The risk-return profile of Contingent Convertible Bonds from an investor’s perspective

(1)

1

J.E. Ligterink

09/02/2015

Bachelor’s Thesis

The risk-return profile of Contingent

Convertible Bonds from an investor’s

perspective

Valérie Dekker

10181113

BSc Economics and Business

Specialization: Finance and Organization

(2)

Index

§1. Introduction………....…2

§2. CoCo’s main features and risks………...…5

§2.1. CoCo’s, the new capital instruments for financial institutions………...…..…..5

§2.2. Investors in CoCos………...….6

§2.3. Trigger and conversion mechanism………..….….6

§2.4. Classification……….…...7

§2.5. Risks of investing in CoCos………9

§2.5.a. Base interest rate……….…...10

§2.5.b. Credit spread for subordinated bonds……….……...10

§2.5.c. Subordination premium……….…….10

§2.5.d. CoCo’s base premium……….………11

§2.5.e. Issuer’s call-right premium……….11

§2.5.f. Possible interest-loss premium………...12

§2.5.g. Regulatory loss absorption premium………..…12

§2.5.h. Contractual loss absorption premium………..……….…..12

§2.6. The spread between CoCos and senior debt……….….…13

§3. Methodology……….….…14

§4. Results………...….16

§5. Conclusion……….…20

§6. References………..…22

§7. Appendix………..…..25

§1. CoCo contractual features………..…..….25

§2. Summary statistics……….…...26

§2.1. Credit Suisse……….26

§2.2. Barclays……….27

§2.3. Credit Agricole………..28

(3)

3 §2.5. Banco Bilbao………..29 §3. Frequency distributions………..…29 §3.1. Credit Suisse………...29 §3.2. Barclays………..32 §3.3. Credit Agricole………...33 §3.4. KBC………...….34 §3.5. Banco Bilbao………..…34 §4. Regression outputs………..35 §4.1. Credit Suisse………35 §4.2. Barclays………..…37 §4.3. Credit Agricole………38 §4.4. KBC……….…...38 §4.5. Banco Bilbao………...38

§5. Regression outputs compared……….…...39

§6. T-statistics……….…...39

§7. Value development of CoCos, bonds and stocks……….40

§7.1. Credit Suisse……….40

§7.2. Barclays………41

§7.3. Credit Agricole……….41

§7.4. KBC………..…41

(4)

§1. Introduction

During the financial crisis, banks found themselves saddled with too much debt, and weren’t able to absorb systemic trading and credit losses (Harvard Law Review, 1991; Bank For International Settlements, 2010). Nowadays, banks are seeking to reorganize their capital structures to avoid financial distress (Harvard Law Review, 1991). In 2009, higher minimum capital requirements were introduced in the new Basel III accord, which focusses on strengthening banks’ capital structure (Bank For International Settlements, 2010).

In addition to the new capital requirements, the Basel Committee and the Financial Stability Board recommend financial institutions to issue Contingent Convertible Bonds to obtain stronger capital buffers (Bank For International Settlements, 2010). Based on already existing ideas, these Contingent Convertibles are debt contracts that are aimed at automatically rearranging capital structures by providing for a mandatory conversion of debt to equity or are written down partially or completely as banks reach a pre-specified distress threshold (Harvard Law Review, 1991).

Contingent convertible bonds, also known as CCBs, CoCo’s or Enhanced Capital Notes, were introduced primarily for implementing prudential bank regulation. As CCBs are a beneficial capital instruments for regulators and banks, the majority of the articles concerning contingent convertibles were written from regulator’s perspective. Articles taking this perspective into account were written by Marquardt & Wiedman (2005), Chan & Wijnbergen, Albul et al (2014), Harvard Law Review Association (2010), Veiteberg et al (2012), Guschin et al (2008), and Roggi et al (2013).

However, CCBs are also interesting to analyse from an investor’s perspective. Since CCBs are recently issued instruments, offering high returns, they are getting increasingly popular among investors. By analysing these CoCos from an investor’s perspective, which is a different approach than adopted in already existing articles, the value of my thesis is to offer potential investors new insights about the risk-return profile of CCBs compared to that of stocks and bonds. This purpose is therefore reflected in my research question: What does the risk-return profile of a CoCo look like, compared to that of stocks and bonds?

In the second section I will point out the main features of a CoCo and the risks that affect its total return. In the third section I will describe the methodology used to find the answer on my research question. In the fourth section I make use of the results I found to give an answer on my research question. Finally, in the fifth section I summarize my findings.

(5)

5

§2. CoCos’ main features and risks

§2. 1. CoCos, the new capital instruments for financial institutions

The financial crisis, that started in 2007, revealed some very significant deficiencies in the Basel I and Basel II Accords. Consequently, Basel III was formed. The new framework was meant to reinforce the stability of the financial system by strengthening its minimum capital requirements (Miu, Ozdemir, & Giesinger, 2010). To be more precise, banks have to keep their total capital at a level that is at least 10,5% of their risk weighted assets According to Basel III, total capital consists of Common Equity Tier 1 (CET1), a Capital Conservation Buffer, Additional Tier 1 (AT1) Capital, and Tier 2 (T2) Capital (Bank For

International Settlements, 2010). Figure 1 depicts the structure of the minimum capital requirements by Basel III. As you can see from the bar chart in the middle and to the right, Common Equity Tier 1 consists of equity capital, profits carried forward, and reserves. The Capital Conservation buffer is an extra 2,5% of CET1 capital on top of the 4,5% minimum requirement. The Additional Tier 1 capital consists of extra capital stock. The Tier 2 capital is formed by loss-absorbing capital, for instance preferred stock, and other subordinated instruments (Bank For International Settlements, 2010; Credit Suisse Asset Management, 2014).

Figure 1

Basel III: minimum capital requirements for financial institutions

(6)

To meet these Basel III requirements, the issuance of a special kind of hybrid bond became popular within the banking sector: the so-called ‘Contingent Convertible Bonds’, also known as CoCos, CCBs, or Enhanced Capital Notes. These hybrid bonds, which were issued by more than 300 firms in the first half of 2004 (Marquardt & Wiedman, 2004), were designed to absorb losses either by converting into common equity or by suffering a principal write down (PWD) when a certain threshold or trigger level is reached (Avdjiev, Bogdanova & Kartasheva, 2013).

§2. 2. Investors in CoCos

Investors in CoCos are primarily asset managers, hedge funds, the insurance industry and employees of financial institutions. Since asset managers have been investing in hybrid bonds before, they have the expertise to manage their large existing Tier 1 and Tier 2 products properly. Hedge funds are attracted by CCBs, because they are more risk-tasking, and are, to some extent, able to hedge away the unwanted risks. Insurance companies favour investing in short dated, high rated debt. Although CoCos are neither classified as short dated, nor as high rated, a minority of insurance companies’ investment portfolio consists of CoCos, since these debt instruments offer high yields, but do not meet the requirements for prudential investing. Finally, there are cases in which financial institutions pay their top-management a salary that partially consists of CoCo bonuses (De Spiegeleer & Schoutens, 2014).

