The announcement effect of contingent convertible bond issuance by financial institutions

(1)

The announcement effect of Contingent Convertible bond issuance by

financial institutions

Bachelor Thesis by Wesley Kuiper Faculty of Economics & Business

Finance and Organization Supervised by Ms. N. Martynova MSc

(2)

One of the underlying reasons of the financial market crisis, which started in 2008, has been the high level of leverage present in financial institutions (Demirguc-Kunt, Detragiache, & Merrouche, 2013, p. 1148). Contingent capital, which is capital that can be used when unexpected losses occur, was considered to provide a buffer during the financial crisis. This appeared not to be the case, resulting in firms not being able to meet their obligations. In order to prevent financial institutions from creating excessive leverage ratios in the future, new capital requirement were put in place with the introduction of the Basel III accord. The introduction of these new capital requirements stimulated the use contingent convertible bonds (CoCo’s). CoCo’s pay coupons just as any ordinary bond, but when the issuing

financial institution experiences difficulties, the bond will convert to equity if a predetermined trigger is hit, lowering the risk of a default. The design of CoCo bonds therefore acts as a safeguard during periods of financial distress. CoCo’s are praised to be transparent, efficient and less costly compared to other instruments to prevent financial distress.

In November 2009 Lloyds Banking Group was the first to issue a series of CoCo’s, worth £8.5 billion. The announcement of this issue series was received positively by the public. However, little is known about the wealth effects of CoCo’s and whether the, at first sight, positive reactions by the market are transformed in positive abnormal returns. Research done by Abhyankar and Dunning (1999) shows that in general the announcement of

convertible bonds has a negative effect on the issuing institution’s abnormal return, whether it is a financial institution or not. This research took place in the UK and is supported by

research in other markets. Although there is an extensive amount of literature available on the issuance of convertible bonds, there is little available on the issuance effect for financial institutions specifically. This research will therefore add value by focusing on the

announcement effect of CoCo’s and specifically, on the announcement effect of CoCo issues by financial institutions in the period 2009-2014.

It is stated that the announcement effect of CoCo issues lead to significant negative abnormal returns for the issuing financial institutions. This statement is based on two arguments. The first is based on the ‘back-door’ theory introduced by Stein (1992). In this theory it is stated that firms favor convertible bond issuance as a method to raise equity since it prevents the risk of costly financial distress and doesn’t provides the negative information signal paired with issuing equity.

(4)

However, CoCo’s also have equity like characteristics, which might cause that the theory doesn’t hold and issuance does cause a negative signal to the market. The second argument is based on the risk shifting theory described by Green (1984), where it is stated that the issuance of convertible bonds is subject to risk shifting by the issuer of the convertible bond. Projecting this theory on the issuance of CoCo’s, investors could be exposed to increased risk when the shareholders of the issuing financial institution move to riskier investments at the cost of the debt holders. Additional hypotheses in this paper state that the conversion trigger and the loss absorption mechanism have a significant effect on the

cumulative abnormal return at the announcement of the issue. These hypotheses are based on the fact that these issue specific characteristics signal information to the market about the firm.

This study will be based on an event study looking at the issues of CoCo’s by financial institutions in the period of November 2009 till January 2014. The data on these issues is collected from Datastream, company quarterly financial statements and the issue

prospectuses. A description will be given on the firm and issue specific characteristics of the financial institutions and issued CoCo’s. The cumulative abnormal returns (CAR’s) are calculated by comparing the three-day event windows of the issuing financial institutions to the Euronext 100 index returns over the same period, since all financial institutions are based in Europe. An univariate regression will be performed on firm and issue specific

characteristics to determine the influence of these variables on the CAR’s.

It is shown that the announcement of CoCo issues leads to negative abnormal returns, although not significant. Furthermore the descriptive statistics show that there is a strong correlation between issue and firm specific characteristics such as the core tier 1 ratio and the size of the CoCo issue. The univariate regression shows us that, for example, the issue size is indeed of significant importance in determining the CAR as already suggested.

The paper is organized as follows. The second section will elaborate on the key features of contingent convertible bonds and compare it to convertible bonds. The third section will cover research on several basic theories in finance and describe the methodology used in this paper. The fourth section elaborates on the data used. The fifth section reports the results on the event study and the regressions. The last section gives a brief summary on the results and makes recommendations for future research.

(5)

2. Literature review

After the financial crisis, the Basel committee set out to implement new regulation to prevent institution from running out of capital and becoming illiquid. Main reforms in the Basel III accord require a bank to hold a Total Capital of 8% of its risk-weighted assets, where total capital is divided in Tier 1 (T1) and Tier 2 (T2) capital. Additionally, Swiss regulators require Swiss banks to have at least 9% of risk-weighted assets in loss-absorbing instruments

(Avdjiev, Kartasheva, & Bogdanova, 2013).Tier 1 capital is defined as “going concern” capital, which means it’s most likely to incur losses in case of bad times. The Tier 1 capital is divided in Common Equity Tier 1 (CET1) and Additional Tier 1 (AT1). CoCo’s with a high-trigger are likely to be classified under AT1 and low-high-triggered CoCo’s as Tier 2 capitals. Bonds with low triggers generally have a lower loss-absorbing capacity, for this reason they don’t qualify as T1 capital according to Basel III. Contingent convertible bonds are classified as debt and will, by means of a trigger, automatically convert into equity upon reaching the predetermined conversion trigger (Avdjiev, Kartasheva, & Bogdanova, 2013). With the implementation of the Basel III accord, the interest in this type of bond increased. The amount of debt outstanding will be lowered and the bank is recapitalized automatically.

2.1 Design and features of CoCo bonds

CoCo bonds have three main characteristics. The first is its conversion trigger, which is essentially the most important feature in its design. A trigger can either be based a mechanical rule or on a discretionary rule set by the supervisors. In case of a mechanical rule, this trigger can be based on the book value or on the market value (Avdjiev, Kartasheva, & Bogdanova, 2013). According to Flannery (2005), a trigger based on a pre-specified equity ratio of the issuing firm is an acceptable trigger. Currently, all issued CoCo’s are based on a trigger that takes the equity ratio from the accounting book value. Avdjiev, et al. (2013) argue that the success of accounting based triggers depend on how frequent the ratios are calculated and published.

A possible explanation for the reliance on accounting based triggers might be the poor confidence in market prices post crisis (Martynova & Perotti, 2013). When using trigger mechanisms based directly upon equity prices, it is necessary to prevent a transfer of value between equity holders and CoCo holders (Sundaresan & Wang, 2010). If this is not the case, multiple equilibria can exist, opening up the possibility to price manipulations.

(6)

Depending on the design of the conversion ratio, it is possible to prevent a value transfer. As stated by Sundaresan and Wang (2010), there won’t be two equilibria as long as the value transfer doesn’t force the stock price to go below the trigger. At this point investors know, for each given value of the asset, whether or not the conversion will take place. If the value transfer does make the stock price go below the trigger the case of two equilibria exists. The first equilibria will be at the point where investors don’t believe a conversion will take place, keeping the stock price at the same level. The other equilibria persist when investors believe a conversion will take place, causing the stock price to drop to the trigger. Since two prices are possible, there is an expected equity range, even close before the conversion. Such situation can only be prevented by making sure the value of the converted shares is equal to the market value of the unconverted CoCo bonds. By arguing the other way around, one can see that it’s also possible to pull the stock price above the trigger by setting the conversion ratio to high. There will be no equilibrium if this is the case.

