The performance of the Greek banking sector during the Greek sovereign debt crisis

Universiteit van Amsterdam

Amsterdam School of Economics

The performance of the Greek banking sector during

the Greek sovereign debt crisis.

Author:

Emmanouil Chatzopoulos (10100091)

Thesis coordinator:

Dr. John Lorié

Msc in Economics

May 2015

ii CONTENTS

1. INTRODUCTION……….……….…….……….1

2. LITERATURE REVIEW………...………...…….…….4

2.1 Background……….……….…...……..4

2.2 Methods of measuring banking performance and measurement indicators...4

2.2.1 Aspects of measuring banking performance………...…....4

2.2.2 The CAMEL model………..…………6

2.2.3 The Data Envelopment Analysis……….…………..7

2.2.4 The Financial Ratio Analysis………..…..7

2.3 Banking performance in financially turbulent times ……….………..8

2.4.1 Sovereign debt and implications on banking performance…………...…9

2.4.2 Banks during times of banking distress………..……….11

2.4.3 Banking crisis and banking distress………....12

2.4 Greece and the Greek banking sector………...13

3. METHODOLOGY……….………...16 3.1 Research questions……….16 3.2 Theoretical framework……….…………..17 3.3 Research design………...20 3.4 Contribution………..………..21 4. CASE STUDY………..………...…24

4.1 The Greek sovereign debt crisis and the channels of contagion towards the Greek banks………...24

4.1.1 Direct losses on the sovereign bond portfolios………...…25

4.1.2 Credit rating contagion from sovereign to bank level and collateral eligibility……….….………28

4.1.3 International spill overs………...………....32

4.1.4 Panic in the banking market due to political risk………33

4.1.5 Economic recession……….36

iii

4.2 Financial Ratio Analysis………40

4.2.1 Technical presentation of the measurements indicators………...40

4.2.2 Analysis of the indicators………...44

4.2.2.1 Capitalization………...….45

4.2.2.2 Profitability………...…50

4.2.2.3 Liquidity………...…….59

4.2.2.4 Asset Quality………64

4.3 The importance of the Greek banking sector in the Greek economy…………...68

5. CONCLUSION….………...…………73

5.1 Positive developments………..………..73

5.2 Negative developments………...………...……74

5.3 The overall performance of the Greek banking sector………...76

5.4 Methodological remarks………..………...77

6. SUMMARY…..………..……….79

7. ABBREVIATIONS LIST………..80

8. APPENDIX……….…………...81

1 1. INTRODUCTION

The global financial crisis that unravelled in 2007-2009, and the Greek sovereign debt crisis that stroke Greece in the end of 2009 had significant repercussions on the Greek banking sector. Since 2001, the year Greece joined the European Monetary Union (EMU), the Greek economy and banks were enjoying the trust of international capital markets and investors. The Greek GDP was steadily outpacing the average of the European Union (EU) for more than a decade, whereas the reported fiscal fundamentals appeared to be good enough in order not to generate any unrest. The Greek banks had almost zero exposure to “toxic” bonds, and were offering attractive deposit rates that were significantly higher than the ones in other EU-countries. The above factors gave an invigorating boost to the banks’ deposit bases, while allowing them to contemplate the challenges that were lying ahead with confidence.

By the end of 2009 the first signs of economic recovery and stabilization for the economies of the USA and North-Western Europe started coming to the surface, and the non-domestic deposits started leaving Greece. This occurred at a time when several incidents of chronic fiscal mismanagement from the side of the Greek Government were disclosed to the public. Simultaneously, several key macroeconomic indexes such as the public deficit, public debt and spread of the Greek ten-year bond yield against the German one were soaring, accentuating the trend of fleeing deposits towards safe havens.

In the beginning of 2010, Greece suddenly found itself with no access to the financial markets, and in order to cover its financing needs had no other choice but to ask for help from its counterparts in the EU and the International Monetary Fund (IMF). The following months were characterized by intense uncertainty and historically low levels of optimism for the future performance of the Greek economy. The scepticism over Greece’s capability to service its debt, and the surging uncertainty deriving from the weak fiscal fundamentals of the countries of Southern Europe and Ireland brought the country’s economy at the epicentre of Eurozone’s worries.

The major rating agencies started warning that the negative debt dynamics in combination with the unwillingness of the government to proceed with direct drastic reforms in order to avoid further fiscal deterioration would eventually lead to a downgrade of Greece’s country rating. All the above generated unrest in the international markets, and panic across the Greek economy. The Greek depositors afraid of a potential sovereign default that would

2

turn the domestic banks insolvent, and thus put their capitals at risk, started transferring their deposits abroad or withdrawing money-bills and storing them in coffers.

The evolving sovereign crisis was sinking the Greek economy deep into recession, and was hampering the financial condition of the Greek banks' clientele that consisted of households and enterprises operating mainly in Greece. But the hurdles the Greek banks had to confront with did not stop there. The continuous downgrades of the country’s credit rating deprived them of the right to post Greek bonds as collateral in order to access funds via the ECB and the Eurosystem. Moreover, there was a significant negative impact in their trading portfolios due to the decrease in the value of the Greek bonds, which came as a result of the sales realised by investors who were trying to minimize their exposure to Greek debt.

In such an adverse economic environment characterized by intense uncertainty and historically low levels of trust, the Greek banks had to overcome a plethora of unprecedented challenges. This is exactly the point where the motivation for this thesis emerges. We will try to assess the performance of the Greek banking sector during the Greek sovereign debt crisis, while taking into consideration the financial developments generated in the aftermath of the global financial crisis. Our main research question, therefore, can be narrowed down to: “What was the performance of the Greek banking sector during the period of the Greek sovereign crisis?" We deem that the lack of precedence of such adverse economic and financial developments in the Greek economy makes the investigation of the performance of the Greek banks during the specific period more intriguing than ever before.

In order to assess the performance of the Greek banking sector as thoroughly as possible, we have chosen to proceed by examining three supportive questions. All three of them are shedding light to critical aspects of banking performance, and directly contribute to answering our main research question.

In the first supportive question, we make a detailed presentation of the channels through which the sovereign debt crisis was transmitted to the Greek banking sector. This analysis highlights the dimensions of banking performance that were put under stress, and allows us to elaborate on why we have chosen to examine them further in the ratio analysis that follows in the next question. Among other topics, we investigate the impact on the volume of the deposits, the quality of the bond and loan portfolio, the cost and availability of funding, and other critical banking volumes.

In the second supportive question, we assess the performance of the Greek banks from a quantitative perspective. Through conducting a financial ratio analysis that is based on six financial indicators -our performance measurement variables- we investigate the dimensions

3

of banking performance that were discussed in the previous question. We present the trajectory of the indicators over a seven-year period, and discuss the factors that caused the most significant changes.

