1
An event study:
The impact of EU sovereign debt crisis on the
banking sectors across EU countries
Amsterdam Business School, UvA
MSc Business Economics, Finance track
Thesis supervisor: Philippe Versijp
Students Name: Liang Tao
2
Contents
1. Introduction ... 3 2. Literature Review ... 5 3. Data ... 8 3.1 Data ... 8 3.2 Events ... 9 4. Methodology ... 124.1 Event study: the abnormal return and cumulative abnormal returns ... 12
4.2 The empirical model ... 15
5. Empirical Results and Discussion ... 18
5.1 Average Abnormal Return ... 18
5.1.1 The PIIGS countries ... 18
5.1.2 The EURO countries ... 20
5.1.3 The Non-EURO countries ... 20
5.2 The fixed-effect of Euro-zone News ... 22
5.2.1 The fixed-effect on every country ... 22
5.2.2 Robustness ... 27
6.0 Conclusion ... 29
Reference ... 31
Appendix 1 Event List ... 33
3
1.
Introduction
Recently the issue of sovereign debt crisis seems to fade away from the public media. However, it has been tough four years for the European Union (EU, henceforth) because of the sovereign debt crisis. During the crisis, the possibilities of the collapse of EU and EU banking systems have drawn considerable attention. This paper aims to exam the effect of sovereign debt crisis on the EU banking sectors in order to offer the investors and policymakers a better understanding of this crisis.
Beetsma, Giuliodori, de Jong and Widijanto (2013) believe that the EU sovereign debt crisis starts in the fall of 2009 when the Greek government announced that it cannot to repay or refinance its debt. According to the Maastricht Treaty, the budget deficit of EU countries should be less than 3% of their GDP. In the case of Greece, the Greek government announced that its budget deficit is almost 12.9% of its GDP. The turbulence in the Greece is only the beginning of EU sovereign debt crisis. Subsequently, Spain, Italy, Ireland and Portugal also announced that their governments’ deficit are not optimistic and may ask for help from European Central Bank (ECB). Even worse, this crisis went beyond these troubled countries and spread to other EU countries afterward. European banks usually hold government debts from other EU member countries, for instance, French banks hold around 10% of Greek government Debt. The debt of troubled countries may default at any moment. Hence the EU banks have to suffer the illiquidity. (Shambaugh, Reis and Rey, 2012). A great amount of doubt and concern about the banking sectors for EU has been presented on major public newspaper, such as Wall Street Journal. Some radical opinions proposed that the EU and EU banking system will collapse because of the EU sovereign debt crisis.
As time goes by, the EU and EU banking system still survive and those radical opinions do not yet come true. After four years of rumours and disturbances, it
4
becomes interesting to see what the impact of the EU sovereign debt crisis on EU banking sectors is. So this paper will investigate the effect of the sovereign debt crisis events on the banking sectors across the EU countries. During the whole crisis, a jumble of events of different types was recorded. Some of them are positive, and some of them are negative. To be more precisely, the objective of this paper seeks how negative news of the EU sovereign debt crisis affect the EU banking sectors. The event study methodology is used to estimate the impact of the EU sovereign debt crisis on EU banking sectors.
Here I formulate a hypothesis for this paper as following
The negative news of EU sovereign debt crisis will have significant adverse effect on the banking sector returns across European countries.
The reason I write this thesis is that I expect to gauge the impact of sovereign debt crisis on banking sectors across EU countries and to help policymakers and investors to obtain a deep insight of the impact of EU sovereign debt crisis. Additionally, I also expect this paper will assist the policymaker to cope the future crisis and help the investors to select an appropriate haven during the crisis.
The remainder of the thesis is arranged as following, section 2 will briefly introduce the related literatures. Section 3 and 4 provide the information of the data and methodology in this paper. Section 5 will offer the results of the event study and empirical model and the analysis. Last but not least, section 6 is the conclusion of this thesis.
5
2.
Literature Review
The EU sovereign debt crisis is a very new topic but has been discussed a lot in the past five years. The section 2 will focus on the literature review in several ways.
First, and foremost, this article relates to exiting literatures that study the relation between the sovereign debt crisis and the banking crisis. Reinhart and Rogoff (2008, 2009) found that the sovereign debt crisis often coincides with the banking crisis and their studies indicate that, on average, if debt-to-GDP of a country increases by 86%, a banking crisis will be imminent for that country. In addition, Reinhart, Rogoff and Savastano (2003) provided the evidence to suggest the sovereign debt crisis affect the cross-border capital flows.
Furthermore, the BIS (2011) reported the main four channels in which the sovereign debt crisis damages the banking sector: asset holding, collateral channel, rating channel, and guarantee channel. The first channel is asset holdings. Usually the balance sheet of banks are directly hurt because of the loss from holding of sovereign debt. The second channel is collateral channel. The sovereign debt risk usually will spread to its national banks and reduce the collateral value of banks. The third channel is related to rating channel. The downgrade of sovereign debt will affect the national banks negatively. Arezki (2011) revealed that the significant spillover effect of sovereign rating downgrades from different market and countries. Lastly, the guarantee channel is referred to the government guarantees. Government usually insure most domestic banks. If the fiscal position of the government is in danger, then the government guarantee on banks will decrease. Brown and Dince (2011) discovered the evidence that the government can support the financial institutions. Demirguc-Kunt and Huizinga (2011) provided the evidence that the market value of large banks decrease significantly when their government is running a large fiscal deficit. Based on the current literatures, the high level of government debt will hurt
6 the banking sector.
In the case of EU sovereign debt crisis, Blundell-Wignall (2012) believes the impact of sovereign debt crisis is not only isolated but also contagious. Meanwhile, they argue that the EU countries, such as France and Germany, with sufficient capital are likely to hold a large sovereign debt of other EU countries, such as Greece, Spain, and Italy. The sharp price fluctuation of government debt will affect collateral value significantly and cause a banking crisis in other EU countries. The aforementioned literatures build a theoretical foundation for this paper and offer historical evidence to support the whole idea of this paper. Based on the literatures, a hypothesis is formulated for this thesis in the introduction section.
Since there is a substantial literatures on the relation between the sovereign debt crisis and banking crisis, empirical evidence is indeed required to support the relation. In the empirical studies, the results are ambiguous. Pisani and Ferry (2012) found that if banks increase the holding of the troubled countries’ debt, these banks would suffer more banking stress than others. Chan-Lau, Liu and M.Schmittmann (2012) also found that equity returns of the banking sector are driven by the sovereign debt crisis. However, they also revealed that better capitalized and less leveraged banks outperformed other banks. On the other hand, Wolff (2011) recognized that the value of the banks that hold the debt of Italy, Spain Portugal and Ireland have not been changed as of July-October 2011. Hence, based these blurred empirical results, this paper intends to offer additional empirical studies to reveal the obscure relation between the EU sovereign debt crisis and the EU banking crisis.
