The impact of Moody’s creditrating up-and downgrades on stock prices in Europe : Has this effect changed after the implementation of EU regulations for CRAs?

The impact of Moody’s credit rating up- and

downgrades on stock prices in Europe.

Has this effect changed after the implementation of EU regulations for CRAs?

G. Degen-Barbosa 5991706 August 15, 2018

Master’s Thesis to obtain the degree in Actuarial Science and Mathematical Finance University of Amsterdam

Faculty of Economics and Business Amsterdam School of Economics Supervisor: Prof. Dr van Gastel

Statement of Originality

This document is written by student Giselle Barbosa Degen, who declares to take full responsibility

for the contents of this document. I declare that the text and the work presented in this document is original and that no sources others than those mentioned in the text and its references have been

used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of

Abstract

The aim of this research is to investigate the impact of credit rating up- and downgrades, issued by

Moody’s, on the European stock market. In addition, there will be examined whether or not this effect changed after the implementation of European regulations for CRAs between the period of

2009-2013.

An event study was carried out to assess the impact of credit rating up- and downgrades

announcements on the European stockmarket. Data from Moody’s is utilised, and the rating up-and downgrades will be gathered within the period before the last regulation (2013) up-and in the

period after the last implemented European regulation (2017) in order to determine whether the European regulation for CRAs has had an impact.

A linear regression (OLS method) was conducted, to determine the abnormal return of the Mar-ket Model. The abnormal stock return associated with credit rating announcements is measured to

test the significance of the events.

Keywords: CRAs, Moody’s, credit ratings, event study, Market Model, European stock market,

Contents

1 Introduction 5

2 Literature review 8

2.1 History of CRAs . . . 8

2.2 The role of CRAs in the financial market . . . 9

2.3 Credit ratings . . . 10

2.4 Moody’s Rating Process and Credit Rating model . . . 12

2.5 Development of regulation for CRAs in Europe . . . 16

2.6 Prior research on the influence of credit ratings . . . 20

3 Methodology 23 3.1 Data Description . . . 23

3.2 Timeline event study . . . 26

3.3 The Market Model . . . 27

4 Results 32 4.1 The Results of 2013 . . . 32

4.2 The Results of 2017 . . . 33

4.3 Evaluation downgrade effects before and after regulation. . . 34

5 Conclusions and recommendations 35 6 Appendix 37 6.1 Data . . . 37

6.2 Results Market model . . . 38

Preface

At first, I would like to thank my husband Michiel and my son Liam, I owe them a huge amount

of gratitude for supporting me. Furthermore, I wanted to thank my dear friends Marine, Anita and Tahira who gave me valuable feedback and encouragement. My family and friends have assisted,

supported, and encouraged me and guided me through this stressful and very busy period. I also wanted to thank my manager for approving the tuition fee of this master, and of course my

em-ployer ING for funding it. Finally, my sincere gratitude to my supervisor, Prof. Dr van Gastel, for his constructive advice on my thesis.

1. Introduction

The aim of this thesis is to investigate the influence of Credit Rating Agencies (hereinafter:

CRAs) on the stock market. CRAs assign credit ratings to entities issuing bonds reflecting their creditworthiness. These ratings give an estimation of the probability of default and the likelihood

that the company will pay back their debts (Moody’s, 2005). The ratings provide a simple norm for measuring risk, therefore they allow an easier comparison of bonds issued by different companies

and governments. Due to new developments in the regulatory framework with regard to the use of credit ratings, CRAs acquired a more prominent role in the financial market (DNB, 2010). Since

CRAs play a key role in the financial market, one might expect that the ratings assigned by the CRAs have a significant impact on the stock prices as well.

After the financial crisis erupted in the U.S. during 2007, it was clear that the CRAs played a crucial role (Jarrow, 2011). CRAs underestimated the credit risk of corporate and structured

financial products. For instance, the misrated Lehman Brothers who went bankrupt, and the near failure of the insurance company AIG. Furthermore, some Collateralised Debt Obligation

(here-inafter: CDO), were assigned a triple A rating and the CRAs failed to adjust them accordingly in a timely matter. During the financial crisis the top-ratings were adjusted to a subsequent junk status.

As a result of the misrating, a lot of investors were misguided, which exposed them to a higher risk investment than anticipated. When the American housing market collapsed and the financial

crisis erupted, the misrating of CRAs were exposed. These misratings were persistent due to both the complexity of the financial products, the conflict of interest, the use of inadequate models to

estimate the risk of default and the lack of a regulatory framework for CRAs (Jarrow, 2011). The financial crisis highlighted the discussion concerning the influence of CRAs, the

overre-liance on CRAs, and the lack of regulation. Before the financial crisis commenced, the CRAs were barely regulated. During the European debt crisis criticism on the CRAs increased. The CRAs

were accused off downgrading credit ratings of European countries such as Portugal, Ireland and Greece, causing it to became more costly for them to attract new fundings as the interest rates

in-creased. The higher interest rates contributed to a downward spiral during the European debt crisis. The credit rating downgrade also contributed to developments in the bond spread of other

Euro-pean countries (De Santis, 2012). The role of the CRAs in the EuroEuro-pean debt crisis highlighted the discussion whether CRAs have too much influence on the market and if they perhaps should be

regulated by the implementation of legislation. After the financial crisis, Europe introduced new regulation concerning CRAs to address the overreliance on credit ratings. The question whether

or not CRAs have a great impact on the financial market by assigning credit ratings is relevant because then it can be determined if the regulation has had an effect.

The underlying reason for the existence of CRAs is that they mitigate information asymmetry with respect to creditworthiness (AFM, 2014). There has been an ongoing debate on whether

CRAs actually produce new information. If credit ratings consist of new information, the financial market should react significantly after a rating up- or downgrade. According to Boot, Milbourn and

Schmeits (2005), the influence of CRAs on investors and entities is explained of the basis of the coordination theory. The coordination theory describes how credit ratings influence the behaviour

of investors and companies. As mentioned earlier, CRAs can influence investors because investors use credit ratings to assess risk and compare bonds. In addition, the credit rating can also influence

corporate entities, because it is important to have a high credit rating to stay attractive among potential investors. It is also possible that credit rating up- and downgrades reflect information

that is already known and processed by the stockmarket and they provide no new information. It is critical to assess the influence of credit ratings, because whether or not credit ratings have a

significant influence on the financial market, the debate regarding the regulation of CRAs to ensure a stable financial market, can deviate.

