1
Amsterdam Business School
The Impact of Credit Rating Changes on Stock
Prices of Brazilian Firms
Name: Santiago Florez Pinto
Student Number: 11035277
Thesis Supervisor: Robin Doëttling
Date: 31 January 2018
Program: Economics and Business
2
Statement of Originality
This document is written by student Santiago Florez Pinto 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 other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
This study examines the impact of Fitch credit ratings announcements on stock returns of Brazilian firms listed on Bovespa over the period 2007 – 2015. An event study was conducted on 15 firms that were downgraded and 20 firms that were upgraded during this period, to examine stock market reaction. Abnormal returns were calculated using the Capital Asset Pricing Model in Excel for a 21-day period (-10, +10) to examine stock market reaction in the period immediately preceding and after the event which was the credit rating announcement. It was found that the Brazilian stock market is highly volatile and speculative in the period immediately preceding an upgrade announcement and relatively cautious in the period preceding a downgrade announcement. This indicates that information leaks into the Brazilian stock market which with relatively low level of disclosure serves as an information source to investors. This validates the information content hypotheses according to which stock prices react to any source of information to investors. No significantly high positive abnormal returns were observed in the period following an upgrade announcement indicating that an upgrade announcement reduces market speculation and prompts less risky behavior amongst investors. Statistically significant, high negative abnormal returns were observed in the period following a downgrade announcement indicating that the Brazilian stock market, like more stock markets in the world, negatively reacts to negative news like a downgrade. This has implications for an emerging market like Brazil that seeks to promote itself as an investor friendly destination which are discussed further in this paper.
4
Table of Contents Page No.
Abstract…...3
1. Chapter 1 –Introduction & Research Question………...……...6
1.1 Research Back ground………...6
1.2 Research Motivation…………...6
1.3 Overview of the Paper………...6
2. Chapter 2 – Literature Review & Hypothesis...8
2.1 Hypothesis Related to Credit Ratings…...8
2.2 Impact of Credit Ratings on Stock Prices…….…...9
2.3 Analyses of the Literature Review………...12
3. Chapter 3 - Methodology…...13
3.1 Statement of Hypothesis……...13
3.2 Methodology Approach….…………...13
3.3 Event Definition ………...…………...14
3.4 Data Collection…….………15
3.5 Estimation of Normal & Abnormal Returns...16
3.6 Testing for Abnormal Returns….…...18
4. Chapter 4 – Analyses & Discussion...19
4.1 Comparison of CAAR values…………...19
4.2 Analyses of Uowngraded Firms…………...22
4.3 Analyses of Downgraded Firms ………...23
4.4 Discussion………..…….25 5. Chapter 5 – Conclusion………..……….28 Reference List……….……….…….…...29 Appendix 1………..……....…..31 Appendix 2……….….….……….32 Appendix 3……….……..……….33
5 List of Figures
Figure 3.1. Event / Estimation Periods 14
List of Tables
Table 3.1. Fitch Ratings 15
Table 4.1. CAAR values for Upgraded Firms 20
Table 4.2. CAAR values for Downgraded Firms 21
Table 4.3. Returns Data for Upgraded Firm 22
6 Chapter 1. Introduction & Research Question
1.1 Research Background - Credit Ratings are defined as the ‘overall creditworthiness of
agencies to pay their debts’. The three most reputed credit rating agencies (CRA’s) are Standard & Poor’s (S&P), Moody’s and Fitch. Their main task is to assign credit ratings for issuers and rate the creditworthiness of countries / firms / financial products. Gonzalez et al., (2004) point out that credit ratings are assigned based on the scrutiny of a firms’ financial statements, quality of management, competitive positioning in the industry and take into account several macroeconomic and credit conditions. Becker and Milbourn (2011) state that credit ratings issued by CRA’s reflect all the latest available information about a firm and hence impact stock prices according to the efficient market hypothesis. The literature on whether or not stock prices react positively or negatively to CRA’s is inconclusive. This paper bridges this gap in the literature specifically for the emerging market of Brazil by investigating the following research question.
• How do stock prices in Brazil react due to credit rating changes?
1.2 Motivation of the Research - This research is important because it aims to establish how
stock market prices of Brazilian firms react to changes in credit ratings. Credit ratings determine if a country is a high potential investment destination or not. Hence the credit ratings assigned to companies operating in developing countries like Brazil determine the level of foreign investment they attract. Brazil forms part of the BRIC block (the others countries being Russia, India and China) which together represent the world’s fastest growing economies. Brazil is today the largest economy in South America and the 8th largest economy in the world (Credit Suisse, 2016). Brazil has one of the world’s most sophisticated stock exchanges called ‘Bovespa’ on which nearly 450 firms are listed (Singh, 2010). All equity / equity derivative trading in Brazil takes place on Bovespa. Neto (2008) says that the Bovespa plays an important role in the economic growth of Brazil as it is used as the mechanism whereby capital from abroad gets invested in Brazil. Since the establishment of Bovespa, Brazil has become an important investment destination for foreign investors.