Third, the literature review shift the attention to conduct whether the sovereign debt crisis events have an effect on the banking sectors. The closest works are Beetsma, Giuliodori, de Jong and Widijanto (2013) and Aizenman et al. (2012). Beetsma, Giuliodori, de Jong and Widijanto (2013) explore the impact of news on the bond market during the crisis. Unlike me, they use consistent news sources from
7
Eurointelligence; they also break all the events into different categories, which relates to technique of this paper. The reason this paper did not chooses Eurointelligence as unique news provider is that it would like to diversify the news resources and includes more relevant news in the data pool. Eurointelligence mainly concentrates the monetary and fiscal policy on a daily basis, hence it lacks the news about politics news. On the other hand, Aizenman et al.(2012) report the effect of the Euro crisis news on the bond market in emerging countries. The methodology employed in Aizenman et al. (2012) is an event study, and their methodology is consistent with this paper. Meanwhile the empirical model in this thesis is also inspired by Aizenman, et al.(2012)
Overall, the innovations of this thesis relative to the current literatures are the following: a large sample period from September 1, 2009 to June 30, 2013 is used, and the banking sectors returns are recorded on a daily basis. The event database contains the important economic, political and financial news, Etc. during the crisis. All the events are divided into 3 different categories according to characteristics of the events, namely, downgrades, politics-related and other news. Hence, this paper can also exam the asymmetric reaction between the different countries. The asymmetric reaction refers to the uneven performance regarding the same events. For example, in the section 5.2, this paper can estimate and compare the fixed-effect of downgrades announcement on the Greece and Italy in terms of direction and magnitude. Therefore, this paper will offer policymakers and investors a unique opportunity to understand the impact of the EU sovereign debt crisis on the banking sectors.
8
3. Data
3.1 Data
The data comes from several primary sources. The banking sector return indices for 11 EU countries, namely Belgium, France, Germany, Greece, Ireland, Italy, Portugal Sweden, Switzerland, Spain, The Netherlands, and United Kingdom are collected. In addition, the database contains the world banking sector return index as well.
All the countries are divided into three groups. The first group includes all the countries that undergo severe issues during the EU sovereign debt crisis and are in the danger of being forced to leave EU, such as Portugal, Italy, Ireland, Greece and Spain. These countries are called PIIGS countries. The second group is the Euro group. Basically speaking, the currency of these countries is euro, such as Belgium, France, Germany, and the Netherlands. Subsequently, the third group is non-Euro group and includes Sweden, Switzerland and United Kingdom.
Data on the banking sector returns are collected from DataStream. DataStream (Thomson Reuters, 2014) is world leading financial and macroeconomics database providing key economic indicators for 175 countries. The sample period ranges from September 1, 2009 to June 30, 2013. In order to assess the target countries’ response to negative events, all the banking sector returns are collected on a daily basis. In the Appendix 2, it reports all the banking sector returns across the target countries over the sample period.
In the Appendix 2, all the banking sector return indices are set equal to1 at the beginning of the sample period. By doing so, it is easy to observe changes of banking
9
sector returns over the sample period and compare the changes between each other. The Appendix 2 reports the movement of banking sector returns of target countries separately. Hence, we can easily find the turbulence of Portugal, Italy, Ireland, Greece, Spain and the thriving of Sweden. The observations from the Appendix 2 is consistent with the idea that is to divide the target countries into 3 groups and estimate the fixed-effect separately.
[Appendix 2 here]
3.2 Events
The events of the database come from various publishers, such as Bloomberg, Wall Street Journal, The Guardian and Eurointelligence. These publishers provide a number of major events for the EU sovereign debt crisis. The event database contains economic, financial and political information that might affect the EU countries. In the study of Asian crisis, Baig and Golfajn (1998) distinguished the ‘bad news’ from ‘good news’ and estimated how ‘bad news’ and ‘good news’ affect the financial market. Inspired by Baig and Golfajn, this paper will study how the negative events affect the banking sector returns. Thus, only the negative events are collected from these publishers.
A strong assumption is required in order to use the negative events to implement the event study. Firstly, all the negative events are independent. Secondly, all the negative events are not distorted by the positive events. In addition, the negative events are assumed to be significant and unanticipated. Last but not least, the negative events must attract the media interests as well (Aizenman, Etc. al.2012). The definition of negative events is in the line of Beetsma, Giuliodori, de Jong and Widijanto (2013). They believed that the negative events are expected to increase the domestic interest rate or tighten the government budget. In other words, the negative events concern the confidence of financial investors and the sustainability of the government debt.
10
Furthermore, the characteristics of the negative events are diverse; therefore, this paper splits all the negative events into 3 sets. The first set includes the events capturing credit rating downgrades (henceforth, downgrade). In the past study, Gande and Parsley (2005) indicated that one country’s sovereign credit rating downgrade had the negative influence on sovereign bonds of other countries. Arezki and Candelon (2011) believed that the sovereign credit rating downgrade will not merely affect the bond market. Kaminsky and Schmukler (2002) confirmed Arezki and Candelon (2011) and found the evidence that sovereign rating downgrade will not only affect the bond market but also the equity market. Therefore, it is necessary to place all downgrade events into one category and test whether the downgrade news has a significant effect on the banking sector returns. The set of downgrade contains all the events of downgrades, being put on watch list and giving a negative outlook over the sample period. The entire announcements are announced by rating agencies, such as Moody, S&P, and Fitch.
The second set includes events capturing political news. Mohl and Sondermann (2013) built their database based on the European politics news and found that the activities of politicians on the ‘restricting’ and ‘bailout’ had a significant impact on the bond yields for PIIGS countries. Hence, inspired by Mohl and Sondermann (2013), this paper constructs a set that includes the politics-related events and test how the political events affect the banking sector returns. The second set contains all the events of disagreements over economic policies between politicians, failure to pass the austerity plan, and Prime Minster resigns, Etc.
The third set includes the other events. Some of them are about social unsettlement, such as demonstrations. Some of them are about disappointing auctions. Therefore, it is still necessary to include these events and test whether the other events affect the banking sector returns.
11
After knowing what kinds of events are required in this thesis, all the events are collected manually. The first set is the downgrade news. I look for the keywords, such as ‘downgrade,’ ’Moody,’ ’S&P’ and ‘Fitch’ in all the news resources. The second set is the politics-related news. I look for the keywords such as ‘minster’ ‘Merkel’ ‘resign,’ Etc. in all the news resources. Meanwhile, the Guardian lists some politics-related news in their website, some of them are relevant to this thesis and are recorded in the Appendix 1. The third set is the other news. I look for the keywords, such as ‘protest,’ ‘demonstration,’ ‘strike,’ and ‘bailout’ to filter the news.
All the events are list in the Appendix 1. In the appendix, 28 downgrade news, 54 politics-related news and 33 other news are recorded.
12
4.