When Moody’s assigns a credit rating, it can lead to a confirmation, a downgrade or an up-grade of a credit rating. An increase in the credit rating means that the creditworthiness of the

company ameliorates while a downgrade in the credit rating signifies that the creditworthiness of a company decreases. According to the research of Li. et al. (2004), conducted on the influence of

CRAs in the Swedish stock market, a downgrade had a significantly negative impact on the stock price and an upgrade a significantly positive effect. Hand (1992) concluded that both credit rating

upgrades as well as credit rating downgrades had a significant effect on the stock price. Freitas and Minardi (2013) studied the effect of credit ratings on stock prices in South America, and

con-cluded otherwise. Their results demonstrated an asymmetric effect between a credit rating down-and upgrade. Hence, it can be concluded that a credit rating downgrade has a significant effect on

the stock price. Contrary, a credit rating upgrade does not have a significant effect on the stock price. Finally, Holthausen and Leftwich (1992) concluded that only credit rating downgrades, and

a downgrade "Watch list" announcement, had a significant effect on the stock price. This research examines the following questions:

1. Do credit rating up- and downgrades, issued by Moody’s, have a significant impact on the Eu-ropean stock market?

2. Do these credit rating announcements, issued by Moody’s, have a lower impact on the European stock market after the implementation of the European regulation concerning CRAs?

This research is based upon rating announcements containing only actual rating changes, both up- and downgrades. The data of up- and downgrades are retrieved from Moody’s and the data

is used to conduct an event study. The research on the effect on stock prices, will be limited to European corporate companies during the years 2013 and 2017, namely before- and after the

implementation of the latest regulation on CRAs.

This research contributes to existing literature in a different way. Research on the effect of

credit rating up- and downgrades on the stock market is not an exception within existing literature. However existing literature has not yet focused on, if this effect has changed after the

implemen-tation of European regulation. In 2016 the European Commission published "the study state of credit market", where they concluded that the CRAs regulation had not reached the desired results

on the effect on the market concentration.

To answer the research questions, the following topics will be discussed. The qualitative part

of the assessment will be elaborated on in the chapter ‘Literature review’. The following subjects will be discussed. First, the background of CRAs will discussed and the role of CRAs will be

elab-orated on. Furthermore, the definition of credit ratings and the model behind how credit ratings actually are determined will be discussed in this chapter. In addition, the criticism against CRAs

and the development of the regulatory framework concerning CRAs will be specified. The quan-titative part of the assessment will be elaborated on in the chapter ‘Methodology’, which contains

2. Literature review

The chapter on literature provides background information on CRAs and their earnings model.

Section 2.2 examines the role that CRAs play in the financial market. In section 2.3 the credit ratings are elaborated on, followed by the model used to estimate credit ratings, depicted in

sec-tion 2.5. Subsequently, the criticism CRAs face and the developments regarding the regulasec-tion of CRAs are discussed in 2.6. Finally, section 2.7 discusses earlier research on the impact of a credit

rating change in the financial market.

2.1 History of CRAs

John Moody published in 1909 the first available bond ratings, for railroad bonds, followed by Poor’s Publishing company in 1916, and the Fitch Publishing company in 1924 (White, 2010). The

credit rating market has evolved during the years, however the three successors, Moody’s Investor Service, Standard & Poor’s (S&P) and Fitch Rating (the big three), still dominate the credit rating

market. "The big three" have obtained more than 90% of the global rating market (OECD, 2010), and they are all based in the United States of America. Besides "the big three", there are around

150 CRAs in the world varying in size, focus and methodologies, but solely the big three operate globally (DNB, 2011). This research will elaborate on one of the largest players in the credit rating

market; Moody’s, due to the data accessibility .

In the beginning, the CRAs were able to generate profit by selling their credit assessments to

the users, this was called "the user model", due to the fact that the users were compensated by the CRAs in the application of the credit ratings. In the 70s, the CRAs alternated from the

subscriber-pay model to an issuer-subscriber-pay model, where the rated entities who issued debt, paid CRAs to provide them with a rating (Lynch, 2008). Numerous factors caused this transition. First, the advanced

technological reproduction made it conceivable that ratings could easily be shared, and the CRAs were not able to prevent the illegal circulation. This illegal circulation created a huge pressure for

the subscriber-pay model. Additionally, the assessments became more complex and a more sophis-ticated analysis was required. Therefore, additional funds were obliged. A vital alternation in the

financial market was that issuers were expected to issue rated securities. These aforementioned de-velopments lead to the origin of the use of the alternate model; the issuer-pay model (Lynch, 2008).

2.2 The role of CRAs in the financial market

CRAs state that credit ratings are not a trade recommendation with regard to a probable buying

or selling of bonds, but they give an opinion with respect to their creditworthiness. For this reason, credit ratings can be visualised as a tool to reduce information asymmetry between issuers and

investors (AFM, 2014). In addition to this "information service", CRAs also have a "monitoring service". This second role, entails a "watch procedure", where CRAs influence companies to take

actions to prevent a credit rating downgrade (IMF, 2010). The third role of CRAs is to commit the determination of an adequate interest rate before launching a financial product (EPRS, 2016). A

high credit rating leads to lower interest rates, vice versa a low rating or a downgrade could lead to higher interest rates. Finally, it can be contemplated that CRAs have obtained an additional role

through the regulatory requirement using credit ratings (Lynch, 2008).

There are multiple reasons why CRAs have obtained such a crucial role in the financial

mar-ket. First of all, there is a need for a comparison standard, especially in times where financial products have become more complex. Second, the growth of usage of the credit ratings is also

explained by the regulatory endorsement of the use of credit ratings. An example is the Basel II framework, where an external credit assessment of CRAs can be applied in order to determine the

capital requirement for banks (BCBS, 2003). Basel III maintained the same approach with regard to the use of credit ratings, however they acknowledged the pitfalls in the application of the credit

ratings (Chiu, 2013). Another example is that when the Central Bank lends new capital funds to a bank, the bank should have provided a collateral with a minimum credit rating in order for

the application to be granted by the Central Bank (DNB, 2011). In U.S. the CRAs obtained an identical role. The CRAs were appointed as National Statistical Rating Organisations (hereinafter:

NSRO), resulting in a regulatory enforcement of the influence of CRAs (DNB, 2011). The big three have all been appointed NSRO status. Securities and Exchange Commission (hereinafter:

SEC) regulations require the use of NSRO ratings in the issuance of debt. Furthermore, pension fund regulations require pension funds to comply to prudent rules, where they are constrained to

2.3 Credit ratings

Credit ratings are denominated on an alphabetic scale (Moody’s). For instance, according to

the Moody scale, as depicted in figure 1, Aaa is the highest rating and D the lowest rating. Higher ratings imply a lower probability of default. The CRAs made a clear differentiation between

in-vestment grade and non inin-vestment grade, also known as speculative grade. To distinguish within a classification, plusses and minuses are utilised by Fitch and S&P, and subsequently numbers are

employed by Moody’s. Moody’s does not solely have the objective to provide an accurate rating, but also to furnish a more stabilised rating in order to ensure more stability and prevent

procycli-cality in the financial market (IMF, 2010). Consequently, CRA ratings are measured throughout the cycle instead of at a point in time. Cantor and Mann (2007) conducted a research about the

trade-off between credit rating accuracy and stability. According to Cantor and Mann, Moody’s achieved the right combination of stability and accuracy, with the combination of credit ratings,

Outlook announcements and Watch list information.

