Supranational Supervision and Bank Risk-Taking: A study on Significant Banks within the Eurozone

(1)

Supranational Supervision and Bank Risk-Taking:

A study on Significant Banks within the Eurozone

MSc International Financial Management Faculty of Economics and Business

James Smith S3250903

Supervisor: Dr. M.A. Martien Lamers

Abstract:

__________________________________________ ________________________________

The paper investigates the transfer of supervisory authority from the national to the supranational level on the risk-taking of significant banks in the Eurozone. Our results, however, provide no evidence that supranational supervision by the ECB has a negative effect on the risk-taking of significant banks within the Eurozone. Although, we do find that significant banks in countries with higher financial freedom, or higher government integrity, that fall under the jurisdiction of the SSM will actually see an increase in their financial stability.

_________

Keywords:

(2)

i

Acknowledgements

I would first like to begin by thanking my supervisor, Martien Lamers, for the unconditional support and expertise over the course of this semester. His constructive feedback, supervision, and professionalism have encouraged me to push myself and go the extra mile. I would also like to thank those lecturers of the IFM programme who have instilled in me valuable knowledge over the past year for which I am forever grateful.

(3)

ii

List of Tables

Table 1 Country List of Banks ... 12

Table 2 Summary of variables ... 16

Table 3 Descriptive Statistics ... 18

Table 4 Correlation Matrix ... 20

Table 5 Regression Outputs ... 24

(4)

1

1. Introduction

When the Euro Area Sovereign Debt Crisis of 2011 erupted across Europe, calling for interventionist activities of country-level governments and central banks for support of Eurozone banks, it questioned the role and effectiveness of regulation and supervision at the national level that was currently in place. In response, this gave rise to the formation of a Banking Union, and its two pillars, within the following year: The Single Supervisory Mechanism (SSM) and the Single Resolution Mechanism (SRM) (De Grauwe, 2016; Popov & Van Horen, 2015). The Single Supervisory Mechanism transfers supervision from the national level to that of the supranational level (to the ECB) and whose aims are threefold: “ensure the safety and soundness of the European banking system, increase financial integration and stability, and ensure consistent supervision” (ECB, 2017a). From this, since its official inception in 2014, the ECB has produced a list annually of banks that it considers to be either very significant (given specific criteria) or less significant within the Eurozone. For those banks considered significant, the ECB takes over supranational supervision, whilst those considered less significant remain to be supervised by their national authority (ECB, 2017a).

In this paper, we study the effects of the formation of the Banking Union, but more specifically that of the Single Supervisory Mechanism as a starting point for research into whether its assigned supranational supervision to that of the ECB is affecting the behaviour of banks within the Eurozone. This can be further looked at from the point of bank risk-taking, and a distinction made between banks that are considered significant, and then less significant according to the Single Supervisory Mechanism criteria. The reason for the choice of bank risk-taking, is to better reflect the aforementioned goals of the SSM which are to increase bank stability in the near and long-term. The distinction between significant and less significant banks is to question whether supranational supervision is more appropriate and indeed more effective than national supervision of banks in affecting bank risk-taking. As a result, emphasis is placed on whether supranational supervision is having more of an effect on bank risk-taking in the years after its implementation, than in the years prior to this where there was predominantly national supervision of both significant and less significant banks.

(5)

2

In light of this research gap within current literature, we have formulated the following research question in which to answer:

1) Does supranational supervision by the ECB affect the risk-taking of significant banks within the Eurozone?

By using the sample of banks considered very significant by the ECB, we can test whether a stronger relationship exists between bank-risk taking and supranational supervision after 2014 until 2015, than in the years prior to this without supranational supervision (2011-2013). We can also compare this relationship to less significant banks not currently supervised by the ECB within the Eurozone, and also to banks that would be considered significant too, according to SSM criteria, in the United States (US). According to De Haan et al. (2015, pg. 350), “the BU [Banking Union] banking system is comparable to that of the US in several ways.” This is emphasized by similar domestic focus of banks, significant banks being directly comparable, and a close to equal closed banking system (De Haan et al., 2015). This comparability provides an international dimension to our research and allows one to better assess the impact of supranational supervision.

From researching the effects of supranational supervision, one can see whether government resources are indeed being used in the most effective way to affect bank stability. This is important, as the costs to the public from the Eurozone Sovereign Debt Crisis were severe and widespread. Thereby, our study answers the question of whether significant banks should be more closely supervised at the supranational level, or left to be supervised by their national authority – freeing up resources to be used in more effective ways to improve bank stability.

Our results show, however, that there is no evidence for the relationship between supranational supervision by the ECB and its effect on the risk-taking of significant banks. Although, once we account for country-level differences, we do find that significant banks in countries with higher financial freedom, or higher government integrity, that fall under the jurisdiction of the SSM will actually see a reduction in their risk-taking.

(6)

3

2. Literature

In this section, we shall highlight existing literature on the Basel II standards of bank regulation and supervision. From this we shall discuss how the second of the three pillars: supervisory activities, is the only one of the three pillars not empirically found to be as influential in Basel II compliance. This leads us to discuss how independent supranational supervision might solve this national supervision dilemma, and how the supervisory activities by the ECB thereby supports our hypotheses stance.

In terms of size, the European Banking System is considered one of the largest in the world – not only given the economy’s strong focus within the banking sector, but because many of its respective banks hold exceedingly large amounts of assets that surpass that of other individual banks in highly developed economies (García-Kuhnert et al., 2015). Thereby, it is not surprising to see many of these banks considered significant given their size relative to their respective country (Bertay et al., 2013), and as a result operating in one of the most highly regulated industries in the world (Pasiouras et al., 2009). This is in further line with what Gonzalez (2005) suggests, that although banks with greater flexibility, could indeed pursue operations that allows diversification of risk and in turn stabilizing the financial system, it could prove otherwise too, prompting banks to exhibit more risk-taking behaviour thereby compromising the stability of the overall banking system. For significant banks in the Eurozone, the latter coupled with other such factors, was one of the reasons for the Euro Area Sovereign Debt Crisis.

2.1. Regulation

(7)

4

In response, and also in the years prior to this, empirical research has been focused on the extent and applicability of overall Basel compliance, and also its individual pillars on several aspects of banks.