The second main feature of a CoCo bond is its conversion ratio. The conversion ratio determines the amount of shares in which the bonds are converted when the trigger is hit. There are two types of conversion ratios; fixed and floating. A fixed conversion ratio will convert the bond into a fixed number of shares or money. The value of the shares can be set at different values for conversion. A par conversion is where the shares after conversion are worth the same as the bond, a premium conversion is where the shares after conversion are worth less than the par value of the bond. With a discount conversion, the shares after

conversion are worth more than the par value of the bond (McDonald, 2009, p. 3). According to McDonald (2009), a fixed conversion ratio creates on average less value for bondholder when the conversion actually takes place. The closing value of the stock price is usually slightly lower than its par value. Knowing this, the market will demand a higher interest rate on the CoCo bond. With a floating conversion rate the conversion will take place against an ex post determined issue price, meaning that the issuer will receive the proceeds immediately (Hillion & Vermaelen, 2004).

Several articles describe how a floating conversion ratio can lead to a downward price spiral. The conversion of the bond will result in more outstanding shares of the issuing firm. In general, having more shares outstanding will reduce the price per share, an effect called share dilution. When the market expects more shares will be sold, the share prices will start to decline on forehand due to this short-selling pressure (Flannery, 2009). According to French and Roll (1986) such initial price decline could lead to the market thinking that there is some kind of bad news event which hasn’t been observed yet.

(7)

Following these argumentations, issuing firms will experience negative abnormal returns around the announcement date. Hillion and Vermaelen (2004) refer to this as the ‘faulty contract hypotheses. Hillion and Vermaelen (2004) also mention the ‘last-resort financing hypothesis, where issuing firms aren’t able to raise capital with fixed conversion prices because the stock is overvalued. The price decline that follows is the correction upon the overvalued stock price, recognized by the market. Kozial and Lawrence (2010) argue that the risk of a death spiral exists when large institutional investors are forced to sell their shares, since these institutions are not allowed to hold equity. Death spirals refer to the situation in which the conversion of bonds into shares causes share dilutions which reduces the price even further. Death spirals can also occur as a consequence of price manipulation. Stated by

Flannery (2009) and Pennachi (2010), the conversion ratio can be set in such a way that it may give rise to price manipulation. Such price manipulation is only likely to happen when the difference between the conversion value and the CoCo bond value is high enough to make the conversion profitable for the bondholders (Albul, Jaffee, & Tchistyi, 2010).

The third key feature of a contingent convertible bond is the loss absorption

mechanism. The mandatory conversion of the CoCo bond is issued to boost the bank’s equity, which can be done in two ways. First of all, the conversion can take place in the form of equity, which increases the Common Equity Tier 1 (CET1) of financial institutions (Avdjiev, Kartasheva, & Bogdanova, 2013). The second conversion method is based on a principal write-down (PWD) where equity is raised by undergoing a write-down. This write-down can be a full or a partial write-down. Looking at the CoCo bond issues so far, most financial institutions make use of a full write-down. A partial write-down has a dangerous drawback to it, upon conversion the issuer forced to fund a cash payout while in financial distress, which could eventually lead to bankruptcy.

2.2 Why firms issue traditional convertible bonds

According to the Modigliani and Miller theorem (1958), the value of a firm is not affected by the ways of financing. However this theorem makes several assumptions, among which perfect information, which is not the case for the markets mentioned here. The pecking order theory states that, companies will prefer debt financing to equity financing. In practice we see that external equity and straight debt are the preferred choice of financing, but convertible debt is also frequently used to raise capital. Especially in Europe, convertible debt issuance has grown significantly since the year 2000.

(8)

In the U.S. the issuance of convertible debt is about ten percent of the total securities issuance in the last 30 years (Duca, Dutordoir, Veld, & Verwijmeren, 2012). Generally, convertible bonds are interesting for long-term oriented investors who are seeking to diversify their portfolio and aim to participate indirectly in equities (Lummer & Riepe, 1993). According to previous studies, convertible bonds are able to mitigate the adverse selection costs related to attracting common equity financing. This is supported by three theories, which will be discussed below.

The first theory is the ‘back-door’ theory by Stein (1992). Following this theory, firms turn to convertible bonds since it is considered an attractive solution between negative

informational signal towards the market on one side and the risk of a costly financial distress paired with a debt issue on the other side. The goal with the convertible bond issues in

general, is to collect common equity financing at a better price than the issue date stock price. Hence, the stock is overvalued. Information asymmetry is one of the main reasons for firms to use convertible bonds as an indirect way to get equity into their capital structures ‘through the backdoor’ (Stein, 1992). This especially holds when regular equity issues aren’t attractive due to the information asymmetry that gives rise to the lemons problem (Akerlof, 1970). Since managers’ incentives are not aligned with those of the shareholders, a discount is required to compensate the shareholders for the information asymmetry. Knowing this, firms with good investment opportunities may decide not to issue equity. Convertible bonds pose an

alternative here since they prevent the adverse selection costs of a direct sale of common equity (Stein, 1992). The backdoor theory is based on the call provision included in convertible bonds. This call provision is exercisable after the expiration of the protection period. By forcing investors to exercise the conversion option early, the bonds will be

exchanged for shares of stock, implementing equity finance in an indirect way. This prevents the downward correction on the stock price in case of information asymmetry (Stein, 1992).

Secondly, the risk-shifting theory introduced by Green (1984) describes several investment incentive problems related to debt financing. Straight debt might lead investing behavior becoming more risky in order increase the value for creditors. When these wealth transfers are large enough, shareholders might agree taking on negative net present value projects, resulting in a decline of bondholders’ wealth. Green’s (1984) ‘risk-shifting’ theory states that convertible bonds can decrease possible agency costs arising from conflicts between bondholders and shareholders. Since convertible debt can be considered a less risky asset, the issuance of convertible bonds might control for distorted incentives. For this reason, firms facing significant risk are most likely to issue convertible bonds.

(9)

Lastly, Mayers (1998) states that the issuance of convertible bonds could solve sequential-financing problems of firms. A sequential-financing problem is defined as investment options with a future maturity date and securities being too costly to issue (Mayers, 1998). The two main factors in making financing decisions are the issue costs and the overinvestment costs. The convertible bond prevents issue costs since conversion keeps the funds in the firm and reduces leverage in the case of a valuable investment option, according to Mayers (1998). He states that managers have control over the funds as long as these funds are not obligated by contract to pay out, resulting in an overinvestment problem. Convertible bonds can alleviate this problem, since bondholders can choose not to exercise the convertible bond and get the fund back when the project isn’t sufficiently profitable (Mayers, 1998). Dutordoir and Van de Gucht (2009) found that Western European companies mainly use convertible debt as an instrument to sweeten debt. This explains why convertible bonds in Western Europe have more debt-like characteristics compared to the U.S. bonds.