§2. 3. Trigger and conversion mechanism

The trigger and conversion mechanism are the two main defining characteristics of a CoCo (Bank For International Settlements, 2010). At first, triggers could be based on balance sheet ratios, share prices, or ‘Point Of Non-Viability’ (PONV) and are further categorized as low or high (Deutsche Bank Research, 2011)

Triggers based on balance sheet ratios, like for instance the Common Equity Tier 1 (CET1) ratio1_,

are highly inefficient, because the accuracy of this ratio depends strongly on the frequency at which these ratios are calculated (Flannery, 2009; Credit Suisse Asset Management, 2014; Bank For International Settlements, 2010), which means they would, in most cases, be reported quarterly (Deutsche Bank Research, 2011).

Share price triggers could offer a solution to inconsistent balance sheet ratios, like the CET1 ratio mentioned in the previous paragraph. Share price triggers are set at a minimum share price, and can be followed and monitored more often by investors as well as by the bank itself (Deutsche Bank Research, 2011). The downside of the share price trigger is that it is subject to the possibility of stock price

(7)

7 manipulation (Schoutens & De Spiegeleer, 2012). For example, an arbitrageur buys a CCB and short-sells shares of the CCB-issuing bank to drive the price down towards the pre-determined share price trigger, which means conversion will take place automatically. The arbitrageurs obtain the benefits from the conversion into shares once the stock recovered to its fair value (Duffie, 2009; MacDonald, 2009).

In addition to numerical triggers, triggers based on the regulator’s judgment are included in the majority of CCB contracts. When the trigger is based on the regulator’s judgement, it is also known as the ‘point of non-viability’ (PONV) trigger. Although this concept is closest to efficiency, because conversion can take place based on the regulator’s insights instead of balance sheet ratios or share prices, danger

conceals in the fact that for investors it is very difficult to assess the probability and timing of conversion. An equivalent to the PONV-trigger is the one based on the issuer’s discretion, which also makes the estimation of the probability of a trigger event highly inaccurate (Bank For International Settlements, 2010)

Whether numerical or discretionary, triggers are set at a high or low level. High triggers are set at a minimum level of 7%, based on the 4,5% CET1 ratio plus the 2,5% capital conservation buffer ratio to Risk Weighted Assets. The lower level trigger is set at 5,125%, which is slightly above the CET1 ratio (Credit Suisse, 2014). The lower the trigger level of conversion, the more chance the bank is in default. Therefore triggers at the 5,125% level are also known as ‘gone-concern’ triggers, and for the higher trigger levels the term ‘going concern’ is used (Bank for International Settlements, 2010).

As stated before, when the trigger level is reached, the CCB converts automatically into equity or is subject to a principal write-down. The conversion to equity takes place at a pre-specified conversion rate, which could be the market price of the stock at the time the trigger is breached, the share price at the time of the CoCo issuance, or a combination of the two (Bank For International Settlements, 2010). The number of shares to be converted in could be fixed or variable (Credit Suisse, 2014).

The principal write-down of a CoCo could be categorized as temporary, permanent, complete, or it is in the form of a partial write-off (Credit Suisse, 2014; Bank For International Settlements, 2010). For

instance, holders of Rabobank CoCos, issued in March 2010, take the risk to lose 75% of the face value and receive 25% of it in cash in the case that their CoCos are written down (Bank For International Settlements, 2010).

§2. 4. Classification

As maturities differ across bonds, this also holds for CCBs. Dependent on their term, ranking and coupon payment rules, they could be classified as (Additional) Tier 1 (AT1) or Tier 2 (T2) CoCos. An Additional Tier 1 Coco is ranked directly ahead of equity capital, profits carried forward, and reserves in a bank’s capital structure. The coupon payments can be cancelled by the issuer at any point, for any reason, and for any length of time. This is independent of whether the trigger level is breached. Besides, cancelled

(8)

payments do not accumulate, and are instead written off (European Securities and Markets Authority, 2014). AT1 CoCos are typical for b eing perpetuals, but the issuer has a call option, which means the CoCo could be redeemed by the issuer at a pre-specified date in the future (Credit Suisse, 2014).

A Tier 2 CoCo, is ranked directly ahead of AT1 CoCos, together with other subordinated financial instruments and extra capital stock. Unlike AT1 CoCos, the coupons of T2 CoCos cannot be cancelled. Generally, Tier 2 CCBs have fixed maturities, and there is no call option for the issuer (Credit Suisse, 2014). As might be confusing, whether a CoCo is in the AT1 or T2 category has nothing to do with the trigger level. For both categories there exist high and low level triggers. Figure 2 represents an example of some CoCos that are structured differently from each other.

Due to the option to convert, a CoCo is a combination of debt and equity and is therefore positioned in between equity and s enior debt, under preference shares, and over convertible bonds and subordinated debt, as shown in Figure 3 (Schoutens & De Spiegeleer, 2012). Mehra (2003) states that ‘the standard deviation of the returns to stocks (about 20% a year historically) is larger than that of the returns to Treasury bills (about 4 percent a year)’, and that stocks are therefore ‘riskier’ than bonds. As Teneberg (2012) states: ‘To compensate for this extra exposure towards losing their invested money, the shareholders demand an overall larger return than the bondholders’. As equity is seen as most risky, and senior debt being almost risk-free, this explains why CoCos are positioned in between debt and equity, since their risk profile is based on the combination of debt and equity . According to Mehra’s theory, the total return on a CoCo should therefore fluctuate in between the total returns for debt and equity. Investors in CoCos, focussing on high expected returns, are therefore willing to accept a higher volatility than the volatility on Treasury Bills (Fama

Figure 2

Different CoCo structures

(9)

9 and French, 2003). In addition, comparing CoCos to senior debt, investors will ask an extra yield on top of the risk-free rate in order to be fairly compensated for the expected losses (Maes& Schoutens, 2010; Schoutens & De Spiegeleer, 2012). Moreover, the higher the risk level on CoCos, dependent on their contractual terms, the higher the yield demanded on them by investors (Pennacchi, 2010).

§2. 5. Risks of investing in CoCo’s

The risks involved with issuing and investing in CCBs still aren’t well-defined and classified, and therefore it is hard for issuers to consistently determine a CoCo’s coupon rate, which depends on the level of risk, while for investors it is hard to determine whether the total return is compensating them enough for the risks they are taking. Credit Suisse’s Asset Management department did specific research on the CoCo risks, and classified the premia where the total return on a CoCo is built upon. In the next sections I used this framework, which is represented in Figure 4, to explain where these different premia are based upon, in order to give a structured explanation of the CoCo’s risk-return profile.