In the third supportive question, we are examining whether the Greek banking sector managed to retain its position as the prevalent source of funding in the Greek economy. We do so by presenting the funding originated by banks, and the funding flows from other competitive sources towards households and enterprises, and show how the market shares changed over the years.

The remainder of this thesis is organised as follows: In chapter 2, we present literature that is relevant to: a) Methods of measuring banking performance, and measurement indicators, b) banking performance during times of financial distress that derives from sovereign debt and other sources, c) the Greek banking sector. In chapter 3, we discuss further our research questions and methodological approach, and also elaborate on the contribution of our thesis. In chapter 4, we answer the three supportive questions, whereas in Conclusion, the fifth chapter, we combine the incremental answers of the three questions and address our main research question. In the sixth charter, we present a summary of our thesis.

4 2. Literature Review

2.1 Background

The analysis that follows in this chapter intends to shed light on literature that discusses theoretical topics, which are relevant to the conduct of our research. In the first section, we present papers that discuss broadly the concept of banking performance, and the usage of measurement indicators in various measurement methods. In the second section, we turn our attention to literature that examines the implications of sovereign debt stress on different dimensions of banking performance, and papers discussing how banking crises affect the performance of banks. In the last section, we present literature which describes key characteristics of the Greek banking sector.

2.2 Methods of measuring banking performance and measurement indicators

In the first subparagraph of this part we present literature that helped as to structure our approach and highlight how significant some of the measurement indicators that we have included in our ratio analysis are. In the following subparagraphs we present papers that examine different models and methods of measuring banking performance, among which is the one that we have chosen to use in our analysis. In the Case Study chapter before conducting the ratio analysis, we make a brief technical presentation of the measurement indicators of our ratio analysis.

2.2.1 Aspects of measuring banking performance

Ewans et al. (2000) suggest that there is no general agreement on a model or set of indicators when examining the soundness of a financial system, and find that the usage of a complex indicator might be simplistic and potentially misleading. In order for an assessment to be constructive several indicators that take into account the characteristics of the economy and its financial sector must be considered. They support that an intertemporal comparison of indicators that focuses on the banking market of a single country might be more insightful than a cross-country comparison. This might be due to differences in the dynamics of the economies under examination, and to structural differences of their respective financial systems.

5

According to the ECB (2010), an effective financial analysis that assesses one of the most important dimensions of banking performance, profitability, should be considering more than one measurement indicators of the bank or group of banks under examination. The ECB report focuses on Return on Equity (RoE), the most commonly used indicator regarding profitability, and discusses its analytical power and weaknesses. The authors stress that high level of RoE might be reflecting a good level of profitability but also a more limited equity capital, and that the analytical capacity of the specific indicator is rather limited when profits are near to zero. Therefore, in order for an assessment to be more effective RoE needs to be accompanied by other indicators that measure profitability (e.g. the Return on Assets [RoA]).

Bolt and Tieman (2004) argue that higher Capital Adequacy Ratios (CARs) lead to lower failure probabilities for banks. This is a result of the buffer function of the extra capital that banks hold on top of the minimum limits prescribed by the regulating authorities, known as prudential capital. This extra funds might extend the expected life span of a bank as, for example, its balance sheet might absorb sudden losses on its loan portfolio (799).

The Dutch Central Bank (DNB), in a manual released in 2012, stresses the importance for banks to maintain sufficient capital and liquidity cushions in order to prevent mismatches between cash inflows and outflows that might occur as a consequence of expected and unexpected events. In the EBA1 report (2011), the Core Tier 1 (CT1) ratio is mentioned as the principal indicator that locates weaknesses and shortfalls in the capitalization of financial institutions.

Cohen (2013) documents the strategies that can be followed by banks in order to increase their risk-weighted capital ratios (e.g. CAR and CT1), and presents how banks increased their capital ratios during the financial crisis by retaining earnings and without restricting their lending. In the Case Study we will see whether the Greek banks applied this strategy for a number of reasons.

The literature presented in this subparagraph informs us that measuring banking performance can be a very challenging task. It is argued that taking into consideration several measurement indicators in the assessment of a dimension of banking performance is a choice that enhances the explanatory power of the results. For example, when RoA and RoE are included in the same set of indicators they deliver more accurate results regarding profitability. When the focus of the research is on a single country, then an intertemporal analysis of the

1 _{EBA is the European Banking Authority, the institution that was delegated by the ECB to assess the resilience}

6

selected indicators might be more suitable than a cross-country analysis. This might be due to structural differences between the countries under examination that do not allow for extracting meaningful conclusions through comparison. Other papers argue that extra capital can be instrumental in helping banks to withstand adverse shocks, and highlight two very commonly used indicators of measuring capital adequacy, CAR and CT1.

2.2.2 The CAMEL model

A significant body of literature that evaluates banking performance is making use of the CAMEL model, which is popular among both researchers and regulators. The CAMEL model is the acronym of six factors: a) Capital Adequacy, b) Asset Quality, c) Management Ability, d) Earnings, e) Liquidity (Greenbaun and Thackor, 1997, 456). It is a portfolio-based assessment method that includes a series of well-established indicators (Ewans, 2001, 16). Given that it was first introduced by regulating authorities, which are delegated with safeguarding the financial stability, the dimensions that it examines mainly focus on the resilience and soundness of banks. As a weakness of the CAMEL model, Greenbaun and Thackor see the tendency of the approach to be more informative during stable periods rather in times of financial turbulence.

There are several studies that are using the CAMEL model. For example, Gonzalez-Hermosillo et al. (1997) examine empirically the bank soundness of the Mexican banking system during the Mexican financial crisis of 1994. Their basic CAMEL-type model consists of several bank specific variables, and some macroeconomic ones. Among those variables we find the Non-Performing Loans (NPLs) for the examination of loan-portfolio quality, and RoA and RoE for measuring profitability. In a later study, González-Hermosillo (1999) shows for five banking system episodes by using empirical evidence that a CAMEL-type assessment is statistically significant only if NPLs and capital adequacy are simultaneously included in the examined variables of the model.

In another study, Hays et al. (2009) assess the performance of a big group of small banks in the USA for the period 2006-2008 by selecting indicators that are covering all the factors of the CAMELS2 approach, but not following it besides the selection of the proxies. For the earnings factor they have chosen the Return on Average Assets ratio. Another research that

2 _{The S is a sixth factor that was added in 1997 to the initial model and stands for Sensitivity to Market Risk.}

7

uses RoA in order to measure the profitability of a group of banks is the one of Boyd and Runkle (1993, 53).