It may be inferred from the above findings that it is important to understand how stock market prices of firms listed in Bovespa react to credit ratings assigned to them. This research is situated within the findings from the literature on the impact of credit ratings on stock prices. From the literature, it was identified that it is not known for sure if there is a directly proportionate relationship between credit ratings and stock price movements. Hence it is
7
important to investigate if negative credit ratings do in fact result in negative abnormal returns as this can foreign investment, reduce growth and create a negative impression of Brazil’s international image. If positive credit ratings result in positive abnormal returns, this can stimulate further investment and growth. If credit ratings have no impact on stock prices, this invalidates the supposition that such ratings incorporate latest information about stock markets.
In 2016, Moody’s along with S&P and Fitch, downgraded the sovereign credit rating of Brazil to Ba2 which meant a further lowering of Brazil’s investment grade status. This downgrading reflected the deterioration in the status in Brazil’s gross debt and skepticism about future GDP growth rate and implementation of structural reforms (Credit Suisse, 2016). Furthermore, the downgrading also reflected the political uncertainties in Brazil and the lack of synchronization between the presidential cabinet and the Congress (Credit Suisse, 2016). In this scenario, it becomes important to understand what can potentially happen to investor wealth in the event of further downgrades by the credit rating agencies.
1.3 Overview of the Paper – There is considerable ambiguity in the literature on stock price
movements on how investors react to credit rating changes. Hence it is not possible to predict investor behavior in Brazil in response to changes in credit ratings. In order to examine how changes in credit ratings impact stock price movements on the Bovespa, an event study is conducted in this research using 15 downgraded and 20 upgraded firms over the period 2007 – 2015 by Fitch. The Capital Asset Pricing Model (CAPM) was used to examine market reaction 10 days prior and 10 days after the announcement of credit rating change for each firm. A key finding is that there is a lack of transparency / information in the Brazilian stock market. Hence stock prices tend to over-react even to rumors of a credit rating change which serves as an information source. This is indicated by high abnormal return accruals in the period before an upgrade announcement and cautious investor behaviour in the period before a downgrade announcement. It was also found that like most financial markets across the world, there are statistically significant negative abnormal return accruals in the period following a downgrade. This means that further downgrades by credit ratings agencies could make Brazil a less attractive investor destination.
8 Chapter 2 – Literature Review and Hypothesis
This chapter will first examine hypothesis related to credit ratings and stock prices followed by a review of literature on impact of credit rating announcements on stock prices.
2.1. Hypotheses related to credit ratings – Various hypotheses have been proposed to
explain the relationship between credit ratings and stock returns. An understanding of these hypotheses will indicate which of them operate in the context of credit rating changes in the Brazilian stock market. The hypotheses include:
Efficient Market Hypothesis – According to Fama (1970), a perfectly efficient market is one
where stock prices reflect all available information at a given point in time. The efficient market hypothesis is based on three underlying principles including (i) investors are all rational beings, (ii) if they behave irrationally, their arbitrary trades will cancel out each other and (iii) full leverage of all arbitrage opportunities. Morana and Beltratti (2008) indicate that there are three types of efficient market hypothesis all of which differ in their interpretation of the phrase ‘all available information’. The ‘Weak Form’ hypothesis states that stock prices incorporate all historical information and that investment decisions based on historical information will not yield above average returns to investors. The ‘Semi-Strong’ efficiency hypothesis states that stock prices incorporate publicly available information and that investors will not be able to earn above average returns by basing investment decisions on publicly available information. The ‘strong’ hypothesis states that stocks reflect both private and publicly available information. The implication arising out of the efficient market hypothesis is that credit ratings that should be based on all available information have a significant impact on stock price movements.
Information Content Hypothesis - According to this hypothesis, a change in credit ratings
indicates to the market that the credit-worthiness of the issuing firms have changed (Mathis, 2009). This results in significant stock price reactions to credit rating announcements. Under this hypothesis, stock prices react on the date on which the credit ratings are announced.
Issuer Hypothesis – According to Welch (2004), the nature of the sector / industry in which a
firm operates impacts stock price movements when credit ratings are announced. Caton and Goh (2003) indicates credit rating announcements impact non-financial sector firms more because they do not have access to the same level of disclosure and to financial information as do firms operating in the financial sector. The issuer hypothesis states that stock prices of
9
non-financial firms react more strongly to credit rating announcements than those of financial firms.
Price Pressure Hypothesis – Financial regulations mandate that investors such as insurance,
pension / mutual funds etc. can invest only in investment grade rated debt. These regulations result in abnormal return accruals post credit rating changes. Steiner and Heinke (2001) observes that downgrades in credit ratings compel such investors to sell while upgraded credit ratings do not lead to a compulsion to buy. The price pressure hypothesis implies that cross-boundary credit rating changes significantly impact stock prices.
