The effect of the creation of the Banking Union on the European banking market : did the creation of the Banking Union produce new valuable information for the European banking market?

(1)

The effect of the creation of the Banking Union

on the European banking market.

Did the creation of the Banking Union produce new valuable information for the

European banking market?

Bachelor thesis

Economics & Business

Student: Kevin van Vliet

Student number: 10735070

Word count: 7622 words

(2)

2 Statement of Originality

This document is written by Student Kevin van Vliet who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Abstract

Unsuccessful regulation and supervision of the financial markets are one of the main reasons the worldwide financial crisis of 2008 took place. After the crisis, the European Central Bank recognized that a Banking Union needed to be created to strengthen and stabilize the whole financial market in the European Union. This thesis examines the effect of the creation of the Banking Union on the European banking market. In particular, this thesis investigates fundamental components of the Banking Union. Using an event study approach, we examine financial market’s reactions to the Comprehensive Assessment, the Single Supervisory Mechanism and the Single Resolution Board (as central authority of the Single Resolution Mechanism). The results of the study show that there is no significant effect of the individual events or all events together. These findings suggest that, at least in the short run, the creation of the Banking Union did not produce new valuable information for the European banking market when we look at banks’ stock market prices.

Keywords: Banking Union; European Central Bank; European Banking Authority; Comprehensive Assessment; Single Supervisory Mechanism; Single Resolution Mechanism; Single Resolution Board

(4)

4

Introduction

Unsuccessful regulation and supervision of the financial markets are one of the main reasons the worldwide financial crisis of 2008 took place (Levine, 2010; The Turner review, 2009). Supervision was carried out subject to different national rules and since most of these national supervisory models were outdated, i.e. lag behind the reality, the existing supervisory measures were not sufficient enough to accommodate to a financial crisis (Junevičius & Puidokas, 2015).

Monitoring the risk taking and other financial information of the banking sector is very important to central banks for making monetary policy. Monitoring the banking sector correctly and efficiently enables central banks to improve market efficiency. One important way to monitor the performance of the banking sector is risk reporting. After the financial crisis, it has been argued that looking at the risk reporting of banks is useful for avoiding a banking crisis. But to be able to monitor correctly, there must be a transparent working central bank which is able to observe the banking sector consistently. The global financial crisis in 2008 triggered policymakers to revise the current micro-prudential approach of banking supervision (Black et al., 2016). With this thesis we will investigate if the creation of the Banking Union, starting with the Comprehensive Assessment and followed by the Single Supervisory Mechanism and the Single Resolution Mechanism, produced new valuable information for the European banking market. As far as we know, there is no research that examines the effect of the Banking Union as a whole, so in this way we will contribute to the existing literature, which investigates only parts of the Banking Union. Below we will explain more about the creation of the Banking Union and its pillars.

After the global financial crisis, the European Central Bank (ECB) recognized that a more system-wide approach was needed and reacted with both standard and non-standard monetary policy to ensure that the confidence in the banking system was restored (Ricci, 2015). Next to that, in 2012, the European Commission proposed to extend the powers of the ECB with relation to the Euro zone banking supervision. A Banking Union needed to be created to strengthen and stabilize the whole financial market in the European Union. A fundamental component of this Banking Union is the Single Supervisory Mechanism (SSM), implemented in November 2014, which is responsible for the direct monitoring of the biggest European banks, by the ECB designated as significant institutions (Carboni et al., 2017). As preparation for the change of banking supervision, the ECB carried out a Comprehensive Assessment (CA) to make the quality of the available information of the status of banks more transparent. Furthermore, the goal was to restore confidence of stakeholders by assuring that the banks subject to the CA were trustworthy and applying necessary corrective actions to the balance sheets of the banks

(6)

6 (European Central Bank (ECB), 2013, p.2). A second step in the creation of the Banking Union is the Single Resolution Mechanism (SRM), which contains a new concept of payment for the rescue of failing banks. No longer these banks are bailed out by the European taxpayer but are now initially saved by the party responsible for the banks’ operations itself, i.e. the investors and depositors. This, with the help of the Single Resolution Fund, contributes to the creation of a valuable regime for the rescue of failing banks.

Previous studies (e.g. Carboni et al., 2017; Sahin & De Haan, 2016) examined the impact of the announcement, on October 23rd_{, 2013, and publication of the results, on October 26}th_{, 2014,}

of the Comprehensive Assessment, prior to the implementation of the SSM. They conclude that for some banks the assessment has led to increased transparency, since the market reacted on the publication of the results, presuming that these results contained new valuable information.

In this thesis, we will study the change in stock price of the listed, by the ECB supervised, banks with total assets more than 500 billion in euro’s. The hypothesis is stated as follows; the creation of the Banking Union has led to more transparency in the European banking market and so, produced new valuable information. The rest of the paper is structured as follows. First, we review previous literature and explain more about the events that have taken place regarding the ECB’s supervision of banks. Then, the choice of data and banks is explained. Next, we describe how the hypothesis is analyzed, by empirical analysis using event studies. For the event studies we elaborate the estimation of normal performance, the choice of event windows and the Abnormal Returns Method, which we use to eventually compute the Cumulative Average Abnormal Returns to test the hypothesis of a market reaction significant different from zero. After this, the main results of the research are discussed. Finally, the conclusions are drawn followed by suggestions for further research. Hereby we will give an answer to the research question: Did the creation of the Banking Union produce new valuable information for the European banking market?

Literature review

Since March 2001, the Lamfalussy process (2001) was the basis of supervisory measures for financial markets. The Lamfalussy process is a four-step approach of financial regulation where each step focuses on a specific part of financial legislation. This four-step structure aimed to make the decision-making process and promoting efficiency of financial markets supervision simpler (Junevičius & Puidokas, 2015). But later this process turned out to have become outdated and had to be reformed. So, The European Commission appointed a group of experts directed by the

(7)

7 French Central Banker J. Larosiere to analyze the current way of financial regulation and supervision. This resulted in a recommendation report with 31 practical improvements of the whole supervisory mechanism (Larosiere, 2009). These recommendations led to the initiation of a new European system of financial supervision (ESFS) which main task is to ensure appropriate and consistent financial supervision in all EU countries (European Central Bank, n.d.). This was an important achievement regarding the developments of a single market of financial services. ESFS is responsible for the financial supervision on both macro-economic and microeconomic level. The aim on the macro-level is to prevent systematic risk which endangers financial stability and to minimize the effect of this risk regarding macro-economic changes. On a micro-economic level, the European system of financial supervision has different layers depending on the sector of regulation and supervision (insurance, securities markets and banking) and the level of regulation and supervision (European and national). But even though banking supervision was executed on a European level, research results showed that the supervision model were not able to prevent disunity of the European financial markets. (Regulations (EU) No 1092/2010, 1093/2010, 1094/2010, 1095/2010 of the European Parliament and of the Council, 2010). Several institutions and researchers claim that the ESFS was not able to properly respond to the financial crisis. A Banking Union needed to be created including 5 important elements: a single rulebook for financial services in the EU, single supervision of banks, a single framework on bank resolution, a deposit guarantee scheme and centralized help for banks with financial problems. All elements to strengthen and stabilize the whole financial market in the European Union (IMF, 2013; Ionnidou, 2012; Schoenmaker, 2012).