2.1.1 Basel II Compliance

According to Demirgüc-Kunt & Detragiache (2011), in a study spanning 3000 banks in 86 countries, they found no statistically significant relationship with compliance to the Basel Core Principles and that of bank risk, measured using the Z-score. They highlight this evidence as a cause of concern regarding the benefits of the Basel II pillars, and the expensive measures taken by the IMF and World Bank in conducting the Financial Sector Assessment Programme (FSAP). Furthermore, Ayadi et al. (2016), also carried out similar research and found no connection between that of banking efficiency and any of the Basel Core Principles. However, in contrast, Pasiouras et al. (2009) when looking at the impact of these Basel II standards on the profit and cost efficiency of banks, found that market discipline and that of the supervisory mechanism were influential. However, they found these effects of each to be rather interdependent amongst each other. Furthermore, it was also evident that more stringent capital requirements were associated with a negative influence on profit efficiency, but a rather positive effect on cost efficiency of banks (Pasiouras et al., 2009). Although these studies provide conflicting results on the benefits of overall Basel II compliance on bank-level aspects, the following section finds support for the first and/or third pillar.

(8)

5

Other empirical findings have considered the degree of constraints of regulations and its effect on bank charter value and risk-taking, and found that increased regulatory restrictions are connected with increased risk-taking, from decreasing their charter value (Gonzalez, 2005). Also, Bertay et al. (2013) concluded that given the potential cost of bank insolvency from systemically large banks, market discipline does exert a positive influence on these banks in turn.

2.1.2 Missing Link: Supervisory Activities

This mix of existing research provides conflicting results on compliance with the Basel II regulatory and supervisory standards. However, what appears is that there is a dominant trend that supports the first and third pillars, more so than the second, highlighting their influential impact, on bank stability, development, performance, and risk-taking worldwide. This lack of support for the second pillar: supervisory activities, highlights that national supervision might not be as effective nor influential at all in enforcing the legislation set out in the Basel II standards, thereby not having the desired effect on bank stability.

However, it may be possible that an independent supervisor and ‘official supervisory power’ could be a more appropriate complement as suggested by (Barth et al., 2013) whereby independence from both political parties and banks would lead to better bank efficiency. Thereby, national supervision not being undertaken by the Central Bank or government. Furthermore, Chortareas et al. (2012) highlights that increasing supervisory power can have a distinctive effect on bank efficiency – through the lower probability of financial default, reduced agency conflicts, and market power. Another study by Klomp & De Haan (2012) also found that the impact of supervision is higher for more high-risk banks, rather than less risky banks, in respect to capital and asset risk, and also liquidity and market risk. Given these insights, an independent supervisory power could complement the already universally applied Basel standards, and Basel III standards (full implementation by 2019) in the supervision of banks, especially banks considered significantly important.

2.2 From Regulation to Supervision

(9)

6

response is the Single Supervisory Mechanism (SSM), and Single Resolution Mechanism (SRM), although currently the addition of the European Insurance Deposit Scheme is being discussed (European Commission, 2015). Whilst the SRM aims at both resolving failing banks without severely affecting taxpayers and that of the economy, in a timely manner, and setting up a fund to help with this resolution, the SSM’s focus is the implementation of supranational supervision (assigned to the ECB) over that of the European Union – although oversight is predominantly over the Eurozone, unless Non-Eurozone member states actively decide to participate (ECB, 2017a).

However, in order to better re-establish trust in the banking system of the Eurozone, the ECB has focused its efforts on a select few banks deemed significant according to specific criteria. Only one of these criteria needs to be satisfied in order for a bank to be considered significant, such as:

“Size, the total value of its assets exceeds €30 billion; Economic importance, for the specific country or the EU economy as a whole; Cross-border activities, the total value of its assets exceeds €5 billion and the ratio of its cross-border assets/liabilities in more than one other participating Member State to its total assets/liabilities is above 20%; or Direct public financial assistance, it has requested or received funding from the European Stability Mechanism or the European Financial Stability Facility” (ECB, 2017b).

All other banks in the Eurozone that do not satisfy any of these criteria are considered ‘less significant’. Furthermore, banks deemed significant (125 banks, updated 1st_{January 2017) are}

(10)

7 2.2.1 Single Supervisory Authority

Doumpos et al. (2015) firstly highlights some arguments that could be made towards a single authority responsible for supervision, such as reduced information asymmetry in communication, more efficient use of human capital and infrastructural systems, better aligned goal congruence, easier collaboration with other regulatory authorities, independence in decision making, entirely liable in their actions, and easier supervision of financial entity groups that have cross-border activities – and that are accountable to multiple regulators. They also emphasize some drawbacks that might include, cultural disagreements, monopolistic opportunities with stricter measures, moral issues and the potential for agency conflicts. However, what Doumpos et al. (2015) found when looking at commercial banks in multiple countries between 2000 to 2011 was that not only does the central bank positively influence the soundness of financial institutions, but so does that of other supervisory authorities. Furthermore, they also find that the design of the supervisory system is influential given the size of the bank, such that that a unified supervisor is more appropriate in alleviating the effects of a crisis for large significant banks, whilst central bank independence is more suited to that of smaller banks. Others, highlight the benefit of an independent supervisor authority exhibiting appropriate judgement and competence in dealing with highly complicated situations, whilst also avoiding pressure from political or economic parties (Nieto, 2015; Masciandaro et al., 2009). Contrasting this, Gaganis & Pasiouras (2013), found the benefits to banks to not be as likely. They concluded that banks in countries which offer stronger consolidation of supervisory authorities have lower profits, whilst the independence of the central banks also showed the same relationship.

From these studies, we can see that there is some evidence for a supervisory authority that is both unified and independent, especially for those banks that can be considered larger in size. However, as aforementioned, in light of the introduction of the SSM, very little research has been carried out regarding supranational supervision progress in improving the Eurozone financial system; nonetheless, a mix of theoretical and empirical literature has arisen and tried to answer this question in the last few years.

2.2.2 Single Supranational Authority: The ECB

(11)

8

only ‘fundamental and day-to-day’ commercial decisions, but that of significant banks risk exposure and foreseeable future success. This is because the ECB is responsible for ensuring that the governance of the credit institution conform to EU legislation in such areas as careful management of the bank, management practices that address internal and external risk, and compensatory procedures, to name a few.