Figure 1: Rating Scale

Credit ratings announcements, such as up- and downgrades and confirmations are published on the website of Moody’s. Further, Moody’s also publishes the rating Outlook. The Outlook is a

view on the likely future direction of a credit rating, it can be stable, negative or positive. Finally, Moody’s also publishes when a company is posted on a "Watch list" to be reviewed for an up- or

downgrade. When the credit quality has suddenly decreased, the credit watch will be negative. When there has been an improvement what could affect the rating, the company is listed on the

positive credit watch. The credit assessment will take place after the listing and at the end of the evaluation the credit rate will be upgraded, downgraded or stay the same (Cantor and Mann, 2007).

The following distinction can be made with respect to the ratings; unsolicited- and solicited ratings. Unsolicited ratings are ratings that are initiated by the CRAs and are not amended based

upon a request of an entity. Equivalent credit assessments are made without contractual agree-ments with the concerning entity. Consequently, the entity does not have to pay any fees for the

credit assessment. Solicited ratings are requested by the entities, which are being assessed. This in contrary to the unsolicited ratings, which do require entities to settle a contractual agreement with

a required underlying compensation with the CRAs. In common practise, the solicited ratings are the most frequently applied (Moody’s).

2.4 Moody’s Rating Process and Credit Rating model

CRAs practise different methodologies and models in order to determine credit ratings. The

steps that are executed within the rating process of these different CRAs are equivalent (Alcubilla & Del Pozo, 2012). After a company request a rating, the process is initiated. The first step is the

preparation of the credit rating, where all information concerning the entity is gathered. The second step is the assessment of the credit rating, in which the appropriate credit rating methodology is

applied (the basics of the model will be discussed later on in further detail). In this step the results are analysed, and a recommendation by the analyst is computed. The third step is the decision

phase, where the committee will contemplate on the findings and will reach upon an agreement. The fourth and consequently last step of the credit rating process is that the credit ratings will be

published and will be made available in the financial market . After the credit rating is published the CRAs will monitor the creditworthiness of the company. At Moody’s all corporate credit ratings

are reviewed at least once a year, and Soevereign ratings at least once a half of year (Moody’s). In figure 2 the comprehensive rating process of Moody’s is depicted.

Figure 2: Moody’s Rating process

source: Moody’s

Credit rating model of Moody’s

KMV was the company owned by Kealhofer, McQuown, and Vasicek, they were the first to

in-troduce the EDF model (Moody’s). In 2002 the company was bought by Moody’s, and was re-named Moody’s KMV. Moody’s uses this EDF model model of KMV (Crosbie and Bohn, 2003).

The KMV’s EDF-model is an extension of the Merton Model (Moody’s Modelling Methodology, 2012). The Merton model is a Structural model, because a default event occurs when the assets

reach a sufficiently low level in comparison to its liabilities, and the company therefore can’t fulfill their legal obligation to their bondholders (McNeil et al, 2015). In the Merton model, the

r_f. Secondly, the firm’s asset-value process (VT) is independent from the way the firm is financed.

Lastly, the asset value (VT) can be traded on a frictionless market (McNeil et al, 2015).

The basic of the Merton model is that the firm’s assets value follows the stochastic Geometric Brownian motion (hereinafter: GBM). The GBM dynamic of the asset value is depicted in the

following equation (where W stands for a Brownian Motion):

dV_t = µvVtdt+ σvVtdWt; (1)

In the Merton model the pay-off for the equity holder as a function of the firm value is the same as the pay-off for a call option on Vt with strike B, as depicted in figure 3. Therefore the

Black-Scholes-Merton option pricing methods are allowed .

Figure 3: Merton Model

source: Website BNP Paribas

The Black-Scholes formula for the value of equity is given by:

S0= CBS(V0, B) = V0Φ(d1) − Be−rTΦ(d2) (2)

Where d1and d2are:

d1= log  V0 B  +  r+σ 2 V 2  T σV √ T and d2= d1− σV √ T (3)

with strike B, the following equation can be derived.

B₀= Br f− PBS(V0, B) = Be−rTΦ(d2) +V0Φ(−d1) (4)

The first procedure in modelling the default probability is determining the value of the asset and the volatility. With the technique successive substitution applied on the Merton model, using

the firms historical stock value St,the firms asset value Vt and the volatility parameter σvare found.

The second step in the credit modelling is to estimate the distance to default (hereinafter: DD).

The DD measures how many standard deviation the company is facing due to the default (Crosbie and Bohn, 2002). The higher the DD, the less likely the company is to default. The DD is

deter-mined by the following formula:

DD= log (V0/B) σV

√

T (5)

Finally, with a database of defaults, a distribution can be generated, which links the DD with a default probability. This empirical DD-to-EDF mapping gives a one-year default probability. The

assigned analysts use the result to make a rating recommendation.

Figure 4: Distance to default

2.5 Development of regulation for CRAs in Europe

The accuracy, objectivity and reliability of credit ratings are crucial given the significant role

CRAs has obtained in the financial market. However, the objectivity of the CRAs has been a topic of debate. Currently, the compensation model of CRAs is titled the "issuer-pay model", in which

entities that issue debts, compensate CRAs for a credit rating (DNB, 2011). The CRAs have been criticised that their issuer-pay model creates a conflict of interest, because it could give an

incen-tive to the CRAs to give the paying entity a favourable rating (White, 2010). The CRAs argue that this is not the case, since they must safeguard their credibility within the financial market or else

their reputation will be damaged. The effect of this market-discipline will encourage a genuine assessment of creditworthiness. Nonetheless, it is uncertain if the market discipline actually is able

to restrain CRAs and can penalise them. The market-discipline does not work properly when the reputation pressure is undermined by a lack of competition (Delimatsis & Herger, 2012). Due to

the absence of competition, the market-discipline can not be relied upon (DNB, 2011). Further-more, some studies suggest that CRAs have a "moral hazard" problem because the compensation

influenced the assigned ratings (Caprio et al., 2008). A second issue of critique is that there is a lack of competition, since the credit rating market is so highly concentrated. Healthy competition

is important because it can stimulate a higher rating quality (Malik, 2014). A third issue is that the CRAs methodology is not transparent and that they have no accountability (Voorhees, 2011).

Finally, a fourth issue is the overreliance on credit ratings. The European debt crisis highlighted the discussion, concerning the influence of CRAs in Europe. Requests were made to tackle the

overreliance on credit ratings through regulations.