In an early stage, the European Parliament already implemented motions for the creation of a system of single financial supervision as a reaction on the many discussions about this topic. But only after the financial crisis in 2008, governments were willing to think about a change of supervision (Larosiere, 2009). One of the main goals of the Banking Union is to restore public confidence in the financial markets, since the ongoing effects of the financial crisis endanger the single currency and the financial position of all EU countries. In this paper, we will research if this change of supervision indeed restored this confidence. Lack of confidence could for instance make private (banking) debt sovereign debt, since the state is the ultimate guarantor. The Banking Union makes sure that private and public debt is better separated by keeping banks under stricter control and avoiding the idea that eventually the state will step in. To accomplish this goal, the ECB has started with stronger regulation and precise supervision towards the banking system. If it is up to the ECB, the presumption of bank directors and shareholders of being “too big to fail”,

(8)

8 and the moral hazard which is connected to this idea, is considered past tense (Wymeersch, 2014).

The ECB chose that the Banking Union originally would be created for the single currency, the Euro. This was the currency which had the risk of being devaluated or even being eliminated when a big financial crisis would occur. The Euro needed to be supported by a strong supervisory regime. But this meant, that initially non-euro countries could not participate. Critics stated that this would split the single internal market into a ‘two-speed Europe’ which is a risk to the integrity of the market following UK publications (House of Lords, 2012; Seyad, 2013). However, the Banking Union aims to reverse the fragmentation of the European financial markets (Wymeersch, 2014). Pisani-Ferry, Sapir, Veronica and Wolff (2012) state that the Banking Union should include all the 27 European Union countries to maintain financial integration in the European single market. Otherwise it could be possible that national regulators of non-participating Member States restrict cross-border operations of banks that are located in their country to avoid the prudential requirements mandated by the ECB. Besides that, the authors mentioned above claim that, given the geographical area of the Banking Union, the European Banking Authority (EBA) will not gain new skills regarding prudential supervision. The EBA, responsible for the regulation of banks in all 27 European Union countries, aims at improving the cooperation between the national supervisors of all EU Member States and continue the process of strengthening the methods of prudential supervision. Within the Banking Union, the EBA will continue to develop rules applicable to all Member States. Also, the EBA exercises stress tests to increase transparency in the European banking market and to identify weaknesses in bank capital structures (Barbu & Boitan, 2013). The EBA stress tests were also part of the Comprehensive Assessment carried out by the ECB prior to the implementation of the Single Supervisory Mechanism.

Comprehensive Assessment

Theoretical literature about regulatory stress tests states that full disclosure of stress tests results is useful for investors since it gives consistent and transparent information about the financial condition of banks, which makes it easy to compare banks (Bernanke, 2013). Other authors also point out these benefits but argue that disclosing too much information leads to the destroying of risk-sharing opportunities and reduces liquidity in the interbank market (Hirtle & Lehnert, 2014). During a crisis, these risk-sharing opportunities are already damaged by the perception that banks are under-capitalized say Goldstein and Leitner (2015). In this situation, disclosure of stress tests could cause a stabilizing effect. To produce this effect, it is important that these regulatory stress tests produce new valuable information to the market and so increase transparency on the

(9)

9 financial conditions of banks. We will investigate if the stress tests executed by the EBA did produce new valuable information. Petrella and Resti (2013) proved that the stress test performed by the EBA in 2011 provided new valuable information for the market and that investors were beforehand not able to anticipate the results. Next to that, they found out that the market did not only reacted to the released detailed historical data but also to the performance of the investigated banks during the stress tests (i.e. the banks vulnerability to simulated crisis scenarios).

But the stress tests were only a part of a much bigger picture. The step before the launching of the SSM was the 2014 Comprehensive Assessment, a financial heath check, which consisted of an Asset Quality Review (AQR), a stress test and a supervisory risk assessment for all significant institutions subject to the direct supervision of the ECB. The goal of the AQR was to boost the transparency of bank exposures alongside with the correctness of asset and collateral valuation and the related provisions. The stress test, performed by the EBA, tested the resilience of banks’ balance sheets (European Central Bank (ECB), 2014). The supervisory risk assessment reviewed the major risks regarding liquidity, funding and leverage. The aim of the CA was to increase transparency on the condition of banks, to restore confidence of stakeholders by assuring that the banks subject to the CA were trustworthy and applying necessary corrective actions to the balance sheets of the banks (European Central Bank (ECB), 2013, p.2). Carboni et al. (2017) show with empirical analysis that investors were able to identify weaknesses in bank capital structures before the CA exercise, but that it still was able to produce new valuable information. This indicates that the CA was successful relating to the aim of increasing transparency.

Single Supervisory Mechanism

A fundamental component of the Banking Union is the Single Supervisory Mechanism (SSM), implemented in November 2014, which is responsible for the direct monitoring of the biggest European banks, by the ECB designated as significant institutions. We will examine if the start of the SSM had a significant effect on the European banking market. The SSM makes sure that these 130 most significant banks in 19 countries (with total assets worth 22 trillion Euro, i.e. 82% of the total banking assets in the Euro area) are subject to the same uniform supervision, risk control, prevention of economic crises and prudential regulations. Furthermore, the SSM causes bigger financial integration between European banks (Carboni et al., 2017). The more large and risky credit institutions are directly supervised by the ECB whereas the smaller credit institutions are supervised by their national regulators. Although, the ECB can take over supervision if they think this is necessary. Furthermore, the national supervisors have to follow guidelines

(10)

10 determined by the ECB. The determination of significant institutions (SIs) depends on three criteria: size, state support and minimal local presence where size is the most important. A credit institution will be labeled as ‘significant’ if it has total assets bigger than 30 billion Euro, the ratio of total assets over GDP in its country is bigger than 20% or if the institution is characterized by its national regulator as of being of significant relevance to the domestic economy. The last element is always examined by the ECB to make sure that the concerned institution truly is significant. Moreover, the ECB can decide that a credit institution is significant if it has banking subsidiaries in more than one Member State and the cross-border assets or liabilities of the institution are a big part of the total assets or liabilities (The Council Regulation, Article 6, 2013).

However, there are exceptions where the ECB believes that the credit institution should remain under national supervision. An example of a case like this is when the volume of assets does not exceed 5 billion Euro’s. This could be possible for a bank which is part of a much larger financial group (supervised by the ECB). You could say that every exception should be based on objective factors like the nature of the business or volume of turnover, excluding discretionary decisions. Exceptions like this allow asset managers and security firms run under the form of a bank to be excluded from the SSM (Wymeersch, 2014).

Next to the domestic part of significant institutions, also the foreign branches located in non-participating member states of the European Union are regulated. This regulation goes as far as that the ECB carries out the whole task of the national regulator consistent with European Union law. Since there is no criterion for a certain volume of these foreign branches, all branches come under direct supervision of the European Central Bank. Because of this extended supervision the Single Supervisory Mechanism affects also the non-euro states of the European Union (The Council Regulation, Article 4, 2013) Most remarkable is the effect on supervision in some EU countries where most banks are either branches or subsidiaries, usually of the largest Euro area banks. With these supervision criteria the SSM covers 80% of all the banking assets in the SSM area (Constâncio, 2012). But aside the institutions that are subject to the SSM determined by the ECB, non-Euro Countries can decide by themselves to voluntarily participate in the SSM. There are various reasons for non-Euro countries to choose to participate. Banking groups which have a lot of operations inside the SSM area might want to participate because of cost saving and simplification. Furthermore, a reason could be that participation goes together with reputational advantages since the market prefers a stable supervisory regime over a regime that could be influenced by national biases. This eventually may lead to more beneficial interest rates, equity prices and credit ratings (Wymeersch, 2014).