In contrast, Ferrarini (2015) questions the extent of the collaboration between the respective national authorities and the ECB, given that in a crisis situation the national authority might be more inclined to serve the interests of their own country, rather than consider the overall Eurozone system. Ongena et al. (2013) also found, although not connected to that of the SSM, that the stricter the supervision and regulation in a country, the more likely it is that banks increase their risk-taking abroad, via lower standards for lending. However, the ECB is responsible for supervision not only for the consolidated form of the significant bank, but that of its group entities that cross-over into member and non-member states (Binder, 2015; Tröger, 2013; Gortsos, 2015). As a result, this can to a certain extent limit cross-border risk-taking of significant banks, thereby reducing the likelihood for contagion to spread between the banks of the financial system. In respect to Ferrarini’s argument, the Comprehensive Assessment conducted at the end of 2013 and throughout 2014, whereby then significant banks had been identified and subjected to stress and asset quality tests, demonstrates the initiative and collaboration by the ECB and respective national authorities to work together (Gren et al., 2015). Furthermore, Carboni et al. (2017) even demonstrates investors’ reactions to the announcement of significant banks being directly supervised by the ECB, rather than by their respective national authorities. They concluded that investors upon hearing this news, penalized those banks (in respect to the stock price) – potentially in light of stricter supervision that would affect the risk behaviour and profitability of banks. From this empirical literature, it appears evident that the SSM has had an effect on significant banks, and reactions by investors believe this to be so too.

(12)

9

ECB (the world’s first supranational authority) on the risk-taking behaviour of Eurozone significant banks. Given this research gap within existing literature, alongside the SSM’s current demonstration of progress and investors’ reaction to this, we can derive our first hypothesis to test:

H (1): ECB Supervision has a negative effect on the risk-taking of significant banks, in light of less significant banks which are not.

(13)

10

3. Methodology

In this section, we shall firstly explain our data source from which we derive our samples for carrying out our research. Following this we shall specify our model which reflects our dependent, independent, and control variables being used, and how the model will be run in order to test our hypotheses. Lastly, we shall outline how our variables are measured and also their sources.

3.1 Data

In order to answer our research question ‘Does Supranational supervision by the ECB affect the risk-taking of significant banks within the Eurozone?’, it is worth considering the most appropriate data source. For this we used the Orbis Bank Focus database by Bureau van Dijk which compiles extensive up-to-date data on more than 42,000 financial institutions from all across the world, and for which is used by many researchers (Cubillas & Gonzalez, 2014; Delis & Kouretas, 2011). For our research, in order to extract the most meaningful bank-level data, a time span between 2011 until 2015 was selected to account for the change in supervision of significant banks. In this respect panel data was the most suitable, as it consists of both a cross-sectional and time dimension, and allows us to see this change, if present, on the risk-taking of significant banks.

Given that the SSM was introduced in 2014, it was important to locate the significant banks within Orbis Bank Focus that were classified as so on both the 2014 and also on the 2015 Annual List of Significant and Less Significant Banks (although the list has also been updated in 2016, and 2017). On this list, the consolidated form of the bank is highlighted, along with its group entities below it for each Eurozone country. As a result, when filtering banks within the Eurozone region through Orbis Bank Focus, we used two consolidation codes: C1, and C2, in order to find the consolidated form of the bank. As highlighted by Cubillas & Gonzalez (2014), by using the consolidated form, and omitting the group entities, you avoid the possibility of double counting – as such, not using data on the consolidated bank twice. Furthermore, the consolidated form of the bank was cross-checked with its respective financial statements to make sure it was indeed the consolidated bank that fulfilled the criteria for being significant as stated on the Annual Lists. This allowed us to form our first sample: Significant Banks in the Eurozone.

(14)

11

financial institutions was either commercial or cooperative banks within the Significant Bank Lists, we used this added criterion to compile a list banks that were either, and that were clearly stated as ‘Less Significant’ within the Lists. We used this approach in order to account for the similarity in deposit-taking between the two types of banks (Delis & Kouretas, 2011). Furthermore, the consolidated form of the bank (C1, and C2) also allowed us to use banks that were as close as possible in size to that of the significant banks, but that were just under the threshold of €30billion in total assets (one of the distinctions between significant and less significant banks). These factors allowed us to improve the comparability between our first sample: Significant Banks in the Eurozone, and our second sample (first control group): Less Significant Banks in the Eurozone.

In our third sample, which reflects our second control group, we used criteria from both our previous two samples to filter ‘significant banks in the United States’: C1 or C2 consolidated, Commercial or Cooperative, and Total Assets in excess of €30 billion. This allowed improved comparability with significant banks in the Eurozone whose similar size could exhibit comparable risk-taking behaviour, and for which the effect of the SSM introduction on Eurozone significant banks could be assessed.

(15)

12 Table 1 Country List of Banks

This table presents the three separate samples of banks used in our study: Significant Banks (Eurozone), Less Significant Banks (Eurozone), and Significant Banks (United States), the number of banks in each sample, and its corresponding country location.

3.2 Model Specification

In light of our dependent, independent, and control variables we have formulated our model as such. This model shall be tested twice when running our regression analyses. The first time, the dataset comprising significant banks in the Eurozone and less significant banks in the Eurozone will be used, whilst the second time our dataset consisting of significant banks in the Eurozone and significant banks in the United States will be used.

𝑍𝑆𝑐𝑜𝑟𝑒_𝑖,𝑡 = 𝛽0_𝑖,𝑡+ 𝛽1𝐵𝑎𝑛𝑘𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒_𝑖,𝑡+ 𝛽2𝑆𝑆𝑀𝐼𝑛𝑡𝑟𝑜_𝑖,𝑡 + 𝛽3𝐵𝑎𝑆𝑖𝑔. 𝑆𝑆𝑀𝐼𝑛𝑡𝑟𝑜_𝑖,𝑡 + 𝛽4𝐶𝑜𝑛𝑡𝑟𝑜𝑙_𝑖,𝑡+ 𝜀_𝑖,𝑡

Here the ZScore represents the dependent variable being tested. The intercept is defined as 𝛽0, whilst 𝛽1, 𝛽2, 𝛽3, and 𝛽4 reflect their respective variable coefficients. Lastly, 𝜀 describes the error term.