Before the financial crisis commenced, CRAs were barely regulated in Europe. European

leg-islators confided upon the fact that the threat of losing their reputation and the market discipline could ensure reliable credit ratings (Utzig, 2010). In 2004 a request was made by the European

Parliament to conduct an assessment for the need of new legislation concerning CRAs. The Com-mittee of European Securities Regulators conducted the assessment and concluded that legislation

concerning CRAs were not needed to tackle the problems and failings of CRAs (Utzig, 2010). They relied on self regulation, enforced by the market-discipline and the International IOSCO

The first IOSCO CRA code was released in 2004, this code is intended to function as a guide-line to ensure an integer rating process (IOSCO, 2015). To meet the requirements of the IOSCO

CRA code, the following three steps were necessary. First, a self assessment of the CRA had to been conducted which indicated how the CRA had complied with the IOSCO CRA Code. This

assessment had to be submitted to the Committee of European Securities Regulators (hereinafter: Committee). The second requirement was that the CRA attend the yearly meeting with the

Com-mittee to discuss the obstacles and execution of the IOSCO CRA code. Finally, the last requirement was that CRAs would be transparent and provide the Committee with information about incidents

that occurred.

After the financial crisis, during the 2009 Summit, the G-20 top reached consensus that CRAs

should have a regulatory supervisory framework (IMF, 2010). In 2009 the first law was imple-mented in Europe: Regulation (EC) No 1060/2009. CRAs were required to register, and the NCA

obtained the responsibility of supervision.

During 2011, this new law was modified, in which the European Securities and Markets

Au-thority (ESMA), were made responsible for the supervision of CRAs operating in the EU. When CRAs do not comply with the regulations, ESMA can take action. This can be in the form of fines

or in more severe cases the withdrawal of registration (ESMA).

During the EU debt crisis, requests were made to create a new public European CRA (EPRS,

2016). The European parliament requested an assessment of the impact of the creation of a fully independent European Credit Rating Foundation. These attempts were cancelled. The main

prob-lems were recognised to be primarly an insufficient reputation, and the lack of financing. Accord-ing to the European Commission: "a new European credit ratAccord-ing agency would add little value to

investors information"(EPRS, 2016).

In the middle of 2013 the latest regulation package on CRAs was introduced: Regulation (EU)

No 462/2013 of the European Parliament. This legislation seeked to tackle several issues; the risk of overreliance on credit ratings, enhancing competition, improvement of the overall credit rating

quality, and the reduction of the conflict of interests with regard to the issuer-pays model (EUR-LEX). To stimulate the competition in the credit rating market that is dominated by "the big three",

new regulation encourages the use of credit rating of small CRAs. Despite all this new legislation, the effects seem to be less than desired with respect to the market concentration. According to

the studies published by the European Commission(2016), the conclusion of "the study state of the credit rating market" , is that the CRA regulation still has not had an effect on the market

concentration.

For this research the most relevant article is "article 5" of the 2013 regulation package. The

goal of this article is to tackle the problem of overreliance, by diminishing the references to credit ratings in EU legislation, and to encourage the use of internal ratings. This legislation should

encourage the use of internal procedures in order to make a autonomous credit risk assessment, instead of relying on primarily external credit ratings. This should result in investments decisions

that not solely depend on credit ratings. The article 5 of Regulation (EC) No 1060/2009, as depicted below, are found on the website of EUR-Lex access to European Union law .

Article 5a

Over-reliance on credit ratings by financial institutions

1. The entitites referred to in the first subparagraph of Article 4(1) shall make their own credit risk assessment and shall not solely or mechanistically rely on credit ratings for assessing the

credit-worthiness of an entity or financial instrument.

2. Sectoral competent authorities in charge of supervising the entities referred to in the first

sub-paragraph of Article 4(1) shall, taking into account the nature, scale and complexity of their activ-ities, monitor the adequacy of their credit risk assessment processes, assess the use of contractual

references to credit ratings and, where appropriate, encourage them to mitigate the impact of such

references, with a view to reducing sole and mechanistic reliance on credit ratings, in line with specific sectoral legislation.

Article 5b

Reliance on credit ratings by the European Supervisory Authorities and the European Systemic Risk Board

1. The European Supervisory Authority (European Banking Authority) (EBA) established by

Regulation (EU) No 1093/2010 of the European Parliament and of the Council (22), the Euro-pean Supervisory Authority (EuroEuro-pean Insurance and Occupational Pensions Authority) (EIOPA)

established by Regulation (EU) No 1094/2010 of the European Parliament and of the Council (23) and ESMA shall not refer to credit ratings in their guidelines, recommendations and draft technical

standards where such references have the potential to trigger sole or mechanistic reliance on credit ratings by the competent authorities, the sectoral competent authorities, the entitites referred to in

the first subparagraph of Article 4(1) or other financial market participants. Accordingly, by 31 December 2013, EBA, EIOPA and ESMA shall review and remove, where appropriate, all such

references to credit ratings in existing guidelines and recommendations.

2. The European Systemic Risk Board (ESRB) established by Regulation (EU) No 1092/2010 of

the European Parliament and of the Council of 24 November 2010 on European Union macro-prudential oversight of the financial system and establishing a European Systemic Risk Board (24)

shall not refer to credit ratings in its warnings and recommendations where such references have the potential to trigger sole or mechanistic reliance on credit ratings.

Article 5c

Over-reliance on credit ratings in Union law

Without prejudice to its right of initiative, the Commission shall continue to review whether ref-erences to credit ratings in Union law trigger or have the potential to trigger sole or mechanistic

reliance on credit ratings by the competent authorities, the sectoral competent authorities, the en-tities referred to in the first subparagraph of Article 4(1) or other financial market participants

with a view to deleting all references to credit ratings in Union law for regulatory purposes by 1 January 2020, provided that appropriate alternatives to credit risk assessment have been identified

and implemented.

The goal of this article is to tackle the problem of overreliance on credit ratings. However, the credit rating references in European legislation, like Basel III and Solvency II, have not been

replaced. For example, Basel III still allows external credit ratings to be used to determine risk weights. Although, the new Basel IV contains a proposition to reduce the overreliance on credit

ratings, through the introduction of standardised risk weights that does not rely on the use of credit ratings (D&P, 2017).

In this research we will investigate if the "overreliance" on credit ratings changed since the introduction of the new regulation measures. In the following section, prior conducted studies

2.6 Prior research on the influence of credit ratings

The underlying reason for the existence of CRAs is that they mitigate information asymmetry

with respect to their creditworthiness. According to the Efficient Markets Theory, all the significant relevant information is processed in the stock price. This entails that stock prices will only react to

credit rating up-and downgrades if the announcements contains new information. When the CRAs do not provide new information and the financial market already anticipated on the developments,

then credit rating announcements will not have a significant impact on the stock market. When credit ratings provide new information, then a change in credit rating can create an impact on the

stock price. Generally, since a credit upgrade means that the creditworthiness has increased, this could in theory lead to a higher stock price. Vice versa, a credit rating downgrade could lead to a

lower stock price.