Figure 3

Position of CoCos on a financial institution’s balance sheet

(10)

1. 5a. Base interest rate

§2.5a. Base interest rate

Starting from the lower left corner to end at the upper right corner, the base interest rate of a CoCo is determined by its currency and term, and is incorporated in all securities issued within the bank. As

explained earlier, terms vary across CCBs. For instance, Tier 2 CCBs have fixed maturities, generally ranging from 5 to 20 years, while AT1 CoCos are generally perpetuals (Schoutens & de Spiegeleer, 2014).

§2. 5b. Credit spread for subordinated bonds

The credit spread for subordinated bonds is the difference between the interest earned on senior bonds over Treasury bonds (Credit Suisse, 2014). This spread depends on a wide range of risks that are firm-specific, as well as systemic, and is incorporated in all returns of the securities issued within the bank, from returns on senior debt to returns on common equity.

§2. 5c. Subordination premium

The subordination premium is the difference between the interest earned on subordinated bonds and senior bonds. This premium is based on all other risks involved with investing in a subordinated bond than

Figure 4

A CoCo’s total return breakdown. The total return on a CoCo can be divided in various components. Total returns can vary per CoCo due to the fact that total returns on CoCos consist of different premia, as shown in the figure. Issuer’s call-right premium applies to AT1 CoCos and some AT2 CoCos, whereas the possible interest-loss premium applies only to AT1 CoCos.

(11)

11 the risks that apply for the senior bonds. I won’t discuss these risks in my thesis, because they are beyond the scope of my researchquestion. This premium is incorporated in all securities’ returns ranging from

unsecured debt returns to common equity returns (Credit Suisse, 2014). §2. 5d. CoCo’s base premium

Although not represented in Figure 4, there should be another risk premium class in between the subordination premium and the issuer’s call-right premium, of which the latter is discussed in the next paragraph. This not depicted premium I call the ‘CoCo’s base premium’. The CoC’os base premium is affected by the sector risk, single risk, and level-of-liquidity risk on a CoCo (Credit Suisse, 2014; European Securities and Markets Authority, 2014).

Sector risk is the risk that exists when the event of a banking crisis takes place, and the valuations of many CoCos will come under heavy pressure, because they will be closely correlated with each other. This increases the probability of making a loss for investors (Credit Suisse, 2014). The sector risk affects all returns of CoCos, and could therefore be classified as a systemic risk (Berk & DeMarzo, 2011).

Single security risk is incorporated in every sort of CCB, since CoCos in general have a higher default risk than senior and unsecured debt with comparable terms, because of their equity- or principal write-down features. This loss absorption mechanism on its own, characterizing the CoCo, makes it more risky to invest in this kind of debt, independent of a CoCos trigger level, which can be high or low. This theory is in line with the theory presented in the research done by Mehra (2003), since there has to be a risk premium that accounts for the fact that a CoCo is a debt instrument with equity features (Schoutens & De Spiegeleer, 2012). Once the CoCo is converted, it acts like stock, which can increase in value, decrease in value, or completely lose its value, of which the last option is similar to a principal write-down.

Finally, the liquidity risk is an important factor affecting the CoCo’s base premium (Schoutens & De Spiegeleer, 2014). Nowadays, CoCos still are a relatively new asset class with a lot of regulatory- and model uncertainty. The amount of CoCos traded in the European market seems to be growing, but still isn’t

comparable to the amount outstanding for senior debt, which makes CoCos less easy to sell, so less liquid (Maes & Schoutens, 2010; European Securities and Markets Authority, 2014).

§2. 5e. Issuer’s call-right premium

The issuer’s call right premium only applies to T2 CoCos and is determined solely by the issuer’s option to redeem the CCB on an early base (Bank for International Settlements, 2010). The issuer’s call option is disadvantageous for the investor, because he or she has to sell its CoCo for a lower price than the CoCo’s market value. This disadvantage has an increasing effect on the call-right premium.

(12)

On the other hand, the extension risk has to be taken into account, as investors cannot blindly assume that the call option will be exercised by the issuer. This means, when assuming the call option will be

exercised and the CoCo will be redeemed, but in reality it doesn’t materialize, the term of the CoCo is ‘extended’ from this point of view (European Securities and Markets Authority, 2014; Schoutens & De Spiegeleer, 2014). This ‘extension risk’ has to betaken into account when calculating the issuer’s call right premium, although it is very hard to predict whether this kind of risk has an increasing or decreasing effect on the call right premium.

§2. 5f. Possible interest-loss premium

The possible interest-loss premium applies only to AT1 CoCos , in some cases also for T2 CoCos, and is determined by the probability that the coupons on a CoCo will be cancelled by the issuer. However, this probability is hardly calculable, since it depends on the bank’s future achievements and capital structure decisions. For example, if the bank is approaching the high trigger levels of AT1 CoCos, it might decide to cancel its coupon payments in order to strengthen its capital structure and prevent these CoCos from triggering.

§2. 5g. Regulatory loss absorption-premium

The regulatory loss absorption-premium is dependent on the regulator’s judgment and the future financial state of the bank. The connection between the factors influencing the regulatory loss-absorption premium is the fact that the bank’s future financial position, if distressed, could be the regulator’s motive to exact the CoCos to convert or to be written down (European Securities and Markets Authority, 2014). As the timing of the CoCo being converted or written down on the regulator’s initiative - if it materializes at all - is uncertain, the investor will demand compensation for this uncertainty. Therefore the probability that the regulator could force a bank’s CCBs to trigger has an increasing effect on the regulatory loss absorption-premium (Deutsche Bank Research, 2011).

§2. 5h. Contractual loss-absorption premium

Finally, the contractual loss-absorption premium is the highest risk premium among the other risk premia added on top of the base interest rate. This premium is determined primarily by the numerical trigger and whether the loss-absorption mechanism is based on conversion or a principal write-down. The higher the trigger level, the higher the probability of conversion or a write-down. At first sight it seems to be more beneficial to invest in low trigger CoCos, because the probability of getting converted or written down is smaller than for high trigger CCBs. However, the downside of a low trigger CoCo, is that the probability of

(13)

13 making a loss is much higher, because at the point when the low trigger level is breached, the bank will be closer to the event of default, which will result in a permanent write down or a very unfavourable timed conversion to equity (European Securities and Markets Authority, 2014). Within the loss absorbency

mechanism, conversion into equity seems to be less severe, because the investor is still partially compensated for his loss with shares. However, fixed income investors will be in trouble when being forced to holdCoCos once they are converted into equity (Deutsche Bank Research, 2011). The partial write-down still gives the investor some compensation for the losses made. However, when written down completely, investors suffer the most (Schoutens & De Spiegeleer, 2014; Deutsche Bank Research, 2011).