3.2.1 Data Analysis

As highlighted above within our model specification, we are using two combined datasets in testing our model to represent two different control groups: less significant banks in the Eurozone, and significant banks in the United States. Each of these control groups are individually combined with significant banks in the Eurozone, to form these datasets. This is important because if all of the data were pooled together during statistical analysis, it would

Significant Banks (Eurozone) Less Significant Banks (Eurozone)

Austria: 2 Austria: 4 Belgium: 2 Belgium: 1 Cyprus: 2 Finland: 3 Estonia: 2 France: 1 Finland: 1 Germany: 2 France: 4 Greece: 1 Germany: 4 Italy: 15 Greece: 4 Netherlands: 3 Ireland: 3 Total: 30 Italy: 11 Latvia: 1

Netherlands: 3 Significant Banks (United States)

Portugal: 2

Slovakia: 2 Total: 34

Slovenia: 1

Spain: 6

(16)

13

produce unreliable results. Therefore, in order to test our hypotheses, it is necessary to take this two-control group approach. The reason for this is due to the differences-in-differences method, using the OLS estimator, we are using within regression analysis in EViews. We are using two control groups – one for each of our models – in order to test the interaction effect for each of them. This allows us to estimate whether there are cause and effect relationships present (Bertrand et al., 2004). Furthermore, it is important to be aware of autocorrelation, when using the OLS estimator as the standard errors might not be accurate and thereby lead to inconsistent t-statistics and significance levels (Bertrand et al., 2004). As a result, White robust standard errors are applied to help alleviate this issue (Brooks, 2014).

3.2.2 Variables

All the variables that are important to conducting our research have been defined and further outlined below. Following this, Table 2 provides a summary of the variables, how they are measured, and from which database they are sourced from.

Dependent:

Z-Score: This score is used as a proxy for risk-taking behaviour of banks, in measuring the likelihood of bank insolvency. Many researchers use various proxies in order determine the Z-score, such as the “ratio of risk assets to total assets… and the ratio of non-performing loans to total loans….” (Delis & Kouretas, 2011, pg. 843), or the return-on-assets plus the capital-asset ratio divided by the standard deviation of the return-on-assets (Laeven & Levine, 2009). However, in this study we use the latter. The reason being, is highlighted by García-Kuhnert et al. (2015, pg. 611), who mention that this Z-score proxy “measures the bank’s risk-taking as riskier investment decisions increase the volatility of profits which in turn raises the probability of default.” A Z-score that has a higher value represents a more stable bank and therefore a reduced probability of failure, whilst a lower Z-score value highlights a higher chance of bank insolvency (Cubillas & Gonzalez, 2014). Furthermore, as data for the Z-Score can be very skewed we use the logarithm of the Z-Score in order to be normally distributed, alongside controlling for outliers (Laeven & Levine, 2009).

Independent:

(17)

14

dummy variable [1,0] whereby, an assigned value of 1 equates to a Eurozone bank that is considered significantly important, and a value of 0 equates to a less significant Eurozone bank. This dummy variable is also used again whereby the value of 1 stays the same, but the value of 0 equates to a significant bank in the United States.

SSM Introduction: This represents the year at which the Single Supervisory Mechanism was introduced, which is considered 2014 – although leading to up to this time, a comprehensive assessment of these significant banks was underway between the ECB and the national authorities. This SSM introduction, therefore, represents the change from national to supranational supervision of significant banks in the Eurozone. Overall, this variable highlights SSM’s increased supervision of these significant banks directly, from 2014 until 2015 (in the case of our study).

Bank Significance * SSM Introduction: In order to determine the extent of ECB supervision on the significant banks in the Eurozone, we use an interaction effect to determine whether this is statistically meaningful, using two control groups: less significant banks in the Eurozone, and significant banks in the United States.

Control Firm-Level:

Size: In order to control for factors that might influence the relationship between our dependent and independent variable at the bank-level, it is worth controlling for these bank characteristics as such. Size reflects the size of the bank given its total assets, so therefore we use the logarithm of total assets. This is the same as Doumpos et al. (2015) who use this in their model with the Z-score as their dependent variable.

Bank Capitalization: We use the capital-asset ratio such as equity capital divided by total assets in order to determine the significance of bank capitalization (Delis & Kouretas, 2011). We use this ratio as a control variable, as do García-Kuhnert et al. (2015), when testing their model using the same Z-score proxy as the dependent variable.

Control Country-Level:

GDP Growth: This reflects annual percentage growth in light of market prices. Gross Domestic Profit can be considered the total price of all goods and services in the respective country.

(18)

15

Investment Freedom: This reflects the ability to invest one’s own capital within their own respective country and across national borders, without being constrained. Factors that are used in order to determine an accurate score, are as such: “national treatment of foreign investment, foreign investment code, restrictions on land ownership, sectoral investment restrictions, foreign exchange controls, and capital controls” (Heritage, 2017).

Financial Freedom: We use this indicator of banking efficiency in order to account for the degree to which there is government intervention in the banking sector, and how much oversight they exhibit. This considers regulatory restrictions within financial services, whether banks are state-owned, the impact on the provision of credit, and the degree of development of the financial and capital markets (Heritage, 2017).

(19)

16 Table 2 Summary of variables

Variable Measurement Source

Dependent:

1) Z-Score Logarithm of (ROA+CAR)/SD(ROA) Orbis Bank Focus ROA: 2011, 2012,…

CAR: 2011, 2012,…

SD(ROA): Standard Deviation of the range of ROA, 2011 to 2015. Independent:

2a) Bank Significance Dummy Variable ECB Significant Bank

1 = Significant Bank (Eurozone) List (2014,2015)

0 = Less Significant Bank (Eurozone)

2b) Bank Significance Dummy Variable ECB Significant Bank

1 = Significant Bank (Eurozone) List (2014, 2015)

0 = Significant Bank (United States) Orbis Bank Focus

Region: US, Assets: $30B+

3) SSM Introduction Dummy Variable ECB

1 = Introduced (2014, 2015)

0 = Not Introduced (2011, 2012, 2013)

4a) Interaction Effect (Significant Bank (Eurozone)*(SSM Introduction 2014, 2015)