(11)

11

Single Resolution Mechanism

Another important part of the Banking Union is the Single Resolution Mechanism (SRM). The SRM is the second step after the SSM in the creation of the Banking Union. If a bank experiences financial problems despite the fact that it is strictly regulated, the SRM will make sure that any necessary reorganizations will cost the taxpayers and the real economy as little as possible (Regulation of the European Parliament and of the Council, 2014). Any reorganizations are funded by the financial sector itself. The idea that the taxpayer should bail should bail out failing banks, which led the European Union in a big financial crisis, is abandoned. In the first place, loses shall be paid by the party responsible for the banks’ operations, i.e. the investors and depositors. Secondly, there is the Single Resolution Fund (SRF) which contains of contributions of the banks raised on the EU level to make sure that reorganization resources are objectively distributed around the Member States. With the SRF, a change has taken place where a system of ‘bail outs’ cleared the way for a system of ‘bail ins’. Resolving financial problems of banks is no longer (fully) paid by the taxpayer but (mostly) paid by the banks’ stakeholders (Junevičius & Puidokas, 2015). The Single Resolution Mechanism contributes to the creation of a valuable regime for the rescue of failing banks.

As central authority of the SRM there is the Single Resolution Board (SRB). In this paper we will test the effect of the implementation of the SRB on the European banking market. The SRB is the institution which has the power of making centralized decisions. They make decisions based on the Bank Recovery and Resolution Directive (BRRD) and the Single Resolution Mechanism Regulation (SRMR). The BRRD gives banks guidelines to prepare recovery plans to overcome any form of financial distress and includes rules regarding the contributions each bank has to make to the SRF. The size of these contributions depend on the institution’s size and risk profile. The SRMR is the basis of a central decision making guideline regarding resolution of financial distress in the Banking Union. With all these tools, the SRB controls everything that has to do with resolution. They make resolution plans for the banks under direct supervision of the SSM, decide on the use of resolution resources and give instructions to national resolution authorities (Single Resolution Board, n.d.).

(12)

12

Methodology

In this section we explain how we will produce answers to the question: Did the creation of the Banking Union produce new valuable information for the European banking market? First, we will explain the choice of events and the choice of banks. With empirical analysis, we will analyze the hypothesis; the creation of the Banking Union has led to more transparency in the European banking market and so, produced new valuable information. To examine the reaction of the investors in the European banking market due to the changes of the European supervision, we use standard event study techniques (Campbell et al., 1997). For the event studies we elaborate the estimation of normal performance, the choice of event windows and the Abnormal Returns Method, which we use to eventually compute the Cumulative Average Abnormal Returns to test the hypothesis of a market reaction significant different from zero.

Choice of events

The implementation of the Banking Union has some big event dates, as shown in Table 1. There are four events that have taken place which we think could have a significant effect on investors’ behavior. The dates are originally collected from the European Central Bank (2014) press releases. These dates represent large announcements and events relating to the implementation of the Banking Union. We focus on these large announcements and events since they probably have a strong effect because of their importance and size.

Table 1: The significant events of the Banking Union process

23/10/2013 The ECB announces the start of the Comprehensive Assessment preceding the launch of the Single Supervisory Mechanism.

26/10/2014 After completing the Comprehensive Assessment, the ECB publishes the results for both countries as a whole and individual banks, together with recommendations for supervisory measures.

04/11/2014 The Single Supervisory Mechanism becomes operational. 01/01/2016 The Single Resolution Board becomes operational.

This table reports the most important ECB press releases related to the Banking Union.

At the first date, October 23th 2013, the ECB releases a statement about the start of the Comprehensive Assessment. The CA, carried out by the ECB together with the national competent authorities (NCAs), starts in November 2013 and takes 12 months. As earlier examined by Carboni et al. (2017) and Sahin and De Haan (2016), the announcement of the CA

(13)

13 could have a significant effect on the European banking market and so we will investigate if the announcement created new valuable information.

At the second date, October 26st 2014, the ECB discloses the definitive results of the CA. The main results state that at 25 participating banks a capital shortfall of €25 billion is detected of which 12 banks already covered this shortfall increasing their capital with €15 billion in 2014. From each bank with capital shortfall a capital plan is demanded within two weeks after the announcement in which banks have to elaborate a plan on how they will cover the capital shortfall within 9 months. Next to the capital shortfall, asset value adjustments of €37 billion are located by the CA. This represents a total impact of €62 billion on banks (European Central Bank, 2014). As with the announcement of the CA, Carboni et al. (2017) and Sahin and De Haan (2016) also conclude that the disclosure of the CA results has had a significant effect on the European banking market, which is why we also will take this date in account.

At the next date, November 4th, 2014, the SSM becomes operational and so the ECB acquires the total responsibility for the supervision of the euro area banks which causes bigger financial integration between European banks. The significant institutions are from now on subject to the same uniform supervision, risk control, prevention of economic crises and prudential regulations (Carboni et al., 2017). We will investigate if investors reacted on the implementation of the SSM.

Finally, the last date, January 1st, 2016, we review if the launce of the Single Resolution Board had a significant effect on the European banking market. Since January 1st, 2015, the SRB has been operational as independent European Union Agency and became fully operational, with all its resolution tools, on January 1st, 2016 (Single Resolution Board, n.d.).

(14)

14

Choice of banks

The daily stock data of 9 of the biggest banks in Europe which are supervised by the European Central Bank are collected (see table 2). I looked at the most recent list of supervised entities publicized by the European Central Bank (2017) and selected the banks with total assets exceeding 500 billion euro. Of these 13 banks, 9 are listed since before the date of the first event (and before the start of the estimation window of the normal return estimation). I collected the daily stock data of these 9 banks from Yahoo Finance from 15-10-2012 until 30-11-2017. The Stoxx Europe 600 Banks Index, consisting of 47 European banks (Stoxx, n.d.), is used as proxy for the market portfolio, which is obtained from DataStream with the same time span as the daily stock data. All the data is collected using only trading days.