There is no consensus on whether CRAs actually provide new information. Numerous

stud-ies were conducted regarding the question if CRAs provide new information. According to Hull (2004), who studied the impact of rating up- and downgrades on CDS prices, spreads anticipate

negative rating announcements. His results suggested that credit ratings provide no new infor-mation. Macey also stated that CRAs provide no new information in the financial market (Rhee,

2016). While Rhee (2015) had a nuanced opinion, he stated that CRAs produce little new infor-mation, but they sort available inforinfor-mation, which is an useful role in the financial market.

According to Herzog (2017), the influence of CRAs on the financial market can be best ex-plained by the regulatory endorsement of the use of credit ratings. De Haan (DNB, 2011) confirms

that the problem with regard to the overreliance on credit ratings, is a result of the use of credit ratings in regulations (for examples see section 2.2), resulting in investment decisions that are

strongly influenced by credit ratings. The New York Times reporter Friedman (1995) stated the following (Herzog, 2017):

"In fact, you could almost say that we live again in a two-superpower world. There’s the U.S. and there’s Moody’s. The U.S. can destroy a country by levelling it with bombs: Moody’s can destroy

a country by downgrading its bonds."

As credit ratings are able to affect the investment decision of investors, credit ratings can induce

wide shifts within the value of stock prices, government- and company debt. Consequently, a credit-rating downgrade can make it more difficult to attract new or additional funds, resulting in

higher expenses. For this reason Friedman concluded that CRAs indeed have the power to "break or make" companies and governments.

Effects of credit rating upgrades and downgrades on stock prices

In this research we will investigate if credit rating announcements, such as upgrades and down-grades, issued by Moody’s have a significant impact on the European stock market. Numerous

studies are conducted regarding the effect of credit rating announcements on the stock prices. Hand (1992) conducted research on the effect of announcements on bond- and stock prices. He

drew an overall conclusion that stock- and bond prices are affected by announcements of actual rat-ing changes as well as announcements of companies that are added on the Watch list. Holthausen

and Leftwich (1986) concluded otherwise. According to their study only a credit rating downgrade and a downgrade "Watch list" announcement have a significant negative effect on the stock price.

In contrast, a credit rating upgrade shows no significant effect. Dichev en Piotroski (2001) had similar results. The first month after a downgrade there was a significantly negative effect, while

no significant result was found in the case of an upgrade. Finally, Barron et al (1997), conducted research on credit rating changes and Watchlist announcements of S&P, in the U.K. stockmarket

between 1984 till 1992. They also found significant negative returns for credit rating downgrades. Below you will find an overview of earlier conducted research regarding the effects of credit

rating announcements on the stock prices (figure 5). The common perception of most academics is that credit rating upgrade announcements show insignificant impact on the stock return. Credit

Figure 5: Overview of earlier conducted research

3. Methodology

This chapter provides the data description, the timeline of the event study and the Market

model. The section data description describes which data is used in this study and how the data was collected and processed. hereafter, the selected timeline of the event study will be explained.

Finally, the Market model that is applied will be discussed in order to answer the main research questions:

1. Do credit rating up- and downgrade announcements, issued by Moody’s have a significant im-pact on the European stock market?

2. Do these announcements have a lower impact on the European stock market after the imple-mentation of the European regulation concerning credit rating agencies?

3.1 Data Description

This research is based upon announcements containing credit rating changes, both upgrades

and downgrades, retrieved from Moody’s. Moody’s was chosen because of their information ac-cessibility. The credit rating up-and downgrade announcements are gathered from the year 2013

and 2017, before- and after the regulations concerning CRAs were implemented. The data of daily stock prices are obtained from Bloomberg and Excel is used to process the data.

By using a period before and period after the new regulation it can be determined whether or not there is a difference between the effect on the stock price before- and after the implementation

of the regulation. This enables us to assess whether the effect of credit rating up- and downgrade has reduced after the new legislation has been implemented. To test if the credit ratings have a

significant effect on the stock price of European companies and to evaluate whether there is a dif-ference between the effect before and after the regulation, the data is divided into four different

groups. Namely, up- and downgrades before- and after the new regulation.

Data gathering process

The first step is to create an account to obtain access to Moody’s rating database (moodys.com).

On the website search engine of Moody’s, all credit rating actions can be gathered. The initial data sample of 2013 contains 104 credit rating up- and downgrades, and 88 for 2017. This number is

rating downgrades (63) occurred more often than upgrades (41). Contrary to 2013, credit rating upgrades (52) occurred more often than downgrades (36) in 2017. The differences in economic

growth could explain why the data sample of 2013 and 2017 are not similar. In 2013, Europe was still recovering from the debt crisis. In 2017 the monetary stimulus policy of the ECB and the low

interest rates resulted in economic growth. The difference between the economic growth in 2013 and 2017 is visualised in figure 6.

Figure 6: GDP annual growth rate

source: tradingeconomics.com

The countries where the companies of 2017 were founded and operate from are depicted in the

chart below:

The initial data is greatly reduced by imposing additional criteria. Since private companies do not have stock prices, only listed companies are selected. This information is gathered from

Bloomberg. When looking up credit rating upgrades and downgrades on the website search engine of Moody’s, the region must be verified. The "domicile" that Moody’s displays is not always

correct. When Europe is selected, it also displays results of Russian-, U.S.- or Arabic companies that are listed in Europe. These observations are also excluded, an example is the Russian oil

company Rosneft. Finally, the data must be checked for irregularities. For example, the company Northland (upgrade of 2013) was removed due to the occurrence of a reverse stock split. Another

example is the removal of the company Kion, because some data in the "estimation window" was missing. The upgrade occurred in July, but the company was listed on June 28th.

The final data-sample of upgrades and downgrades in Europe are presented in table 1. As depicted in tabel 1, the data-sample includes 12 downgrades and 25 upgrades in 2017 and 28

downgrades and 18 upgrades in 2013. An example of the dataset of credit rating downgrades of European companies is added in the Appendix.

Events Upgrades Downgrades

2013 18 28

2017 25 12

Table 1: Amount of events of listed companies

Finally, national holidays must be taken into account when they occur in the estimation window of 50 days. For example, when the stock price of a Dutch company is missing on april 27 (King’s

3.2 Timeline event study

An event study is a statistical method to assess the impact of an event on the valuation of a

corporate entity (MacKinlay, 1997). In order to carry out an event study, it must first be determined what the event is and what the selected timeline will be. Within this research an upgrade and

downgrade of a credit rating is considered an event. The timeline consists of the estimation window and the event window. The event window is the period in which the event occurs. The estimation

window is the period before the event, when the OLS-parameters are estimated to determine the "normal" return during the event window. The "normal" return is the estimated return without the

occurrence of the credit rating event.