2.2.3 The Data Envelopment Analysis

Another approach that combines several input variables in order to examine the performance of a bank or group of banks is the Data Envelopment Analysis (DEA). It is a non-statistical method that makes use of linear programming. Some papers that make use of the DEA are: a) Halkos and Salamouris (2004) who are measuring the performance of the Greek banking sector for the period 1997-1999 by looking at a group of 15-18 banks. As variables in their suggested model they are using six standard measurement indicators, among which are RoA and RoE. b) Siriopoulos and Tziogkidis (2010) who are also studying the Greek banking system after the occurrence of significant events that had an impact on banking performance. RoE and RoA are again in the group of variables that are being used in their model.

2.2.4 The Financial Ratio Analysis

A third approach is the financial ratio analysis, which encompasses several indicators that illustrate descriptively specific areas of banking performance. Under this approach, the researcher uses financial ratios in order to locate banks’ strengths and weaknesses, and to provide detailed information around profitability, credit quality and liquidity (Dietrich, 1996). Sinkey (2002) mentions as disadvantages of this approach: a) that it is based on the financial statements of the banks, which are backward-looking, and b) the danger of the data manipulation by the management of the banks. On the other hand, he mentions that the financial statements are the only source of banks’ achievements that researchers can use.

There are numerous studies that are examining banking performance through a financial ratio analysis. Below we present some of them: Kumbirai and Webb (2010) are measuring the performance of the South African banking system during the period 2005-2009 -two years before and after the global financial crisis. They pay attention to the five biggest banks, which account for more than 85% of the market, and base their analysis on seven ratios among which are RoA and RoE for profitability, the Net Loans to Deposit and Borrowing ratio for liquidity, whereas for describing the credit quality trend they are using the NPLs.

Strojwas (2010) examines the performance of the Polish banking sector through focusing on several indicators such as RoE, NPLs, the Loan-deposit ratio etc. Besides

8

discussing the indicators, Strojwas in order to make his examination more accurate he incorporates information regarding several key banking volumes such as: a) the liquidity conditions in the Polish banking market, b) the level of deposits, c) the funding costs of the banks, and d) the volume of credit they were providing towards the Polish households and corporate clients.

Mathuva (2009) investigates the performance of the Kenyan banking sector by focusing on how the capital adequacy and cost to income ratio relate to the profitability of the banks over a ten-year period. He is using RoE and RoA in order to measure profitability, and the CAR and CT1 for capital adequacy. Guy and Lowe (2011) study the NPLs during the global financial crisis in the banking system of Barbados by examining how a series of common bank variables and macroeconomic factors affected their trajectory. Their study is both at aggregate and at individual bank level.

Above we have presented literature that discusses three methods of measuring banking performance: i) the CAMEL method, which is portfolio-based and focuses on capital adequacy, asset quality, management ability, earnings and liquidity. The CAMEL method has been criticized as yielding limited information during times of financial distress. ii) The Data Envelopment Analysis, which is a non-statistical method that can handle numerous variables as input through making use of linear programming. iii) The Financial Ratio Analysis, a straight-forward method that through the examination of financial ratios assesses various dimensions of banking performance. As disadvantages of this method theorists mention the backward-looking nature of the input data -as the source is the financial statements of the banks under examination- and the danger of this data having been manipulated. The majority of the presented papers mention several financial ratios included in the sets of indicators with which they conduct their assessments on various dimensions of banking performance. Such ratios are the: RoA, RoE, CAR, CT1, Loans to Deposits and NPLs.

2.3 Banking performance in financially turbulent times

In this section we present papers that: a) discuss the implications of sovereign debt on banking performance, b) highlight the banking volumes, and factors of banking performance that are being tested during times of financial distress, and c) describe the terms banking crisis and banking distress.

9

2.3.1 Sovereign debt and implications on banking performance

The credit risk that derives from sovereign debt can have a severe impact on banking performance. Banks with exposures to debt issued by countries under financial distress might see an increase in their funding costs, and even find themselves excluded from the wholesale capital markets.

According to the Committee of Global Financial Stability3 _{(CGFS) (2011), there are}

four main channels through which sovereign risk can affect the availability of funding and the funding costs of banks: First, banks’ balance sheets are weakened through direct losses4 at their sovereign bond holdings. These damages increase the banks’ riskiness, and as a result the risk premiums they have to pay in order to access funding rise. Second, the steep decrease in the sovereign’s rating reduces the value of the collateral that banks can use in order to reach wholesale funding. Third, funding benefits that banks receive from government through implicit or explicit guarantees diminish. Fourth, the funding costs of banks when issuing corporate bonds increase due to the direct contagion of the credit rate downgrade from sovereign to bank level.

In addition to the four basic channels, the authors of the CGFS report mention several others that were found to be inconclusive. Some of them worth mentioning are: a) international spill overs, which are by-products of the close interconnection of banking systems. b) The increased risk aversion, which is triggered by fiscal mismanagement that makes depositors/investors to ask for higher risk premiums when depositing their savings or buying bank securities. c) The negative impact on banks’ revenues from fees and trading after an increase in sovereign risk; and d) the chance that an increase on sovereign financing needs will overshadow the demand for debt issued by commercial banks. We will revisit some of those channels of impact in the Case Study chapter, and see if they held an important role in the Greek banking market during the examined period.

Demirguc-Kunt et al. (2006, 712) support that: “lower asset prices may reduce the value of collateral, a phenomenon sometimes referred to as the collateral crunch”. An argument that is in agreement with the second channel of impact of the CGFS (2011, 17): “when the price of

3 _{The CGFS is a central bank forum that monitors issues regarding financial markets and systems, it is supported}

by the Bank of International Settlements (BIS).

4 _{The extent of losses is dependent on how banks are holding the bonds (i.e. at market value or at amortised}

10

a sovereign bond falls, the value of the collateral pool for institutions holding that asset automatically shrinks”. In the most extreme case: “a downgrade could even exclude a government’s bonds from the pool of collateral eligible for specific operations.” A more detailed presentation of how commercial banks use sovereign securities as collateral in order to access funding from the central bank they report to, can be found in CGFS (2001). In the same report, there is also an analysis of the trends in the usage of collateral in the wholesale financial markets.

Demirguc-Kunt and Huizinga (2010) argue that there is an important positive correlation between the worsening of a country’s public finances and the increase in CDS premiums for the banks operating in that country. An occurrence that eventually increases the funding costs of banks. Borensztein and Panizza (2008) examine the implications of a sovereign default, the most severe form of sovereign distress, on the domestic banking sector. Through regressions they estimate a high probability of a banking crisis occurring in the aftermath of a sovereign default. The underlying reason for this is the banks’ damaged balance sheets that could potentially trigger a depositor run. The direct balance sheet damages could either be the direct result of: a) a haircut on the sovereign debt related assets (i.e. bonds), b) the increased provisions the banks had to take in order to deal with the rising NPLs that come as a result of the economic downturn.

Borensztein and Panizza support that “recent default events suggest that domestic residents tend to account for a sizable portion of the (domestic debt) holdings” (2008, 17). A statement which implies that the bigger the proportion of the sovereign debt held by the domestic banking sector the bigger the size of the losses it will suffer in the occurrence of a haircut.