Differential Information Hypothesis – According to this hypothesis, the reaction of stock
prices to credit rating upgrades or downgrades is dependent on firm size. The smaller the firm, the more the volatility of the reaction to credit rating announcements. Lopes (2012) explained this by saying that smaller firms have limited disclosure and are not analyzed by financial experts. Whereas, investors typically like to know more about the firm that they invest in and prefer firms that disseminate more market information. Market capitalization of the firm is used to determine the extent of information asymmetries between the firm and capital markets. Bigger firms attract more investment as they reveal more information to the market and are better investigated by market analysts as compared to smaller firms. This is in accordance with the research of Dichev and Piotroski (2011) which found that a large firm capitalization is related to less information asymmetry and where abnormal returns are lower for larger firms.
The findings of this research will be examined in relation to these hypotheses.
2.2. Impact of Credit Ratings on Stock Prices - This section examines the impact of credit
ratings on stock price movements. The various views will indicate whether stock markets across the world react uniformly to credit rating changes in which case it will be possible to predict how the Bovespa will react either to an upgrade or a downgrade. A lack of consensus of views will indicate that it is not possible to predict investor behaviour in Brazil with any degree of certainty and it is necessary to empirically investigate impact of credit rating changes on the Bovespa.
The research of Steiner and Heinke (2001) which is based on German Eurobond information revealed that a negative / downgraded credit rating announcements resulted in abnormally negative stock price returns on the day of the announcement and on subsequent days. However, upgrades and positive reviews did not result in abnormal returns. Abnormal stock price movements are larger for speculative grade downgrades. Abnormal returns for bank
10
bonds were much lower than that of corporate / government bonds. Steiner and Heinke (2001) explained this by saying banks have better access to credit information and are more subject to prudential regulation. Elayan et al., (2003) examined the impact of credit rating announcements on New Zealand firms and found that both upgrades and downgrades resulted in significantly abnormal returns. This finding suggested that credit ratings serve as an information source to investors and this results in stock price movements. Linciano (2004) analyzed impact of changes in credit ratings on stock prices of listed Italian firms and found that there was no significant reaction to credit rating announcements. This research indicated that expected rating announcements have larger impact on stock prices than unexpected rating announcements. Poon and Chan (2008) examined the impact of credit rating announcements on stock price movements in China and found that availability of information, firm size and sector contribute to negative abnormal returns to credit rating announcements. Haan et al., (2011) found that in emerging markets, stock market reactions reacted abnormally to both positive and negative credit ratings and attributed this to the lack of financial disclosure in less developed markets. Given that Brazil is also a BRIC country like China, the findings from the Chinese markets can potentially indicate how the Bovespa will react to stock rating changes.
Abdeldayem and Nekhili (2016) conducted an event based study to examine the impact of changes in credit ratings on stock price movements in Bahrain. The specific event considered was the downgrading of Bahrain’s banking, services and industrial sectors in March 2016 by Moody’s. It was found that credit ratings serve as an information source to investors and that downgrades that reflect a negative outlook on the economy adversely impact long term equity returns. However, such downgrades have little impact on short term equity returns. Timmermans (2012) investigated the impact of changes in credit ratings on stock prices in the European Market over the period 1997-2012. It was found that downgrades resulted in significantly negative abnormal returns. Upgrades do not significantly impact abnormal returns on the date of announcement and post the announcements. Smaller firms and those in the financial sector react more strongly to credit rating downgrades.
Archana and Jayanna (2015) examined the impact of credit rating changes on stock prices in India. It was found that in there is no statistically significant impact of credit rating changes on stock price movements. This is so because the credit rating announcements contain no special data, do not surprise the market in any way and merely summarize information that is publicly available. However, the research of Tripathi (2017) suggests that credit ratings are an indication of the socio-economic and political climate prevailing in India currently. It was
11
found that the low credit ratings given to India by S&P reflects political apathy to economic policy making, high budget deficit, lack of tax reform, of infrastructure and very high corruption levels. A downgraded credit rating was found to be positively correlated with threat of foreign investor exit, transition from investment grade credit rating to speculative / junk and exit of pension and long-term funds. Rao (2013) states that downgraded rating announcements are characterized by significantly lower investor returns as such ratings are considered to be providing new information. On the other hand, upgraded credit rating announcements do not provide any new information other than what has already been considered in stock prices and hence the latter do not display much change in response. India is also a part of the BRIC block and from the literature seems to have a similar socio-political and economic climate like Brazil. The market reactions in India to credit rating changes can indicate how the Brazilian stock market also reacts to credit rating changes
The findings of Miyamoto (2016) on the impact of credit rating announcement in the Tokyo stock market suggest that the markets react positively to positive credit announcements. Stock prices were found to react even before the rating change announcements were made. This indicates that investors act on rumors of rating changes. They buy stocks before the announcement of credit rating in anticipation of an upgrade and sell the shares as soon as the upgrade announcements are made. It was concluded that an upgrade rating announcement is associated with significantly positive stock returns and downgrade rating announcements are associated with significantly negative stock returns.