As depicted in figure 8, a visual representation of the time line where τ = 0 is the rating

an-nouncement date, the estimation window (L1) starts at τ = T0, and ends at τ = T1. Furthermore,

the event window(L2) starts from τ = T1till τ = T2. The post event window, can be neglected since

it is not relevant for this research.

Figure 8: Timeline for the event study

source: Eventstudymetrics

On the basis of the stock data during the estimation window, the "normal" return of the stock price will be estimated in order to capture the behaviour without the occurrence of the credit rating

event. The purpose of an event study is to calculate the abnormal return in response to its event. The ’normal’ return, is estimated with the use of the market index. In this thesis the Stoxx 600 Europe

is used as the market index. This index consist of the following 17 European countries, Austria, the Netherlands, Belgium, Czech Republic, Denmark, Finland, France, Germany, Ireland, Italy,

In the event study conducted, an event window of (T1, T2) = (-2,+2) days is chosen. The

esti-mation window has been set to (T0, T1) = (-55,-5) days. The estimation window has been selected

well before the credit rating event occurred, such that the credit rating event has no influence on this period and the estimated Market Model parameters.

3.3 The Market Model

MacKinlay (1997) discusses the different methodologies of the event study in detail in his

paper. There are two different types of approaches to carry out an event study, the statistical and economic model. In this research the statistical approach to carry out an event study is selected.

The most common statistical models to estimate normal performance are the Constant Mean Return Model and the Market Model. The Market Model is selected to conduct the event study because

it is more sophisticated and accurate than the Constant Mean Return Model. Finally, the Market Model has been used by multiple researchers (Hand,1992).

The following notations are used in the Methodology. Stock prices are denoted with S, Returns with R, companies are denoted with i, time is denoted with τ. Where τ = 0 is the time when the

credit rating is upgraded or downgraded. As depicted in graph 7, the estimation window L1starts

at τ = T0, and ends at τ = T1. Furthermore, L2 represents the event window, which occurs from

τ = T1+ 1 till τ = T2. The daily stockprices are retrieved from Boomberg to derive the returns

that are used in the Market Model. The returns of the corporate company R_i,τ and the return of

the corresponding market index, i.e. the STOXX 600 index, are determined in the estimation and event window. The log stock return is calculated as follows:

R_i,τ = log Si,τ

S_i,τ−1 (6)

Where i indicates the company, τ the date and S the stock price.

For example, Ri,τ is the return of company i at date τ. These retrieved returns described in the

following equation are used in the Market Model (MacKinlay, 1997) :

R_i,τ = αi+ βiRm,τ+ εi,τ (7)

and βi is the sensitivity of the stock return to the market return. εi,τ is the disturbance term,

where E[ε_i,τ] = 0 and var[εi,τ] = σ2. The market model is estimated using Ordinary Least squares

(hereinafter: OLS). The OLS-estimators are derived with the data during the estimation window (L1). The formulas for the OLS-estimators are depicted below:

ˆ βi= ∑T_{τ =T}1 ₀₊₁(Ri,τ− ˆµi)(Rm,τ− ˆµm) ∑T_{τ =T}1 ₀₊₁(Rm,τ− ˆµm)2 (8) ˆ αi= ˆµi− ˆβiµˆm (9) ˆ σε i 2 = 1 L1− 2 T1

∑

Where ˆµmand ˆµiare the mean of the real return and market return during the estimation window,

L₁are the number of days in the estimation window.

Abnormal Return AR_i,τ

After the above mentioned, the abnormal returns are calculated. The abnormal return is the differ-ence between the return conditional on the credit rating event and the normal performance without

the event. The abnormal return is equal to the actual observed return Ri,t minus the estimated

"nor-mal return" ˆαi+ ˆβiRm,t. The formula to calculate the Abnormal Return for each event observation

is shown in the formula below:

AR_i,τ= R_i,τ− ( ˆαi+ ˆβiRm,τ) (11)

Cumulative Abnormal Return CARi

MacKinlay denoted the cumulative abnormal return as CARi(τ1, τ2) (hereinafter: CARi). The CARi,

during the event window, for each company is obtained by:

CARi= T2

∑

τ =T1

AR_i,τ (12)

with the the distribution of the CAR under H0, CARi∼ N(0, σi(CARi)2). This test statistic, derived

from the distribution, measures the significance for every individual upgrade or downgrade.

How-ever this is not useful, since multiple observations of credit rating up- and downgrades need to be analysed. In this thesis the "general" effect of up- and downgrades will be analysed, therefore the

observations must be aggregated.

The Average Aggregated Abnormal Return ARτ

The Average Aggregated Abnormal Return (herinafter:ARτ) of more than one observation can be

calculated with the following formula:

AR_τ = 1 N N

∑

where τ = T1+ 1, ..., T2given N companies.

The Cumulative Average Abnormal Return CAR

In this event study, multiple observations of credit ratings are applied, which means that the

Cu-mulative Average Abnormal Return (hereinafter: CAR) must be calculated. The CAR is calculated according the formula:

CAR=

T2

∑

τ =T1

AR_τ (14)

On the basis of this CAR it can be determined whether the event (the change of the credit rating)

has had a significant effect on the return of the company. The null hypothesis is when the mean of

CARis zero, the test statistic of the basic approach can be derived from the distribution:

CAR∼ N(0, var[CARi,t]) (15)

This gives us the "basic" test statistic (MarcKinlay,1997):

θ = CAR

The estimator for the variance is given by: var(CAR) = 1 N N

∑

As described by MacKinlay a modification on this basic approach was selected, the Patell test, to ensure a "more powerfull test". The Patell test standardise the AR_i,τ with an estimator of their

standard deviation.

James Patell test

The James Patell test is used to determine the significance of aggregated events (the upgrades and

downgrades). The Patell test standardise the ARi,τ (Patell, 1976). The Standardised Abnormal

Return (hereinafter: SAR_i,τ) for each company is calculated by standardising the Abnormal return

with a corrected estimated standard deviation:

SAR_i,τ= ARi,τ

S_ARi,τ (18)

Where SARi,t is the error corrected standard deviation obtained by the formula:

S2_ARi,t = ˆσ_AR2 _i,∗ (1 + 1 L₁+ (R_m,τ− µm,τ)2 T1 ∑ τ =T0 (R_m,τ− µm,τ)2 ) (19) ˆ σ_AR2 _i, = 1 L₁− 2 T1

∑

Furthermore, the Cumulative Standardised Abnormal (hereinafter: CSARi) is given by the

equa-tion: CSAR_i= T2

∑

The Patell statistic for testing H0: mean(CAR) = 0 is given by (Muller):

Z_p= √1 N N ∑ i=1 CSAR_i q L₂(L1−2) L1−4 ∼ N(0, 1) (22)

The test statistic Zp follows the standard normal distribution (MacKinlay, 1997). The significance

level for this research is set at 5%, this means that the critical values are -1.96 and 1.96.