2.3 Effects of issuing convertible bonds and CoCo’s

Bonds

As mentioned by Green (1984), issuing contingent bonds could give rise to risk shifting. The risk shifting theory is based upon two agency problems. There is an information asymmetry between the issuing bank and the bond and shareholders since the banks are better informed as a whole. As a consequence banks can influence the decisions of their borrower and determine how much risk they will take on by lending to certain businesses at higher yields. This opens up the possibility of ex post risk shifting since shareholders could move to risky loans at the cost of the debt holders’ interest. As seen in the recent financial crisis, risk-taking by financial institutions isn’t limited to the institution itself, but could have a much larger macro-economic effect. Since the cost paired with bankruptcy of large financial institutions are significantly high, regulators are not willing to let this happen, giving rise to the too-big-too-fail situation.

Contrary to the findings of Modigliani and Miller, Lewis et al. (1999) found that corporate activity does affect the trading activities, and therefore the value, of common stock. These changes in trading activities are a consequence of market signaling, a phenomenon described by Spence (1973) and applied to the financing policy by Ross (1977). Since debt requires the issuing company to make fixed payments over the duration of the outstanding debt, certain costs are paired with not meeting these payments.

(10)

Issuing more debt is a credible signal of positive future expectations, since managers have inside information (Barclay & Smith, 2005). Still, Roon (1995) only found an insignificant positive abnormal return for the announcement of convertible bond issues. According to this study, announcements in the U.S. are more negatively received by the market than in the Netherlands, where the announcement where packed together with other announcements of firm specific news. Dutordoir and Van de Gucht (2004) explain this result by characterizing the European convertible bonds as more debt-like due to regulatory differences. Dann and Mikkelson (1984) and Eckbo (1986) also show significant negative effects following the announcement of convertible bonds. Conversion prices can be one of the causes for these negative abnormal returns since low conversion prices are perceived as a more negative signal (Davidson III, Glascock, & Schwartz, 1995).

Supported by recent research, hedge fund arbitrage is seen as one of the forces behind the negative abnormal returns following convertible bond issues. When convertible bonds are underpriced or better techniques for managing convertible risk are available to a select group, arbitrage opportunities will arise (Agarwal et al., 2007). The short selling techniques used by the arbitrageurs will impose a downward pressure on the stock value, possibly leading to the downward spiral discussed before. Convertible bond arbitrageurs will take a long in

convertible securities and a short in the stock of the same company. When the price of the stock goes down, the hedge fund will make a profit on the short position it has, whereas an increase in the stock price leads to a profit on the conversion and sale of the stock at market value (Choi, Getmansky, & Tookes, 2009).

CoCo’s

Just as is the case with regular convertible bonds, Skinner (2011) argues that the risk shifting can also take place with financial institution issuing CoCo’s. Once the CoCo is issued, the bank is assured of a financial buffer, which can be seen as a ‘bailout package’. With this prospect, banks might shift their preferences to riskier loans with higher margins. When the financial institutions have problems meeting their obligations, the CoCo’s will assure the bank of a recapitalization through conversion. The use of CoCo’s could thus actually cause the financial problems that it tries to prevent by nature. This is a classical moral hazard problem.

Research also shows that in markets where there is imperfect information, regulated capital standards may lead to a decline in the stock price of the firms’ concerned (Besanko & Kanatas, 1996).

(11)

Since CoCo issues can be used to meet capital requirements under the Basel III regulation, the announcement of CoCo’s could lead to a downward correction of the share prices of the issuing firm. For the same reason the issuance of CoCo’s could signal that companies have not met their capital requirements yet, indicating a possible lemon’s problem on the CoCo market.

Contingent convertible bonds deliver leverage in good times and provide an equity buffer in bad times. This automatic conversion could prevent the holdout problem, in which regular bondholders have an incentive to not convert their swap of debt for equity when the firm is performing a restructuring. All bondholders together would be better off to swap all their debt for equity according to Pennachi (2010). For banks in financial distress it is difficult to raise new capital, CoCo’s may pose a solution for these companies (Kashyap, Rajan, & Stein, 2008). Following its design, CoCo bonds not only automatically raise equity but also lower debt when the conversion takes place, since the outstanding debt is converted. The chances of a costly default will hereby decrease according to Flannery (2010). Since contingent capital is seen as debt, a tax-shield advantage is also present here when the regulatory environment automatically classifies CoCo’s as debt, which is the case in Europe (Pennachi, 2010). When the bonds are designed in such a way that they act ‘dilutive’ for the existing shareholders, which means that upon conversion one dollar of par value will be swapped for more than one dollar of equity, banks will operate with a larger equity buffer. If they don’t, a negative capital shock will expose the bank to an increased risk of share dilution due to the automated conversion (Himmelberg & Tsyplakov, 2012). Hereby, CoCo bonds can incentivize companies to focus on a more conventional, less risky, capital structure.

Issuance of CoCo bonds might also reduce the debt overhang problem. As discussed before, a bank financed with straight debt and a focus on shareholder value has no incentive to raise new capital through equity issues and therefore is left with too much debt in its capital structure. Debt overhang also has a side effect according to Himmelberg and Tsyplakov (2012). Firms suffering from a debt overhang have the intention to take on high-risk, but low-return investments, which even further reduces the asset quality during periods of financial distress. However, when bonds are not designed to act dilutive, the incentive effect works the opposite way since the CoCo is written down and creates value for the shareholder. This could further increase the conflict of interest and lead to an earlier conversion and a fast deterioration of wealth for the bondholders, even more than for straight debt. Himmelberg and Tsyplakov (2012) show by numerical calculations that indeed CoCo’s can have strong

(12)

However, these incentive effects can be lowered or even become negative due to the costs of issuing equity and the “jump-to-default” risk. Jump-to-default risk denotes the risk of the actual market value of equity suddenly ‘jumping’ to zero due to the risk of rapid

deterioration, for example when a default is looming. Trigger-uncertainty is another possible cause of disappearing incentives. Uncertainties may arise due to delayed or premature

triggering. Delay in triggering could arise from a market with weak supervision whereas premature triggering could arise due to manipulation.

Company characteristics might influence the announcement effect of bonds and CoCo’s. In existing literature, generally six main characteristics are used to research this effect: firm size, profitability, growth opportunities, business sector, managerial ownership and asset tangibility (Chin & Abdullah, 2013). Company size can act as a determinant of the degree of information asymmetry problem since larger companies tend to be under heavier regulation compared to smaller companies (Stein, 1992). Contradicting literature however, for example by Abhyankar and Dunning (1999) show that there is no correlation between firm size and investor reactions.

(13)

3. Methodology

The main goal of this study is to give a clear description of the CoCo issues thus far and to see if the announcement of CoCo’s has a significant effect on the cumulative abnormal return around the announcement

3.1 Hypothesis

Previous research on the announcement effect of regular convertible bonds generally found that the announcement of convertible bonds are followed by negative abnormal returns due to hedge fund activities, information asymmetry and negative signaling from low conversion prices. As such, the null hypothesis is formulated as follows:

Ho: The announcement of CoCo issues is followed by significantly negative cumulative

abnormal returns.

This hypothesis follows from the asymmetric information in the market and risk shifting incentives created after the CoCo’s are set in place. The ex-post risk shifting can be anticipated by the market and therefore included in a downward correction on the firm value at the announcement of the CoCo issues. The asymmetric information between the issuing firm and the market will also lead to a negative correction on the market value of the issuing company.