Frietas and Minardi (2013) examined the impact of credit rating announcements in the Latin American market and found a significantly negative impact for rating downgrades but not much for rating upgrades. It was concluded that credit ratings are a source of relevant information for Latin American investors. Sabel (2012) did a similar study for the Scandinavian market and found that both upgrades and downgrades yield significantly abnormal returns on the pre-announcement, the announcement and the post-announcement days. According to Li et al., (2015) who analyzed the response of the Swedish stock markets to credit rating announcements, credit rating announcements do not impact stock price in the long term. However, positive ratings positively impact stock prices and negative ratings negatively impact stock prices in the short term.
12
The analysis of the hypotheses in section 2.1 suggests that credit ratings significantly impact stock price movements and that sector and firm size are important parameters that determine reaction to credit rating announcements. However, they do not indicate whether stock price movements move up or down in response to credit rating announcements. The literature on stock price movements in different countries is inconclusive. It is not clear whether stock prices move upwards or downwards in response to credit upgrades or downgrades or whether stock prices remain unaffected by changes in credit ratings. There is no research on how stock prices behave in response to changes in credit ratings given to Brazil. Hence it may be inferred that there from the literature review alone it is not possible to predict how the Bovespa reacts to credit rating announcements and this phenomenon needs to be empirically investigated. In doing so, this paper will contribute to the literature on impact of credit rating changes on stock price movements.
13 Chapter 3. Methodology
This section examines the methodology that is used in this research to validate the hypotheses derived in the literature review.
3.1. Statement of Hypotheses
Given the inconclusive nature of the findings from the literature on stock price movements, Hypothesis 1 is stated as follows:
H0 – There are no abnormal returns in stock prices of Brazilian companies after the announcement of a credit rating announcement.
H1 – There are significantly large abnormal returns in stock prices of Brazilian companies after
the announcement of a credit rating announcement.
3.2. Methodology Approach
For this, the event study methodology of De Jong (2007) is used. According to Miyamoto (2016), an event study methodology is the most appropriate tool with which to investigate the relationship between credit rating changes announcement and the corresponding stock price reactions. This methodology is based on the principle that an event such as the announcement of a credit rating change must be reflected immediately in stock prices due to the efficient market hypotheses. The objective of this methodology is to examine whether there is an abnormal change in the stock prices of Brazilian companies before and after the credit rating announcement event. There are four steps used in this method.
1. Event Definition – The first step is to determine the event for which the research is going to be conducted and the period for which the associated stock prices are to be collated.
2. Estimation of normal and abnormal returns – are the predicted returns after the announcement, while abnormal returns are the differences between actual and predicted returns.
3. Test for Abnormal Returns – The abnormal returns are aggregated and they are tested for statistical significance
14
4. Analyze results – Interpret results to (i) observe if rating changes impact stock prices and (ii) attempt an explanation for any observed changes based on literature review.
3.3. Event Definition
In this research, the credit rating announcement is an event and the day on which the announcement is made is called the credit rating announcement day or ‘Day 0’. The event window refers to the period surrounding the event. There is no consensus in the literature on event windows to analyze market reactions to credit rating changes Calderoni et al., (2009) point out that the choice of the event window is arbitrary and should not be too lengthy that would lead to the incorporation of other events or so small that would risk not capturing fully any abnormality in stock prices. Gonzalez et al., (2004) stated that the event window can range from 3 to 6 months and even up to 3 years from the day of the credit rating announcement. The literature indicates event window sizes of 1 year, 1 day, 3 days, 50 days, 26 days and even 25 days before and after the announcement (Norden and Weber, 2014).
I chose to analyze abnormalities in stock returns from 10 days prior to 10 days after the announcement date. I believe that this window is not so far from ‘Day 0’ that it can be impacted by other news but is long enough to examine reactions if any only to the credit rating announcement. Figure 3.1 indicates the event / estimation periods used in this research.
Figure 3.1. Event / Estimation Periods
Monthly index returns will be used as benchmark to check for abnormal stock returns, based on an estimation window period of 20 days preceding event date as indicated in figure 1. This research uses Fitch Ratings given to Brazilian companies. Table 3.1 gives the Fitch ratings, their numerical code equivalents expressed in points and the interpretations.
0
T1 T2 T3
T0
Event Window
(-10, +10 Days) Post Event Window(20 Days) Estimation Window
15
Table 3.1. Fitch Ratings
From table 3.1 it is observed that ratings of BBB- upwards means that a firm is investment grade because it has the capacity to meet its financial obligations. Else the firm is speculative grade and investments in these firms are risky as they may not be able to fulfill their financial obligations.
3.4. Data Collection
I collected daily closing prices and financial information from a sample of 35 Brazilian firms. The firms were chosen according to the following inclusion criteria.