Other factors influencing the contractual premium are the volatility of a bank’s earnings and the bank’s total capital ratio2_{. The lower the total capital ratio, due to high leverage or low total capital, the}

higher the risk of the CoCo getting triggered (European Securities and Markets Authority, 2014). Many European banks do not yet appear to have formed sufficient core capital to provide CoCo investors with a sufficiently comfortable buffer above the trigger level (Deutsche Bank, 2011). When concerning the volatility of a bank’s earnings, Deutsche Bank (2011) states that the larger the spread in earnings, the more inconsistent they are, which increases the chance that the capital ratio will decrease under the trigger level, resulting in a conversion or a principal write-down of the CCB.

§2.6. The spread between CoCos and Senior Debt

CoCos will generally carry a higher coupon then the traditional senior debt. From Figure 5, one can see that, on average, CoCos offer a credit spread of 400 basis points over senior debt (Credit Suisse, 2014).

2_{Total capital ratio = Total capital / RWA. With RWA = Risk Weighted Assets.}

Figure 5

The spread between CoCos and Senior Debt. On average, CoCos offer a 4 times spread pickup over senior debt, or over 400 basis points in absolute terms.

(14)

§3. Methodology

To collect data on different CoCos from different banks, I made use of a list of recently issued CoCos published by Barclays (2014). In the period of December 2009 until March 2014, CoCos have been issued by the following banks: Lloyds, Credit Suisse, UBS, Barclays, Banco Popular Espanol, Credit Agricole, KBC Group, Banco Bilbao Vizcaya Argentaria, Santander, and Danske Bank (Barclays, 2014). In order to run regressions, data from equity and bonds had to be selected too. For each CoCo on the list published by Barclays, I selected a bond, which is similar in term to the CCB, with an issue date as recent as possible, and which is ranked as senior debt.

The historical data from the CoCos, equity and debt is downloaded from Bloomberg, during a visit to the Rabobank office at Croeselaan in Utrecht. I have been supervised very well by Marc van Rooijen and Quirijn Reusch while downloading the data and processing it in Excel. I also want to thank Rogier

Raaijmakers for his permission on this. From the data available in Bloomberg, a selection of CoCos and their accompanying equity and debt is made. Not all CoCos were suitable to analyze, because for some banks, data on or the CoCo itself, or on the accompanying equity or debt was not or only partly available, which makes running regressions on these data inaccurate or even impossible. As a result, three different CoCos from Credit Suisse, one from Barclays, one from Credit Agricole, one from KBC and one from Banco Bilbao Vizcaya Argentaria were suitable to analyse on the basis of the research question. Daily data on these CoCos was available from February 2011 or later.

First, I calculated the value development of a CoCo and its accompanying stock and bond is. To give a clear graphical description of this value development, I made use of the logical-test function in Excel. With this function the relative increase or decrease in the value of the CoCo, stock or bond, in comparison with the previous value is measured by assuming that we invest 100 Euro/Dollar/Pound in each CoCo and the same amount in its accompanying stock and bond.

Secondly, I calculated the total return on the CoCos, stocks and bonds selected in order to test if they are normally distributed. I started with calculating their skewness factor and testing for significance of this skewness factor. In addition, I ranked the total returns on CoCo’s from low to high. Then I calculated their cumulative normal distribution, and from this I constructed the inverse of the cumulative distribution, given the standard deviation and average calculated from the CoCos’ total returns. This inverse of the cumulative distribution is the expected value of the CoCo values based on a normal distribution. Finally, the Z-values were calculated by taking the inverse of the standard normal distribution. By setting off the historical total return values of the CoCos versus the accompanying Z-values, and doing the same for the expected values of the total return on a CoCo, two plots are created in one graph to see whether their shape is deviating strongly or not. When not deviating strongly from each other, we can assume the data to be normally distributed. These excel outputs can be found in the appendix.

(15)

15 On top of that, frequency distributions were constructed from data on each CoCo and their

accompanying equity and debt. From these distributions one could assume the data to be normally distributed. The frequency distributions discussed here are represented in the appendix.

By assuming normality of the population means, and robust standard errors, I run regressions and T-tests, which I discuss in more detail below.

For each CoCo, I run two regressions. The first regression is according to the most simple model, with the CoCo’s total return as the dependent variable, and the total returns on equity and debt as the independent variables.

RCoCo= α + βRBond·RBondi + βRStock·RStocki + ei

With RCoCo being the total return of the CoCo being analysed, RBond and RStock representing the total returns on the accompanying bond and stock, and e as the residual.

To the second regression I added the VIX index as a control variable. This index measures the market risk and shows the market’s expectation of 30-day volatility of a wide range of S&P 500 options. For each CoCo being analysed, the second regression is as follows:

RCoCoi = α + βRBond·RBondi + βRStock·Rstocki + βVIX·VIX + ei

With RCoCo being the total return of the CoCo being analysed, RBond and RStock representing the total returns on the accompanying bond and stock, VIX as the market’s expectation of volatility on S&P 500 options, and e as the residual.

Finally, I tested whether the average total return on a Coco is different from the average total return on its accompanying bond. I also executed this test for the difference in average total returns on the CoCo and its accompanying stock. For this testing procedure I used the T-test, as depicted below.

(!"#$%&'&'!!"#$%&'$()

√!" and

(!"#$%&'&'!!"#$%&'()* √!"

With MeanRCoCo as the average total return on a CoCo, RBond and RStock representing the average total returns on the accompanying bond and stock, SD representing the standard deviation of the CoCo’s total returns.

(16)

§4. Results

As the equity holders are the residual claimants, bearing the highest risk among securities, they should receive the highest compensation for this risk in terms of risk premia. Therefore they should, according to Teneberg (2012) obtain the highest total return across securities. As the holders of senior debt are the first claimants, their risk premium is the lowest, or equal to zero. Holders of senior debt should therefore obtain the lowest total return on their investment. All holders of securities in between senior debt and equity, should receive a risk premium in between zero and the highest risk premium on equity. This means that investors in securities ranked in between senior debt and equity, should obtain a total return that is higher than the total return on senior debt and lower than the total return on common equity.

Holders of CoCos are categorized as the group of investors that hold securities ranked in between senior debt and equity, and should therefore obtain a total return in between equity and debt. In addition, Pennacchi (2010) states that the yields on CoCos should rise above default free yields since they are subject to the fluctuations in a bank’s asset returns, or generally, they are riskier than senior and unsecured debt. It would therefore be strange if the value you obtain from investing 100 in a CoCo is higher than the value of equity, or lower than the value of debt when you invest the same amount in these securities. The reasoning that a CoCo’s value should fluctuate in between the value of debt and equity, since a CoCo has a risk level that is in between that for debt and equity, I refer to as ‘economic theory’. This logic hails from the reasoning discussed at the end of section 1.4, where I point out several economic reasonings stated by Mehra (2003), Teneberg (2012), Fama & French (2003), Pennacchi (2010), etc.