Correa et al. (2012) find that for the period between 1995 and 2001 downgrades of sovereign ratings had a crucial effect on the costs of banks’ equity financing for financial institutions located in both advanced and emerging economies. In the same spirit with the third suggested channel (state guarantees) of CGFS (2011), Panetta et al. (2009) examine the side effects of explicit and implicit state-guarantees provided to banks. They find that guarantees generate distortions that hamper international competition. This is because banks without solid fundamentals might be enjoying the benefits (i.e. cheap funding) of being backed by a “strong” sovereign; whereas other competitors, which might be well-capitalized banks with prudent management domiciled in countries with “weak” fiscal status, have to pay more in order to acquire access to funding.

11

According to the literature presented in this subparagraph there are four main channels through which sovereign stress is transmitted to the banking sector. First, through direct losses on the sovereign bond portfolio occurring after a debt-restructuring. Second, the credit rating downgrades of the sovereign are reflected in the titles it has issued. The banks that hold titles of a country facing financial troubles might not be able to post the titles as collateral in order to access funding through the interbank money market. Third, funding that banks were securing through government guarantees diminishes. Four, the downgrades at sovereign level are directly passed along to banks, which in order to secure funding in capital markets through issuing corporate bonds need to pay more due to increased risk premiums. Data also suggest that the chances of a banking crisis occurring after a sovereign default are very high as banks are hit hard from the increasing NPLs, and the diminished liquidity.

2.3.2 Banks during times of banking distress

The main source of banks’ funding is a key element in assessing their capability of withstanding adverse shocks. Banks which are dependent on short-term market funding and securitization activities seem to be stricken in a more severe way. Gambacorta and Ibanez (2011) observe that banks with higher proportions of more profitable –and more volatile- non-interest activities decrease their lending to borrowers at a greater extent than their peers with less non-interest activities.

Hutchinson (2002) examines the resilience of several EU banking sectors against systemic risk. Besides from four variables that comprise the canonical model of banking sector distress: real GDP, inflation, exchange rate turbulence and financial liberalization (368); he takes into consideration some extra ones in order to capture any institutional characteristics that might be influencing financial stability. Such factors are: supervision indicators, regulatory guidelines and the restrictions on banks’ portfolios set by monitoring authorities. The Greek banking system, with regards to its regulatory environment, was found second from the bottom, right above the Portuguese one (373).

Demirguc-Kunt et al. (2006, 715) find that banking crises do not necessarily generate steep declines in bank deposits on national level, as we would expect. The authors rather see depositor runs as “sideshows of banking crises at best” that have less dynamic than suggested by literature. This is probably happening due to the explicit guarantees on deposits. On bank-level though, depositors are “flying to safety” (p.713-4) by moving their deposits to better capitalized banks with more sound financial fundamentals.

12

Other theorists support that bank runs occur due to reductions in the value of the assets of a bank (Allen and Gale, 1998, 1247), and that insufficient or distorted information reaching depositors might have the same unpleasant results (Chari and Jagannathan, 1988). In Chapter 4, we will see how some of the factors mentioned above (i.e. insufficient or even distorted information reaching depositors, deposit-flee, decreasing asset value etc.) affected the performance of the Greek banks during the Greek sovereign crisis.

In this section we have seen that during adverse financial situations banks, which are not dependent on short-term funding and stay away from securitization activities, are coping better with difficulties than peers who do otherwise. It is argued that banking crises do not always result in big reductions on the level of deposits due to state guarantees, and that a transfer of deposits from banks with weak financials to ones with stronger financials is observed. Causes of bank runs might be a steep reduction in the value of a bank’s assets or even inaccurate information over a bank’s financial condition.

2.3.3 Banking crisis and banking distress

In the remainder of this thesis we will be using the terms banking crisis and banking distress often. Therefore, we present three descriptive definitions that we believe are capturing the essence of this economic phenomenon. Demirguc-Kunt et al. (2006, 703) define banking crisis as “a period in which significant segments of the banking sector become illiquid or insolvent”, whereas Hutchinson (2002, 367) sees it as “a banking problem posing a systemic risk to the entire financial sector”. According to Boyd et al. (2004, 489), “a banking crisis is a situation during which banks exhaust their reserve assets and might start liquidating storage assets”.

As banking distress, we see any intermediate condition between a banking crisis and a business as usual situation. During such periods, banks continue with their everyday operations, but are struggling with adverse conditions such as higher funding costs, increasing NPLs and diminishing demand for their services. These negative developments “affect isolated banks but are not systemic in nature” Laeven and Valencia (2008, 5).

13 2.4 Greece and the Greek banking sector

The majority of the existing literature concerning the Greek banking sector is focusing on the productivity and management structure of the Greek banks. Both these topics are not at the focal point of this thesis, and therefore will not be examined. Instead, in this section we focus on papers that: i) Discuss key characteristics of the Greek banking sector such as the structural developments that occurred during the two decades before the sovereign debt crisis; ii) present the Greek banks’ exposure to the Greek debt and other economies; iii) examine the condition of the Greek economy during the period of the debt crisis in order to better illustrate the adverse economic environment in which the Greek banks were operating during the examined period.

During the past two decades several key changes occurred in the Greek banking market among which were the interest rate liberalization, the relaxation and gradually the complete liberalization of capital movements, and the establishment of free entrance of foreign financial institutions (Noulas, 1999). All these reforms were realised in order for Greece to comply with the relevant directives and regulations of the EU so that it would be eligible for entering the EMU in 2001.

In the same vain, Hondroyiannis et al. (1999) provide an overview of the competitive conditions of the Greek banking system for the period 1993-1995. They find that the gradual abolishment of foreign exchange controls, the implementation of the Second Banking European Directive of the EU, and the liberalization of capital movements boosted the competitiveness of the Greek banks on international level. The gradual adoption of the above reforms made the transition to the new competitive European banking market smoother by providing the Greek banks with time in order to make use of the advantages (e.g. lower funding costs) that came along.

Kosmidou and Zoupounidis (2008) assess the performance of commercial and cooperative banks active in Greece for the period 2003-2004. They find that commercial banks: a) tend to increase their accounts and profits, b) improve their financial ratios, and c) augment their competitiveness. On the contrary, the results for cooperative banks varied as there were banks that improved their profitability and market share, whereas others failed to do so.

Siriopoulos and Tziogkidis (2010) find that Greek banks need a period of two to three years to regain the pre-shock levels of efficiency after the occurrence of significant economic events (endogenous and exogenous shocks). As endogenous shocks they see Mergers and Acquisitions (M&As), privatizations, organizational restructurings etc., while as exogenous

14

shocks they consider stock market crises, sovereign crises, economic recessions etc. Through the examined period of 1995-2003 there was an apparent tendency for the post shock efficiency levels to be higher than the old ones. The shocks that occurred in the Greek economy during the examined period were: a) a wave of M&As and privatizations (endogenous), and b) the Athens Stock Exchange crisis in 1999 (exogenous) (p.295).