• The firm was a publicly traded company with stocks listed in Bovespa since 2006 • The firm had to experienced changes in Fitch ratings over the period 2007 – 2015 • Must have opening and closing price data for a 30-day period before and after the
event date
The Fitch ratings for each firm was obtained from the Datastream database. The corresponding stock price information each firm for the event window was also obtained from the Datastream and from Wharton database. Based on the analysis of the Fitch credit
16
rating changes, I found 20firms which had experienced a Fitch upgrade over the period 2007 – 2015 and 15 firms which were downgraded by Fitch over the same period. The list of upgraded and downgraded firms along with the date of their Fitch Ratings are indicated in Appendices 1 and 2 respectively.
3.5. Estimation of Normal & Abnormal Returns
After establishing the event window length, the next step was to identify an appropriate model to determine the normal returns or expected returns and then calculate the abnormal returns.
3.5.1. Normal Returns
There are a variety of models including the mean - adjusted returns, the market – adjusted returns and the Capital Asset Pricing Model (CAPM). The main difference between these models is the choice of benchmark return model and estimation intervals.
Mean-Adjusted Returns Model – In this model, the benchmark refers to the average return
over the period T1 and T2. The normal returns are obtained using the formula:
…where T = T2-T1+1 refers to the number of time periods that can be used to evaluate the average return or length of estimation period. The disadvantage of this model is its arbitrary choice of benchmark period. It ignores market-wide stock price movements related to benchmark returns. This is particularly so if the events for different companies occur at the same point in time. This results in biased results if the entire market goes up or down during the event period. This results in the detection of significant abnormal returns which may occur not due to the event itself but rather to price movements across the markets. In order to adjust for this, the market adjusted returns model is used to determine the normal returns.
Market-Adjusted Returns Model – uses a return on market index (Rmt) as the benchmark and
given by the formula: NRit = Rmt
The abnormal returns identified from this formulate are termed as market adjusted returns. The disadvantage of this method is that it assumes that the ‘beta’ value of every stock is equal to 1 when in reality this is not the case. It is more appropriate to consider variations in ‘beta’
17
whilst determining abnormal returns. Hence the market adjusted returns model determines abnormal returns as residuals of the market model according to the formula:
The abnormal returns are derived in terms of residuals of the market adjusted returns model according to the formula:
The market model uses an estimation period just preceding the event period or around the event period but does not include the event period. One of the disadvantages of the market adjusted returns model is the impact of calendar time effects. It is well known that the stock price returns on Monday are less than other days while those of Friday and marginally higher than those of other days. Thus, if a credit rating announcement occurs on one of these days, there may be a bias in accrual of abnormal returns. To avoid this problem the CAPM model is used.
Capital Asset Pricing Model – models excess returns in terms of the formula:
The normal returns calculated using the CAPM model is given by the formula:
3.5.2. Abnormal Returns
Having identified the normal returns, the next step is to ascertain the abnormal returns. Abnormal returns are calculated by the difference between the actual observed returns and the expected returns (De Jong, 2007,2). The formula for the abnormal return is given as:
ARi = Ri – NRit.
…where Abnormal returns (ARi) are defined as the return (Ri) minus a benchmark or normal
return (NRit).
De Jong (2007) points out, the analysis of each firm’s returns in response to an event is not very informative, as stock price movements of individual firms can be caused by information unrelated to the event being considered. Therefore, the unweighted cross-sectional average of abnormal return is calculated using the equation:
18 …where is the number of events in the sample.
It is also necessary to look at longer periods surrounding the event. For that reason, cumulative abnormal returns are calculated (De Jong, 2007). The cumulative abnormal return is the sum of all abnormal returns. Aggregating the abnormal returns from the start (-10) up to the end (10) of the event the CAR may be calculated using the formula:
CAR = ARt1 + ARt2 + …. AR tN = ∑ 𝑨𝑨𝑨𝑨
…where the abnormal returns are aggregated from the start of the event period t1, up to time
tN .
In event studies the CARs are collected over the cross-section of events to get cumulative average abnormal returns (CAAR).
CAAR = ∑ 𝐀𝐀𝐀𝐀𝐀𝐀 or CAAR = 𝟏𝟏
𝐍𝐍∑ 𝐂𝐂𝐀𝐀𝐀𝐀
3.6. Testing for Abnormal Returns
The observed abnormal returns are aggregated and a statistical significance test implemented. In this research a simple t-test of significance will be used. The equation used is:
T = √𝑵𝑵 𝑨𝑨𝑨𝑨𝑨𝑨
𝑺𝑺 ≈ N(0,1)
with s = � 𝟏𝟏
𝑵𝑵−𝟏𝟏 �∑(𝑨𝑨𝑨𝑨 − 𝑨𝑨𝑨𝑨𝑨𝑨)^𝟐𝟐
The significance of the cumulative average abnormal return (sum of average abnormal return) can be tested using the equation:
T =√𝑵𝑵 𝑪𝑪𝑨𝑨𝑨𝑨𝑨𝑨
𝒔𝒔 ≈ N(0,1)
with s = � 𝟏𝟏
𝑵𝑵−𝟏𝟏 �∑(𝑪𝑪𝑨𝑨𝑨𝑨 − 𝑪𝑪𝑨𝑨𝑨𝑨𝑨𝑨)^𝟐𝟐
Microsoft Excel was used to calculate the above formulae. The above process will first yield abnormal returns (ARs) as an outcome. These are them cumulated over time to cumulative abnormal returns (CARs) which are then averaged over different observations of similar events to averaged abnormal returns (AARs). The cumulative averaged abnormal returns CAARs are then obtained. Regression tests were conducted to obtain the CAR Beta over the 21-day window (-10, +10) and the co-efficient of significance ‘p’ for each value of CAR
19
Beta. T tests were used to evaluate if the daily returns 10 days before and after the event date were significantly different from the returns on the event date.