The analyses of the values of CoCos and their corresponding stocks and bonds, I put in graphs, to see how the relationship between the values obtained from investing the same amount in a CoCo, a stock and a bond looks like when time goes on. From graph 1 you can see that, from the end of 2012 until February 2014, the value obtained from investing 100 in a CoCo is lower than the value obtained from investing 100 in a stock, and is higher than the value obtained by investing 100 in a bond. In this period, the CoCo behaves according economic theory, because it’s value fluctuates in between the values of equity and debt. However, graph 2 shows a very distorted situation. From this graph it becomes clear that one cannot assume the CoCo values to fluctuate consistently between debt and equity values, therefore this is a deviation from economic theory. In the appendix you can find more graphs of CoCos that behave differently than predicted by economic theory.

Since there are graphs that prove that the CoCo values are outside the range of equity and debt values, there is a deviation from economic theory (Teneberg, 2012). These deviations from the theory could exist for many reasons. It would be very interesting to do some research on the reasons for these deviations, but that is unfortunately beyond the scope of this Bachelor’s thesis.

(17)

17 The regressions I ran, show that a CoCo is generally dependent on stock movement, as shown in Table 1. However, if comparing regression model 1 with regression model 2, we see that the effects of stocks on CoCos become less significant when a control variable is added. Of course I could add many other control variables, since adding only one control variable could end up in omitted variable bias. For further research it could be interesting to examine whether stocks still have a significant effect on CoCos when more variables are added to the regression.

Finally, I tested whether the average total return on a CoCo is significantly different from the average total returns on stocks and bonds. As becomes clear from the Table 2, in most cases the average return on a CoCo is not significantly different from the average total returns on stocks and bonds, despite the fact that CoCos offer a 4 times spread pickup over senior debt, or over 400 basis points in absolute terms (see Figure 5) (Credit Suisse, 2014).

Graph 1

Value development of Barclay’s CoCo, equity and senior debt. Barclay’s CoCo values behave according to economic theory, since the CoCo values fluctuate in between the values for equity and debt.

(18)

Table 1

Regression models 1 and 2: Do the total returns on stocks and bonds have a significant impact on the total return of the CoCo? Model 1 tests whether the total return on stocks and bonds have a significant effect on the CoCo’s total return. Model 2 is similar to model 1, except for the fact that a control variable, the VIX index, is added. When comparing model 2 with model 1, we see that the effects from stocks’ total returns on the CoCo’s total return become less significant, and in the case of Barclay’s CoCo, not significant anymore.

Regression Model 1 CoCo 1 Credit Suisse CoCo 2 Credit Suisse CoCo 3 Credit Suisse CoCo Barclays CoCo Credit Agricole CoCo KBC CoCo Banco Bilbao

Rbond No No Yes* No No No Yes

Rstock Yes ** Yes** No Yes*** Yes* Yes** Yes*

Regression model 2 CoCo 1 Credit Suisse CoCo 2 Credit Suisse CoCo 3 Credit Suisse CoCo Barclays CoCo Credit Agricole CoCo KBC CoCo Banco Bilbao Rbond No No No No No No Yes*

Rstock Yes*** Yes*** Yes* No Yes* Yes*** Yes**

VIX Yes** No No No No No No

*Significant at 1% level, ** Significant at 5% level, *** Significant at 10% level Graph 2

Value development of Credit Suisse CoCo 3, equity and senior debt. Credit Suisses CoCo values do not behave according to economic theory, since the CoCo values do not fluctuate in between the values of debt and equity.

(19)

19

Table 2

T-tests on the difference in average total returns on CoCos and their stocks and bonds: Is the average total return on a CoCo significantly different from the average total returns on the accompanying stock and bond? Generally, this is not the case.

Difference between average total return CoCo and: CoCo 1 Credit Suisse CoCo 2 Credit Suisse CoCo 3 Credit Suisse CoCo Barclays CoCo Credit Agricole CoCo KBC CoCo Banco Bilbao Average total return Bond No Yes No No No No No Average total return Stock Yes Yes No No No No No

(20)

§5. Conclusion

This is the first paper in which the risk-return profile of a CoCo is compared to that of stocks and bonds, from an investor’s perspective. In this paper, I analysed the risks involved with investing in a CoCo according to the total return breakdown framework introduced in Credit Suisse’s research on Contingent Convertibles.

From other research by Schoutens & De Spiegeleer, the Bank for International Settlements, Deutsche Bank, and the European Securities and Markets Authority, I found other important risks involved with CoCo’s, which weren’t included in Credit Suisse’s analysis. Therefore I suggested to add the ‘CoCo’s base premium’, which consummates the framework for the risks not discussed by Credit Suisse. From the reasonings by Mehra, Schoutens & De Spiegeleer, Teneberg, Maes & Schoutens and Pennacchi, I

constructed an ‘economic theory’, which implies that a CoCo’s value should fluctuate in between the value of debt and equity, since its risk profile is in between that for senior debt and equity. I applied this theory in my CoCo analysis in order to find the answer on my research question.

I started my analysis by putting the value developments of the selected CoCos, bonds and stocks in a graphical representation, to find out if CoCos act according to the ‘economic theory’. While I found CoCos behaving according to this theory, I also found CoCos that show a deviating behaviour. Given the fact that the graphs show a CoCo to behave different from economic theory, it is therefore incorrect to assume that the risk-return profile of a CoCo is in between those of senior debt and equity.

From the regressions I found that a CoCo is most affected by the stock’s movements, and in only one of the seven CoCo analyses the bond showed a significant effect.

When I examined whether the average total return on a CoCo is different from the average total returns of stocks and bonds, I found hardly any significant differences. Although research from Credit Suisse showed that the credit spread on CoCos is on average 400 basis points over senior debt, the T-tests in my analysis showed that the average total return on CoCos was not significantly different from debt and equity. The graphs showed that the value development of CoCos can be very disturbed, which makes it hard to define the fair return for investors.

Since CoCos are capital instruments that recently started to get incorporated in the capital structure of financial institutions, there isn’t yet done enough research on the risk-return profile of CoCos from an investor’s perspective. We simply need more data on these CoCos. Although CoCos are called contingent convertible debt instruments, they behave different from most debt instruments and their behaviour can deviate strongly from the ‘economic theory’ applied on them.

From my thesis I can conclude that each CoCo can have a very different risk-return structure, since their contractual features are not standardized yet. Although their position on the balance sheet is in between

(21)

21 equity and debt, their value development shows that the CoCo in many cases behaves outside the range of equity and debt values. In addition, stocks have more influence on CoCos than bonds, but the CoCo’s

average total return is not significantly different from that of stocks and bonds. In the future there needs to be done more research on the risks of investing in CoCos, in order to define the fair total return for investors.

(22)

§6. References

Albul, B., Jaffee, D. & Tchistyi, A. (2010). Contingent convertible bonds and capital structure decisions, working paper, University of California. (http://faculty.haas.berkeley.edu/Tchistyi/CCB.pdf) Avdjiev, S., Bogdanova, B., & Karthasheva, A., (2013, September). CoCos: a primer. Bis Quarterly Review.