3.2 Event study

This research will be performed by means of an event study. Event studies are used to measure the effect of a certain event on the value of the firm. In general, this is done by looking at the difference in stock prices in the period close to the event. These abnormal returns will be tested on significance to determine if there is indeed an announcement effect. Furthermore the cumulative abnormal returns (CAR) will be regressed on several firm and issue specific characteristics in an attempt to determine the drivers behind the abnormal returns.

In this study, t=0 indicates the announcement day of the issue. In an efficient market theory it is stated that the price adjustment happens at t=0. However, since this is not an efficient market situation, direct price adjustment at t=0 is not guaranteed.

(14)

In order to determine the announcement effect we therefore look at the common stock returns using a three-day event window. A three-day event study looks at the stock price one day prior to the announcement, the day of the announcement and the day after the

announcement (-1,0,1). This methodology is described by De Roon and Veld (1998) and commonly used in determining the announcement effect of events.

An event study, firstly introduced by Fisher, Jensen and Roll (1969), is a statistical method to define the impact of a certain event on the value of the firm. Their proposed model is as follows;

(1)

Rat = return on stock i

αi = excess return on stock i

βi = measure of systematic risk on stock i

Rmt = market return on day t for stock i

= zero mean disturbance term (normally distributed)

Here, Rit is the return on the stock i due to the event where they control for the normal relation

between the return on I at day t, Rit, and the return on the stock market index chosen. The

impact is measured by looking at the abnormal return, which is defied in the model as the residual. To explain this, imagine an event during month s affecting stock i. When the event is announced (t=0) the residual is used as the estimator for the abnormal return of stock i during the event, hereby excluding the effect of other factors in the market on the return of stock i. The model of Fisher, Jensen and Roll (1969) is often modified with two adjustments. These adjustments are made to improve the stationarity (Gonedes, 1973)of the market model parameters and to prevent a bias on the announcement effect by the event itself (Ball & Brown, 1968, pp. 164-164).

In this research the change in daily returns are determined by the closing prices of the stocks and the market index for the consecutive days of the estimated time period. The following formula is used to find the daily return for stock i at day t.

(15)

The next step is to determine the daily abnormal return for stock i at day t. This is done by the formula described below

̂ ̂ (3)

In this model, called the adjusted market model, ARit stands for the abnormal returns for stock

i at day t. The daily market index return is represented by Rmt. The beta coefficient in this

model is assumed to be equal to 1 and the intercept is stated to be zero. According to Marsh (1979) and Liljeblom (1989), the market adjusted model is equally reliable compared to the market risk adjusted model, where the beta coefficient and intercept are determined

separately. The next step is to calculate the daily average abnormal return as well as the cumulative abnormal returns for the event window.

∑

(5)

∑ (6)

In this formula, ARt describes the average abnormal return for N on day t. CART is defined as

the cumulative abnormal return for the event window used, indicated by t.

When the cumulative abnormal returns are determined, these values will be tested on significance using the t-test described below.

_√

(7)

Here t stands for the days surrounding the event. As said, a 3-day event window will be used. In this case t=0 will be the event day on which the issue is announced. The standard deviation is represented by and is calculated over the period of the event window.

3.3 Regression analysis

A regression analysis is used to further the influence of firm and issue specific characteristics on the cumulative abnormal returns.

(16)

The dependent variable in this model is the cumulative abnormal return. Each variable is regressed separately on the dependent variable CARi. The basic model is as follows:

CARi= βo + β1VARIABLE+ ei

The variables that will be tested on significant influence on CARi are denoted below.

Assets describes the amount of assets owned by the issuing financial institutions at the announcement of the issue. The amount is stated in billion Euros

Debt describes the amount of debt held by the issuing financial institutions at the announcement of the issue. The amount is denoted in billion Euros.

Debt to Asset ratio describes the ratio of the amount of assets divided by the amount of outstanding debt of the issuing financial institutions.

Market to book ratio describes the ratio of the share price at the moment of announcement to the book value at that moment.

Coupon rate describes the coupon rate paid on the CoCo issued. Issue size describes the size of the CoCo issued, measured in Euros

Issue size to Assets ratio describes the amount issued relative to the assets owned by the financial institution at the moment of announcement.

Trigger describes the value of the core tier 1 ratio at which conversion of the CoCo is triggered.

Core tier 1 ratio gives the value of the core tier 1 ratio at the moment of announcement.

Core tier 1 cushion describes the difference between the core tier 1 ratio and the trigger of the announced CoCo.

Perpetual is a dummy variable equal to ‘1’ when the CoCo is issued as a perpetuity Callable is a dummy variable equal to ‘1’ when the CoCo is callable.

Seniority is a dummy variable indicating the seniority ranking in case for the CoCo issued. The value of this dummy is equal to ‘1’ when the CoCo is classified as Tier 1 capital and 0 otherwise.

Swiss issue is a dummy variable indicating if the CoCo issue is done by a financial institution based in Switzerland.

(17)

4. Data

Firm related data and data on the issuance of contingent convertible bonds will be collected from Datastream. The announcement dates used are the publication dates of the prospectus of the issues. The prospectus contains all details considering the issue and can therefore be seen as an official announcement of the issue. Furthermore, the financial statements of the issuing firms were used to retrieve the necessary accounting data.

4.1 Event study data

The issues looked at in this research took place from November 2009 till January 2014 and are executed by financial institutions. In total this results in 50 issues of contingent

convertible bonds by financial institutions. As mentioned earlier, the first series of issues were done by Lloyds Banking Group, announced on November 3rd, 2009. Lloyds offered its

existing debt holders the chance to exchange outstanding hybrid debt for the new enhanced capital notes (ECNs). This set of issues took place at the height of the financial crisis and was set up to prevent Lloyds from entering the UK Asset Protection Scheme (Maes & Schoutens, 2012). For these reasons, sub samples excluding these first series of issues by Lloyds are also presented.

Since all issuing companies are European, the market returns used are the returns from the Euronext 100 index. This pan-European index is blue chip index, which means that it holds companies with a national reputation for quality, reliability and stability over time. The index represents over 80% of the total market capitalization of the Euronext stock exchange. In order to compare the issues, all issues are converted into Euro’s in Datastream.

The data for the regression includes the calculated CAR’s and the variables described earlier. For the regression the abnormal returns of the first series of issues by Lloyds are seen as one event, resulting in 22 events to calculate the CAR’s on. Again, several sub samples are also described.

4.2 Descriptive statistics

In order to give a clear view on the CoCo market, an overview of issued amount of CoCo’s is given in Appendix 1. As can be seen from the graph, after the first initial series of CoCo issues by Lloyds, there were barely any issues in 2010. As of 2011, the amount of issues started to increase. For 2014 only a small amount is reported since only issues until January 2014 are included.

(18)

In Appendix 2 the amount of issued CoCo’s is displayed for each company. Lloyds Banking Group issued most contingent convertible bonds so far. The number of bonds issued is 30, of which 29 issues were announced at the same date. The total amount issued by Lloyds is 10.076 million euro, which is by far the largest issue compared to the other events. Credit Suisse Group had an amount outstanding of 7.003 million euro after six issues in the entire period from 2009 till beginning of 2014. The issues by other companies ranged between 5.500 million Euro and 500 million Euro as can be seen from the graph.