Table 2: List of supervised banks with total assets > EUR 500 billion

Event Studies

Regression

The parameters of the normal performance are estimated with an estimation window of 255 trading days which is a sufficient window for an event study using daily data (MacKinlay, 1997). During this estimation window the risk-free rate is always equal to 0.00 when we look at the Fama

Size Bank Country Listed*

Total assets > EUR 500-1,000 billion

COMMERZBANK Aktiengesellschaft Germany Yes

Banco Bilbao Vizcaya Argentaria, S.A. Spain Yes Confédération Nationale du Crédit Mutuel France No

Intesa Sanpaolo S.p.A. Italy Yes

UniCredit S.p.A. Italy Yes

Coöperatieve Rabobank U.A. Netherlands No

ING Groep N.V. Netherlands Yes

Total assets > EUR 500-1,000 billion

Deutsche Bank AG Germany Yes

Banco Santander, S.A. Spain Yes

BNP Paribas France Yes

BPCE S.A. France No

Crédit Agricole S.A France Yes

Société générale S.A France No

(15)

15 and French Europe 3 factors daily data. Due to the lack of variance on daily data we use the Fama and French Europe 3 factors monthly data, where there is some difference in the level of the risk-free rate and compute this data on a daily basis (French, n.d.). Following Sahin and de Haan (2016) the interval of (-265, -10) is used, where t=0 is the date of the first event to determine the normal return of the banks. The event periods itself are not included in the estimation window to make sure that the events do not influence the parameter estimates of the normal return model (Campbell et al., 1997). For the estimation of the parameters of normal performance we use the market model (MacKinlay, 1997), which relates the return of a security to the return of the market. The following equation is used:

𝑅𝑖𝑡− 𝑅𝑓 = 𝛼𝑖+ 𝛽𝑖(𝑅𝑚𝑡− 𝑅𝑓) (1)

We calculate the excess return of the banks by deducting the normal return (total stock return in the above specified estimation window) by the risk-free rate determined through the Fama and French Europe 3 factors monthly data. The excess return of the market is determined deducting the Stoxx Europe 600 Banks Index with the same risk-free rate as mentioned above. Now, we can do a regression in Stata with dependent variable the excess normal return of a bank and independent variable the excess market return. Results of an Ordinary Least Squares (OLS) regression give 𝛼 as constant and 𝛽 as effect of the market return on the bank’s return.

Choice of windows

Following Morgan et al. (2015) and other papers measuring market reaction to big events and/or policy announcements (e.g. Carboni et al., 2017; Onali et al., 2016; Fiordelisi and Ricci, 2016), short event windows of (-1,+1) and (0,+1) should be used in order to make sure that there are no overlapping events. With these event windows the effect of the events as a whole but also the effect which occurs after the stock market closes on the day of the event can be examined. Flammer (2011) claims that it is also of importance to consider bigger time intervals prior to and after these event windows to examine the impact of the event before or after the two- or three-day event window. Within this research we actually have one announcement of a new phase of the development of the Banking Union, one publication of results and two starts of phases of the development of the Banking Union. Regarding the announcement of the CA and the publication of the results we expect that an effect could be seen in a very short time window and so we include an event window of (-1,+1). Some information could have leaked before the announcement/publication, which is why we take the day beforehand into account. For the two starts of new phases, bigger time intervals might be more useful. We consider the time interval (-9,0) since we used (-265,-10) as estimation window for normal performance and so we cannot go

(16)

16 any further before the event date. We consider time interval (0;14) since each bank with capital shortfall was obliged to come up with an on how they will cover their capital shortfall after the publication of the CA results and we think that the effects of the start of new phases should be visible within two weeks.

Abnormal Returns Method

A common way to calculate the normal and abnormal returns of the daily stock data (DeLong and DeYoung, 2007) of the 9 biggest, listed banks, designated as significant institutions is used. Following the market model (MacKinlay, 1997), which relates the return of a security to the return of the market, the normal returns are estimated:

𝑅𝑖𝑡 = 𝛼𝑖+ 𝛽𝑖𝑅𝑚𝑡+ 𝜀𝑖𝑡 (2)

Where 𝑅𝑖𝑡 and 𝑅𝑚𝑡 are the time dependent returns, respectively on security i and the

market portfolio, and 𝜀𝑖𝑡 is the zero-mean disturbance term. 𝛼𝑖 and 𝛽𝑖 are the parameters of the

market model. So, every observation (𝑅𝑖𝑡) is a function of the market portfolio return (𝑅𝑚𝑡).

Now, the abnormal returns (ARs) are estimated as the difference between the actual returns and the expected returns. The expected returns are the returns which would have occurred if the event did not take place. The ARs meant by the market model are given by:

𝐴𝑅𝑖𝑡 = 𝑅𝑖𝑡− (𝛼̂𝑖+ 𝛽̂𝑖𝑅𝑚𝑡) (3)

The hat tells us that the parameters 𝛼𝑖 and 𝛽𝑖 are estimated. The next step, following

Morgan et al. (2014), is to sum up all the ARs to compute the Cumulative Abnormal Returns (CAR) on a certain time t (time of the event). After this the mean for every event window, the Cumulative Average Abnormal Return (CAAR), can be calculated. Once all the CAARs are calculated, the hypothesis of a market reaction significant different from zero is tested by computing a test statistic:

𝑇𝑡𝑒𝑠𝑡𝑖𝑤= 1 √𝑁𝑤 ×

𝐶𝐴𝐴𝑅_𝑖𝑤

𝑆𝐷_𝐴𝑅𝑤 (4)

Where 𝑁𝑤 is the number of days in a certain event window, 𝐶𝐴𝐴𝑅𝑖𝑤 is the Cumulative

Average Abnormal Return of bank 𝑖 in a certain event window and 𝑆𝐷𝐴𝑅 is the Standard Deviation

of the Abnormal Returns in a certain event window. If the absolute value of Ttest is greater than 1.96, then the Cumulative Average Abnormal Return for that stock is significantly different from zero at the 5% level.

(17)

17

Results

The first thing we do is preparing the data for an event study in Stata. We have to format the dates of the stock data and the event dates so that Stata recognizes these as dates. After that we have to combine the stock data and the event data and duplicate the stock data for the different events. Now we generate an identifier to be able to determine every unique combination of bank and event date. Then we label the event days and set the event- and estimation windows. Now we can do an OLS regression to estimate normal performance.

The main results of the OLS regression for the market model parameters 𝛼 and 𝛽 are presented in Table 3 and the complete results in Appendix 1. As we can see in Appendix 1, looking at the adjusted R-squared the excess market returns explain much about the excess returns of the banks (58,85%-74,41%), except for the Commerzbank AG (21,99%). However, Commerzbank AG is the only bank of the sample which has an alpha significant at the 5% level but still is quite small. The other alpha’s are small and insignificant, which I think is because of the size and importance of these banks, they are mostly dependent on and move along with the market.

After the estimation of the parameters we can calculate the Abnormal Returns using equation (3), presented in the Methodology. We compute the Cumulative Abnormal Returns simply by cumulating the ARs within the different event windows (-7,+7), (-6,0) and (0,+14). We use these CARs to calculate the Cumulative Average Abnormal Returns by dividing the CARs by the number of days within a certain event window. Eventually, to test whether the CAARs are statistically different from zero, we want to calculate the test statistic as shown as equation (4). To do this, we first have to compute the standard deviation of the Abnormal Returns, which can be done in Stata. The results of the event studies with different event windows are shown in tables 4, 5, 6 and 7 below.