Methodology
4.1 Event study: the abnormal return and cumulative abnormal returns
The Event study method is a useful tool to estimate the financial effect of changes. Applying this method, I need to test whether there is a significant “abnormal returns” associated with an unexpected event. Usually the “abnormal returns” is considered to be an abnormal reaction to the arrival of new information. The reason this paper can implement event study is the market efficient hypothesis. The market efficiency declares that the stock prices and indices consist all the relevant information that is available to the market. Therefore, under such circumstance, any financially relevant news or events will be incorporated into prices of stock and indices quickly.
(McWilliams and Siegel, 1997).
Since the event date and sample period have been identified in the section 3, subsequently it is important to determine the expected return, abnormal return and cumulative abnormal return. Here the market model is employed to calculate the expected return. McWilliams and Siegel (1997) proposed the market model as following:
𝑹𝑹𝒊𝒊,𝒕𝒕= 𝒂𝒂𝒊𝒊+ 𝒃𝒃𝒊𝒊∗ 𝑹𝑹𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊,𝒕𝒕+ 𝒊𝒊𝒊𝒊,𝒕𝒕
Where
Ri,t :the return of banking sector of country I on day t, Rindex,t :the return of world banking sector on day t ei,t : the idiosyncratic error term.
In fact, there are many models can be used to calculate the expected return, such as CAPM, multi-factor models and market return model. Unfortunately there is no best model to compute the expected return. Here this paper selects the market model because the market model is simple and the most popular model in the practise. In the
13
market model, it assumes the constant and liner relationship between individual banking sector return and the return of world banking sector.
Subsequently, the estimation window and event window are required. The estimation window, referring to a short period before the event, is used to estimate the expected return. In this paper, the estimation window is a period from day t-35 to day t-5 (30 trading day observations). The event window refers to the period 2 days before and after the event date. A short event window is important because Brown and Warner (1980, 1985) found the long event window decrease the power of test statistic damaging the significance of an event. Moreover, the empirical studies also indicate that a short event window provides the significant influence of an event (Ringers and Netter, 1990).
After obtaining the parameters in the market model, we can calculate the expected return during the event window. The expected return is the return in the absence of the event.
𝑬𝑬𝒊𝒊𝑬𝑬𝒊𝒊𝑬𝑬𝒕𝒕𝒊𝒊𝒊𝒊𝒕𝒕 = 𝒂𝒂𝒊𝒊+ 𝒃𝒃𝒊𝒊∗ 𝑹𝑹𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊,𝒕𝒕
Where ai and bi are the ordinary least squares (OLS) coefficients obtained in the
previous market model. R index,i are world banking return on day t during the event
window. Expectedt is the expected return based on the market model.
The daily abnormal return for each event is the difference between expected return and actual return:
𝑨𝑨𝑹𝑹𝒊𝒊.𝒕𝒕= 𝑹𝑹𝒊𝒊,𝒕𝒕− 𝑬𝑬𝒊𝒊𝑬𝑬𝒊𝒊𝑬𝑬𝒕𝒕𝒊𝒊𝒊𝒊𝒕𝒕
𝑪𝑪𝑨𝑨𝑹𝑹𝒊𝒊= � 𝑨𝑨𝑹𝑹𝒊𝒊 𝒊𝒊 𝟏𝟏
Where the abnormal return (ARi,t) represents the adjusted return which is the
difference between the actual return and expected return. The cumulative abnormal return (CARi) is the sum of all the abnormal return for each country.
14
The test-of-significance approach is applied to test whether the abnormal returns for every country and every group are statistically different from zero. The T-test is defined as following:
T-test =
(𝑪𝑪𝑨𝑨𝑹𝑹𝒊𝒊⁄ )𝒊𝒊 (𝑨𝑨𝑹𝑹_𝒔𝒔𝒊𝒊 𝒔𝒔𝒔𝒔𝒔𝒔𝒕𝒕(𝒊𝒊)⁄ )Where AR_sd is the standard deviation of the abnormal return, n is the number of total events. Meanwhile, the hypothesis for this test is as following:
H0: The average abnormal return is undistinguished from zero
H1: The null hypothesis is not true.
Here the significance level is equal to 1%, 5% and 10% hence the critical value is -2.335, -1.645 and -1.285 respectively. If the absolute value of T-test is greater than 2.335, 1.645 and 1.285 then it means there is sufficient statistical evidence to reject the null hypothesis and the average abnormal return is statistically distinguished from zero at 1%, 5% and 10 % level.
The results of test-of-significance approach indicate whether all events in the database has a significant effect on the banking sector returns. The section 4.2 will discuss abnormal return in further and estimate the fixed-effect of different events.
15 4.2 The empirical model
The section 4.1 states the methodology of event study and offers the test-of-significance approach to test the significance of average abnormal return for every country and every group. The events database is filled with all kinds of events. The test-of-significance approach did not elaborate what kind of events contributes the largest abnormal return and what kind of events has the least influence on abnormal return. In addition, the test-of-significance approach cannot test whether politics-related, downgrade and others news have a significant impact on abnormal return. Hence, this section will concern the contribution of abnormal return and estimate the fixed-effect of different events.
In this sections, the regression model aims to analyse how the abnormal return is affected by the events. The events are not the well-defined quantitative variables but qualitative variables (also called dummy variables). The dummy variables can discover the absence or presence of a ‘quality’, therefore it is useful to establish artificial variables that only have the value of 0 and 1. 0 indicates the absence of a ‘quality’ and 1 represents the presence of a ‘quality’.
In fact, the dummy variables in the regression model are as same as the well-fined quantitative variables. This section aims to estimate and test the impact of politics-related, downgrade and other news respectively, hence analysis-of-variance models1 are employed. However, it is necessary to emphasis the dummy variable trap before the ANOVA models are constructed. The dummy variable trap is a situation in which the explanatory variables are perfect multicollinearity. In the case of perfect multicollinearity, the OLS is invalid and the coefficients of the regression model are biased and inconsistent. There are many solutions to avoid the perfect multicollinerity. These solutions are different from each other and provide the different parameters of dummy variable but the results are equivalent.
1
The analysis-of-variance (ANOVA) models are the regression models only include the dummy variables as explanatory variables.