Test the difference of downgrade effects before- and after regulation.

To evaluate the effects before and after the regulation, we will conduct a simple test to analyse if there is a significant difference between the years with the credit rating downgrades. This test is

solely the difference of the obtained results of 2013 and 2017 divided by the standard deviation of this difference. The hypothesis are defined as follows:

4. Results

First off the results of 2013 are discussed. To test whether the credit ratings have a significant

effect on the stock price of European companies, the data has been divided into two distinct groups, namely "downgrade" and "upgrade". An event study was carried out on each group. The

signifi-cance level for this research is set at 5%, this means that the critical values for the test statistics are -1.96 and 1.96. Next, the results of 2017 are discussed for the second research question. Finally,

the difference between the results of 2013 and 2017 are compared and analysed. This allows the impact of credit rating changes to be evaluated over time.

Saens & Sandoval (2005) conducted a study to examine how different data sizes ( 10-50 ob-servations) can influence the performance of an event study. They concluded that the power of the

test is very sensitive to data size. While discussing the results we should keep in mind that smaller sample sizes give less reliable results.

4.1 The Results of 2013

The results of 2013 are depicted in the table below.

2013 Upgrade Downgrade Observations 18 28 CAR 0.0191 -0.0201 Zp 1.5983 -2.2132 P-value 0.110 0.027 Table 2: Results 2013

For the downgrades, the null hypothesis is rejected. Subsequently, the effect of a credit rating downgrade on the stock price is considered to be significantly negative. Although the effect of

credit rating upgrades is positive, this result is considered to be insignificant. The results of the downgrades clearly suggest that downgrades have a significant negative effect on the stock price.

According to earlier conducted research, which was discussed in section 2.6, the common percep-tion is that only credit rating downgrades show a significant negative effect on stock prices. The

The reliance of the results also depend strongly on the sample size of the credit rating upgrades and downgrades. The data sample for the downgrades contain 28 observations and the data sample

of upgrades contain 18. Therefore the results of the upgrades are less reliable than the results of the downgrades. 4.2 The Results of 2017 2017 Upgrade Downgrade Observations 25 12 CAR 0.6834 -0.6484 Z_p 0.9439 -5,1778 P-value 0.345 2.24277E-07 Table 3: Results 2017

The results suggests that an upgrade has a positive effect on the stock prices and a downgrade

has a negative effect. But only the effect of the downgrades on the stock prices are significant, since the null-hypothesis is rejected. It should be taken into consideration, that the results for downgrades

are less reliable since the data sample only contains 12 observations. The null-hypothesis of the upgrades is not rejected, hence the effect of credit rating upgrades are insignificant. The results

of the aggregated observations of 2017 are similar with 2013, only the credit rating downgrades have a significant effect. These results correspond with the previous conducted studies (Frietas an

4.3 Evaluation downgrade effects before and after regulation.

The results of the differences in downgrades are depicted in the table below.

Difference between downgrades results

Observations 40

Z_{di f f} 1.482

P-value 0.069

Table 4: Results Difference

In both 2013 and 2017 the downgrades had a significant negative effect on the stock prices.

The effect of credit rating downgrades in 2017 seem to be more severe than in 2013. It should be noted that the result of 2017 is less reliable, since the number of observations in 2017(12) are less

than 2013(28). A test is conducted, described in section 3.3, to analyse the difference in the effect of downgrades between those years. The test did not yield significant results, since the p-value was

0.069. Hence, it can be concluded that there is no difference between the effects of downgrades in 2013 and 2017, although, for a more trustworthy assessment the sample size of 2017 should have

5. Conclusions and recommendations

This thesis examines whether the credit rating announcements of Moody’s have a significant

influence on the European stock market. The main research questions of this thesis are:

1. Do credit rating up- and downgrade announcements, issued by Moody’s, have a significant

impact on the European stock market?

2. Do credit rating announcements issued by Moody’s have a lower impact on the European stock

market after the implementation of the European regulation concerning credit rating agencies? Numerous studies have been conducted regarding the effects of credit rating announcements on

the stock prices. Prior research is not consensual on whether or not credit ratings impact the stock market. The common perception of most academics (section 2.6) is that credit rating downgrades

have shown a significant negative effect. Credit rating upgrade announcements have no significant effect on the stock return.

To make an assessment of the effect of credit rating up- and downgrades on stock prices, an event study is conducted with data of 2013. The results suggest that credit rating downgrades have

a significant negative effect on the stock price, credit rating upgrades have an insignificant effect. The results are in line with earlier conducted research.

For the second research question, an assessment is conducted to test if there is a difference before- and after the latest regulation. An event study on the data of 2017 is conducted, analysed,

and compared to the results of 2013. The results are similar: credit rating downgrades result in a significant negative return, and credit rating upgrades have an insignificant effect. Since the

effects of the upgrades in both years are insignificant, only the difference in the effect of credit rating downgrades are analysed. It is noticeable that the effect of credit rating downgrades in 2017

seems to be more severe than in 2013, but the results of 2017 are deemed less reliable than those of 2013. This is because the number of observations in 2013(28) are more than double the amount

of 2017(12). A test has been performed to analyse if there was a difference in the effect between those years. The results show that there is no difference in the impact of credit rating downgrades

in 2013 and 2017. Although, for a more reliable assessment, the sample size of 2017 should have been higher. Further research is recommended, because the results of 2017 are not deemed as

reliable as those of 2013. A good option could be to include the data of 2018, in order to increase the trustworthiness of the results of 2017 regarding the effect of the downgrades.

In 2013 and 2017 credit rating downgrades resulted in significant negative returns, therefore the desired effect is not demonstrated in our study. Consequently, it can not be concluded that the

implementation of the latest regulation on CRAs have the desired effect to tackle the overreliance on CRAs. A possible explanation could be the fact that article 5, that tackles the overreliance of

CRAs, is very complicated to carry out in practice. The strategy to tackle the before mentioned overreliance, is to diminish the references to credit ratings in EU legislations and to encourage

the use of internal ratings (As was mentioned in section 2.5). Up to this day the credit rating references in European legislation, such as Basel III, have not been replaced. Yet, there are some

developments, for example Basel IV, which introduces the use of standardised risk weights that does not rely on the use of credit ratings (DP, 2017).