(http://www.bis.org/publ/qtrpdf/r_qt1309f.htm)

Bank for International Settlements, Basel Committee on Banking Supervision (2010): Proposal to ensure the

loss absorbency of regulatory capital at the point of non-viability, August.

(http://www.bis.org/publ/bcbs174.htm)

Bank for International Settlements, Basel Committee on Banking Supervision (2011): Basel III: A global

regulatory framework for more resilient banks and banking systems, June.

(http://www.bis.org/publ/bcbs189.htm)

Barclays, European Banks Credit Research (2014). The CoCo Handbook, 3.

(http://www.scribd.com/doc/226738351/Barclays-European-Banks-the-CoCo-Handbook-Vol-3#scribd)

Berk, J., & DeMarzo, P. (2011). Corporate Finance. London, England: Edinburgh Gate.

Chan, S., & Wijnbergen, S. (2014). CoCos, Contagion, and Systemic Risk. University of Amsterdam, Faculty of Economics and Business & Tinbergen Institute, the Netherlands.

(http://papers.tinbergen.nl/14110.pdf)

Duffie, D. (2009). Contractual methods for out-of-court restructuring of systemically important financial institutions, working paper, Stanford University.

(http://www.darrellduffie.com/uploads/policy/DuffieRestructuringOutOfCourt2009.pdf)

European Securities and Markets Authority (2014). Potential Risks Associated with Investing in Contingent Convertible Instruments. (http://www.esma.europa.eu/content/Potential-Risks-Associated-Investing-Contingent-Convertible-Instruments)

Fama, E. F., & French, K. R. (2004). The Capital Asset Pricing Model: Theory and Evidence. Illinois: University of Chicago.

(http://www1.american.edu/academic.depts/ksb/finance_realestate/mrobe/Library/capm_Fama_Frenc h_JEP04.pdf)

Flannery, M.J. (2009). Stabilizing large financial institutions with contingent capital certificates, working paper, University of

(23)

23 Guschin, V., & Curien, E. (2008). The Pricing of Convertible Bonds within the Tsiveriotis and Fernandes

Framework with Exogenous Credit Spread: Empirical Analysis. Journal of Derivatives, 14(1), pp. 50-65.

(http://www1.american.edu/academic.depts/ksb/finance_realestate/mrobe/Library/capm_Fama_Fren ch_JEP04.pdf)

Harvard Law Review Association. (1991). Distress-Contingent Convertible Bonds: A Proposed Solution to the Excess Debt Problem. Harvard Law Review, 104(8), pp. 1857-1877.

(http://www.jstor.org/discover/10.2307/1341621?sid=21105292046041&uid=4&uid=3738736&uid= 2)

Maes, S. & Schoutens, W. (2010). Contingent capital: an in-depth discussion, working paper, Catholic University of Leuven. (https://perswww.kuleuven.be/~u0009713/ContingentCapital.pdf) Marquardt, C., & Wiedman, C. (2005, May). Earnings Management through Transaction Structuring:

Contingent Convertible Debt and Diluted Earnings per Share. Journal of Accounting Research,

43(2), pp. 205-243. (http://onlinelibrary.wiley.com/doi/10.1111/j.1475-679x.2005.00168.x/abstract)

McDonald, R.L. (2009). Contingent capital with a dual price trigger, working paper, Northwestern University.

(http://business.nd.edu/uploadedFiles/Academic_Centers/Study_of_Financial_Regulation/pdf_and_d ocuments/dualtriggercc02.pdf)

Mehra, R. (2003). The Equity Premium: Why Is It a Puzzle?, working paper, National Bureau of Economic Research. (http://econpapers.repec.org/paper/nbrnberwo/9512.htm)

Miu. P., Ozdemir, B., & Giesinger, M. (2010). Can Basel III work?: Examining the new Capital Stability Rules by the Basel Committee – A Theoretical and Empirical Study of Capital Buffers. BMO Financial Group and McMaster University.

(http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1556446)

Pennacchi, G. (2010). A structural model of contingent bank capital, Federal Reserve of Cleveland, working paper, University of Illinois. (http://business.illinois.edu/gpennacc/ConCap030211.pdf)

Roggi, O., Giannozzi, A., & Mibelli, L. (2013). CoCo Bonds, Conversion Prices and Risk Shifting

Incentives. How does the Conversion Ratio Affect Management’s Behaviour? New York University Salomon Center and Wiley Periodicals, Inc.

(24)

Schmid, M. (2014). Investing in Contingent Convertibles. Credit Suisse Asset Management.

(https://www.credit-suisse.com/media/am/docs/asset_management/events/2014/fits2014-program/4-2-schmid-contingent-convertibles.pdf)

Schoutens, W., & De Spiegeleer, J. (2012) Pricing Contingent Convertibles: A Derivatives Approach. Journal of Derivatives. Vol. 20, No. 2: pp. 27-36.

(http://www.iijournals.com/doi/abs/10.3905/jod.2012.20.2.027)

Schoutens, W., & De Spiegeleer, J.(2014) The World Of CoCos. [Powerpoint Presentation for Reacfin's 10th anniversary event].

(http://www.reacfin.com/en/sites/default/files/documents/Reacfin_CoCos_Final.pdf)

Teneberg, Hendrik. (2012) Pricing Contingent Convertibles using an Equity Derivatives Jump Diffusion Approach. Stockholm, KTH.

(http://www.math.kth.se/matstat/seminarier/reports/M-exjobb12/120125.pdf)

Veiteberg, V. G., Bysveen, F. T., & Rosef, B. H. (2012). Pricing Contingent Convertible Capital: An Empirical Approach. Norwegian University of Science and Technology, Trondheim.

(http://ntnu.diva-portal.org/smash/record.jsf?pid=diva2%3A624626&dswid=5143)

Zähres, M. (2011). Contingent Convertibles: Bank bonds take on a new look. Deutsche Bank Research.

(https://www.dbresearch.com/PROD/DBR_INTERNET_EN-PROD/PROD0000000000273597/Contingent+Convertibles%3A+Bank+bonds+take+on+a+new+.pd f)

(25)

25

§7. Appendix

1. CoCo contractual features

CoCo Features CoCo 1 Credit Suisse CoCo 2 Credit Suisse CoCo 3 Credit Suisse CoCo Barclays CoCo Credit Agricole CoCo KBC CoCo Banco Bilbao

T1 or T2 Tier 2 Tier 2 Tier 1 Tier 2 Tier 2 Tier 2 Tier 1

Trigger Tier1 capital 7% of RWA Tier1 capital 7% of RWA Tier1 capital 7% of RWA Tier1 capital 7% of RWA Tier1 capital 7% of RWA Tier1 capital 7% of RWA Perpetual Loss absorption mechanism Conversion into equity Conversion into equity Conversion into equity Principal write-down Principal write-down Principal write-down Conversion into equity