In 2008, the Greek banks were controlling 13% of the Balkan market and were expanding further by establishing new branches in countries of the South Eastern European region (Kosmidou and Zoupounidis, 2008, 81). During that period, the banking authorities of these countries were implementing a series of modernization and liberalization changes (Mamatzakis et al., 2005, 195), creating a more welcoming and friendly environment for the Greek banks that were looking for new business.

Pascual and Ghezzi (2011) elaborate on the debt dynamics of Greece and support that a reduction in the stock of the debt was inevitable. Their main focus is on assessing the amount of the debt reduction that would restore solvency at sovereign level. According to their calculations the needed “haircut”, estimated to 50% of the total debt, would have catastrophic implications on banks’ viability, as it would eradicate their entire capital. The only solution for their survival then would be a hefty recapitalization of probably a similar amount.

Blundell-Wignall and Slovik (2010) observe that banks tend to be heavily exposed to sovereign debt of their own country, and that large cross-border exposures to Greece are present for Germany, France, Belgium, Cyprus and Portugal. The Greek banks’ exposure to the Greek debt was rising to 226% of their CT1, whereas the exposure to debt of other countries that faced similar troubles with their public finances was almost non-existent.

The papers presented in this section tell us that during the preparatory phase before Greece’s entrance into the EMU the Greek banks underwent a series of important structural reforms that boosted their competitiveness. The Greek banks then next to their activities in Greece started expanding in the SEE region. As underlying reasons that led Greece to its current dire economic condition authors see the chronic fiscal mismanagement, the lack of a long-term strategic plan, the idleness of the main European organisations, and the exposure of Greece to countries that were hit by the international financial crisis. The Greek banks need a relatively small amount of time to overcome disruptive shocks, with their performance levels after the manifestation of such events being higher than before. Regarding the extent of exposure to domestic debt, the Greek banks are no exception to the rule, and tend to be significantly overexposed. Finally, authors argue that a potential haircut of 50% at the nominal value of

15

Greek bonds would generate huge problems for the Greek banks as it would wipe out their entire capital base.

16 3. METHODOLOGY

3.1 Research questions

As we mentioned in the Introduction, our basic research question examines the performance of the Greek banking sector during the period of the Greek sovereign debt crisis. In order for our assessment to be as comprehensive as possible, we first address three supportive questions, and then, in Conclusion, we consolidate or conclusions, and address the main research question in a more concise manner. By answering the supportive questions we provide relevant information regarding the conditions in the Greek banking market during the examined period, and illustrate the implications of the sovereign debt crisis on the financial institutions. All the three supportive questions are shedding light onto different aspects of banking performance in Greece.

In the first supportive question, we examine how the Greek sovereign debt crisis affected the domestic banks. We describe the main channels of impact on both the assets and liabilities side of their balance sheets, and investigate the repercussions on their loan and bond portfolios. Among other issues we discuss the impact of: a) the restructuring of the Greek debt; b) the economic contraction in the Greek economy; c) the tumbling country rating; d) the relationship between the volume of total deposits and the fear of a sovereign default; and e) the funding originated from the Eurosystem and the European Liquidity Assistance (ELA). This analysis indicates the dimensions of banking performance that were put under strain during the sovereign debt crisis, and highlights the most vulnerable characteristics of the banks. All these aspects comprise the basis for the detailed analysis that follows next.

After the presentation of the channels of impact, we proceed with the second supportive question in which we are addressing the issue of measuring banking performance in a quantitative manner. The approach that we have opted for is the financial ratio analysis, a well-established method, which as we have presented in Literature review, provides straight-forward indications over the selected dimensions of banking performance. Our analysis examines the dimensions of banking performance that were put under stress during the Greek sovereign debt crisis by focusing on a set of six measurement indicators.

In the third supportive question we examine the importance of the banking sector in the Greek economy by focusing on the volumes of three different sources of financing. We look at funding that was originated by: a) the banking sector (i.e. loans), b) issuances of corporate

17

bonds, and c) the stock market. The data we present show the extent of the Greek banks’ exposure to the Greek economy, and illustrate whether the discussed competitive sources managed to substitute the decreasing funding that was originating from the Greek banking sector. At this point we also present the trajectory of the ASE General Index and Bank Index, and briefly discuss the market’s reaction to the occurring economic developments.

3.2 Theoretical framework

The idea behind examining the performance of the Greek banking sector on retaining its position as the prevalent source of funding was two generic remarks found in the papers of Gorton (2009) and Wensveen (2008). We have found no additional literature that examines the competition among different sources of funding, and more specifically in the Greek economy. Our research is therefore exploring a new area in that respect. After the presentation and analysis of how the market shares evolved over time, we briefly check whether the claim of Gambacorta and Ibanez (2011) applies to the case of Greece. Namely, we will see whether it took the Greek enterprises a considerable amount of time in order to access the corporate bond markets after the Greek sovereign crisis.

Our choice to examine the channels of impact of the sovereign debt crisis towards the Greek banking sector is based to a big extent on the analysis of the CGFS report (2011). The authors discuss in length the ramifications of the global financial crisis, which eventually reached the banks’ funding costs and the availability of funding. The focus of the report though, is much broader than the focus of our thesis as it examines several banking markets besides the Greek one (e.g. the Portuguese, the Irish).

Given the significant exposure of the Greek banks to the Greek government debt5_{, we}

examine how the debt restructuring affected their sovereign bond portfolios. The idea to examine further this exposure came from the conclusion of Blundell-Wignall and Slovik (2010) that banks tend to be positively predisposed towards debt issued by their own country6. The significant exposure of the Greek banks did not recede until the moment of the debt restructuring, when the Greek banks suffered great losses.

5 _{This is a finding of our examination of their assets.}

6 _{The research of Blundell-Wingnall and Slovik (2010) covers only two years (2010 and 2011) of the period that}

18

In the Literature Review chapter, we saw that banking performance can be examined in several different ways, and that there are several related strands of literature that discuss different methods of measuring it. The extensive study of Ewans et al. (2010) has held a very important role in shaping our methodological approach, and in selecting our measurement indicators.

Specifically, as Ewans et al. (p.30) suggest, in our financial ratio analysis we use aggregated indicators in order to describe the condition of the whole sector, and at the same time we do not skip presenting the ratios of the four participating banks at individual bank level7. Moreover, in order not to miss any underlying information we briefly discuss any important interrelationships between the banks in the Greek banking market and those of our sample. For example, some banks of our sample after the manifestation of the crisis acquired parts of other smaller banks that turned insolvent. These acquisitions caused important changes at certain individual ratios, and eventually influenced the aggregated ones.