It would be interesting to extend my research by including Outlook- and Watchlist announce-ments, in order to evaluate the effects on stock prices. It would also be interesting to assess if

there is a difference between a sudden credit rating downgrade, and a downgrade anticipated by an Outlook- or Watchlist announcements. Finally, another recommendation for further research

is to examine the effects of credit ratings on CDS spreads. A CDS is an agreement in which the buyer pays a premium, and in exchange the buyer will be compensated when the issuer of a loan

defaults (McNeil, 2005). The premium that the buyer pays for a CDS is called the "spread". Hull et al. (2003) studied the impact of rating up- and downgrades on CDS spreads. They concluded

that spreads anticipate negative rating announcements. Micu et al. (2006) concluded that rating Outlook, Watchlist announcements and actual rating changes have a significant effect on the CDS

spread. Finally, the IMF (2012) conducted a study on the impact of sovereign credit ratings. They concluded that a negative Outlook impact the CDS spread significantly and that the impact of credit

6. Appendix

6.1 Data

This table presents the sample of credit rating downgrades of 2013 which is obtained from

Moody’s and used in the event study.

6.2 Results Market model

This table presents the estimated αiand βifor each company with a credit rating downgrade in

2013.

7. References

AFM. (2014). Exploratory study on the use of credit ratings in the netherlands.

Alcubilla, R. G., & Del Pozo, J. R. (2012). Credit rating agencies on the watch list: Analysis of european regulation. OUP Oxford.

Barron, M. J., Clare, A. D., & Thomas, S. H. (1997). The effect of bond rating changes and new ratings on uk stock returns. Journal of Business Finance & Accounting, 24(3), 497–509.

Boot, A. W., Milbourn, T. T., & Schmeits, A. (2005). Credit ratings as coordination mechanisms. The Review of Financial Studies, 19(1), 81–118.

Cantor, R., & Mann, C. (2007). Analyzing the tradeoff between ratings accuracy and stability.

Caprio, G. J., Demirguc-Kunt, A., & Kane, E. J. (2008). The 2007 meltdown in structured

securi-tization: searching for lessons, not scapegoats. The World Bank.

Chiu, I. H. (2013). Regulatory governance of credit rating agencies in the eu: the perils of pursuing

the holy grail of rating accuracy. European Journal of Risk Regulation, 4(2), 209–226.

Comission, E. (2016). Study on the State of the Credit Rating Market Final Report. Retrieved from

https://ec.europa.eu/info/system/files/state-of-credit-rating-market-study-01012016en.pd f

(accessed June 2018)

Crosbie, P., & Bohn, J. (2003). Modelling default risk.

Delimatsis, P., & Herger, N. (2011). Financial regulation at the crossroads: Implications for

supervision, institutional design and trade(Vol. 12). Kluwer Law International.

De Nederlandsche Bank (DNB), J., De Haan, & Amtenbrink, F. (2011). Credit rating agencies.

De Santis, R. (2012). The euro area sovereign debt crisis: safe haven, credit rating agencies and

the spread of the fever from greece, ireland and portugal.

Dichev, I. D., & Piotroski, J. D. (2001). The long-run stock returns following bond ratings changes.

ESMA. (n.d.). supervision Credit Rating Agencies. Retrieved from

https://www.esma.europa.eu/supervision/credit-rating-agencies/supervision (accessed June 2018)

EUR-LEX. (n.d.). Regulation (EU) No 462/2013. Retrieved from

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32013R0462 (ac-cessed June 2018)

European Parliamentary Research Service, E. (2016). The case for a european public credit rating agency.

Eventstudymetrics. (n.d.). Event study methodology. Retrieved from

https://eventstudymetrics.com/index.php/event-study-methodology (accessed June 2018)

Freitas, A. d. P. N., & Minardi, A. M. A. F. (2013). The impact of credit rating changes in latin american stock markets. BAR-Brazilian Administration Review, 10(4), 439–461.

Friedman, T. L. (1995). Donât mess with moodys. New York Times, 22(February), A15.

Hand, J. R., Holthausen, R. W., & LEFTWICH*, R. W. (1992). The effect of bond rating agency announcements on bond and stock prices. The journal of finance, 47(2), 733–752.

Herzog, L. (2017). Just financial markets?: Finance in a just society. Oxford University Press.

Holthausen, R. W., & Leftwich, R. W. (1986). The effect of bond rating changes on common stock

prices. Journal of Financial Economics, 17(1), 57–89.

Hull, J., Predescu, M., & White, A. (2004). The relationship between credit default swap spreads,

bond yields, and credit rating announcements. Journal of Banking & Finance, 28(11), 2789– 2811.

IMF, J., Kiff, Holland, A., Kisser, M., Nowak, S., Saab, S., Schumacher, L., . . . Westin, A. (2010).

The uses and abuses of sovereign credit ratings. Global Financial Stability Report, 10.

Jarrow, R. A. (2011). The role of abs, cds and cdos in the credit crisis and the economy. Rethinking

Li, H., Visaltanachoti, N., & Kesayan, P. (2004). Effects of credit rating announcements: The swedish stock market. International Journal of Finance, 16(1).

Lynch, T. E. (2008). Deeply persistently conflicted: Credit rating agencies in the current regulatory environment. Case W. Res. L. Rev., 59, 227.

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

Malik, K. (2014). Is competition the right answer? a case of credit rating agencies. The Public Sphere.

McNeil, A. J., Frey, R., & Embrechts, P. (2015). Quantitative risk management: Concepts, techniques and tools-revised edition. Princeton university press.

Micu, M., Remolona, E. M., & Wooldridge, P. D. (2006). The price impact of rating announce-ments: which announcements matter?

Moody’s. (2018a). Moody’s Research & Ratings. Retrieved from

https://www.moodys.com/researchandratings/ (accessed June 2018)

Moody’s. (2018b). Moody’s Website. Retrieved from https://www.moodys.com/ (accessed

June 2018)

Muller, S. (n.d.). Significance Tests. Retrieved from

https://www.eventstudytools.com/significance-tests (accessed June 2018)

Norden, L., & Weber, M. (2004). Informational efficiency of credit default swap and stock markets:

The impact of credit rating announcements. Journal of Banking & Finance, 28(11), 2813–2843.

OECD. (2010). Competition and credit rating agencies.

on Banking Supervision (BCBS), B. C. (2003). The new basel capital accord.

Patell, J. M. (1976). Corporate forecasts of earnings per share and stock price behavior: Empirical

Rhee, R. J. (2015). Why credit rating agencies exist. Economic Notes: Review of Banking, Finance and Monetary Economics, 44(2), 161–176.

Saens, R., & Sandoval, E. (2005). Measuring security price performance using chilean daily stock returns: the event study method. Cuadernos de economía, 42(126), 307–328.

The International Organization of Securities Commissions, I. (2015). IOSCO. Retrieved from

https://www.iosco.org/news/pdf/IOSCONEWS375.pdf (accessed June 2018)

Utzig, S. (2010). The financial crisis and the regulation of credit rating agencies: A european banking perspective.

Voorhees, R. (2011). Rating the raters: Restoring confidence and accountability in credit rating agencies. Case W. Res. J. Int’l L., 44, 875.

White, L. J. (2010). Markets: The credit rating agencies. Journal of Economic Perspectives, 24(2), 211–26.