Bank Alpha Beta

CBK -0.0341724 0.0685906 BBVA 0.0025059 0.0367633 ISP -0.0018337 0.0095650 UCG 0.0093561 0.1738768 INGA -0.0011764 0.0432093 DBK -0.0203880 0.1916177 SAN -0.0021765 0.0266912 BNP 0.0111903 0.2686789 ACA -0.0000584 0.0495826

(18)

18

Table 4 Financial market reaction to the announcement of the Comprehensive Assessment

_CAAR _T-statistic 23/10/2013 (-7,+7) (-6, 0) (0,+14) (-7,+7) (-6, 0) (0,+14) ACA -0.007 -0.007 0.014 -0.023 -0.056 0.034 BBVA -0.021 -0.001 -0.034 -0.079 -0.006 -0.142 BNP 0.019 -0.113 0.048 0.012 -0.201 0.031 CBK 0.046 0.051 0.101 0.085 0.170 0.096 DBK 0.010 0.026 -0.024 0.012 0.045 -0.035 INGA 0.032 0.041 0.023 0.139 0.243 0.057 ISP 0.004 0.004 0.002 0.060 0.083 0.025 SAN 0.006 0.011 -0.002 0.034 0.100 -0.009 UCG 0.022 -0.055 0.024 0.014 -0.090 0.013

Table 5 Financial market reaction to the publication of the Comprehensive Assessment results

CAAR T-statistic 27/10/2014 (-7,+7) (-6, 0) (0,+14) (-7,+7) (-6, 0) (0,+14) ACA 0.015 0.046 -0.064 0.051 0.438 -0.098 BBVA -0.037 -0.022 -0.022 -0.159 -0.170 -0.093 BNP -0.003 -0.005 0.005 -0.003 -0.007 0.005 CBK 0.028 0.050 0.006 0.046 0.149 0.008 DBK -0.034 -0.022 0.013 -0.052 -0.070 0.021 INGA 0.023 0.024 0.009 0.080 0.131 0.033 ISP -0.006 -0.005 -0.004 -0.090 -0.138 -0.065 SAN -0.020 -0.022 -0.011 -0.151 -0.411 -0.073 UCG -0.086 -0.080 -0.117 -0.100 -0.241 -0.079

(19)

19

Table 6 Financial market reaction to the start of the Single Supervisory Model

_CAAR _T-statistic 04/11/2014 (-7,+7) (-6, 0) (0,+14) (-7,+7) (-6, 0) (0,+14) ACA -0.062 0.000 -0.067 -0.093 0.001 -0.112 BBVA -0.020 -0.050 -0.022 -0.086 -0.254 -0.071 BNP -0.006 0.076 0.035 -0.006 0.145 0.036 CBK 0.028 0.050 0.006 0.046 0.149 0.008 DBK -0.034 -0.022 0.013 -0.052 -0.070 0.021 INGA 0.008 0.020 -0.005 0.031 0.098 -0.020 ISP -0.004 -0.012 0.003 -0.061 -0.246 0.055 SAN -0.012 -0.018 -0.001 -0.080 -0.145 -0.003 UCG -0.152 -0.106 -0.005 -0.111 -0.143 -0.003

Table 7 Financial market reaction to the start of the Single Supervisory Board

CAAR T-statistic 04/01/2016 (-7,+7) (-6, 0) (0,+14) (-7,+7) (-6, 0) (0,+14) ACA 0.019 0.005 -0.011 0.067 0.025 -0.024 BBVA -0.001 -0.016 0.015 -0.009 -0.150 0.126 BNP -0.068 -0.149 -0.093 -0.063 -0.192 -0.066 CBK -0.012 0.002 -0.028 -0.071 0.020 -0.139 DBK 0.058 0.024 -0.072 0.092 0.054 -0.057 INGA -0.003 -0.004 -0.014 -0.018 -0.032 -0.061 ISP -0.004 -0.012 0.003 -0.061 -0.246 0.055 SAN -0.001 -0.007 0.014 -0.006 -0.073 0.089 UCG -0.034 -0.059 -0.088 -0.029 -0.076 -0.043

(20)

20 First, we have a remark about the event dates: two of the original event days (26/10/2014 and 01/01/2016) are replaced by two other days (27/10/2016 and 04/01/2016). The reason for this is that there was no data available in the sample on these days, we used the first date with available data instead. A second remark is that in case the data extracted from Yahoo Finance stated a NULL value or the date (and so stock data) did not exist, we assumed that the certain stock price stayed the same as the day before, giving an Abnormal Return of minus the risk-free rate.

The results of the event studies show that Cumulative Average Abnormal Returns never exceed the absolute value of 0.152 and the test statistics show that every CAAR is insignificant. In Appendix 3, the linear regressions of the Cumulative Average Abnormal Returns of all the 9 banks from the sample for the three chosen time windows are show. Although these results show a better significance, even the total CAARs are still insignificant.

Conclusion

In 2012, the European Commission proposed to extend the powers of the ECB with relation to the Euro zone banking supervision and stated the essential pillars for the creation of a Banking Union. In November 2013, the ECB carried out the Comprehensive Assessment which took 12 months. The aim of the CA was to increase transparency on the condition of banks, to restore confidence of stakeholders by assuring that the banks subject to the CA were trustworthy and applying necessary corrective actions to the balance sheets of the banks. Next, on November 4th_,

2014, the Single Supervisory Mechanism becomes operational and so the ECB acquires the total responsibility for the supervision of the euro area banks which causes bigger financial integration between European banks. Finally, on January 1st_{, 2016, the Single Resolution Board becomes}

fully operational as central authority of the Single Resolution Mechanism.

In our empirical analysis we focus on financial market’s reactions to the Comprehensive Assessment, the Single Supervisory Mechanism and the Single Resolution Board. When we look at the Cumulative Average Abnormal Returns within the time windows that we have chosen, the test statistics show that every CAAR is insignificant. This indicates that, when looking at the separate banks and time windows, we cannot confirm our hypothesis (; the creation of the Banking Union has led to more transparency in the European banking market and so, produced new valuable information). When we focus on the total CAARs of all banks in the sample of the different time windows, the results show a better significance but still insignificant CAARs. This suggest that there are practically no immediate market effects, concluding that the creation of the Banking

(21)

21 Union did not produce new valuable information for the European banking market. Since the results do not confirm the initial thoughts and previous findings, this thesis does not have any policy implications as it is now.

Limitations

There are many reasons why this research gives this conclusion. Firstly, one of the reasons could be that we used a too small sample. Carboni et al. (2017) and Sahin and De Haan (2016) used much larger samples. Another reason could be that use of only or especially the largest banks in the EU-area is not right. It could be that smaller banks react much more and faster to the chosen events. A third reason could be that we used too small event windows since banks take more time to react to the events. A last argument could be that the European banking market reacted earlier than the chosen events, at the date of the proposal of the Banking Union as a whole in 2012.

Further research

Suggestions for further research are that we might better exclude the days of where data gives NULL value or data doesn’t exist instead of assuming that the stock price stayed the same. A next suggestion is that to use an estimation window for normal performance before more than ten days before the first event, so you can research longer time periods before the event. At last, varying with sample size, bank size, size of event windows and choice of event dates (as stated above under Limitations) could also be useful in further research.

(22)

22

References

Barbu, T. C., & Boitan, I. A. (2013). Implications of the single supervisory mechanism on ECB's functions and on credit institutions' activity. Theoretical & Applied Economics, 20(3).