(Bank_Sig*SSM_Intro) (Significant Bank (Eurozone)*(No SSM Intro 2011, 2012, 2013)

(Less Significant Bank (Eurozone)*(SSM Introduction 2014, 2015) (Less Significant Bank (Eurozone)*(No SSM Intro 2011, 2012, 2013)

4b) Interaction Effect (Significant Bank (Eurozone)*(SSM Introduction 2014, 2015)

(Bank_Sig*SSM_Intro) (Significant Bank (Eurozone)*(No SSM Intro 2011, 2012, 2013)

(Significant Bank (United States)*(SSM Introduction 2014, 2015) (Significant Bank (United States)*(No SSM Intro 2011, 2012, 2013) Control Firm-Level:

5) Bank Size The logarithm of Total Assets Orbis Bank Focus

6) Bank Capitalization Equity Capital/ Total Assets Orbis Bank Focus Control Country-Level:

7) GDP Growth Annual Percentage Growth The World Bank

of GDP (World Development)

8) Business Freedom Scored between 0-100

Heritage Foundation

for each country Database

9) Investment Freedom Scored between 0-100 Heritage Foundation

10) Financial Freedom Scored between 0-100

Heritage Foundation

11) Government Integrity Scored between 0-100

Heritage Foundation

(20)

17

4. Results and Analysis

In this section, we shall firstly discuss the descriptive statistics of our datasets, and our correlation matrix of all the variables being used in our model. Following this we shall explain and discuss our regression outputs.

4.1 Descriptive Statistics

(21)

18 Table 3 Descriptive Statistics

Variable Mean Std. Dev. Median Min Max Observations

Sig. Banks (Eurozone) & Less Sig. (Eurozone)

Z-Score 2.89 1.35 2.98 -2.67 5.36 384 Bank Significance 0.62 0.49 1.00 0.00 1.00 384 SSM Introduction 0.41 0.49 0.00 0.00 1.00 384 Ba_Sig*SSM_Intro 0.26 0.44 0.00 0.00 1.00 384 Bank Size 10.29 2.21 10.31 4.50 14.59 384 Bank Capitalization 7.55 3.83 6.64 0.93 29.67 384 GDP Growth 0.53 3.02 0.58 -5.95 26.28 384 Business Freedom 80.39 6.63 77.40 67.00 95.00 384 Investment Freedom 80.49 8.63 80.00 55.00 90.00 384 Financial Freedom 66.43 8.64 70.00 50.00 80.00 384 Government Integrity 60.58 18.70 62.00 33.20 94.00 384

Sig. Banks (Eurozone) & Sig. Banks (US)

Z-Score 3.39 1.43 3.83 -2.26 5.72 407 Bank Significance 0.58 0.49 1.00 0.00 1.00 407 SSM Introduction 0.41 0.49 0.00 0.00 1.00 407 Ba_Sig*SSM_Intro 0.25 0.43 0.00 0.00 1.00 407 Bank Size 11.59 1.46 11.40 7.84 14.59 407 Bank Capitalization 9.13 4.23 8.17 0.93 21.75 407 GDP Growth 1.35 2.85 1.60 -5.95 26.28 407 Business Freedom 84.57 6.71 88.80 67.00 95.00 407 Investment Freedom 75.96 8.34 75.00 55.00 90.00 407 Financial Freedom 67.96 6.98 70.00 50.00 80.00 407 Government Integrity 65.74 14.03 71.00 33.20 94.00 407

(22)

19 4.2 Correlation Matrix

In Table 4 below, we show our correlation matrix whereby we pool together our three samples used in our research: significant banks in the Eurozone, less significant banks in the Eurozone, and significant banks in the United States. The reason why we pooled the data is that we do not find it appropriate to separate between the two different control groups’ cross-section and time-period correlations.

(23)

20 Table 4 Correlation Matrix

Bank SSM Ba_Sig* Bank Bank GDP Business Investment Financial Government

Z-Score Significance Introduction SSM_Intro Size Capitalization Growth Freedom Freedom Freedom Integrity

Z-Score 1 Bank Significance -0.38 1 SSM Introduction -0.01 0.02 1 Ba_Sig*SSM_Intro -0.22 0.54 0.56 1 Bank Size 0.15 0.35 0.03 0.18 1 Bank Capitalization 0.35 -0.43 0.03 -0.18 -0.15 1 GDP Growth 0.18 -0.06 0.28 0.21 0.11 0.22 1 Business Freedom 0.29 -0.34 -0.12 -0.25 0.26 0.20 0.22 1 Investment Freedom -0.08 0.20 0.14 0.20 -0.28 -0.17 0.04 -0.18 1 Financial Freedom 0.43 -0.12 -0.05 -0.14 0.20 0.12 0.22 0.47 0.29 1 Government Integrity 0.31 -0.17 -0.03 -0.12 0.21 0.09 0.31 0.72 0.16 0.77 1

(24)

21 4.3 Empirical Analysis

As outlined below in Table 5, we have performed our regression analyses on two different datasets, represented by either Z-Score (1) or Z-Score (2) which is our dependent variable proxy for bank risk-taking. Our first regression outputs from 1 to 4 is on our dataset of banks that are significant in the Eurozone and less significant in the Eurozone, whilst our second regression outputs from 5 to 8 are all on our second dataset of banks that are significant in the Eurozone and significant in the United States. Each of the regression analyses for each of the datasets shall be discussed, and emphasis on whether the hypotheses are supported.

(25)

22

regression analyses, we can see that the r-square gets larger with the inclusion of control variables, which contribute to the variance in our dependent variable: Z-score.

For our fifth regression analysis from our second dataset, we again tested our interaction term with the Z-score, which produced a negative value. This coefficient was not significant at any level, and for which does not support our second hypothesis with our control group: significant banks in the United States. Therefore, we do not find any support that supervision by the ECB has a negative effect on the risk-taking of significant banks in the Eurozone, for both our control groups. One variable though that was significant at the 1% level with a negative value, was that of Bank Significance which indicates that significant banks in the Eurozone have a lower Z-score than significant banks in the United States. For our sixth, seventh, and eighth regression analysis, we found our dummy variable Bank Significance to be significant at the 1% level, with a negative coefficient. We then found our bank-level control variable, Bank Size, to be significant at the 10% level in our sixth regression analysis. Regarding our seventh and eighth regression outputs, our country-level control variable, Financial Freedom, was significant at the 1% level with a positive coefficient, but Bank Size failed to remain significant in our regression analysis with all variables included. Overall, the coefficient from our interaction term did not change from being insignificant with the addition of control variables.