The financial ratio analysis that we conduct could either be done intertemporally and focus on a single bank market in order to see how the indicators evolve over the years, or across several countries for a certain period. In this dilemma we have adopted Ewans et al. (2010) suggestion, and opted for an intertemporal comparison within a single country and its banking sector, Greece. We have done so because we consider the conditions in the Greek economy and banking market during the examined period unparalleled to the conditions of any other economy of a similar size and characteristics. Therefore, a comparison with the aggregate ratios of any other country or group of countries should be done carefully. When examining our six indicators, we will not skip referring to relative benchmarks. In that way our assessment also considers the prescribed indicator levels by monitoring authorities (e.g. for CAR and CT1), and the performance of other banking sectors (e.g for L/D and ROE).

As almost any method the financial ratio analysis comes with disadvantages. One of them is that it is based on financial indicators. The researcher in order to calculate every single financial indicator included in his set takes into consideration only a few volumes8 of every

7 _{The tables with the ratios at individual bank level are presented in the Appendix, Appendix table 10. For the very}

few cases that the values at bank level were significantly different from the ones at aggregate level we are making explanatory remarks in the respective sections of the ratio analysis.

8 _{For example, a financial ratio analysis that is based on two financial ratios (e.g. NPLs and RoE) is only taking}

into consideration the volumes that are in the nominator and denominator of the two ratios. Namely, for the calculation of NPLs the researcher needs the amount of Total Loans and the amount of loans that are in a 90 days breach since the date of the last payment. For RoE, the researcher would need the Net Profits and the Equity of the bank under examination. These four volumes do not adequately illustrate the performance of a financial

19

bank. A restriction that minimizes the explanatory potential of every indicator. On that regard, ECB (2010) mentions that when some indicators are accompanied by others, which are also describing the same dimension of banking performance, deliver better explanatory results (e.g. when discussing profitability RoE can provide much safer information when it is accompanied by RoA). In parallel with ECB (2010), Ewans et al. (2010) suggest to take into consideration as many indicators as possible when assessing the performance of a group of financial institutions. This way, more banking volumes are examined, and the dimensions of performance under examination are covered more efficiently. In our effort we consider several financial ratios and banking volumes in order to illustrate the characteristics, and performance of the Greek banking sector as accurately as possible.

As we will see in the Case Study, the conclusion of our analysis regarding the channels of impact of the Greek sovereign debt crisis to domestic banks is that the dimensions that suffered the most were capitalization, profitability, liquidity and portfolio quality. Therefore, these four dimensions were placed at the epicentre of our quantitative analysis. The measurement indicators with which we have chosen to examine them are: the Capital Adequacy Ratio (CAR) and the Core Tier I Ratio (CT1) for capitalization; the Return on Assets (RoA) and the Return on Equity (RoE) for profitability; the Non-Performing Loans (NPLs) for portfolio quality, and finally for liquidity the Loan-to-Deposit ratio.

The six selected indicators, as we saw in section 2.3.1 of the Literature Review, where we presented methods of measuring banking performance and indicators that are included in these models, were very commonly used. Specifically, in many cases that researchers were focusing on dimensions of banking performance similar to ours they were using one or more of the indicators that we have also selected for our analysis. We also saw that some of these indicators are mentioned in papers of monitoring authorities (i.e. the ECB and EBA) as indicators used to monitor the performance of a bank or group of banks. Moreover, these indicators seem to be among the ones used by banks’ management teams to monitor and communicate their performance to the public9. The usage of our six indicators by theorists, monitoring authorities, and the banks themselves is showing clearly the general agreement over

institution as a lot of crucial information regarding several dimensions of banking performance has obviously been omitted. This methodological weakness is effective only when the researcher uses a few measurement indicators.

9 _{This conclusion derives from the examination of the Annual Reports of the four banks that comprise our}

20

their analytical prowess, and their capability of helping the researcher to draw insightful conclusions.

In the Literature Review we have mentioned several papers that examine the performance of several banking sectors through using the financial ratio analysis (e.g. Mathuva, 2009; Kumbirai and Webb, 2010). Our thesis is closer to the approach of Strojwas (2010) who examines the performance of the Polish banking system in the aftermath of a severe crisis (i.e. the global financial crisis). His analysis is on aggregate level, and is based on a series of financial indexes such as RoE, NPLs, Loans to Deposits ratio, and other fundamental banking variables, such as the level of the aggregate deposits, the funding costs and the amount of credit towards the domestic economy. The presentation of all these data is made through graphs and charts.

3.3 Research design

Our research is supported by quantitative data, which are the foundation of our effort to assess the performance of the Greek banks, and to describe their instrumental role in the Greek economy. Our principal analysis, the analysis of the six performance indicators, is made for a seven-year period that starts in 2007 and stretches until 2013. The choice for the data presentation to start three years before and finish three years after the occurrence of the Greek debt crisis (2010) was made in order for our analysis to illustrate the shape of the banking sector at all stages of the crisis. Furthermore, this way we have the chance to show how the banks reacted when the adverse financial conditions started to recede and the economic sentiment to modestly improve.

Our approach is at aggregate bank level, and the analysis and presentation of the data is made through using tables and charts. The data that we use in our ratio-analysis, and draw conclusions from are the averages of the individual data of the four core banks in Greece, as these are reported in their annual reports and financial statements10. The four selected banks are the biggest banks in the Greek banking market in terms of assets, and are considered systemic by all the organisations that monitor the Greek economy and banking sector (e.g. ECB, EBA and IMF).

10 _{For RoE, and RoA, whenever the indicator was not provided in the Annual Report or the Financial Statements}

of the banks, we have proceeded by calculating the ratios using the Net Profits after taxes since we do not make comparisons with banking systems based in other countries that impose different tax rates.

21

In 2010, the EBA during its stress test assessment included the then six core banks that accounted for 90% of the Greek banking market. Three years later, the number of participating banks was reduced to four, as two of the previously participating banks were resolved, and parts of them were absorbed by other banks. The market share of the sample though, increased marginally to 91%. Kumbirai and Webb (2010) are using a similar approach when deciding for the sample of their research.

The selected banks are pretty homogeneous with respect to their offered services, risk appetite, and sources of income as is the vast majority of banks active in Greece. Finally, all the four selected banks have a domestically oriented strategy, and are therefore sensitive to the macroeconomic developments that hold a key role in our analysis. Given all the above and their significant market share in the domestic banking market, we deem that our sample illustrates satisfactorily the trends in the whole Greek banking system and highlights its strengths and weaknesses.

3.4 Contribution

We expect that the contribution of this thesis will be a comprehensive presentation and analysis of how the Greek banking sector performed during the period of the Greek sovereign debt crisis. To the best of our knowledge, our methodological approach and the analysis that takes place in the three supportive questions diversify our effort significantly from other studies. We consider our approach to be different mainly in two ways.