Bernanke, B. S. (2013). Stress Testing Banks: What Have We Learned?: a speech at the" Maintaining Financial Stability: Holding a Tiger by the Tail" financial markets conference sponsored by the Federal Reserve Bank of Atlanta, Stone Mountain, Georgia, April 8, 2013 (No. 624). Board of Governors of the Federal Reserve System (US).

Black, L., Correa, R., Huang, X., & Zhou, H. (2016). The systemic risk of European banks during the financial and sovereign debt crises. Journal of Banking & Finance, 63, 107-125. Campbell, J. Y., Lo, A. W. C., & MacKinlay, A. C. (1997). The econometrics of financial markets. Princeton University press.

Carboni, M., Fiordelisi, F., Ricci, O., & Lopes, F. S. S. (2017). Surprised or not surprised? The investors’ reaction to the comprehensive assessment preceding the launch of the banking union. Journal of Banking & Finance, 74, 122-132.

Constâncio, V. (2012). Towards a European banking union. Lecture held at the start of the academic year of the Duisenberg School of Finance, Amsterdam, 7.

DeLong, G., & DeYoung, R. (2007). Learning by observing: Information spillovers in the execution and valuation of commercial bank M&As. The journal of Finance, 62(1), 181-216.

European Central Bank (ECB) (2014). Results of the 2014 comprehensive assessment. Retrieved January 14, 2018, at https://www.bankingsupervision.europa.eu/banking/tasks/

comprehensive_assessment/2014/html/index.en.html

European Central Bank (ECB). (2013). Note on the Comprehensive Assessment, October 2013. Retrieved January 11, 2018, at https://www.ecb.europa.eu/pub/pdf/other/

(23)

23 European Central Bank (ECB). (2014). ECB’s in-depth review shows banks need to take further action, 26 October, 2014 [press release]. Retrieved January 15, 2018, at http://www.ecb.europa.eu/press/pr/date/2014/html/pr141026.en.html

European Central Bank (ECB). (2017). List of supervised entities. Retrieved January 15, 2018, at https://www.bankingsupervision.europa.eu/ecb/pub/pdf/

ssm.list_of_supervised_entities_201712.en.pdf

European Central Bank (ECB). (n.d.). European System of Financial Supervision. Retrieved January 13, 2018, at https://www.bankingsupervision.europa.eu/

about/esfs/html/index.en.html

Fiordelisi, F., & Ricci, O. (2015). “Whatever it takes”: An Empirical Assessment of the Value of Policy Actions in Banking. Review of Finance, 20(6), 2321-2347.

House of Lords, European Union Committee, European Banking Union: Key issues and challenges, 7th_{Report of Session 2012–13, 12 December 2012}

Hirtle, B., & Lehnert, A. (2014). Supervisory stress tests. Federal Reserve Bank of New York Staff Reports 696/2014.

IMF (International Monetary Fund). (2013). A Banking Union for the Euro Area: Technical Background Notes, 13 February, Retrieved January 26, 2018, at http://www.imf.org/external/ pubs/ft/sdn/2013/ sdn1301technt.pdf

Ioannidou, V. (2012) ‘A first step towards a banking union’. In Beck, T., Banking Union for

Europe: Risks and Challenges, (London: CEPR) pp. 85-94.

Junevičius, A., & Puidokas, M. (2015). Institutional Change of the Supervisory System of the EU Financial Market. Procedia-Social and Behavioral Sciences, 213, 55-60.

French, K.R. (n.d.). Data Library. Retrieved January 15, 2018, at http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html

(24)

24 Lamfalussy, A. (2001). The Regulation of European Securities Markets. The Committee of Wise Men. Final Report, Brussels. Retrieved Januari 13, 2018 at http://ec.europa.eu/

finance/securities/docs/lamfalussy/wisemen/final-report-wise-men_en.pdf

Larosiere, J. (2009). The High-Level Group on Financial Supervision in the EU. Report, Brussel. Retrieved Januari 13, 2018 at http://www.esrb.europa.eu/shared/pdf/ de_larosiere_report_en.pdf?a855185299bf976712dc13c7644ccc8b

Levine, R. (2010). The governance of financial regulation: reform lessons from the recent crisis. BIS Working Papers, No 329, p.1-24.

MacKinlay, A. C. (1997). Event studies in economics and finance. Journal of economic literature, 35(1), 13-39.

Morgan, D. P., Peristiani, S., & Savino, V. (2010). The information value of the stress test and bank opacity. Federal Reserve Bank of New York Staff Reports, 460.

Onali, E., Schaeck, K., & Bruno, B. (2016). Market reaction to bank liquidity regulation. Journal of Financial and Quantitative Analysis.

Petrella, G., & Resti, A. (2013). Supervisors as information producers: do stress tests reduce bank opaqueness?. Journal of Banking & Finance, 37(12), 5406-5420.

Goldstein, I., & Leitner, Y. (2015). Stress tests and information disclosure. Federal Reserve Bank of Philadelphia, Research Department Working Paper no. 15-10.

Regulation (EU) No 806/2014 of the European Parliament and of the Council (2014).

Official Journal of the European Union. L 225 Volume 57. Retrieved January 14, 2018, at

http://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=OJ:L:2014:225:FULL%20&from=EN.

Regulations (EU) No 1092/2010, 1093/2010, 1094/2010, 1095/2010 of the European Parliament and of the Council (2010). Official Journal of the European Union, L 331/1.

(25)

25 Ricci, O. (2015). The impact of monetary policy announcements on the stock price of large European banks during the financial crisis. Journal of Banking & Finance, 52, 245-255.

Sahin, C., & De Haan, J. (2016) Market reactions to the ECB’s Comprehensive Assessment. Economics Letters, 140, 1-5.

Schoenmaker, D. (2012) ‘Banking union: Where we’re going wrong’. In Beck, T., Banking

Union for Europe: Risks and Challenges, (London: CEPR) pp. 95-102.

Seyad, S. (2013). The impact of the proposed Banking Union on the unity and integrity of the EU’s single market. Journal of International Banking Law and Regulation, 28(3), 99-108.

The Council Regulation. (2013). No 1024/2013 of 15 October 2013 conferring specific tasks on the European Central Bank concerning policies relating to the prudential supervision of credit institutions, OJEU, 29 October 2013, L 287/63.

The Turner Review. (2009). A Regulatory Response to the Global Banking Crisis. In the Turner Review. Financial Services Authority, London.

Single Resolution Board. (n.d.). Single Resolution Mechanism (SRM). Retrieved January 15, 2018, at https://srb.europa.eu/en/content/single-resolution-mechanism-srm

Stoxx. (n.d.). STOXX® Europe 600 Banks. Retrieved January 15, 2018, at https://www.stoxx.com/index-details?symbol=SX7P

Wymeersch, E. (2014). The Single Supervisory Mechanism or 'SSM', Part One of the Banking Union.

(26)

26

Appendix 1: Ordinary Least Squares (OLS) Regression; estimation of the parameters 𝜶

and 𝜷.

Commerzbank AG

Banco Bilbao Vizcaya Argentaria S.A.

Intesa Sanpaolo S.p.A.