16
The first ANOVA model drops a dummy variable, other news, to estimate the fixed-effect of politics-related downgrade and other news. The model looks like as following:
𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝒂𝒂𝒕𝒕𝒊𝒊𝑪𝑪𝒊𝒊 𝑨𝑨𝒃𝒃𝒊𝒊𝑨𝑨𝒔𝒔𝑪𝑪𝒂𝒂𝑪𝑪 𝑹𝑹𝒊𝒊𝒕𝒕𝑪𝑪𝒔𝒔𝒊𝒊𝑰𝑰,𝒕𝒕(−𝟐𝟐,𝟐𝟐)
= 𝜶𝜶𝟏𝟏+ 𝜷𝜷𝟏𝟏(𝑬𝑬𝑨𝑨𝑪𝑪𝒊𝒊𝒕𝒕𝒊𝒊𝑬𝑬𝒔𝒔 − 𝒔𝒔𝒊𝒊𝑪𝑪𝒂𝒂𝒕𝒕𝒊𝒊𝒊𝒊, 𝟎𝟎 𝑨𝑨𝒔𝒔 𝟏𝟏 𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝑬𝑬𝒂𝒂𝒕𝒕𝑨𝑨𝒔𝒔)
+ 𝜷𝜷𝟐𝟐(𝒊𝒊𝑨𝑨𝒅𝒅𝒊𝒊𝒅𝒅𝒔𝒔𝒂𝒂𝒊𝒊𝒊𝒊, 𝟎𝟎 𝑨𝑨𝒔𝒔 𝟏𝟏 𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝑬𝑬𝒂𝒂𝒕𝒕𝑨𝑨𝒔𝒔) + 𝜺𝜺𝟏𝟏 (1)
Where
Cumulative Abnormal ReturnI,t(−2,2): the cumulative abnormal returns (2-day before/after event date, t) for country I
𝜶𝜶𝟏𝟏+ 𝜷𝜷𝟏𝟏: the fixed-effect of the politics-related events 𝜶𝜶𝟏𝟏+ 𝜷𝜷𝟐𝟐: the fixed-effect of downgrade events
𝜶𝜶𝟏𝟏: the fixed-effect of others events
In addition, the second ANOVA model includes all the dummy variable but suppresses the constant intercept. The second ANOVA model is as following:
𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝑪𝒂𝒂𝒕𝒕𝒊𝒊𝑪𝑪𝒊𝒊 𝑨𝑨𝒃𝒃𝒊𝒊𝑨𝑨𝒔𝒔𝑪𝑪𝒂𝒂𝑪𝑪 𝑹𝑹𝒊𝒊𝒕𝒕𝑪𝑪𝒔𝒔𝒊𝒊𝑰𝑰,𝒕𝒕(−𝟐𝟐,𝟐𝟐)
= 𝜸𝜸𝟏𝟏(𝑬𝑬𝑨𝑨𝑪𝑪𝒊𝒊𝒕𝒕𝒊𝒊𝑬𝑬𝒔𝒔 − 𝒔𝒔𝒊𝒊𝑪𝑪𝒂𝒂𝒕𝒕𝒊𝒊𝒊𝒊, 𝟎𝟎 𝑨𝑨𝒔𝒔 𝟏𝟏 𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝑬𝑬𝒂𝒂𝒕𝒕𝑨𝑨𝒔𝒔)
+ 𝜸𝜸𝟐𝟐(𝒊𝒊𝑨𝑨𝒅𝒅𝒊𝒊𝒅𝒅𝒔𝒔𝒂𝒂𝒊𝒊𝒊𝒊, 𝟎𝟎 𝑨𝑨𝒔𝒔 𝟏𝟏 𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝑬𝑬𝒂𝒂𝒕𝒕𝑨𝑨𝒔𝒔)
+ 𝜸𝜸𝟑𝟑(𝑨𝑨𝒕𝒕𝒐𝒐𝒊𝒊𝒔𝒔𝒔𝒔, 𝟎𝟎 𝑨𝑨𝒔𝒔 𝟏𝟏 𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝑬𝑬𝒂𝒂𝒕𝒕𝑨𝑨𝒔𝒔) + 𝜺𝜺𝟐𝟐 (𝟐𝟐)
Where
Cumulative Abnormal ReturnI,t(−2,2): the cumulative abnormal returns (2-day before/after event date, t) for country I
𝜸𝜸𝟏𝟏: the fixed-effect of the politics-related events 𝜸𝜸𝟐𝟐: the fixed-effect of downgrade events
𝜸𝜸𝟑𝟑: the fixed-effect of others events
17
from the meaning of parameter in the second ANOVA model except 𝛼𝛼1 and 𝛾𝛾3 . For instance, 𝜷𝜷𝟏𝟏 indicates the extent to which the fixed-effect of politics-related news deviates from the fixed-effect of other news. Hence, the T-test aims to test whether the deviation from others news is zero. However, 𝜸𝜸𝟏𝟏 provides the actual fixed-effect on politics-related news and the second ANOVA model will also test whether fixed-effect of politics-related news is undistinguished from zero directly.
The ANOVA models above can both estimate the fixed-effect of the politics-related, downgrade and other news respectively. However, the second ANOVA model also can help us to find out whether the fixed-effect of politics-related, downgrade and others news is undistinguished from zero. Hence we will mainly focus on the second ANOVA model and the hypothesis of T-test for the second ANOVA model is formulated as following:
Hypothesis 1
H0: The fixed-effect of the politics-related events is zero or 𝜸𝜸𝟏𝟏=0
H1: The null hypothesis is not true.
Similar to the hypothesis1, we can formulate two more hypothesises to test whether the fixed-effect of downgrade and other news have a significantly influence on abnormal return.
Hypothesis 2
H0: the fixed-effect of downgrade events
is zero or 𝜸𝜸𝟐𝟐=0
H1: The null hypothesis is not true.
Hypothesis 3
H0: The fixed-effect of other events is
zero or 𝜸𝜸𝟑𝟑=0
18
5.
Empirical Results and DiscussionThe analysis of Section 5.1 provides the results of the average abnormal return for every country and group. In further, Section 5.2 discusses the results of the fixed-effect of Eurozone news.
5.1 Average Abnormal Return
Table 1 summarizes the abnormal return of the event study during the sample period from January 2008 to June 2013. The coefficients refer to the average abnormal return for each country over the event window. The standard star notation represents the significance levels, and the standard errors are in the parenthesis. In this table, the test-of-significance approach is applied to explain the significance of the average abnormal return. Additionally, there are no dummy variables included. The test-of-significance approach intends to exhibit whether the average abnormal return for each country and group is statistically distinguished from zero.
5.1.1 The PIIGS countries
The first group contains PIIGS countries that are known as the ‘worst-affected countries.’ The results in Table 1 also confirm the description of the ‘worst-affected countries.’ The average abnormal return for PIIGS is -12.823 and its T-test is -12.831. The T-test value is much less than the critical vale -2.33 (1% level).
Subsequently, the average abnormal return of Portugal, Italy, Ireland, Greece and Spain are all statistically different from zero at 1% significance level. The results show that all the events in the data pool have the significant impact on Portugal, Italy, Ireland, Greece and Spain at 1% significance level. Greece and Italy face the largest and second largest loss from the events (-35.112 and -16.927, respectively), and Portugal suffer the smallest impact (-2.436).
19 Table 1
This table reports the results of abnormal return for every country and every group. This table includes the mean, standard deviation (s.e) and T-test results. The T-test reports whether the average abnormal is statistically distinguished from zero. The standard errors are in parentheses. ***, **, * represent statistical significance at 1%, 5% and 10%.
