable Inc Comcast Corp
Telefonica SA Brasilcel NV
Deutsche Boerse AG NYSE Euronext
Alpha Natural Resources Inc Massey Energy Co
Johnson & Johnson Synthes Inc
Sigma-Aldrich Corp Merck KGaA
Medtronic Inc Covidien PLC
Merck & Co Inc Cubist Pharmaceuticals Inc
Reynolds American Inc Lorillard Inc
Schlumberger Ltd Smith International Inc
Williams Partners LP Access Midstream Partners LP
Time Warner Inc Time Warner Inc
BP PLC Devon Energy Corp-Assets
United Technologies Corp Goodrich Corp
Schlumberger Ltd Schlumberger Ltd
Thermo Fisher Scientific Inc Life Technologies Corp
Caterpillar Inc Bucyrus International Inc
Novartis AG GlaxoSmithKline PLC-Oncology
Amgen Inc Onyx Pharmaceuticals Inc
RSA Insurance Group PLC Aviva PLC-Insurance Business
Halliburton Co Baker Hughes Inc
Glencore International PLC Xstrata PLC
Reliance Industries Ltd-21 Oil BP PLC
21st Century Fox Inc Time Warner Inc
Allergan Inc Actavis PLC
Time Warner Cable Inc Charter Communications Inc
Allergan Inc Valeant Pharmaceuticals Intl
NK Rosneft' TNK-BP Ltd
Shire PLC AbbVie Inc
Iliad SA T-Mobile US Inc
SoftBank Corp Sprint Nextel Corp
Holcim Ltd Lafarge SA
RHB Capital Bhd CIMB Group Holdings Bhd
Liberty Media Corp Sirius XM Radio Inc
America Movil SAB de CV Koninklijke KPN NV
Numericable Group SA SFR
Liberty Global Inc Virgin Media Inc
Facebook Inc WhatsApp Inc
Omnicom Group Inc Publicis Groupe SA
BT Group PLC EE Ltd
BHP Billiton PLC Petrohawk Energy Corp
Burger King Worldwide Inc Tim Hortons Inc
Volkswagen AG Dr Ing hcF Porsche AG
Zimmer Holdings Inc Biomet Inc
Loblaw Cos Ltd Shoppers Drug Mart Corp
Icahn Enterprises LP The Clorox Co
CSR Corp Ltd China CNR Corp Ltd
E-Plus Mobilfunk GmbH & Co KG Telefonica Deutschland Holding
Shanghai Jinfeng Investment Co Greenland Holding Group Co Ltd
IntercontinentalExchange Inc NYSE Euronext
Telefonica Brasil SA GVT Participacoes SA
Family Dollar Stores Inc Dollar General Corp
Deutsche Annington Immobilien Gagfah SA
Wisconsin Energy Corp Integrys Energy Group Inc
Family Dollar Stores Inc Dollar Tree Inc
Google Inc Motorola Mobility Holdings Inc
US Foods Inc Sysco Corp