Term 30 years 10 years perpetual 10 years 20 years 10 years Perpetual

Issue date 24-02-2011 22/03/2012 31/07/2012 21/11/2012 19/09/2013 25/01/2013 09/05/2013

Coupon rate 7,875% per year 7,125% per year 9,5% per year 7,625% 8,125% per year 8% per year 9% per year

Call option 24-08-2016 22/03/2017 Non-callable Non-callable 19/09/2018 25/01/2018 09/05/2018

Rating BBB+ _BBB- _BB+ _{Not rated} _{Not rated} _BB+ _BB-

Bloomberg code

XS0595225318 CH0181115681 XS0810846617 US06740L8C27 USF22797QT87 BE6248510610 XS0926832907

(26)

2. Summary statistics

2.1 Credit Suisse

Total return CoCo 2 Total return Bond Total return Stock

Gemiddelde -0,000117459 Gemiddelde 0,00068309 Gemiddelde -0,0016665

Standaardfout 0,000264548 Standaardfout 0,000802114 Standaardfout 0,0008832

Mediaan 1,39641E-05 Mediaan 0,000149076 Mediaan -0,0017469

Modus Modus Modus 0

Standaarddeviatie 0,003665689 Standaarddeviatie 0,011114413 Standaarddeviatie 0,0122378 Steekproefvariantie 1,34373E-05 Steekproefvariantie 0,00012353 Steekproefvariantie 0,0001498

Kurtosis 79,95239805 Kurtosis 27,1578101 Kurtosis 0,0599846

Scheefheid -5,256939748 Scheefheid 3,063787801 Scheefheid 0,0719445

Bereik 0,063206751 Bereik 0,134584375 Bereik 0,068013

Minimum -0,039701087 Minimum -0,055050525 Minimum -0,0332856

Maximum 0,023505664 Maximum 0,07953385 Maximum 0,0347274

Som -0,022552171 Som 0,1311532 Som -0,31996

Aantal 192 Aantal 192 Aantal 192

Gemiddelde 6,94705E-05 Gemiddelde 0,000292284 Gemiddelde -0,0003252

Mediaan 0 Mediaan 4,5936E-05 Mediaan -0,0003749

Modus 0 Modus 0 Modus 0

Standaarddeviatie 0,004858288 Standaarddeviatie 0,008619856 Standaarddeviatie 0,0218666 Steekproefvariantie 2,3603E-05 Steekproefvariantie 7,43019E-05 Steekproefvariantie 0,0004781

Scheefheid -1,185086894 Scheefheid 1,937476613 Scheefheid -0,0410982

Som 0,070304167 Som 0,295791365 Som -0,329077

(27)

27

Gemiddelde -1,85031E-06 Gemiddelde 0,000354191 Gemiddelde -0,000292137

Mediaan 0 Mediaan -1,89338E-05 Mediaan -0,000651042

Modus 0 Modus / Modus 0

Standaarddeviatie 0,009638028 Standaarddeviatie 0,010629714 Standaarddeviatie 0,01451286 Steekproefvariantie 9,28916E-05 Steekproefvariantie 0,000112991 Steekproefvariantie 0,000210623

Scheefheid -7,172340811 Scheefheid 2,012753369 Scheefheid -0,255411551

Som -0,000882596 Som 0,168948953 Som -0,139349434

2.2 Barclays

Total return CoCo Total return Bond Total return Stock

Gemiddelde 0,000394613 Gemiddelde 7,46438E-07 Gemiddelde 0,000204656

Mediaan 0,001005025 Mediaan 9,44709E-05 Mediaan 0

Standaarddeviatie 0,015513667 Standaarddeviatie 0,002947516 Standaarddeviatie 0,017397154 Steekproefvariantie 0,000240674 Steekproefvariantie 8,68785E-06 Steekproefvariantie 0,000302661

Scheefheid -0,318303391 Scheefheid -0,340758527 Scheefheid 0,032150557

Som 0,175602662 Som 0,000396359 Som 0,090867243

(28)

2.3 Credit Agricole

Gemiddelde 0,000520697 Gemiddelde 0,00026195 Gemiddelde 0,000440783

Mediaan 0,000440635 Mediaan 0,000200734 Mediaan 0,002107503

Scheefheid 1,772052704 Scheefheid 0,806103559 Scheefheid -0,08946304

Som 0,119760402 Som 0,060248592 Som 0,101380173

2.4 KBC

Mediaan 6,51205E-05 Mediaan 0,000518026 Mediaan 0,001380338

Modus 0 Modus / Modus 0,002289942

Scheefheid 0,734309489 Scheefheid -0,465718581 Scheefheid 0,397153809

Som 0,143987055 Som 0,126776653 Som 0,460177545

(29)

29 0 100 200 300 400 500 -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 F re que nt ie Verzamelbereik

Bond

Frequentie 2.5 Banco Bilbao

Standaardfout 0,000669536 Standaardfout 8,63192E-05 Standaardfout 0,00082142

Mediaan 0 Mediaan 9,39867E-05 Mediaan 0

Modus 0 Modus Modus 0

Scheefheid 0,328612765 Scheefheid -0,191985888 Scheefheid -0,127216337

Som 0,081365152 Som 0,048204318 Som 0,247842177

3. Frequency distributions

3.1.a Credit Suisse CoCo 1

0 500 1000 -0,060 -0,055 -0,050 -0,045 -0,040 -0,035 -0,030 -0,025 -0,020 -0,015 -0,010 -0,005 0,000 0,100 0,200 0,300 0,400 0,500 F re que nt ie Verzamelbereik

CoCo 1

Frequentie

(30)

0 10 20 30 40 50 60 70 80 90 100 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 F re que nt ie Verzamelbereik

Bond

3.1.b Credit Suisse CoCo 2

0 100 200 300 400 500 600 -0,2 -0,1 0 0,1 0,2 Meer F re que nt ie Verzamelbereik

Stock

Frequentie 0 20 40 60 80 100 120 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 F re que nt ie Verzamelbereik

CoCo 2

Frequentie

(31)

31 3.1.c Credit Suisse CoCo 3

0 20 40 60 80 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 F re que nt ie Verzamelbereik

Stock

Frequentie 0 100 200 300 400 -0,2 _-0,04 _-0,03 _-0,02 _-0,01 0 _0,01 _0,02 _0,03 _0,04 _0,05 _Mee r F re q u en ti e Verzamelbereik

CoCo 3

Frequentie 0 50 100 150 200 250 -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 _Mee r F re que nt ie Verzamelbereik

Bond

Frequentie

(32)

3.2 Barclays 0 50 100 150 -0,08 -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 _Mee r F re que nt ie Verzamelbereik

Stock

Frequentie 0 50 100 150 -0,08 -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 _Mee r F re que nt ie Verzamelbereik

CoCo

Frequentie 0 50 100 150 200 250 300 -0,02 -0,01 0 0,01 0,02 Meer F re que nt ie Verzamelbereik

Bond

Frequentie

(33)