4.3.1 Discussion

(26)

23

However, in our third and fourth regression analyses of our first dataset, and also our seventh and eighth regression analyses of our second dataset, our interaction term takes on a positive value although still not significant at the 1%, 5% or 10% level. In all these regression outputs, country-level control variables have been included. One variable that stands out in each of these outputs is that of Financial Freedom which is significant at the 1% level with a positive value, although Business Freedom is significant at the 5% and 10% level respectively in our third and fourth regression analyses with a positive coefficient. Business Freedom refers to the environment in which regulatory and infrastructural aspects do not restrict the effect operation of business, whilst Financial Freedom “is an indicator of banking efficiency as well as a measure of independence from government control and interference in the financial sector” (Heritage Foundation, 2017). Although Cubillas & Gonzalez (2014), found financial liberalization to result in increased bank competition and therefore an increase in bank risk-taking behaviour, they did determine that for banks in developed countries with more strict capital requirements the opposite can hold true. As aforementioned, Basel II requirements are applied within each of the developed countries within our study, with one of their three pillars being adequate capital requirements. This explanation might highlight why our interaction term has a negative value until we control for country-level variables, whereby it then has a positive value.

(27)

24 Table 5Regression Outputs

Z-Score (1) Z-Score (2) 1 2 3 4 5 6 7 8 Constant 3.119*** 0.611 0.892 -0.6338 4.330*** 2.027* 0.212 -0269 (14.036) (0.581) (0.500) (-0.334) (31.589) (1.712) (0.074) (-0.094) Bank Significance -0.383 -0.914** -0.422 -0.709* -1.594*** -1.358*** -1.623*** -1.689*** (-1.317) (-2.207) (-1.588) (-1.875) (-6.857) (-5.443) (-4.445) (-4.589) SSM Introduction 0.098 0.033 -0.010 -0.049 -0.019 -0.055** -0.045 -0.068 (1.075) (0.405) (-0.088) (-0.421) (-1.560) (-2.034) -0.651 (-0.990) Ba_Sig*SSM_Intro -0.139 -0.140 0.020 0.009 -0.021 -0.012 0.104 0.095 (-1.183) (-1.217) (0.156) (0.079) (-0.289) (-0.142) (0.729) (0.681) Bank Size 0.211** 0.120 0.155* 0.071 (2.129) (1.391) (1.831) (1.028) Bank Capitalization 0.091*** 0.055 0.041 0.001 (2.681) (1.59) (1.201) (0.032) GDP Growth 0.002 0.001 0.000 0.004 (0.129) (0.061) (0.023) (0.191) Business Freedom -0.042** -0.036* -0.018 -0.021 (-2.075) (-1.922) -(0.592) (-0.708) Investment Freedom 0.005 0.008 0.004 0.008 (0.421) (0.565) (0.307) (0.563) Financial Freedom 0.082*** 0.072*** 0.081*** 0.077*** (3.958) (3.588) (3.835) (3.594) Government Integrity -0.004 -0.003 -0.004 -0.005 (-0.460) (-0.331) (-0.298) (-0.352) R-Squared 0.025 0.097 0.238 0.261 0.308 0.330 0.442 0.446 F-Stat 3.349** 8.206*** 14.684*** 13.215*** 59.882*** 39.633*** 39.409*** 31.929*** Observations 384 384 384 384 407 407 407 407

(28)

25 4.4 Further Empirical Analysis

In Table 6 below, we show the results from our regression analyses which incorporate country-level determinants of heterogeneity in our relationship between our interaction term (Bank Significance*SSM Introduction) and the Z-score. As a result, we have used each of our country-level control variables with our previous interaction term (Bank Significance*SSM Introduction) in order to test for a triple interaction effect.

From our regression analyses, we can see that our triple-interaction term (Bank Sig*SSM Intro*Inv Freedom) is statistically significant at the 5% level, with a positive coefficient in our second dataset. However, our triple-interaction term (Bank Sig*SSM Intro*Fin Freedom) is significant at the 5% and 1% level, with a positive value in both datasets, and our other triple-interaction term (Bank Sig*SSM Intro*Gov Integrity) is significant at the 10% and 1% level, with a positive coefficient in both our datasets too. What this means it that significant banks in countries with higher financial freedom, or higher government integrity, that fall under the jurisdiction of the SSM will actually see a respective increase in their Z-score – thereby be more stable. In comparison to our baseline regression outputs in Table 5, this might better explain why our respective coefficient values of our interaction terms (Bank Sig*SSM Intro) in both our datasets are actually insignificant, the effect (as evident in Table 6) only emerges in countries characterized by higher financial freedom or higher government integrity.

This appears to also be evident in a study looking at the heterogeneity from cross-country comparisons on the relationship between bank competition and bank stability, whereby Beck et al. (2013) found bank instability to be stronger in countries where activity restrictions were more stringent, and also in a study by Gonzalez (2005), whereby banks in countries with more strict regulations had lower charter values which did not provide less risk-taking incentives. Each of the factors in both of these studies would fall under factors that affect Financial Freedom, such as regulatory activities of financial services, and governmental control on capital allocation (Heritage, 2017). In other words, financial freedom played a moderating role in each of these cases. Therefore, this might explain why our triple-interaction term is positive.