Mean (s.e) T-test
PIIGS -12.823 (-47.926) -12.831 *** Portugal -2.436 (5.347) -9.771 *** Ireland -5.134 (40.791) -2.699 *** Italy -16.927 (56.766) -6.395 *** Greece -35.112 (75.082) -10.030 *** Spain -4.504 (14.174) -6.816 *** Euro -3.210 (16.442) -8.376 *** Germany -1.796 (12.564) -3.066 *** the Netherlands -3.952 (11.526) -7.353 *** France -3.448 (21.317) -3.469 *** Belgium -3.086 (52.614) -1.258 Non-euro 13.309 (80.385) 6.151 *** UK 15.950 (109.887) 3.113 *** Sweden 24.743 (81.275) 6.529 *** Switzerland -0.765 (19.905) -0.824 Whole sample -3.086 (52.614) -4.357 ***
20 5.1.2 The EURO countries
The second group includes EURO countries. The results of EURO countries are not as consistent as the results of PIIGS group. The average abnormal return of Germany, the Netherlands, and France are statistically different from zero at 1% significance level. However, an abnormal return of Belgium is insignificant.
In EURO group, Germany faces the smallest loss from the negative news (-1.796). The loss of the Netherlands, France and Belgium is close to each other. Overall, the EURO group is significantly affected by the negative news and the average abnormal return of EURO group is -3.210.
5.1.3 The Non-EURO countries
The third group consists of Non-EURO countries. The results of the Non-EURO countries are more intriguing than the results of PIIFGS and EURO countries. In the case of Sweden, during the EU sovereign debt crisis, all the negative events have significantly positive impact on the banking sector of Sweden. Therefore, the Swedish banking sector is in favour of the EU sovereign debt crisis. Reuters (2012) reported some explanations to such results. First, the Swedish government started to pursue banking sector reforms after the 1992 economic implosion. Second, after Euro crisis, Sweden became a haven for most investors. The yield on Swedish 10-year government bond reached the record low during the Euro crisis.
In the case of UK, the average abnormal return (15.950) is also positive but less than the Sweden. The average abnormal return of Switzerland (-0.765) is negative but insignificant. In general, the banking sectors of Non-EURO countries are not only insulated from the negative news, but also stimulated by the negative news (13.309).
21
Last but not least, the average abnormal return of the whole sample is -3.086 and significantly different from zero at 1% significance level. This result indicate that all the negative news have a significant impact on the whole sample.
The test-for-significance approach provides little information about what kind of news has the largest impact and what kind of news has the least impact. Subsequently, the news is divided into three categories to test the fixed-effect of them.
22 5.2 The fixed-effect of Euro-zone News
This section provides the empirical results and discussion for the previous ANOVA models. In addition, this section also analyses the problems regarding the ANOVA models, such as heteroskedasticity and autocorrelation. Hence the “robust standard error” is employed to correct model misspecifications.
5.2.1 The fixed-effect on every country
The table 2 indicates the results of the first ANOVA model and table 3 represents the results of the second ANOVA model. Both regression models report the fixed-effect of political-related, downgrade and other events. The column (1), (5) and (7) in table 2 and the column (1), (4) and (7) in the table 3 indicate the magnitude of the fixed-effect. As we can see in the both tables, the magnitude of the fixed-effect for politics-related, downgrade and other news is the same. This result confirms the previous discussion in section 4.2 which states both models provides the different coefficients but the results are equivalent.
Subsequently, this section will focus on Table 3 because the T-test of the second ANOVA model aims to test whether the fixed-effect of different news is undistinguished from zero. In the Table 3, the standard star notation represents the significance levels, and the standard errors are in the parenthesis.
In Table 3, column (1) and (2) report the fixed-effect of politics-related events and standard errors for each country respectively. Column (3) indicates the significance level at 1%, 5% and 10%. In the PIIGS group, the politics-related events significantly affect the banking sectors of Portugal, Ireland, and Greece. The fixed-effect of Italy and Spain is negative, but the impact is not significant. In the EURO group, the fixed-effect of the politics-related events are all negative. But T-test indicates all the parameters are not significant. Unlike the PIIGS and EURO group, the Non-euro group has all positive fixed-effect from the Politics-related news. However none of
23
the parameters is significant. Overall, based on the analysis, the politics-related news has limited impact across countries. The PIIGS countries are the most likely affected by the politics-related news than EURO and Non-Euro group. Meanwhile the influence of politics-related news on EURO and Non-Euro group is not significant.
Column (4) reports the fixed-effect of downgrade events and column (5) and (6) provides the corresponding standard deviation and significance level. The impact of the downgrade is tremendous since 11 out of 12 countries are negatively affected by the downgrade news. In the PIIGS group, the banking sectors of 5 countries are all significantly affected by downgrade. The significance level for Portugal, Italy and Greece is at 1% and the significance level of Ireland and Spain is at 5%. Additionally, the banking sector of Greece suffers the largest negative influence (-75.620) among all the countries. In the EURO group, the banking sectors of all the 4 countries are also negatively and significantly affected by the downgrade news. Furthermore, the banking sector of France undergoes the greatest loss among the 4 countries (-11.501). In the Non-EURO group, only the banking sector of Switzerland is negatively and significantly affected by the downgrade news. In summary, the downgrade events have consistent negative impact within PIIGS and Euro countries. For Non-euro countries, Switzerland is only country which is negatively and significantly affected by the downgrade news.
Column (7) presents the fixed-effect of other events, and the fixed-effect of other events varies from country to country. The banking sectors of some countries are significantly hurt by other news, such as Portugal, Greece, Spain and the Netherlands. However, the banking sector of Sweden benefits from other events. Again, Greece suffers the greatest loss on the other events among all the countries (-25.458). The Sweden benefits from the other events enormously (33.263). Hence, the other news have the least impact comparing to politics-related and downgrade news
24
downgrade news have the consistent and adverse impact on the banking sectors of most countries consistently. However, the influence of politics-related and other events differ from country to country in terms of magnitude and direction. In a comparison, the other events have the least effect on the banking sector returns than politics-related and downgrade news.
The sovereign credit rating is an indicator to access the ability of a sovereign obligator to repay or refinance its debt. This section finds the downgrade news have the greatest chance to impact both the domestic banking sector and foreign banking sectors. But the contagion is limited and only within PIIGS and Euro countries. The Non-euro countries are immune to downgrade news except Switzerland.
Based on the results, the contagion of politics-related news is only within the PIIGS countries. These troubled countries are more likely significantly affected by the politics-related news than Euro countries and Non-Euro countries. By contrast, the other news, such as demonstrations, have the least likelihood to contaminate other countries than downgrade and politics-related news. The power of other news is limited and small and the impact of other news is only within a few countries.