33 0 10 20 30 40 50 -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 0,07 _Mee r F re que nt ie Verzamelbereik

Stock

Frequentie 3.3 Credit Agricole 0 50 100 150 F re que nt ie Verzamelbereik

Stock

Frequentie 0 20 40 60 80 100 120 -0,1 -0,02 -0,01 0 0,01 0,02 0,03 0,11 Meer F re que nt ie Verzamelbereik

CoCo

Frequentie 0 10 20 30 40 50 60 -0,009 -0,007 -0,005 -0,003 -0,001 0,001 0,003 0,005 0,007 0,009 0,02 F re que nt ie Verzamelbereik

Bond

Frequentie

(34)

3.4 KBC 3.5 Banco Bilbao 0 50 100 150 F re que nt ie Verzamelbereik

CoCo

Frequentie 0 20 40 60 -0,02 _-0,009 _-0,007 _-0,005 _-0,003 _-0,001 0,001 0,003 0,005 0,007 0,009 Mee r F re que nt ie Verzamelbereik

Bond

Frequentie 0 10 20 30 40 50 -0,07 -0,05 -0,03 -0,01 0,01 0,03 0,05 0,07 0,09 0,1 1 F re que nt ie Verzamelbereik

Stock

Frequentie 0 50 100 150 200 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 _Mee r F re que nt ie Verzamelbereik

CoCo

Frequentie

(35)

35 4. Regression outputs: The effect of a stock and bond’s total return on the CoCo’s total return

*Significant at 1% level, ** Significant at 5% level, *** Significant at 10% level

0 50 100 150 F re que nt ie Verzamelbereik

Bond

Frequentie 0 50 100 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 0,05 0,06 _Mee r F re que nt ie Verzamelbereik

Stock

Frequentie

Credit Suisse CoCo 1 Model 1

excluding control variable Model 2 Including control variable Constant

Total return on Bond

Total return on Stock

VIX index 0,000069 0,020607 (0,008615) 0,017564** (0,021855) 0,001173* 0,012254 (0,008725) 0,011724*** (0,022102) -0,000059** (6,194783) Observations Adjusted R2 1012 0,005204 976 0,008117

(36)

excluding control variable

Model 2

Including control variable Constant

VIX index -0,000052 0,0086986 (0,011085432) 0,0426729** (0,012206) 0,000593 0,010466 (0,011155) 0,039766*** (0,0123168) -0,000046 (2,920004) Observations Adjusted R2 192 0,010441 188 0,006002

Model 2

VIX index -0,000037 0,127068696* (0,010619) (0,000354) 0,0426729 (0,012206) (-0,001666) 0,007416 0,238390 (0,010619) (0,000128) 0,039766* (0,0123168) (-0,001660) -0,0005189 (1,865043) (14,419777) Observations Adjusted R2 477 _0,017898 269 _0,039947

(37)

37 4.2 Barclays

Barclays CoCo Model 1

Model 2

VIX index 0,000372 -0,175918 (0,002972) (7,4644E-07) 0,076070*** (0,0173776) (0,000205) 0,004090 -0,188903 (0,002972) (-8,8566E-05) 0,070076 (0,017378) (0,000205) -0,000264 (1,960766) (14,128828) Observations Adjusted R2 444 _0,004750 269 _0,003365

CoCo Credit Agricole Model 1

Model 2

VIX index 0,000535059 -0,226811487 (0,002623) 0,076070* (0,017378) 0,002909 -0,189805 (0,002623) 0,100045* (0,017378) -0,000164 (2,584115) Observations Adjusted R2 230 _0,030555 230 _0,027360

(38)

4.4 KBC

4.5 Banco Bilbao

CoCo KBC Model 1

Model 2

VIX index 0,000413624 0,149068853 (0,003209) (0,000395) 0,059212852** (0,024049) (0,001730) 0,001732895 0,146984664 (0,003209) (0,000395) 0,057338655*** (0,024049) (0,001730) -9,3809E-05 (1,933422) (14,129323) Observations Adjusted R2 266 0,008595 266 0,004928

CoCo Banco Bilbao Model 1

Model 2

VIX index -9,4475E-05 1,560313* (0,001847) (0,000122) 1,560313* (0,001847) ( 0,000740) 0,002048121 1,568367473* (0,001847146) ( 0,000122036) 0,104732823** (0,015012) (0,000740) -0,000154187 (1,960583) (14,038866) Observations Adjusted R2 335 _0,064348 335 _0,063048

(39)

39 5. Regression outputs compared

6. T-statistics about the difference in average total returns on CoCos versus stocks and bonds Difference

between average total return CoCo and: CoCo 1 Credit Suisse CoCo 2 Credit Suisse CoCo 3 Credit Suisse CoCo Barclays CoCo Credit Agricole CoCo KBC CoCo Banco Bilbao Average total return Bond -1,4597 -3,0340 -0,8077 0,5350 0,3492 0,0770 0,1154 Average total return Stock 2,5854 5,8706 0,6585 0,2580 0,1078 -1,8397 -0,7338

The T-values in grey are not significant.

Regression Model 1 CoCo 1 Cr. Suisse CoCo 2 Cr. Suisse CoCo 3 Cr. Suisse CoCo Barclays CoCo Cr. Agri. CoCo KBC CoCo Banco B. Rbond 0,02061 (0,0086) 0,0087 (0,0111) 0,1271* (0,0106) -0,1759 (0,0030) -0,2268 (0,0026) 0,1491 (0,0032) 1,5603* (0,0018) Rstock 0,0176** (0,0219) 0,0427** (0,0122) 0,0427 (0,0122) 0,0761*** (0,0174) 0,0761* (0,0174) 0,05921** (0,0240) 1,5603* (0,0018) Regression model 2 CoCo 1 Cr. Suisse CoCo 2 Cr. Suisse CoCo 3 Cr. Suisse CoCo Barclays CoCo Cr. Agri. CoCo KBC CoCo Banco B. Rbond 0,0123 (0,0087) 0,0105 (0,0112) 0,2384 (0,0106) -0,1889 (0,0030) -0,1898 (0,0026) 0,1470 (0,0032) 1,5684* (0,0018) Rstock 0,0117*** (0,0221) 0,0398*** (0,0123168) 0,0398* (0,0123) 0,0701 (0,0174) 0,1000* (0,0174) 0,05733*** (0,0240) 0,1047** (0,0150) VIX -0,0001** (6,1948) -0,0001 (2,9200) -0,0005 (1,8650) -0,0003 (1,9608) -0,0002 (2,5841) -9,4E-05 (1,9334) -0,0002 (1,9606) *Significant at 1% level, ** Significant at 5% level, *** Significant at 10% level

(40)

7. Value developments of CoCos, and their accompanying stock and bond. 7.1 Credit Suisse 7.1.a CoCo 1 7.1.b CoCo 2 7.1.c CoCo 3

(41)

41 7.2 Barclays

(42)