(29)

26

(30)

27

Table 6 Country-level determinants of heterogeneity in regression results

Z-Score (1) Z-Score (2) 1 2 3 4 5 6 7 8 9 10 Constant 3.148*** 3.137*** 3.190*** 3.236*** 3.244*** 4.332*** 4.345*** 4.390*** 4.430*** 4.434*** (14.576) (14.091) (14.255) (14.414) (14.308) (31.251) (31.499) (31.420) (31.449) (30.518) Bank Significance -0.431 -0.413 -0.499* -0.574* -0.588* -1.597*** -1.621*** -1.699*** -1.767*** -1.774*** (-1.506) (-1.414) (-1.692) (-1.930) (-1.922) (-6.734) (-6.921) (-7.144) (-7.337) (-7.052) SSM Introduction 0.024 0.052 -0.078 -0.192 -0.212 -0.023 -0.057 -0.169** -0.267*** -0.277** (0.336 (0.520) (-0.660) (-1.486) (-1.331) (-0.398) (-1.406) (-2.454) (-3.258) (-2.285) Ba_Sig*SSM_Intro*GDP_Growth -0.009 -0.007 (-0.567) (-0.412) Ba_Sig*SSM_Intro*Bus_Freedom -0.000 0.000 (-0.551) (0.553) Ba_Sig*SSM_Intro*Inv_Freedom 0.001 0.002** (1.100) (2.545) Ba_Sig*SSM_Intro*Fin_Freedom 0.005** 0.006*** (2.316) (3.897) Ba_Sig*SSM_Intro*Gov_Integrity 0.006* 0.006*** (1.939) (2.654) R-Squared 0.025 0.025 0.025 0.028 0.030 0.308 0.308 .309 0.313 0.314 F-Stat 3.306** 3.286** 3.360** 3.759** 3.998*** 59.915*** 59.89*** 60.349*** 61.302*** 61.760*** Observations 384 384 384 384 384 407 407 407 407 407 Countries 16 16 16 16 16 17 17 17 17 17

(31)

28

5. Conclusion

In light of the most recent Euro Area Sovereign Debt Crisis, this paper investigates whether the transfer of supervisory power from the national to the supranational level actually impacts the risk-taking of significant banks. This is important because in the wake of the crisis, the response was to improve the financial stability of the Eurozone banking system. This responsibility was assigned to the ECB, thereby making it the world’s first supranational supervisory authority. As a result, little empirical research had been conducted on the extent of its supervisory influence on significant banks. In this paper, however, we do not find any evidence that supranational supervision by the ECB has a negative effect on the risk-taking of significant banks in the Eurozone. Our results are insignificant for both our control groups that we use: less significant banks in the Eurozone, and significant banks in the United States. However, after we ran further empirical analysis of our data and accounted for country-level differences, we did find that significant banks in countries with higher financial freedom, or higher government integrity, that fall under the jurisdiction of the SSM will actually see a respective increase in their Z-score.

From this, some notable contributions can be made. Firstly, our current research adds to the field of empirical literature on the effect of supervision on financial institutions. Given that a lot of research has focused on national regulatory and supervisory authorities such as central banks, little to no research exists on supranational supervisory authorities – especially from a non-theoretical standpoint. Furthermore, emphasis has been placed on the most significant banks that appear on the annual list of significant and less significant banks in the Eurozone. To the best of my knowledge, little to no research has ever focused on purely this sample of banks. Lastly, the time frame for this research has been very recent given that a lot of previous literature have focused on time frames for their data from the early 2000s until just after the Global Financial Crisis of 2007/2008.

5.1 Limitations

(32)

29

ECB is making in achieving their mandate of improving the safety and soundness of the Eurozone banking system. This may, to some degree, highlight why our results were insignificant in both our datasets. In this respect, it could be advisable to conduct similar research in the foreseeable future thereby again looking at the risk-taking of significant banks.

Another limitation is that data from the Orbis Bank Focus database for the years 2010 and most recent 2016 were very limited. The initial time frame for the study was intended to be from 2010 until 2016 (instead of 2011 to 2015) in order to better account for the implementation of the SSM, thereby adding more weight to our results. It was also intended to include all of the 19 Eurozone countries in our research; however, only 16 could be used given the lack of data for the remaining three. Furthermore, many banks in the Orbis Bank Focus database had missing values over many years for a lot of the financial data – especially so for the components of our bank risk-taking variable. This in turn reduced the number of banks in both our datasets, and as a result the number of observations we could make. Also, although many of the banks that are considered significant in the Eurozone fall under the bank specialization of either commercial or cooperative banks, which are similar in deposit-taking behaviour, some banks do not fall under any of these two types which might exhibit certain types of risk-taking behaviour. This as a result could potentially skew the results if the data from different specialized banks are pooled together. However, a small number of these banks are still included in our research in order to increase the number of banks in our datasets.

5.2 Implications

Although our baseline regression results are insignificant, a number of implications can be discussed from this. An important implication is that the use of public resources in supporting the ECB in its supranational supervision might not be necessary, or influential to any degree in improving the stability of the Eurozone banking system. As a result, supervision at the national level may be a more appropriate and alternative solution in the near and long-term for both significant and less significant banks. This is due to the fact that there is neither a significant negative nor positive relationship between supranational supervision and the risk-taking of significant banks. However, it may also be that the channels through which the ECB affects the stability of banks may not provide the most optimal nor desirable outcome (however as mentioned in the ‘limitations’ section, it may be too early to tell).

(33)

30

be relatively low, it may still be appropriate to implement policies that better facilitates banking efficiency. Further steps could be made towards reducing the interference in regulating financial services and influencing capital allocation. This would allow significant banks to better diversify their risk in the Eurozone.

An important implication is also that there may be some underlying agency conflict between the management of the significant banks and the ECB. Given that one of the ECB’s goals is to increase the stability of the Eurozone financial system, this remains to be seen as our results show no evidence of this. Thereby, this could highlight that the ECB and the management of the significant bank needs to work more closely together. From this better decision-making can be made that not only increases the banks’ stability, but provides a more sustainable business plan.

5.3 Future Research

A potential area for future research would be the channels through which supranational supervision by the ECB affects the risk-taking of significant banks. In this study, although we highlighted some ways in which the ECB exerts its supervision over significant banks, we take a more generalized approach to whether the ECB overall is having the desired effect on bank risk-taking. Therefore, a more specific focus such as how the ECB affects the governance of these credit institutions might serve to highlight the dynamics between the ECB and the management of banks and any immediate results that may follow from this. This would allow one to gauge whether each of the channels are aligned and indeed effective in supporting the ECB’s goal of a more stable Eurozone financial system.

Another area would be to consider further the different country-level characteristics that can influence the relationship between bank risk-taking and that of supranational or national supervision. In our study, although our baseline regression results were insignificant, we did find that significant banks in countries with higher financial freedom, or higher government integrity, that fall under the jurisdiction of the SSM will actually see a corresponding increase in their Z-score. As a result, it could be worth carrying out more comprehensive research into these cases, and also others, and expanding the time frame and countries accounted for in the study.