25
Table 2 This table reports the results of the first ANOVA model. The independent variables are Politics Related news and Downgrade news respectively. The dependent variable is the cumulative abnormal returns over the event window. The standard errors are in parentheses. ***, **, * represent statistical significance at 1%, 5% and 10%.
Politics-related Downgrade Intercept (Other)
s.e s.e s.e
PIIGS 1 2 3 4 5 6 7 8 9 10 11 Portugal -1.326 0.929 (1.249) -5.100 -2.844 (1.662) *** -2.256 (0.933) ** Ireland -7.835 -9.357 (7.540) -7.269 -8.791 (12.623) 1.522 (4.841) Italy -14.084 -1.470 (14.511) -27.793 -15.129 (14.694) -12.614 (10.415) Greece -25.791 -0.333 (14.703) -75.620 -50.162 (26.571) -25.458 (9.797) ** Spain -3.677 1.928 (3.876) -5.291 0.315 (3.798) -5.605 (2.902) ** Euro Germany -0.933 0.295 (3.282) -3.937 2.709 (3.481) -1.228 (2.644) Netherland -1.806 4.950 (2.791) -4.091 2.665 (3.072) -6.756 (2.071) * France -4.803 -6.149 (5.345) -11.501 -12.847 (6.138) ** 1.346 (3.847) Belgium -1.041 2.700 (5.137) -9.199 -5.459 (4.857) -3.740 (4.065) Non-euro UK 0.876 -26.061 (28.116) 6.751 -20.186 (30.207) 26.936 (21.344) Sweden 17.333 -15.929 (21.982) -13.543 -19.719 (21.379) 33.263 (17.204) ** Switzerland 1.441 -2.735 (4.696) -10.076 -14.252 (5.122) * 4.176 (3.625)
26 Table 3
This table reports the results of the second ANOVA model. The independent variables are Politics Related news, Downgrade news and other news respectively. The dependent variable is the cumulative abnormal returns over the event window. The standard errors are in parentheses. ***, **, * represent statistical significance at 1%, 5% and 10%.
Politics-related Downgrade Others
𝜸𝜸𝟏𝟏 s.e 𝜸𝜸𝟐𝟐 s.e 𝜸𝜸𝟑𝟑 s.e
PIIGS 1 2 3 4 5 6 7 8 9 Portugal -1.326 (0.831) *** -5.100 (1.375) * -2.256 (0.933) ** Ireland -7.835 (5.781) *** -7.269 (4.658) ** 1.522 (4.841) Italy -14.084 (10.104) -27.793 (10.365) * -12.614 (10.415) Greece -25.791 (10.964) ** -75.620 (24.699) * -25.458 (9.797) ** Spain -3.677 (2.570) -5.291 (2.451) ** -5.605 (2.902) ** Euro Germany -0.933 (1.945) -3.937 (2.265) *** -1.228 (2.644) Netherland -1.806 (1.870) -4.091 (2.268) ** -6.756 (2.071) * France -4.803 (3.711) -11.501 (4.783) * 1.346 (3.847) Belgium -1.041 (3.141) -9.199 (2.660) * -3.740 (4.065) Non-euro UK 0.876 (18.302) 6.751 (21.375) 26.936 (21.344) Sweden 17.333 (13.684) -13.543 (12.692) 33.263 (17.204) ** Switzerland 1.441 (2.985) -10.076 (3.620) * 4.176 (3.625)
27 5.2.2 Robustness
In order to achieve an unbiased and consistent estimators in the linear regression model, a bunch of assumptions are made. This section mainly focuses on the assumption of no heteroscedasticity and no autocorrelation.
Heteroscedasticity is a situation in which the variance of error term depends on the independent variables. If error term of a regression model depends on its independent variables, the T-test and F-test are both invalid. Hence it is important to detect the presence of heteroscedasticity and correct it. In this paper, the Breusch-Pagan test (BP test henceforth) is deployed to find the heteroscedasticity and the results of BP test are given in the Table 4.
The table 4 shows there are only 3 countries have the heteroscedasticity in the 1st ANOVA model and all the countries have the heteroscedasticity in the 2nd ANOVA model. Without an adjustment to these models, the T-test is invalid and we cannot interpret whether the fixed-effect of different news is distinguished from zero. All the results in the Table 2 and Table 3 are actually corrected by the Stata based on the detection of heterosecdasticity. So all the observations in the section 5.2 are still valid.
In addition, the assumption of no autocorrelation means the error term is not serially correlated. The consequence of autocorrelation gives both inconsistent estimators and invalid T-test and F-test. The Breusch-Godfrey test (BG test henceforth) is used to detect the autocorrelation in the regression model. This section skips the BG test because the Stata corrects both the heterosecdasticity and autocorrelation of a regression model simultaneously. So the results in the Table 2 and Table 3 are adjusted by correcting heteroscedasticity and autocorrelation.
28 Table 4
This table reports the results of Breushch-Pagan test for 1st and 2nd ANOVA model. Column (1) and Column (2) present the value of F-test and p-value respectively. In Table 4, the critical level is 10%. Column (3) reports the presence of heterosedasticity.
1st ANOVA model 2nd ANOVA model
F-test p-value Heteroscedasticity F-test p-value Heteroscedasticity
1 2 3 4 5 6
PIIGS
Portugal 3.42 3.70% YES 9.89 0.00% YES
Ireland 4.47 1.42% YES 10.31 0.00% YES
Italy 0.34 71.05% 11.31 0.00% YES
Greece 6.92 0.16% YES 13.33 0.00% YES
Spain 1.1 33.83% 16.72 0.00% YES Euro Germany 1.28 28.29% 14.77 0.00% YES Netherland 0.02 98.49% 12.54 0.00% YES France 0.06 94.50% 7.81 0.00% YES Belgium 2 14.10% 10.69 0.00% YES Non-euro UK 0.24 78.95% 16.94 0.00% YES Sweden 1.42 24.61% 11.81 0.00% YES Switzerland 0.27 76.31% 14.2 0.00% YES
29
6.0 Conclusion
The EU sovereign debt crisis roots in the unsustainable public debt and heavy exposure of Eurozone banks to such debt. The heavy exposure transforms the sovereign debt crisis into Europe-wide banking sector crisis rapidly. This thesis highlights the fragility of banking sectors in EU to major impact emanating from the PIIGS countries. Indeed there are considerable difference among these countries in terms of the magnitude of the effect, but the EU sovereign debt crisis has a pronounced influence for most countries. The figures in Appendix 2 meet the idea and indicate that the banking sector indices of PIIGS countries drop substantially since the 4th quarter of 2009.
In this paper, the event study methodology is employed to investigate the influence of negative news from Eurozone area. The results of the event study are provided in Table 1. Subsequently, the test-of-significance approach is applied to test whether the average abnormal return for every country and group is statistically different from zero. The negative average abnormal returns in Table 1 suggest the most banking sectors are significantly damaged by negative news. Nonetheless, the banking sectors of UK and Sweden are not only largely insulated from the negative news but also stimulated by the negative news.