In Appendix 3 a summary is given on the firm and issue specific characteristics of all issues. As can be seen, there are only 41 observations for the variable maturity, which indicates the years until maturity at the moment the CoCo’s were issued. The missing values for this variable will be winsorized to the 99th percentile in the following statistics, unless mentioned differently. Appendix 4 shows the effect of winsorizing on the summary statistics. These missing values are generated for the CoCo’s classified as perpetuities. Appendix 5 describes the same variables for the subsample of the first series of issues by Lloyds Banking Group and Appendix 6 does the same for the remainder of the issues. Looking at the firm specific characteristics such as total assets and debt, it is shown that there is a high standard deviation between the amount of assets and debt held by the issuing financial institution. The standard deviation in assets and debt are 302.53 million Euro and 130.08 million Euro for the full sample. This amount is even larger for the sample excluding the first series of issues by Lloyds, namely 463.60 million Euro and 143.37 million Euro respectively. The large deviations are probably related to the small sample size and the fact that at CoCo’s are relatively new instruments, not yet adopted by a specific type of companies. Contrary, the debt-to-capital ratio is relatively constant for all issuing firms. The Core Tier 1 capital ratio requirement of 4,5% with the introduction of Basel III, is met by all companies although de lowest value is reported for Lloyds Banking Group at the time of their first CoCo issues.

Appendix 7 describes the correlations between the variables for the entire sample and Appendix 8 does this for the sub sample excluding the first series of issues by Lloyds. From Appendix 9 it can be seen that there is a moderate negative correlation between the issue size and the amount of debt held on the firm’s balance sheet. However, this is only the case for the sample including the first series of issues by Lloyds. The moderate negative correlation could be related to the weak financial position of Lloyds at the time of the issue, where it tried to prevent entering the UK Asset Protection Scheme. The debt-to- capital and market-to-book ratio, are relatively the same for both samples. Both samples show a strong positive

(19)

the trigger is hit. The company’s core tier 1 capital ratio also seems to be correlated to the size of the issue scaled on the assets size of the issuing institution, as can be seen in appendix 9. Most values, however, represent the 6,3% core tier 1 ratio of Lloyds Banking Group in November 2009. A strong negative correlation also exists between the debt/asset ratio and the Core Tier 1 cushion, which describes the difference between the Core Tier 1 ratio of the firm and the trigger ratio of the CoCo issued. This correlation is shown in a scatterplot in appendix 10.

The higher the debt-to-asset ratio, the more likely the issue has an equity conversion feature instead of a write-down. For the issues classifying as a perpetual, there is a strong positive correlation with being classified as tier 1 capital in the debt seniority ranking of the issuing firm. Overall, the correlations for the sub sample, shown in Appendix 8, show quite some differences compared to the full sample. The sub sample can be seen as a more

diversified group since these issues took place over a longer period of time and were done by several firms with different characteristics.

(20)

5. Results

In this section the results of the study are described in relation to the stated hypotheses. First the event study will be analyzed based on the descriptive statistics of the variables. Next, t the results on the cumulative abnormal returns and the regression will be discussed, followed by an elaboration on the results all-together.

5.1 The effect of issue announcement on abnormal returns

Since the first series of CoCo issues took place at the same date and was also announced at the same date, this is seen as one event when looking at the abnormal returns. The total sample consists of 50 issues, of which 29 issues were announced on the same date. Together with the remainder of the issues, a sample of 22 issues is used to evaluate the abnormal returns.

Appendix 11 gives a summary of the abnormal returns found for the 22 issues over the three-day event window. On the announcement day itself, on average, a positive abnormal return is found. Both the day before the announcement and the day after the announcement have negative abnormal returns on average. The same holds for the average abnormal return over the entire event window.

Appendix 12 describes the same variables, however it excludes the first series of issues by Lloyds entirely instead of merging it to one event. Excluding this event results in a slightly less negative average abnormal return for the three-day event window. Also, the skewness increased, indicating there are more abnormal returns in the right-tail of the

distribution. This can also be seen in the observed minimum abnormal return, which became less negative.

The sample group contains three events where the announcement date equals the issue date of the CoCo. Leaving out these three events shows that average abnormal return further approaches zero, this is displayed in Appendix 13. Appendix 14 shows the same variables for the events where the announcement date equals the issue date, the abnormal returns tend to be more negative on average for these observations.

In appendix 15 the cumulative abnormal returns (CAR) over the events are shown. The CAR’s are tested on the sample of events to determine if the abnormal returns due to the announcement are significant. Again, these tables describes the entire sample as well as a sub sample excluding Lloyds’ first issue, a sub sample excluding the issues for which the

(21)

None of the CAR’s appears to be significantly negative according to the t-tests performed. Taking the sub sample of 21 events and a sub sample of 18 events only further reduces the test statistics for the CAR’s. For the first issue of Lloyds it’s known that it was paired with an announcement on coupon payments, which might have caused a negative abnormal effect as well. For the sub sample of events where the announcement date equals the issue date a higher test statistic is shown. The abnormal returns are more concentrated due to the fact that these dates are the same. The leakage effect, where abnormal returns occur earlier due to information leakage, and the adjustment effect, where abnormal returns occur after the event are thus smaller.

According to the results of the t-test it can be concluded that the first hypothesis, which states that at the issue announcement significant negative CAR’s occur, will be rejected. However, non-significant negative CAR’s are observed.

5.2 Determinants abnormal return

Although there is no significant effect on the returns due to the announcement of an issue, there are negative abnormal returns in the event window surrounding the announcement. In order to find out what the possible drivers behind the abnormal returns are, an univariate regression was performed on the different variables. For regression on the complete sample, the first series of Lloyds Banking Group is again considered as one issue. The variables in this series are averaged, although most of them are constant. The averaging has effect on the years to maturity, the issue size and the issue size scaled by asset. The results of the univariate regression for the complete sample and subsample A and B are shown in Appendix 16 through 18.

Firstly, the issue size scaled on assets is significant at a 5% level in all samples. The coefficients for the main sample and subsample A are relatively close to each other. In the subsample B the significance drops a little but stays significant at a 5 % level. This result is contradicting with the propositions by Modigliani and Miller (1985) and Fama and French (1988) stating that the way of financing is irrelevant for the firm value. Secondly, maturity in years is also significant in every sample, but at different levels. In both the main sample and the subsample A the variable is significant at an 1% level, in subsample B it’s significant at a 10% level.

(22)

Although the coefficient for the coupon rate is not significant in the main sample, it does have a significant influence on the CAR’s in subsample A and B shown in appendix 17 and appendix 18. In both cases it’s significant at a 5% level. The coefficient indicates the market wants to be compensated for the risk of information asymmetry paired with the issuance of CoCo’s. A lower coupon rate, hence less compensation for information asymmetry and risk, has a negative effect on the CAR’s.

In the main sample the variable for the Core Tier 1 ratio is significant at a 10% level, indicating that it has meaningful influence in determining the CAR. This also holds for sub sample B where the coefficient of the core tier 1 ratio is significant at the same level. Additionally, the coefficient for the core tier 1 cushion, the difference between core tier 1 ratio and the trigger, is meaningfully in determining the CAR for this sample. Both

coefficients are significantly positive, indicating that in the samples considered, a lower core tier 1 ratio and core tier 1 cushion are perceived to be a bad signal and lead to a decline in CAR’s. It should be mentioned though that, as can be seen in Appendix 18, the values of these two coefficients are on the edge of the 10% significance level.