Chapter 4 – Analysis & Discussion
This chapter presents and analyses the key findings of this research.
4.1. Comparison of CAAR Values
Tables 4.1 and 4.2 give the CAAR values for upgraded and downgraded firms respectively. From both tables, it is observed that the most significant difference between upgraded and downgraded firms is that the CAAR after the event of announcement of a downgrade is negative indicating statistically significantly negative abnormal returns. This is indicated in the ‘Average 10 Day CAAR’ column in table 4.2 where the CAAR is indicating negative values after the event date. In comparison, the CAAR for firms which are upgraded which show positive abnormal returns which are however not statistically significant.
20
Table 4.1. CAAR values for Upgraded Firms
Day Ambev ON Banco Bradesco PN Banco Do Brasil Banco ITAU 4 Banco Sandnder Brasil BR Malls Participacoes BR Properties ON Bradesco On Braskem PN Series Companhia Vale Do Rio CPAD Sanmt Basico De Engie Brasil Energia Fibria Cellulose ON Hypermarc as ON ITAU Unibanco Holding PN JBS On Petrobras PET4 Petroleo Brasileiro PN Sabesa ON Sul American Units Average 10 Day CAAR -10 1.214 3.117 8.938 1.876 3.183 2.123 3.014 2.365 3.171 5.549 -0.006 8.783 2.365 0.501 5.476 0.277 2.333 9.993 2.136 5.528 3.597 -9 2.429 6.020 17.901 3.616 6.784 4.332 5.679 4.730 6.293 11.079 -0.003 17.612 4.620 1.417 11.060 0.563 4.226 17.297 3.831 11.301 7.039 -8 3.641 9.171 27.009 5.474 10.385 6.877 8.275 7.167 9.134 16.682 -0.011 26.425 6.865 2.383 17.073 1.000 6.559 25.075 5.115 17.394 10.585 -7 4.878 12.044 36.112 7.331 13.618 9.692 10.773 9.328 12.431 21.989 -0.017 35.724 9.230 3.389 23.259 1.416 8.592 32.378 6.345 23.257 14.088 -6 6.103 14.977 45.120 9.107 16.623 12.667 13.201 11.156 15.662 27.139 -0.035 45.487 10.690 4.374 28.844 1.958 9.574 42.997 7.303 29.285 17.612 -5 7.316 18.008 54.048 11.147 19.684 15.589 15.707 12.826 18.813 32.661 -0.060 55.160 12.600 5.340 34.969 2.394 10.625 52.030 8.012 35.644 21.126 -4 8.569 20.775 63.132 13.337 22.798 18.695 18.088 14.100 22.100 37.624 -0.084 64.794 14.061 6.226 40.723 2.771 11.606 59.388 8.698 42.292 24.485 -3 9.804 23.964 72.695 15.404 25.767 21.895 20.315 15.478 25.466 42.644 -0.103 74.207 15.456 7.082 46.173 3.162 13.939 63.687 9.341 48.695 27.753 -2 10.939 27.592 82.473 17.471 28.896 25.031 22.543 16.780 28.402 47.707 -0.131 83.515 16.941 7.938 51.427 3.479 15.352 67.370 9.964 55.093 30.939 -1 12.099 31.062 92.186 19.692 31.547 28.127 24.657 17.985 31.164 52.936 -0.167 92.799 18.661 8.954 56.497 3.920 16.605 71.194 10.622 61.686 34.111 0 1.293 3.500 10.123 2.221 2.447 2.905 1.850 1.219 2.466 5.181 -0.037 8.918 1.480 1.091 5.149 0.352 0.703 3.078 0.842 6.358 3.057 1 2.608 4.511 20.251 4.443 4.893 5.771 3.555 2.362 6.038 10.342 -0.089 18.217 2.915 2.192 9.825 0.728 0.946 5.402 1.884 13.251 6.002 2 3.863 4.646 30.089 6.664 7.340 8.616 5.321 3.663 9.324 15.335 -0.123 27.870 4.490 3.208 14.504 1.090 1.350 6.060 2.718 20.184 8.811 3 5.050 4.372 39.797 8.886 9.786 11.257 7.255 4.964 12.096 20.564 -0.150 37.839 6.165 4.379 19.600 1.496 1.802 3.803 3.534 27.717 11.511 4 6.159 3.729 48.610 11.376 12.187 13.242 9.188 6.265 14.767 25.855 -0.180 48.242 7.785 5.499 24.595 1.978 2.241 2.312 4.369 34.855 14.154 5 7.265 4.169 56.488 14.034 14.534 14.857 11.312 7.567 17.568 30.883 -0.190 58.690 9.435 6.575 29.414 2.669 2.568 2.935 5.242 40.954 16.848 6 8.337 4.154 64.332 16.691 16.880 16.334 13.526 9.232 20.705 36.068 -0.199 