The basic test-of-significance approach contributes little information to the impact of the different negative events. Therefore, all the negative events are divided into 3 groups (politics-related, downgrade and other news) and two ANOVA models are deployed to estimate the fixed-effect of each group. A more comprehensive results are added in the Table 2 and 3. In the Table 2 and 3, the results indicate that the downgrade news consistently and negatively affect all the countries except UK and Sweden. As to the politics-related news, they have limited influence and only
30
significantly and negatively affect countries within PIIGS countries. The other news has the least impact among all the negative news. The fixed-effect of other news substantially differ from each other in terms of both direction and magnitude. For example, the banking sector of Greece is significantly and negatively affected by the other news (-25.458), but the banking sector of Sweden is significantly and positively affected by the other news (33.263).
Indeed, there are some limitations regarding to the results of this thesis. At first, in the event study part, this thesis only chooses 5 days as event window to calculate the abnormal return. By doing so, the thesis may provide inaccurate abnormal return. The length of events window is quite flexible, in future the author can construct more event windows simultaneously, such as 1 days, 3 days, 5 days and 10 days and compare abnormal returns across different event window. Secondly, the section 4.2 only employs the dummy variables (politics-related, downgrade and other news) as independent variables to explain the abnormal return. The limited independent variables may reduce the accuracy of estimators. Hence more indicators can be included in the future model, such as the extent of exposure to PIIGS countries debt. Last but not least, this thesis only divides all the negative news into 3 groups. In the future, the author can split the 3 groups into more sub-groups and estimate the fixed-effect of each sub-group. For example, downgrade news can be divided based on the rating agency, such as S&P, Moody and Fitch.
In this thesis, the empirical results will help the policymakers to gain insight into the EU sovereign debt crisis and to address the risks in the future crisis. First, as long as some countries are downgraded, policymakers should step out and react actively to reduce the concern and doubt of the market and to calm the market. By doing so, the policymakers can minimize the negative consequence. Second, policymakers also should be well-prepared to address the contingent event, such as local protest or strike. As to the investors, they should be cautious with the downgrade news because the downgrades news have the consistent negative impact on the banking sectors. Further,
31
during the crisis, it is easy to observe that Sweden is a good haven for European investors. Meanwhile, investors also should be prudent about their investment on PIIGS countries in the future.
Reference
Arezki, R., Candelon, B., & Sy, A. (2011). Sovereign rating news and financial markets spillovers:
Evidence from the European debt crisis. IMF working papers, 1-27.
Aizenman, J., Jinjarak, Y., Lee, M., & Park, D. (2012). Developing countries’ financial vulnerability to
the euro crisis: An event study of equity and bond markets (No. w18028). National Bureau of
Economic Research.
Arteta, C., & Hale, G. (2008). Sovereign debt crises and credit to the private sector. Journal of
International Economics, 74(1), 53-69.
Baig, T., & Goldfajn, I. (1998). Financial market contagion in the Asian crisis.
Beetsma, R., Giuliodori, M., de Jong, F., & Widijanto, D. (2013). Spread the news: The impact of news on the European sovereign bond markets during the crisis. Journal of International Money and Finance,
34, 83-101.
BIS, 2011. The impact of sovereign credit risk on bank funding conditions. Committee on the Global
Financial System Papers 43.
Brown, S. J., & Warner, J. B. (1980). Measuring security price performance.Journal of financial
economics, 8(3), 205-258.
Brown, C. O., & Dinc, I. S. (2011). Too many to fail? Evidence of regulatory forbearance when the banking sector is weak. Review of Financial Studies,24(4), 1378-1405.
Brown, S. J., & Warner, J. B. (1985). Using daily stock returns: The case of event studies. Journal of
financial economics, 14(1), 3-31.
Blundell-Wignall, A. (2012). Solving the financial and sovereign debt crisis in Europe. OECD Journal, 201.
Chan-Lau, J. A., Liu, E. X., & Schmittmann, J. M. (2012). Equity returns in the banking sector in the
32
Datastream , Thomson Reuters., 2014.
http://www.eui.eu/Research/Library/ResearchGuides/Economics/Statistics/DataPortal/datastream.aspx
Demirgüç-Kunt, A., & Huizinga, H. (2011). Do we need big banks? Evidence on performance, strategy and market discipline.
Gande, A., & David, C. Parsley, 2005, “News Spillovers in the Sovereign Debt Market.”. Journal of
Financial Economics, 75, 691-734.
s
Kaminsky, G., & Schmukler, S. L. (2002). Emerging market instability: do sovereign ratings affect country risk and stock returns?. The World Bank Economic Review, 16(2), 171-195.
Mackinlay, A.C.,(1997) Event Studies in Economics and Finance, Journal of Economics Literature Vol.
XXXV (March 1997(,pp 13-39)
McWilliams, A., & Siegel, D. (1997). Event studies in management research: Theoretical and empirical issues. Academy of management journal, 40(3), 626-657.
Mohl, P., & Sondermann, D. (2013). Has political communication during the crisis impacted sovereign bond spreads in the euro area?. Applied Economics Letters, 20(1), 48-61.
Pisani-Ferry, J. (2012). The euro crisis and the new impossible trinity (No. 2012/01). Bruegel Policy Contribution.
Reinhart, C. M., Rogoff, K. S., & Savastano, M. A. (2003). Debt intolerance (No. w9908). National Bureau of Economic Research.
Reinhart, C. M., & Rogoff, K. S. (2009). The aftermath of financial crises (No. w14656). National Bureau of Economic Research.
Reinhart, C. M., & Rogoff, K. S. (2008). Banking crises: an equal opportunity menace (No. w14587). National Bureau of Economic Research.
Ryngaert, M., & Netter, J. (1990). Shareholder wealth effects of the 1986 Ohio antitakeover law revisited: Its real effects. Journal of Law, Economics, & Organization, 253-262.
Shambaugh, J. C., Reis, R., & Rey, H. (2012). The Euro's Three Crises [with Comments and Discussion]. Brookings Papers on Economic Activity, 157-231.
Thomson Reuters., 2012
http://www.reuters.com/article/2012/06/13/us-sweden-eurocrisis-idUSBRE85C0GQ20120613
Wolff, G. B. (2011). Is recent bank stress really driven by the sovereign debt crisis? (No. 2011/12). Bruegel Policy Contribution.