(23)

6. Conclusion and summary

Following the recent financial crisis new capital regulations were put in place in the form of Basel III to prevent financial institutions from running out of capital and becoming insolvable. These regulations were shaped in the form of minimum requirements for several capital ratios and additional requirements. Contingent convertible bonds made their introduction in 2009 with the first series of issues by Lloyds Banking Group. CoCo’s could be used to raise debt as well as meeting the new requirements set under the Basel III regulation. Initially issued as debt, CoCo’s will convert into equity upon hitting a predetermined trigger, providing a capital buffer for the issuing firm. CoCo’s are relatively new instruments, therefore little is known about the market perceptions on the issuance of CoCo’s by financial institutions. As such, this paper focuses on determining whether there is an announcement effect when issuing CoCo’s.

Previous research found that the announcement of convertible bond issues in general led to significant abnormal returns for the issuing financial institution mainly due to

information asymmetries. Modigliani and Miller propose that the financing structure of firms doesn’t influence the firm value under the assumption of perfect markets. Since information asymmetries exist with the issuance of CoCo’s there are no perfect markets. It is expected that due to information asymmetry and risk shifting incentive, the announcement of CoCo’s will lead to significant negative CAR’s.

This study is based on the issuance of CoCo’s from November 2009 till January 2014 by financial institutions. Performing an event study, the announcement effect and the

influence of firm and issue specific characteristics are tested on the CAR’s. The abnormal returns are calculated by comparing the returns in a three-day event window surrounding the announcement to the returns of the Euronext stock market index. For the test several sub samples were created to determine the announcement effects when leaving out certain announcements.

The empirical results found in this study indicate that there are no significantly negative CAR’s associated with the announcement of CoCo issues. Although negative cumulative abnormal returns were found for the announcements in the period 2009-2014, none of them appeared to be significant. The CAR for the subsample of issues where the announcement date equals the issue date shows the highest relevance, but is still not significant.

Secondly, the descriptive statistics and the correlations show that there is strong negative relation between the amount of debt the financial institution has and the core tier 1

(24)

cushion. However, this is not confirmed by the regression on the CAR. The same holds for the correlation between the core tier 1 cushion and the issue specific characteristics such the CoCo being callable, having an equity conversion and being issued by a Swiss bank.

Thirdly, the issue size scaled by the assets is significant in all samples. This indicates that the market considers large issues of debt, in the form of CoCo’s, to be a negative signal for future performance.

Lastly, The coupon rate has a significant effect on the CAR’s in the subsamples, showing that CoCo’s with lower coupon rates signal negative expectations to the

Several assumptions have been made during this research. In determining the

significance of the CAR’s a three-day event window was assumed around the announcement date of the CoCo issue. Future research could focus on extending this three-day event window and make use of the market adjusted risk model to see if the results found still persist when a robustness check is performed. Limited by the number of CoCo issues so at this moment, a valuable contribution to this research topic can be made when a larger sample of CoCo issues is available.

(25)

7. Bibliography

Abhyankar, A., & Dunning, A. (1999). Wealth effects of convertible bond and convertible preference share issues. Journal of Banking & Finance(23), 1043-1065.

Akerlof, A. (1970). The market for 'Lemons': Quality uncertainty and the market mechanism. Quarterly Journal of Economics, 84, 488-500.

Albul, B., Jaffee, D., & Tchistyi, A. (2010). Contingent convertible bonds and capital structure decisions, working paper. University of Califonia.

Ammann, M., Fehr, M., & Seiz, R. (2006). New evidence on the announcement effect of convertible and exchangeable bonds. Journal of Multinational Financial

Management, 16(1), 43-63.

Avdjiev, S., Kartasheva, A., & Bogdanova, B. (2013). CoCos: a primer. BIS Quarterly Review, 3, 43-56.

Baker, M., & Wurgler, J. (2000). The Equity Share in New Issues and Aggregate Stock Returns. The Journal of Finance, 55, 2219-2254.

Ball, R., & Brown, P. (1968). An empirical evaluation of accounting income numbers. Journal of Accountin Research, 6, 159-178.

Barclay, M., & Smith, C. (2005). The capital structure puzzle: The evidence revisited. Journal of Applied Corporate Finance, 17(1), 8-18.

Besanko, D., & Kanatas, G. (1996). The regulation of bank capital: Do capital standards promote bank safety? 5, 160-183.

Blume, M. (1971). On the assessment of risk. Journal of Finance, 26, 1-10.

Brown, S., & Warner, J. (1980). Measuring Security Price Performance. Journal of Financial Economics, 32, 57-66.

Chin, S., & Abdullah, N. (2013). Announcement effect of corporate bond issuance and its determinants. Contemporary Economics, 7, 5-18.

Choi, D., Getmansky, M., & Tookes, H. (2009). Convertible bond arbitrage, liquidity externalities, and stock prices. Journal of Financial Economics, 91, 227-225. Credit Suisse. (2011). Retrieved June 11, 2014, from

https://www.credit-suisse.com/news/en/media_release.jsp?ns=41709

Davidson III, W., Glascock, J., & Schwartz, T. (1995). Signalling with convertible debt. Journal of Financial and Quantitative Analysis, 30, 425-440.

De Roon, F., & Veld, C. (1998). Announcement effects of convertible bond loans and warrant bond loans: An empirical analysis for the Dutch market. Journal of Banking & Finance, 22, 1481-1506.

Demirguc-Kunt, A., Detragiache, E., & Merrouche, O. (2013). Bank Capital: Lessons from the Financial Crisis. Journal of Money, Credit and Baning, 45(45), 1148-1164. Duca, E., Dutordoir, M., Veld, C., & Verwijmeren, P. (2012). Why are convertible bond

announcements associated with increasingly negative abnormal stock return? Journal of banking and Finance, 36, 2884-2899.

Dutordoir, M., & Van de Gucht, L. (2004). Are european convertibles more debt-like than the US issues? An empirical analysis. Tijdschrift voor Economie en Management, 49, 533-568.

Fisher, L., Jensen, M., & Roll, R. (1969). The adjustment of stock prices to new information. International Economic Review, 10, 1-21.

Flannery, M. (2009). Stabilizing large financial institutions with contingent capital certificates, working paper. 1-30.

Flannery, M. (n.d.). No pain, No gain? Effecting Market Discipline via "Reverse

Convertible Debentures". In H. Scott, Capital Adequacy Beyond basel: Banking, Securities, and Insurance. Oxford: Oxford University Press.

(26)

Gonedes, N. (1973). Evidence on the information content of accounting numbers: Accounting-based and market-based estimates of systematic risk. Journal of Financial and Quantitative Analysis, 8, 407-444.

Green, C. (1984). Investment incentives, debt, and warrants. Journal of Financial Economics, 13, 115-136.

Hillion, P., & Vermaelen, T. (2004). Death spiral convertibles. Journal of Economics(71), 382-415.

Himmelberg, C., & Tsyplakov, S. (2012). Incentive Effects of Contingent Capital, working paper. Goldman Sachs & Co and University of South Carolina.