Firstly, the chronological period of our analysis is more extensive than analyses in other papers that we have reviewed. More specifically, in our analysis we incorporate the debt restructuring, which occurred between Q4 2011 and Q1 2012. The repercussions of this credit event damaged severely the balance sheets of the Greek banks and hindered their performance. Furthermore, we examine the financial implications during the years that followed the crisis (i.e. 2011, 2012 and 2013), and pay attention to the actions taken by the banks’ management teams to remedy the inflicted damages (e.g. decreasing lending, looking for other sources of funding, offering higher deposit rates, strengthening their capital buffers etc.)

Secondly, the method that we have chosen to measure banking performance, and the set of banking ratios that we use in it differ from the ones in other studies. All the papers presented in the Literature Review that examine the performance of the Greek banks are using the Data Envelopment Analysis. The ones using either the CAMEL method or the financial ratio analysis focus on banking sectors of other countries. On top of that, the set of the six

22

banking performance ratios that we use as measurement indicators11 is more extensive and differs from the ones in other studies.

We make use of a popular methodological approach (the financial ratio analysis), and apply it to the conditions and special characteristics of the Greek economic and financial environment. As we described in the Introduction, both the Greek economy and banking sector were going through unprecedented difficulties during the period that we examine. We therefore try to grasp the implications on the performance of the Greek banks through: a) Selecting a set of measurement indicators that examines the most heavily inflicted dimensions12 _{of banking}

performance -something which is not done by others. b) Examining the implications of the weak sovereign fiscal condition and the debt restructuring on the assets and liabilities of the Greek banks, and also on the funding costs of banks.

A supplementary factor of our analysis that diversifies it further is that next to the measurement indicators we present and discuss several additional banking volumes that illustrate the condition of the Greek banking market more accurately. We examine how these volumes fluctuated during the examined period in the supportive questions. Some of them are: a) the credit ratings of the Greek banks, b) the amount of central bank funding extended towards them through the Eurosystem and the ELA mechanism, c) the amount of total deposits in the Greek banking system etc. All these auxiliary variables are also included in the extensive table of banking indicators of Ewans et al. (2010, 4).

Regarding the part of our research in which we examine the channels of impact of the Greek debt crisis to the Greek banks, as we have noted, we consider closer to our approach the report of CGFS (2011). In several points though, our approach is distinctively different: a) we are focusing exclusively on the banking system of a single country (Greece) rather on a group of countries that faced difficulties similar to the ones of Greece. This choice allows us to delve more into depth. b) We are examining the underlying reasons of banks’ financial problems instead of focusing exclusively on the increased funding costs and the availability of liquidity. c) We enrich the analysis of CGFS regarding the contagion of banking distress by incorporating arguments from the existing literature. For example, the Demirguc-Kunt and Huizinga’s (2010) suggestion over the correlation of banks’ funding costs and the fiscal condition of a sovereign. d) We enrich the presentation of data from various sources such as the Greek Ministry of

11 _{In the CAMEL approach they are called factors of performance.} 12 _{These dimensions are examined in the second supportive question.}

23

Finance, Bank of Greece, Hellenic Statistical Authority etc. e) We consider the impact of the recession in the Greek economy on the loan-portfolios of the banks, a channel that it is not under the radar of the authors of the CGFS report. We do so by including a relative indicator in our ratio analysis.

Another contribution of this thesis is the detailed presentation of both the latest literature and available empirical data regarding the performance of the Greek banking sector. A key contribution when studying the developments that occurred during that period. Furthermore, in some instances we present what the prevalent theory suggests over certain aspects that might affect banking performance (e.g. the exposure of banks to domestic debt), and then check whether it indeed explains what occurred in the Greek banking sector.

24 4. CASE STUDY

In this chapter we are addressing the three supportive questions that have been set in Introduction and Methodology. The following analyses comprise the basis of the analysis that follows in the next chapter, where we consolidate the incremental conclusions of these three questions and conclude by answering our main research question.

4.1 The Greek sovereign debt crisis and the channels of contagion to the Greek banks

In our first supportive question, we analyse potential channels through which the Greek sovereign debt crisis might have affected the performance of the Greek banks. Our analysis emphasizes the dimensions of banking performance that were put under strain, and allows us to elaborate on why we have chosen to examine these dimensions further in the financial ratio analysis that follows in the next question.

As we have noted in Literature Review, this part of our analysis is influenced by the approach in the CGFS report (2011). Besides from the channels examined in that report, we present some additional ones that consider economic developments that were occurring in Greece at that time and had direct or indirect implications on the Greek banking sector. In this question our approach is mainly supported by the presentation of stylised facts and relative information that help the reader understand if and how certain dimensions of banking performance were affected by the debt crisis. These dimensions are examined further in the next supportive question with the use of financial ratios.

The channels of impact we examine are: i) the direct losses in the sovereign bond portfolios of the Greek banks; ii) the credit rating contagion from sovereign to bank level and its implications on the collateral the Greek banks could post to the ECB in order to access wholesale funding; iii) spill overs from exposures of the Greek banking sector to other economies, a channel which will be examined mainly by looking at the economic return of the Greek banks from their international operations ; iv) the uncertainty in the Greek political environment, and v) the recession that had stricken the Greek economy.

25

4.1.1 Direct losses on the sovereign bond portfolios

In this subparagraph we examine the direct implications of the restructuring of the Greek debt on the sovereign bond portfolios of the Greek banks. The Greek government in February 201213 _{announced a programme of debt exchange under which the owners of Greek}

bonds had the opportunity to exchange their old titles with new ones at a nominal face value decreased by 53.5%, and significantly extended maturities, BIS (2012, 67). The programme was called Private Sector Involvement (PSI). Even though the creditor-base of Greece was very diversified and heterogenic the participation arose to 85%. The main holders of the Greek debt at the time were banks and private investors from Greece and other countries (mainly from Europe), and Greek pension funds. The bonds that ECB had bought were exempted from the debt restructuring14.

What took place in the case of Greece’s debt restructuring is explained very accurately by Detragiache and Garella (1996). Namely, in cases of multiple creditors the debt restructuring can be succeeded through an exchange offer of bonds. In such an offer, bondholders are offered to trade their old claims for a new security, such as a bond with a lower face value and probably with a different tenor. The bondholders are free to exchange any fraction of their portfolio. The restructuring process might be cumbersome and does not happen without objections raised by creditors who remain unsatisfied with the exchange agreement that they are being offered. Bulow and Shoven (1978), and White (1980) argue that the larger the number of creditors the more difficult is the renegotiation of the debt. A valid point, which in the case of the Greek debt restructuring, was dealt with the involvement in the exchange procedure of very influential and highly respected organisations and institutions such as the ECB and the European Commission (EC).