33
Appendix 1 Event List
Category 1 Politics-related events
Date Title
1/12/2010 Angela Merkel has changed her mind on policy co-ordination 2/15/2010 Greece under massive pressure to step up austerity package 2/22/2010 Dutch government collapses
3/4/2010 Greece is now putting pressure on the Eurozone-if you don’t help, we will turn to the IMF 3/22/2010 Germans think they would be better off outside the euro area
4/4/2010 Francois Hollande tones down his proposals for treaty revision 4/27/2010 Merkel about to push the Eurozone over the brink
5/7/2010 Instead of solving the problem European leaders play the blame game 5/19/2010 The arrival of regulatory revenge
5/21/2010 Merkel's comments trigger a mass sell-off of global stocks
6/1/2010 Axel Weber blames stimulus for crises, criticizes ECB decision to buy bonds, and rejects common bond 9/16/2010 Greek finance minister says Greek default will break the Eurozone
10/14/2010 A political crisis looms large in Portugal 10/26/2010 Merkel's Eurozone proposals likely to fail
11/2/2010 EU leaders trigger another bond market crisis
12/9/2010 The European council is once again at each other's throat 1/19/2011 Dutch finance minister rattles the markets
34
2/11/2011 Weber's withdrawal triggers another speculative attack 3/3/2011 Merkel says no to Ireland
3/23/2011 Merkel humiliates Schauble by reopening discussions on ESM
5/31/2011 A political landslide in Italy-and a lame-duck government for two more years 6/11/2011 Greece PM resigns
7/15/2011 Merkel hides behind procedures, leaving Eurozone on the brink of collapse 7/21/2011 Sarlozy call Merkel 'Criminal'
8/21/2011 Another day in the Eurozone: Merkel says definitely No to Eurobonds and five small countries endanger the EFSF 10/18/2011 Don’t expect too much on Sunday, say Merkel and Schauble
11/8/2011 Silvio Berlusconi vows to resign as Italy's PM 1/18/2012 A German end to the Euro vision
1/30/2012 Greek government calls German proposal,' a product of a sick imagination' 2/23/2012 Germany is now opposed to merger of EFSF and ESM
6/27/2012 Merkel: Not over my dead body; Monti throws down gauntlet 8/28/2012 Mariano Rajoy may delay Spain's EFSF application further 9/18/2012 Rajoy's dithering sends Spanish yields creeping up again 9/19/2012 Political tensions mount in Portugal
12/3/2012 Bersani will lead the Centre-left in Italian general elections 1/16/2013 Germany want to delay, and delay, and delay a decision on Cyprus
2/4/2013 Political worries
2/19/2013 Mediobanca says Italy won’t have a stable government
35 Category 2 Downgrade events
Date Tittle
12/8/2009 Greece downgrade by Fitch 12/9/2009 Spain outlook revision by S&P 12/16/2009 Greece downgrade by S&P 12/22/2009 Greece downgrade by Moody's
3/24/2010 Portugal downgrade by Fitch 4/22/2010 Greece downgrade by Moody's
4/27/2010 Greece and Portugal downgrade by S&P 4/28/2010 Spain downgrade by S&P
5/5/2010 Portugal review by Moody's
7/19/2010 Ireland's debt downgraded by credit rating agency(Moody) 8/25/2010 Ireland's debt downgraded by S&P
12/14/2010 S&P gives Belgium credit rating warning amid political uncertainty 7/5/2011 Moody's downgrades Portugal to Ba2 with negative outlook
7/6/2011 Portugal's credit rating downgraded to junk status 7/13/2011 Ireland cut to junk by Moody's
9/20/2011 Italy downgraded
9/20/2011 Italy's rating cut one level by S&P
1/15/2012 Portugal's credit rating downgraded to junk status by S&P 2/13/2012 Moody's downgrades Italy, Portugal and Spain
2/27/2012 S&P downgrades Greece to selective default 3/3/2012 Moody's downgrades Greece to lowest level 4/26/2012 Spain downgrade by S&P
36
10/11/2012 S&P cuts Spain credit rating to near junk
11/10/2012 S&P says it has lowered the rating on Spain's debt 11/24/2012 S&P downgrades Ireland on debt worries
1/10/2013 Cyprus downgrade by Moody's 3/21/2013 Cyprus downgrade by S&P
Category 3 other events
Date Tittle
2/4/2010 Greek strike threaten plans to tackle debt crisis 2/5/2010 Spain seeks to calm fears it is 'next Greece'
2/21/2010 Greece not looking for EU handouts, says Greek prime minster George Papadreou 4/28/2010 Portugal announces tougher austerity measures
8/12/2010 Greece's economy deeper in recession than forecast 9/20/2010 Greece delays bank stress tests
11/27/2010 Thousands protest against Irish bailout 1/7/2011 Portuguese borrowing costs hit record
3/10/2011 Spanish banks fail stress tests and debt is downgraded 4/7/2011 Portugal's PM calls on EU for bailout
9/6/2011 Italian workers striker over austerity measure 10/5/2011 General strike brings Greece to a standstill 10/19/2011 Greece Erupts
11/9/2011 Italy's borrowing costs hit new high 11/24/2011 Portugal paralyzed by strike over cuts
37
12/5/2011 Ireland, Italy and Greece face more cuts and tax rises 1/27/2012 Spanish unemployment tops 5.3m and set to get worse
4/10/2012 Eurozone crisis reignites as Spanish bond yields hit four-month high 5/7/2012 Greece: leaders fail to form a coalition
5/14/2012 Greece euro exit fears rattle stock market
6/25/2012 Ticking off by Troika heightens fears of Greek exit from euro 7/23/2012 Spain in crisis talks with Germany over 300bn euro bailout 9/26/2012 Athen descends into violence as 200,000 march against austerity 11/14/2012 Europe unites in austerity protests against cuts and job losses 11/23/2012 Cyprus becomes fourth EU country to ask for bailout
1/9/2013 German politicians threaten to block Cyprus bailout 1/21/2013 Cyprus bailout delayed until spring
2/4/2013 Investors alarmed by corruption scandal around Spanish PM Mariano Rajoy 2/22/2013 Eurozone recession set to continue
3/20/2013 Cyprus rejects bailout deal leaving Eurozone facing fresh crisis
6/5/2013 IMF admits: we failed to realize the damage austerity would do to Greece 7/31/2013 IMF finds $11bn black hole in Greek finances and warns of new write-off
38 Appendix 2 The banking sector return indices
PIIGS countries 0,00 0,50 1,00 1,50 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy
Portugal
0,00 0,20 0,40 0,60 0,80 1,00 1,20 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yyItaly
39 0,00 0,50 1,00 1,50 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy
Ireland
0,00 0,50 1,00 1,50 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yyGreece
0,00 0,20 0,40 0,60 0,80 1,00 1,20 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yySpain
40 EURO countries 0,00 0,50 1,00 1,50 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy
Germany
0,00 0,50 1,00 1,50 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yyBelgium
0,00 0,20 0,40 0,60 0,80 1,00 1,20 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yyThe Netherlands
0,00 0,20 0,40 0,60 0,80 1,00 1,20 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yyFrance
41 Non-Euro countries 0,00 0,20 0,40 0,60 0,80 1,00 1,20 yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy yy