Hughes, J., & Jenkins, P. (2011, February 14). Credit Suisse to test coco market with $7bn issue. Retrieved June 12, 2014, from Financial Times:

http://www.ft.com/intl/cms/s/0/aaa9e006-380a-11e0-8257-00144feabdc0.html#axzz34MfAnUA2

Ineke, J., Street, L., Simpson, F., & Blackman, N. (2011). European Banks: CoCo Compendium. Morgan Stanley.

Jensen, M. (1986). Agency costs of free cash flow, corporate fiannce and takeovers. American Economic Review, 76(2), 323-329.

Kashyap, A., Rajan, R., & Stein, J. (2008). Rethinking the capital regulations, working paper. University of Chicago and Harvard University .

Kenneth, R., & Roll, R. (n.d.). Stock return variances: The arrival information and the reaction of traders. Journal of Financial Economics, 17(1), 5-26.

Koziol, C., & Lawrenz, J. (2012). Contingent convertibles. Solving or seeding the next banking crisis? Journal of Banking & Finance(36), 90-104.

Lummer, S., & Riepe, M. (1993). Convertible bonds as an asset class; 1957 - 1992. Journal of Fixed Income, 3, 47-57.

Maes, S., & Schoutens, W. (2012). Contingent capital: an in-depth discussion. Economic Notes, 41(1/22), 59-79.

Martel, M., & Padron, Y. (2006). Debt and information content: Evidence in the Spanish stock market. International research Journal of Finance & Economics, 202-208. Martynova, N., & Perotti, E. (2013). Convertible bonds and bank risk-taking. Working

paper, 1-26.

Mayers, D. (1998). Wy firms issue convertible bonds: the matching of financial and real investment options. Journal of Financial Economics, 47, 83-102.

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

Pennachi, G. (2010). A strucutral model of contingent bank capital, working paper. Federal Reserve of Cleveland. University of Illinois.

Post, J. (2011). Contingent Convertible Bonds. University of Amsterdam.

Ross, S. (1977). The determination of financial structure: The incentive-signaling approach. Bell Journal of Economics, 8(1), 23-40.

Skinner, F., & Loannides, M. (2012). Contingent Capital Securities: Problems and

Solutions, in Jonathan A. Batten, Niklas Wagner (ed). Derivative Securities Pricing and Modelling (Contemporary Studies in Economic and Financial Analysis), 94, 71-92.

Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87, 296-332. Stein, C. (1992). Convertible bonds as backdoor equity financing. Journal of Financial

(27)

Sundaresan, S., & Wang, Z. (2010). Design of contingent capital with a stock price trigger for mandatory conversion. Federal Reserve Bank of New York Staff Report no. 448.

Yu, P., & van Luu, B. (2009). Lessons from the collapse in Hybrid Bank Capital Securities. International Journal of Management Practice(5), 3-26.

(28)

Appendix 1: Issue size per company in million €

Appendix 2: Total issue size per year in million €

0 2,000 4,000 6,000 8,000 10,000 12,000 Issue size (m il lio n € ) Company 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 2009 2010 2011 2012 2013 2014 Amount iss u ed p er yea r (m il lio n € ) Year

(29)

Appendix 3: Summary statistics on firm- and issue characteristics

Variable Obs Mean Std. Dev. Min Max

Total Assets (€ million) 50 1089,82 302,53 144,71 1895,98

Total Debt (€ million) 50 350,35 130,08 40,54 560,34

Debt/Asset (%) 50 31,69 8,47 12,01 38,98

Debt/Capital (%) 50 85,68 5,13 71,61 90,80

Market-to-Book ratio 50 0,92 0,19 0,44 1,58

Coupon rate (%) 50 9,02 2,59 4,75 16,13

Maturity in years 41 12,68 4,26 10,00 30,00

Issue size-to-Asset ratio (%) 50 8.39 1.35 3,00 867,06

Trigger rate (%) 50 5,34 0,73 5,00 7,00 CET 1 rate (%) 50 8,67 3,19 6,30 16,30 CET 1 buffer (%) 50 3,33 2,94 1,30 11,18 Perpetuity (dummy) 50 0,18 0,39 0,00 1,00 Callable (D) 50 0,34 0,48 0,00 1,00 Equity conversion (D) 50 0,74 0,44 0,00 1,00 Debt seniorty (D) 50 0,18 0,39 0,00 1,00 Swiss issue (D) 50 0,20 0,40 0,00 1,00

Appendix 4: Summary statistics on winsorized variable maturity (years)

Maturity (years) 50 15,42 7,05 10,00 30,00

Appendix 5: Summary statistics on firm- and issue characteristics, sub-sample Lloyds

Debt/Asset (%) 29 37,50 0,00 37,50 37,50

Debt/Capital (%) 29 88,67 0,00 88,67 88,67

Coupon rate (%) 29 10,01 2,85 6,39 16,13

Maturity (years) 29 12,66 3,33 10,00 23,00

Issue size-to-assets ratio (%) 29 29,52 27,62 3,00 85,11

(30)

Appendix 6: Summary statistics on firm- and issue characteristics, sub sample Other

Debt/Asset (%) 21 23,67 7,69 12,01 38,98

Debt/Capital (%) 21 81,55 5,80 71,61 90,80

Coupon rate (%) 21 7,64 1,31 4,75 11,50

Maturity (years) 12 12,75 6,14 10,00 30,00

Issue size-to-asset ratio (%) 21 159,03 182.07 27.31 867,06

(31)

Appendix 7: Correlation firm and issue specific variables Issues = 50 Assets Debt Debt/Assets Debt/capital Market/book

ratio Coupon rate Maturity (years) Issue size/Assets Trigger Core tier 1 ratio Core tier 1

cushion Perpetual Callable

Equity conversion Seniority Swiss issue Assets 1,000 Debt 0,731 1,000 Debt/Assets 0,197 0,801 1,000 Debt/capital 0,483 0,893 0,895 1,000 Market/book ratio -0,347 -0,013 0,274 0,153 1,000 Coupon rate 0,080 0,345 0,444 0,365 -0,051 1,000 Maturity (years) -0,092 -0,368 -0,365 -0,355 -0,160 -0,040 1,000 Issue size/Assets -0,631 -0.661 -0.458 -0.571 -0.177 -0.208 0.1055 1,000 Trigger 0,036 -0,178 -0,254 -0,086 -0,118 -0,208 0,230 0,376 1,000 Tier 1 ratio -0,312 -0,744 -0,773 -0,726 -0,021 -0,430 0,402 0,693 0,443 1,000 Core tier 1 cushion -0,347 -0,763 -0,776 -0,767 0,007 -0,415 0,379 0,658 0,231 0,975 1,000 Perpetual -0,108 -0,476 -0,531 -0,548 -0,352 -0,140 0,838 0,327 0,132 0,466 0,473 1,000 Callable -0,156 -0,641 -0,712 -0,582 -0,052 -0,369 0,617 0,477 0,531 0,773 0,706 0,653 1,000 Equity conversion -0,009 0,569 0,855 0,651 0,309 0,424 -0,089 -0,548 -0,258 -0,607 -0,594 -0,197 -0,537 1,000 Seniority -0,160 -0,516 -0,613 -0,594 -0,085 -0,277 0,584 0,358 0,114 0,397 0,402 0,729 0,653 -0,316 1,000 Swiss issue -0,205 -0,452 -0,487 -0,356 0,266 -0,427 0,149 0,516 0,069 0,594 0,627 0,156 0,486 -0,502 0,417 1,000