There are several sources of data regarding the exposure of the Greek banking sector to the Greek debt. During our data-mining efforts we have witnessed several discrepancies among data presented in various papers, reports and databases that in some cases were substantial. In

13 _{For further details see the Timeline in the Appendix. In that table we present: i) financial milestones, and}

decisions of the Greek Government, ECB, IMF and EU that had direct or indirect impact on the Greek banking sector; ii) political events and decisions of the three major credit rating agencies concerning Greece.

14 _{As the economic situation of Greece was deteriorating the ECB temporarily waived its policy of not buying}

26

order to deal with this hindrance we have chosen to work with the most recently published data that we have managed to find15.

In Table 1 of Blundell-Wignall and Slovik (2010), we see the five countries with the largest exposures to the Greek debt. In the second column we see the total value of the bonds held by the banking sector of every country, and in the third column the exposure as a percentage of the Core Tier 1 capital of every banking sector. The Greek banking sector holds by far the largest portion (56.1 billion EUR) with the French and German ones following at distance. The aggregate exposure of the Greek banks expressed as a percentage of their Core Tier 1 capitals rises to an eye-catching 226%16. A percentage which indicates that the value of the Greek bonds held in the portfolios of the Greek banks was worth more than twice the value of their top tier capitals.

Table 1

Exposures to Greek debt (country-level) as per August 2010:

Amounts in (000.000s EUR)

Country Amount Exposure/CT1

Greece 56,148 226%

Germany 18,718 12%

France 11,624 6%

Cyprus 4,837 109%

Belgium 4,656 14%

Source: Blundel-Wignall and Slovik, p.8

Table 2 focuses on the then six largest banks in the Greek banking market. Similarly to the previous table, we see the value of the exposure of every bank to the Greek debt, and its Exposure/CT1 ratio in August 2010 and June 2011. What is eye-catching in this table is the enormous percentages of Exposure/CT1 on the third and fifth column. There is only one bank that its exposure is valued less than its CT1 capital worth, Alpha Bank at 86% in August 2010. Eurobank is second-best at 139% again in August 2010. As it can be seen in the table, there are

15 _{When we had the chance to select data from various sources we were always opting for data presented in reports}

of organizations that were actively involved in the monitoring of the Greek banks and economy (i.e., the ECB, the Bank of Greece, the Hellenic Bank Association etc.) or from the websites of the banks under examination. This approach has been adopted for the entire thesis.

16 _{As we will explain in sections 4.2.1 and 4.2.2, Core Tier 1 capital is bank capital of the best quality. It has a}

27

no significant changes in the exposure of the banks during the period between August 2010 and June 2011.

Table 2

Exposures to Greek debt:

Amounts in (000.000s EUR)

August 2010 June 2011

Bank Amount Exposure/ CT1 Amount Exposure/ CT1

NBG 19,756 260% 19,400 218%

Agrotiki Bank 10,187 807% 10,000 n.a.

Piraeus Bank 8,306 244% 8,700 248%

Eurobank 7,458 139% 7,900 196%

TT Postbank 5,371 418% 5,371 590%

Alpha Bank 5,070 86% 4,600 110%

Source: Blundel-Wignall and Slovik (p.8), and UBS

In March 2012, when the debt restructuring occurred, the Greek banks were holding Greek bonds worth 50 billion EUR, an amount that was equivalent to 25% of the total 200 billion EUR held by private investors17_{. Their losses after the haircut rose to 37 billion EUR,}

and were reported in the financial results of 2011. The international organisations involved in the restructuring of the Greek debt (i.e. EC, IMF and ECB) were expecting that the Greek banks would need help in order to maintain their capitalizations at sustainable levels, and for this reason they established the Hellenic Financial Stability Fund (HFSF)18.

The magnitude of the losses triggered important developments in the Greek banking market. Specifically in May 2012, Agrotiki Bank and TT Postbank, the two banks with the biggest Exposure/CT1 ratios, turned insolvent and it was decided that they should get resolved. The exposure for Agrotiki Bank was circa eight times bigger than its CT1 capitals for August 2010, whereas for TT Postbank the exposure from circa four times its CT1 capital in August 2010 rose to six times in June 2011.

On the contrary, the remaining four banks (NBG, Piraeus Bank, Eurobank and Alpha Bank) received 18 billion EUR aid from the HFSF. The allocated amounts were expected to

17 _{The Greek banks’ holdings of domestic sovereign debt were exceeding the average of the EU-countries,}

indicating a strong bias towards the Greek debt (CGFS, 2011, 14).

18 _{The HFSF was funded by the EU and the IMF and its mandate was to support the capitalization, and enhance}

the soundness of the Greek banking sector. The available funds under its control were 50 billion EUR. In return for these funds the HFSF was acquiring preference stocks of the banks it was assisting. The HFSF was also supported by the European Financial Stability Fund (EFSF).

28

augment their shuttered capitalizations and help them stay solvent in order to continue channelling loans towards the real economy. After the PSI, the exposure of the Greek banks to Greek debt was ranging between 3% and 7% of their total assets19.

In this section, we have seen that the debt restructuring that was an offspring of the Greek sovereign debt crisis had a substantially negative impact on the asset side of the Greek banks. The enormous write-downs in the values of the Greek bonds generated very big losses that eventually damaged the capitalization and profitability of the Greek banks. For some of them the impact on their capitalization was so extensive that turned them insolvent, whereas the rest were forced to make use of the supporting funding mechanisms in order to continue with their operations.

4.1.2 Credit rating contagion from sovereign to bank level and collateral eligibility

The Greek banks after the admittance of Greece to the EMU were subject to the control and regulations of the ECB, which together with the national central banks of the other EMU countries comprise the Eurosystem. All commercial banks domiciled in EMU countries in order to borrow from the Eurosystem need to post assets as collateral to the ECB. The most commonly used category of assets is sovereign bonds due to their inherently low credit and liquidity risk. The ECB has in place a minimum credit rating threshold for assets to be considered eligible as collateral20. This collateral framework protects the ECB from incurring losses, and simultaneously acts as an incentive for member-countries to follow disciplined fiscal policies in order for their banking sectors to be able to use their sovereign bond portfolios in order to enhance their liquidity.

In the previous section, we saw that the Greek banks were heavily exposed to debt issued by the Greek Government. Therefore, they were highly dependent on the performance of the Greek economy, and the condition of Greece’s macroeconomic performance. This is because the owners of a bond receive the coupon-payments and the principal amount upon expiration from the issuer, and thus the credit rating of a sovereign bond essentially represents the credit rating of the country-issuer. In our case, of course, the owners of the bonds under

19 _{The data in this paragraph and the previous are from the website of the BoG.}

20 _{For a detailed presentation of the importance of collateral in payment and settlement procedures see the report}