_cons -.0341724 .0169367 -2.02 0.045 -.0675272 -.0008175 r_market_index .0685906 .0080489 8.52 0.000 .0527392 .0844421 r_cbk Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 23.6830171 254 .093240225 Root MSE = .26969 Adj R-squared = 0.2199 Residual 18.4012511 253 .072732218 R-squared = 0.2230 Model 5.28176595 1 5.28176595 Prob > F = 0.0000 F(1, 253) = 72.62 Source SS df MS Number of obs = 255

_cons .0025059 .0031682 0.79 0.430 -.0037336 .0087454 r_market_index .0367633 .0015057 24.42 0.000 .0337981 .0397286 r_bbva Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 2.1612411 254 .008508823 Root MSE = .05045 Adj R-squared = 0.7009 Residual .643915569 253 .002545121 R-squared = 0.7021 Model 1.51732553 1 1.51732553 Prob > F = 0.0000 F(1, 253) = 596.17 Source SS df MS Number of obs = 255

_cons -.0018337 .0010302 -1.78 0.076 -.0038627 .0001952 r_market_index .009565 .0004896 19.54 0.000 .0086007 .0105292 r_isp Coef. Std. Err. t P>|t| [95% Conf. Interval] Total .170799278 254 .000672438 Root MSE = .0164 Adj R-squared = 0.5998 Residual .068088231 253 .000269123 R-squared = 0.6014 Model .102711048 1 .102711048 Prob > F = 0.0000 F(1, 253) = 381.65 Source SS df MS Number of obs = 255

(27)

27 UniCredit S.p.A. ING Groep N.V. Deutsche Bank AG _cons .0093561 .0179469 0.52 0.603 -.0259883 .0447006 r_market_index .1738768 .008529 20.39 0.000 .1570798 .1906737 r_ucg Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 54.6035471 254 .214974595 Root MSE = .28578 Adj R-squared = 0.6201 Residual 20.6619298 253 .081667707 R-squared = 0.6216 Model 33.9416173 1 33.9416173 Prob > F = 0.0000 F(1, 253) = 415.61 Source SS df MS Number of obs = 255

_cons -.0011764 .0047637 -0.25 0.805 -.0105579 .0082051 r_market_index .0432093 .0022639 19.09 0.000 .0387508 .0476677 r_inga Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 3.55177087 254 .01398335 Root MSE = .07585 Adj R-squared = 0.5885 Residual 1.45571175 253 .005753801 R-squared = 0.5901 Model 2.09605912 1 2.09605912 Prob > F = 0.0000 F(1, 253) = 364.29 Source SS df MS Number of obs = 255

_cons -.020388 .0180585 -1.13 0.260 -.0559522 .0151761 r_market_index .1916177 .0085821 22.33 0.000 .1747163 .2085191 r_dkb Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 62.1407928 254 .244648791 Root MSE = .28755 Adj R-squared = 0.6620 Residual 20.9195992 253 .082686163 R-squared = 0.6634 Model 41.2211936 1 41.2211936 Prob > F = 0.0000 F(1, 253) = 498.53 Source SS df MS Number of obs = 255

(28)

28 Banco Santander S.A.

BNP Paribas S.A.

Crédit Agricole S.A.

_cons -.0021765 .0023456 -0.93 0.354 -.0067958 .0024429 r_market_index .0266912 .0011147 23.94 0.000 .0244959 .0288865 r_san Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 1.15274097 254 .00453835 Root MSE = .03735 Adj R-squared = 0.6926 Residual .352933098 253 .001394992 R-squared = 0.6938 Model .799807876 1 .799807876 Prob > F = 0.0000 F(1, 253) = 573.34 Source SS df MS Number of obs = 255

_cons .0111903 .0207916 0.54 0.591 -.0297562 .0521369 r_market_index .2686789 .0098809 27.19 0.000 .2492196 .2881383 r_bnp Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 108.774143 254 .428244659 Root MSE = .33107 Adj R-squared = 0.7441 Residual 27.7309288 253 .109608414 R-squared = 0.7451 Model 81.0432146 1 81.0432146 Prob > F = 0.0000 F(1, 253) = 739.39 Source SS df MS Number of obs = 255

_cons -.0000584 .0047268 -0.01 0.990 -.0093673 .0092504 r_market_index .0495826 .0022463 22.07 0.000 .0451587 .0540065 r_asa Coef. Std. Err. t P>|t| [95% Conf. Interval] Total 4.19324913 254 .016508855 Root MSE = .07527 Adj R-squared = 0.6568 Residual 1.43325847 253 .005665053 R-squared = 0.6582 Model 2.75999067 1 2.75999067 Prob > F = 0.0000 F(1, 253) = 487.20 Source SS df MS Number of obs = 255

(29)

29

Appendix 2: Cumulative Average Abnormal Returns and Test Statistic

This appendix reports the Statistic, grouped by bank windows (-7,+7), (-6,0) and (0,14). 46776. UCG 04/01/2016 45167. UCG 04/11/2014 43848. UCG 27/10/2014 42278. UCG 23/10/2013 41524. SAN 04/01/2016 39915. SAN 04/11/2014 38596. SAN 27/10/2014 37026. SAN 23/10/2013 36272. ISP 04/01/2016 34663. ISP 04/11/2014 33344. ISP 27/10/2014 31774. ISP 23/10/2013 31020. INGA 04/01/2016 29411. INGA 04/11/2014 28092. INGA 27/10/2014 26522. INGA 23/10/2013 25768. DBK 04/01/2016 24159. DBK 04/11/2014 22840. DBK 27/10/2014 21270. DBK 23/10/2013 20516. CBK 04/01/2016 -.0119948 18907. CBK 04/11/2014 17588. CBK 27/10/2014 16018. CBK 23/10/2013 15264. BNP 04/01/2016 13655. BNP 04/11/2014 -.0057785 12336. BNP 27/10/2014 10766. BNP 23/10/2013 10012. BBVA 04/01/2016 8403. BBVA 04/11/2014 7084. BBVA 27/10/2014 5514. BBVA 23/10/2013 4760. ACA 04/01/2016 .0194499 3151. ACA 04/11/2014 -.0617237 1832. ACA 27/10/2014 262. ACA 23/10/2013 company_id event_date total results of the Cumulative Average Abnormal Returns and Test and event date. The (1), (2) and (3) represent respectively the event -.0342249 -.0588277 -.0878204 -.0293312 -.0761759 -.0432058 -.1515083 -.1058916 -.0047625 -.1111019 -.1425928 -.0030763 -.0859815 -.0796327 -.1169268 -.0995267 -.2405695 -.0789815 .0215611 -.054774 .0235456 .0144952 -.0904238 .0128951 -.0007913 -.0070566 .0142158 -.005966 -.0732759 .0894547 -.0118092 -.0183963 -.0005595 -.0797759 -.1446491 -.0030918 -.0195441 -.0217848 -.0107248 -.1512686 -.4110299 -.0727921 .0064435 .0109479 -.0015381 .0337937 .1003561 -.0093815 -.007901 -.0220913 -.0073939 -.0580373 -.2366803 -.0502163 -.0038229 -.0120164 .0027753 -.0605204 -.2459778 .0547766 -.0058478 -.0050851 -.0040319 -.090163 -.1376067 -.0645848 .0043359 .0041639 .0018874 .0595383 .0830601 .0246045 -.0027342 -.003857 -.0141943 -.017537 -.0323624 -.060763 .0084308 .0200572 -.0049701 .0309006 .0981837 -.0198514 .0225374 .024092 .0088675 .0797549 .1308532 .0325715 .0319407 .0409395 .0227363 .1392324 .2426881 .0568187 .0581955 .0239244 -.0723779 .0921341 .0537398 -.0569629 .0056236 .0252287 .0513587 .009113 .0455291 .125645 -.0344964 -.0224045 .0130602 -.0520476 -.0700937 .0212252 .0095722 .0262362 -.0241955 .0116805 .0454915 -.0345099 .0024338 -.0276096 -.0710734 .0204993 -.1393027 .0100438 .0514095 -.0314346 .0134285 .0906578 -.053573 .0275723 .0500662 .0059268 .0457074 .1487497 .0078704 .0462578 .0505272 .1006044 .0848012 .1701407 .0957736 -.0681123 -.1486241 -.0928437 -.0625812 -.191551 -.065743 .0764316 .0346559 -.0062083 .1448949 .0359681 -.0028149 -.0051355 .0052639 -.003407 -.0071595 .0054408 .0187262 -.1133008 .0482774 .011702 -.2005908 .0312947 -.0012761 -.0156092 .0149081 -.0086413 -.1497126 .1261373 -.02047 -.050273 -.022401 -.0857264 -.2537956 -.0705261 -.0365726 -.0218737 -.0218099 -.1592574 -.1697679 -.0932581 -.0208807 -.0011242 -.0336711 -.0790553 -.0062307 -.1424256 .00535 -.011368 .0668312 .0246641 -.0241202 .000312 -.0674711 -.0928479 .0010926 -.1119242 .0154497 .0461667 -.064448 .0509159 .4376315 -.0980132 -.0068802 -.0073305 .014356 -.0225147 -.0556528 .0336159 caar1 caar2 caar3 ttest1 ttest2 ttest3