(32)

Appendix 8: Correlation firm and issue specific variables, sub sample

Issues = 21 Assets Debt Debt/Assets Debt/Capital Market/book _ratio Coupon _rate Maturity _(years) _size/AssetsIssue Trigger Core tier 1 _ratio Core tier 1 _cushion Perpetual Callable _conversionEquity Seniority Swiss _issue Assets 1,000 Debt 0,851 1,000 Debt/Assets 0,055 0,546 1,000 Debt/Capital 0,484 0,791 0,791 1,000 Market/book ratio -0,408 -0,238 0,178 0,009 1,000 Coupon rate -0,041 0,094 0,391 0,211 -0,468 1,000 Maturity (years) 0,003 -0,065 0,031 -0,056 -0,077 0,443 1,000 Issue size/Assets -0,631 -0.528 -0.134 -0.384 -0.089 0.133 -0.133 1,000 Trigger 0,179 0,355 0,391 0,481 -0,003 0,149 -0,037 -0,001 1,000

Core tier 1 ratio -0,285 -0,355 -0,205 -0,343 0,358 -0,191 -0,022 0,308 -0,095 1,000 Core tier 1 cushion -0,321 -0,448 -0,327 -0,483 0,322 -0,225 -0,007 0,277 -0,450 0,932 1,000 Perpetual 0,005 -0,144 -0,171 -0,277 -0,289 0,410 0,860 -0,109 -0,241 -0,047 0,046 1,000 Callable 0,030 -0,112 -0,080 0,005 0,241 0,090 0,514 -0,334 0,157 0,118 0,049 0,420 1,000 Equity conversion -0,215 0,147 0,690 0,327 0,230 0,457 0,406 -0,159 0,203 0,020 -0,056 0,311 0,131 1,000 Seniority -0,059 -0,213 -0,339 -0,353 0,039 -0,096 0,484 -0,046 -0,267 -0,222 -0,101 0,611 0,420 0,113 1,000 Swiss issue -0,108 -0,061 -0,018 0,087 0,494 -0,608 -0,190 0,233 -0,371 0,200 0,315 -0,248 -0,023 -0,159 0,138 1,000

(33)

Appendix 9: Scatterplot issues size on core tier 1 ratio

(34)

Appendix 11: Summary statistics abnormal returns for t (n=22)

t mean median sd variance min max skewness kurtosis

-1 -0,004 -0,003 0,013 0,000 -0,038 0,023 -0,595 4,740

0 0,002 0,000 0,017 0,000 -0,028 0,044 0,935 3,818

1 -0,003 -0,003 0,015 0,000 -0,037 0,026 -0,293 2,917

Total -0,002 -0,003 0,015 0,000 -0,038 0,044 0,317 4,361

-1 -0,003 -0,003 0,010 0,000 -0,026 0,023 0,229 4,583

0 0,000 0,000 0,014 0,000 -0,028 0,035 0,718 4,017

1 -0,002 -0,002 0,013 0,000 -0,030 0,026 -0,002 2,760

Total -0,001 -0,003 0,013 0,000 -0,030 0,035 0,425 3,964

-1 -0,003 -0,003 0,011 0,000 -0,026 0,023 0,284 4,138

0 0,002 0,000 0,013 0,000 -0,016 0,035 1,277 4,339

1 0,000 -0,001 0,014 0,000 -0,030 0,026 -0,151 2,844

Total -0,001 -0,003 0,013 0,000 -0,030 0,035 0,515 3,994

-1 -0,001 -0,001 0,005 0,000 -0,006 0,004 -0,053 1,500

0 -0,007 -0,007 0,020 0,000 -0,028 0,013 -0,012 1,500

1 -0,008 -0,012 0,011 0,000 -0,015 0,004 0,632 1,500

Total -0,005 -0,006 0,012 0,000 -0,028 0,013 -0,365 2,511

Appendix 15: T-test on CAR’s

n Mean Std. Err. Std. Dev t-test

22 -.0049018 .0034396 .016133 -1.4521

21 -.0036682 .0033675 .0154316 -1.0893

(35)

Appendix16: Regression of CAR on firm and issue specific variables Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) Assets -0.012 0.001 Debt -0.004 (0.003) Debt/Asset -0.008 (0.044) Debt/Capital -0.040 (0.058) Market-to-Book ratio -0.008 (0.011) Coupon rate 0.361 (0.291) Maturity in years 0.096** (0.032)

Issue size-to-assets ratio 0.004**

(0.002)

Trigger 0.372

(0.335)

Core Tier 1 ratio 0.234*

(0.115)

Core Tier 1 cushion 0.162

(0.110) Perpetual 0.895 (0.670) Callable 0.670 (0.931) Equity conversion -0.074 (0.780) Seniority -0.056 (0.725) Swiss issue -0.186 (0.692) R-squared 0.115 0.104 0.001 0.021 0.024 0.210 0.136 0.164 0.048 0.143 0.078 0.078 0.032 0.001 0.000 0.004

(36)

Appendix 17: Regression of CAR on firm and issue specific variables, subsample A Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) Assets -0.001 (0.001) Debt 0.003 (0.002) Debt/Asset 0.021 (0.040) Debt/Capital -0.016 (0.053) Market-to-Book ratio -0.008 (0.011) Coupon rate 0.594*** (0.148) Maturity in years 0.096*** (0.032)

Issue size-to-assets ratio 0.003**

(0.0013)

Trigger 0.272

(0.338)

Core Tier 1 ratio 0.167

(0.136)

Core Tier 1 cushion 0.099

(0.114) Perpetual 0.709 (0.670) Callable 0.151 (0.955) Equity conversion 0.244 (0.783) Seniority -0.273 (0.711) Swiss issue -0.429 (0.680) R-squared 0.117 0.063 0.012 0.004 0.021 0.257 0.135 0.154 0.029 0.066 0.029 0.054 0.002 0.006 0.008 0.020

(37)

Appendix 18: Regression of CAR on firm and issue specific variables, subsample B Variable (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) Assets -0.002** (0.001) Debt -0.005* (0.003) Debt/Asset -0.005 (0.043) Debt/Capital -0.046 (0.057) Market-to-Book ratio 0.002 (0.011) Coupon rate 0.552** (0.203) Maturity in years 0.076* (0.040)

Issue size-to-assets ratio (0.003)**

(0.001)

Trigger 0.093

(0.374)

Core Tier 1 ratio 0.224*

(0.125)

Core Tier 1 cushion 0.170*

(0.969) Perpetual 0.393 (0.715) Callable 0.621 (1.214) Equity conversion -0.104 (0.775) Seniority -0.318 (0.743) Swiss issue 0.011 (0.689) R-squared 0.231 0.217 0.001 0.030 0.002 0.207 0.112 0.187 0.004 0.133 0.094 0.0185 0.0257 0.0013 0.0120 0.0000

The announcement effect of contingent convertible bond issuance by financial institutions

The announcement effect of Contingent Convertible bond issuance by

financial institutions

Table of Contents