(34)

31

(35)

32

References

Ayadi, R., Naceur, S.B., Casu, B., Quinn, B., 2016. Does Basel compliance matter for bank performance? Journal of Financial Stability 23, 15-32.

Barth, J.R., Caprio, G., Levine, R., 2004. Bank regulation and supervision: What works best? Journal of Financial Intermediation 13, 205-248.

Barth, J.R., Lin, C., Ma, Y., Seade, J., Song, F.M., 2013. Do bank regulation, supervision and monitoring enhance or impede bank efficiency? Journal of Banking and Finance 37, 2879-2892.

Beck, T., Demirgüç-Kunt, A., Levine, R., 2006. Bank supervision and corruption in lending. Journal of Monetary Economics 53, 2131-2163.

Beck, T., De Jonghe, O., Schepens, G., 2013. Bank competition and stability: Cross-country heterogeneity. Journal of Financial Intermediation 22, 218-244.

Bertay, A.C., Demirgüç-Kunt, A., Huizinga, H., 2013. Do we need big banks? Evidence on performance, strategy and market discipline. Journal of Financial Intermediation 22, 532-558.

Bertrand, M., Duflo, E., Mullainathan, S., 2004. How much should we trust differences-in-differences estimates? The Quarterly Journal of Economics 119, 249-275.

Binder, J-H., 2015. The Banking Union and the governance of credit institutions: A legal perspective. European Business Organization Law Review 16, 467-490.

Brooks, C., 2014. Introductory Econometrics for Finance. Cambridge University Press, United Kingdom.

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 and Finance 74, 122-132.

Chortareas, G.E., Girardone, C., Ventouri, A., 2012. Bank supervision, regulation, and efficiency: Evidence from the European Union. Journal of Financial Stability 8, 292-302.

Cubillas, E., Gonzalez, F., 2014. Financial liberalization and bank risk-taking: International Evidence. Journal of Financial Stability 11, 32-48.

(36)

33

De Grauwe, P., 2016. Economics of Monetary Union. Oxford University Press, New York.

Delis, M.D., Kouretas, G.P., 2011. Interest Rates and bank risk-taking. Journal of Banking and Finance 35, 840-855.

Demirgüç-Kunt, A., Detragiache, E., 2011. Basel Core Principles and bank soundness: Does compliance matter? Journal of Financial Stability 7, 179-190.

Doumpos, M., Gaganis, C., Pasiouras, F., 2015. Central bank independence, financial

supervision structure and bank soundness: An empirical analysis around the crisis. Journal of Banking and Finance 61, S69-S83.

ECB, 2017a., https://www.bankingsupervision.europa.eu/about/thessm/html/index.en.html. [Accessed, 15-02-2017].

ECB, 2017b.,

https://www.bankingsupervision.europa.eu/banking/list/criteria/html/index.en.html. [Accessed 15-02-2017].

European Commission, 2017., https://ec.europa.eu/info/business-economy-euro/banking-and-finance/banking-union/european-deposit-insurance-scheme_en. [Accessed 15-02-2017]. Ferrarini, G., 2015. Single Supervision and the Governance of Banking Markets: Will the SSM Deliver the Expected Benefits? European Business Organization Law Review 16, 513-537.

Gaganis, D., Pasiouras, F., 2013. Financial supervision regimes and bank efficiency: International evidence. Journal of Banking and Finance 37, 5463-5475.

Garcia, G.G.H., 2009. Sovereignty versus soundness: Cross Border/interstate banking in the European Union and in the United States: Similarities, differences and policy issues.

Contemporary Economic Policy 27, 109-129.

García-Kuhnert, Y., Marchica, M-T., Mura, R., 2015. Shareholder diversification and bank risk-taking. Journal of Financial Intermediation 24, 602-635.

(37)

34

Gortsos, C.V., 2015. Competence Sharing Between the ECB and the National Competent Supervisory Authorities Within the Single Supervisory Mechanism (SSM). European Business Organization Law Review 16, 401-420.

Gren, J., Howarth, D., Quaglia, L., 2015. Supranational Banking Supervision in Europe: The Construction of a Credible Watchdog. Journal of Common Market Studies 53, 181-199.

Heritage, 2017. http://www.heritage.org/index/book/methodology#rule-of-law. [Accessed 15-02-2017].

Howarth, D., Quaglia, L., 2014. The steep road to European Banking Union: Constructing the Single Resolution Mechanism. Journal of Common Market Studies 52, 125-140.

Klomp, J., De Haan, J., 2012. Banking risk and regulation: Does one size fit all? Journal of Banking and Finance 36, 3197-3212.

Laeven, L., Levine, R., 2009. Bank Governance, regulation and risk-taking. Journal of Financial Economics 93, 259-279.

Lee, K., Lu, W., 2015. Do bank regulation and supervision matter? International evidence from the recent financial crisis. Journal of Economic Financial Policy 7, 275-288.

Masciandaro, D., Nieto, M., Quintyn, M., 2009. Financial supervision in the EU: is there convergence in the national architectures? Journal of Financial Regulation and Compliance 17, 86-95.

Nieto, M., 2015. Banking on Single Supervision in the Eurozone: Scepticism and a Reform Proposal. European Business Organization Law Review 16, 539-546.

Ongena, S., Popov, A., Udell, G.F., 2013. ‘‘When the cat’s away the mice will play’’: Does regulation at home affect bank risk-taking abroad? Journal of Financial Economics 108, 727-750.

Pasiouras, F., Gaganis, C., Zopounidis, C., 2006. The impact of bank regulations,

supervision, market structure, and bank characteristics on individual bank ratings: A cross-country analysis. Review of Quantitative Finance and Accounting 27, 403-438.

(38)

35

Popov, A., Van Horen, N., 2015. Exporting sovereign stress: evidence from syndicated bank lending during the euro area sovereign debt crisis. Review of Finance 19, 1825-1866.

Shehzad, C.T., De Haan, J., 2015. Supervisory powers and bank risk taking. Journal of Internal Financial Markets, Institutions and Money 39, 15-24.

Supranational Supervision and Bank Risk-Taking: A study on Significant Banks within the Eurozone