(30)

30

_cons -.0100031 .0071089 -1.41 0.168 -.024435 .0044287 caar3 Coef. Std. Err. t P>|t| [95% Conf. Interval] Robust Root MSE = .04265 R-squared = 0.0000 Prob > F = . F(0, 35) = 0.00 Linear regression Number of obs = 36 . regress caar3 if dif==0, robust

Appendix 3: Regression of the total Cumulative Average Abnormal Returns

This appendix reports the linear regressions of the Cumulative Average Abnormal Returns of all the 9 banks from the sample for the three chosen time windows. (1) represents the linear regression of the CAAR for time window (-7,+7), (2) represents the linear regression of the CAAR for time window (-6,0) and (3) represents the linear regression of the CAAR for time window (0,14).

(1)

(2)

(31)

31

Appendix 4 Stata commands (do-file)

// Prepare dates for event study generate date1 = date(date,"DMY"), after(date)

format %tdDD/NN/CCYY date1 drop date

rename date1 date save stockdata, clear

// Prepare event_dates for event study generate event_date1 =

date(event_date,"DMY")

format %tdDD/NN/CCYY event_date1 drop event_date

rename event_date1 event_date save eventdates

// Combining event and stock data sort company_id event_date

by company_id: gen eventcount=_N by company_id: keep if _n==1 sort company_id

keep company_id eventcount save eventcount

use stockdata, clear sort company_id

merge company_id using eventcount tab _merge

keep if _merge==3 drop _merge

// Duplicate stockdata for different events expand eventcount

drop eventcount sort company_id date

by company_id date: gen set=_n sort company_id set

save stockdata2

// Create matching set variable eventdata use eventdates, clear

sort company_id event_date by company_id: gen set=_n sort company_id set

save eventdates2

// Merge data and generate company_id-event_date identifier

use stockdata2, clear

merge company_id set using eventdates2 tab _merge

keep if _merge==3 drop _merge

egen group_id = group(company_id set) save totaldata

// Label event days sort group_id date

by group_id: gen datenum=_n by group_id: gen target=datenum if date==event_date

egen td=min(target), by(group_id) drop target

gen dif=datenum-td

(32)

32 by group_id: gen event_window1=1 if

dif>=-7 & dif<=dif>=-7

by group_id: gen event_window2=1 if dif>=-6 & dif<=0

by group_id: gen event_window3=1 if dif>=0 & dif<=14

by group_id: gen estimation_window=1 if dif<=-7 & dif>=-261

egen count_event_obs1=count(event_window1), by(group_id) egen count_event_obs2= count(event_window2), by(group_id) egen count_event_obs3= count(event_window3), by(group_id) egen count_est_obs= count(estimation_window), by(group_id) replace event_window1=0 if event_window1==. replace event_window2=0 if event_window2==. replace event_window3=0 if event_window3==. replace estimation_window=0 if estimation_window==.

// Estimate Normal Performance gen predicted_return1=.

gen predicted_return2=. gen predicted_return3=. egen id=group(group_id)

forvalues i=1(1)36 {

l id group_id if id==`i' & dif==0 reg ret market_return if id==`i' & estimation_window==1

predict p if id==`i'

replace predicted_return1 = p if id==`i' & event_window1==1

drop p }

// Generate Abnormal Returns and Cumulative Abnormal Returns sort id date gen abnormal_return1=ret-predicted_return1 if event_window1==1 gen abnormal_return2=ret-predicted_return2 if event_window2==1 gen abnormal_return3=ret-predicted_return3 if event_window3==1

by id: egen cumulative_abnormal_return1 = sum(abnormal_return1)

// Calculate Standard deviation sort id date

by id: egen ar_sd1 = sd(abnormal_return1) by id: egen ar_sd2 = sd(abnormal_return2) by id: egen ar_sd3 = sd(abnormal_return3)

(33)

33 // Calculate Cumulative Average Abnormal

Returns and teststatistic by id: egen caar1 =

sum(abnormal_return1/count_event_obs1) by id: egen caar2 =

sum(abnormal_return2/count_event_obs2) by id: egen caar3 =

sum(abnormal_return3/count_event_obs3)

gen ttest1 =(1/sqrt(count_event_obs1)) * ( caar1 /ar_sd1)

// Generate output

list company_id event_date caar1 caar2 caar3 ttest1 ttest2 ttest3 if dif==0,

abbreviate(10) divider sepby(company_id) export excel company_id event_date caar1 caar2 caar3 ttest1 ttest2 ttest3 using

"stats.xls" if dif==0, firstrow(variables) replace

regress caar1 if dif==0, robust regress caar2 if dif==0, robust regress caar3 if dif==0, robust

// ttest Abnormal Returns and teststatistic CAR (not used)

sort group_id

by group_id: ttest abnormal_return1==0 by group_id: ttest abnormal_return2==0 by group_id: ttest abnormal_return3==0

gen test1 =(1/sqrt(15)) * ( cumulative_abnormal_return1 /ar_sd1) gen test2 =(1/sqrt(7)) * ( cumulative_abnormal_return2 /ar_sd2) gen test3 =(1/sqrt(15)) * ( cumulative_abnormal_return3 /ar_sd3) list company_id cumulative_abnormal_return1 cumulative_abnormal_return2

cumulative_abnormal_return3 test1 test2 test3 if dif==0, sepby(company_id)

The effect of the creation of the Banking Union on the European banking market : did the creation of the Banking Union produce new valuable information for the European banking market?