69.009 11.161 7.726 33.702 3.476 2.177 4.128 6.129 46.662 19.511 7 9.401 3.823 72.595 19.349 18.890 17.624 15.563 10.897 23.716 41.300 -0.219 79.367 12.501 8.682 38.355 4.297 1.780 6.467 6.955 52.350 22.185 8 10.480 3.992 80.693 22.007 20.900 19.088 17.282 12.561 26.937 46.501 -0.225 88.685 13.571 9.618 42.546 4.879 2.031 8.775 8.140 57.943 24.820 9 11.557 4.098 88.446 23.665 22.911 20.702 18.794 14.226 29.974 52.001 -0.223 97.764 14.471 10.594 46.546 5.680 2.824 8.889 9.142 63.206 27.263 10 12.587 3.643 95.599 25.596 24.921 22.153 19.858 15.891 33.110 57.507 -0.221 106.762 15.691 11.520 50.504 6.517 4.168 8.697 10.311 68.239 29.653
21
Table 4.2. CAAR values for Downgraded Firms
Day Brasil Telcom Brookfield On Centrais Eletr Bras Companhia Siderurgica Cosan Industria Eletrobras Ele5 Gol Linhas Aereas Marfrig OGX Petroleo E Gas OI PN Redecard ON Sid Nacional PNA Sider Nacional PNB 1000 Telefonica DR 3 Usiminas-Usinas Siderurg Average -10 3.02942857 1.34594561 1.81687084 1.64176055 -0.131052 1.2638793 -0.4634789 0.417 34.96683348 -0.1952325 0.96541951 0.82857606 1.393091095 5.62965843 0.164073366 3.511518 -9 5.44285714 2.82801722 3.48874169 3.1835211 -0.582104 2.6327586 -0.8069569 0.909 72.43366697 -0.1554644 1.93413902 1.55715213 2.32718219 11.0893179 0.328146732 7.107332 -8 6.90878571 4.41469383 5.06561253 4.52528165 -1.1781561 4.10663791 -0.5704358 1.216 126.9005005 0.0343036 2.91417353 2.03572819 3.373273286 16.7939763 0.492220098 11.80217 -7 8.42171429 6.04377743 6.59248338 5.62204219 -1.1842081 5.59551721 0.03608576 1.563 177.8673339 0.31407064 3.89892305 2.35930425 4.326364381 22.6261352 0.656293463 16.31592 -6 11.0246429 7.90185904 8.16435422 6.56880274 -1.1052601 7.08939651 0.78260683 1.805 228.3341674 0.89383767 4.88885756 2.71288032 5.232455476 28.4057932 0.693700829 20.89287 -5 12.7750714 9.80517465 9.61122506 7.64056329 -0.6413121 8.61827581 1.76912889 2.012 269.3010009 2.04360521 5.88303707 3.03645638 6.120546571 34.2229521 0.677774195 24.85837 -4 17.653 11.4483933 10.8330959 8.64232384 -0.0973642 10.2521551 4.10564996 2.369 321.7678344 3.36337174 6.88476158 3.53003244 6.869637667 39.907611 0.661847561 29.87942 -3 21.3464286 12.9530824 11.5499668 9.58408439 0.69158381 11.9760344 6.53217152 2.781 365.7346679 4.85313878 7.87894109 4.27360851 7.626728762 45.5247699 1.179254927 34.29903 -2 23.0493571 14.5680295 12.0518376 10.5208449 1.78553179 13.6249137 8.36869259 3.188 375.2015014 6.88790581 8.8778356 5.29218457 8.323819857 51.0844289 1.696662293 36.30144 -1 23.8042857 16.1575316 12.6137084 11.4976055 2.89447976 15.398793 10.0852132 3.635 367.1683348 1.89476753 9.89605512 6.42576063 9.108910952 56.6965873 2.214069659 36.63274 0 0.28092857 1.63191011 0.62187084 0.79176055 1.36894798 1.8038793 1.03652107 0.432 -30.53316652 3.58453557 1.02670951 1.17357606 0.737091095 5.71465893 0.517407366 -0.65409 1 -2.7551429 3.40800422 1.23374169 1.3085211 2.53789595 3.4977586 1.20304263 -0.048 -84.56633303 4.8893031 2.05247402 2.44215213 1.38918219 11.2543174 1.201480732 -3.39677 2 -7.7347143 5.15299983 1.80061253 1.86528165 3.81184393 5.04663791 0.7195642 -0.148 -147.0994996 6.86907064 3.08342353 4.10572819 2.129773286 16.7714763 2.082221098 -6.76957 3 -13.851786 6.90647643 2.32748338 2.30704219 5.5307919 6.50051721 -0.2839147 -0.343 -197.1326661 8.38383817 4.10588805 6.53930425 2.851364381 22.4411357 3.136294963 -9.37208 4 -24.233857 8.65712654 2.26935422 2.62380274 7.43973988 7.90439651 -1.4673937 -0.423 -264.1658326 9.80360521 5.12458256 8.66288032 3.691455476 28.4482952 4.310368329 -13.4236 5 -33.763429 10.1561612 2.36622506 3.21556329 9.43868786 9.35327581 -3.3308721 -0.488 -350.6989991 10.8033722 6.13337707 10.6614564 4.510546571 34.3979541 5.567775695 -18.7785 6 -33.577 11.4799123 2.63809591 3.86732384 11.4176358 10.6771551 -5.394351 -0.543 -471.7321656 11.9281393 7.15018658 12.6650324 5.390637667 40.1651125 6.781849061 -25.8057 7 -30.121071 13.0326624 2.90996675 4.73408439 13.2665838 11.7760344 -7.6478285 -0.328 -578.2653321 13.3379073 8.16605609 14.5586085 6.341728762 45.6647709 8.232589427 -31.6227 8 -31.498643 14.610857 3.14183759 5.56584494 15.0605318 12.3699137 -9.8813064 -0.058 -688.2984986 14.9526753 9.1833356 16.5121846 7.149819857 51.1019304 9.683329793 -38.0269 9 -34.724714 16.3077916 3.34870844 6.21760548 17.3044798 12.748793 -11.674785 0.007 -828.8316652 16.2174429 10.2006151 18.2957606 8.033910952 56.5490893 11.13407016 -46.5911 10 -37.287286 18.2591682 3.48557928 7.03436603 19.5684277 13.1876723 -13.738264 -0.188 -990.3648317 11.2178946 19.6943367 8.864002048 62.1462487 12.40147702 -57.7146
22 4.2. Analysis of Upgraded Firms
Tables 4.3 present the summaries of the of the analysis conducted on the data obtained from Datastream and Wharton database including assessment of return on event date, average returns 10 days before and after the event date, the value of CAR Beta 10 days before and after event date and the corresponding significance ‘p’ values for 20 firms upgraded by Fitch. An example of how this analysis was done in excel is indicated in appendix 3.
Table 4.3. Returns Data for Upgraded Firms
Note: A regression analysis was done on the stock return values for each firm for the period (+10, -10) around the event date and values of Beta Coefficient and significance value of ‘p’ summarized in table 4.3. The Beta coefficient indicates the amount by which the stock return changes in response to an event announcement. The higher the value of beta (either positive or negative) the more the market reaction to an event such as credit rating announcement. If the value of ‘p’ is less than 0.05, this change is statistically significant. If the value of p is more than 0.05, changes in Beta are not statistically significant. Statistically significant values of ‘p’ are indicted through a * sign.
The key findings from table 4.3 are indicated as follows:
Stock Name Upgrades From
Return on Date of Event Average Returns 10 Days before Event Average Returns 10 Days after Event 10 Days Before Beta CAR
P Value for 10 Days Before Beta CAR
10 Days After Beta CAR P Value for 10 Days After Beta CAR Ambev ON BBB to A 4.30 4.24 4.17 1.63 0.0057* 0.15 0.8136 Banco Bradesco PN BB- to BB+ 9.92 9.52 10.23 1.02 0.0016 0.82 0.00032* Banco Do Brasil BB- to BB+ 19.02 18.31 17.73 0.65 0.1268 0.88 0.00284* Banco ITAU 4 BBB to BBB+ 10.34 9.84 10.18 0.81 0.004992* 0.95 0.0196
Banco Sandnder Brasil B to B+ 10.47 10.99 10.31 1.13 0.002032* 0.85 0.000429*
BR Malls Participacoes BB to BB+ 17.13 17.11 16.32 0.85 0.2450 1.04 0.0514
BR Properties ON BB to BB+ 14.63 15.14 14.60 -0.88 0.1239 0.65 0.1197
Bradesco On A- to A 0.00 -0.01 0.00 1.31 0.0709 0.73 0.00479*
Braskem PN Series A+ to AA- 10.79 11.11 11.15 1.33 0.000173* 0.93 0.000045*
Companhia Vale Do Rio BBB to BB+ 4.21 4.10 4.10 0.61 0.0076 0.92 0.0857
CPAD Sanmt Basico De A- to A 1.23 1.25 1.25 1.12 0.000384* 1.00 0.1353
Engie Brasil Energia BBB- to BB+ 35.80 36.00 36.40 1.23 0.0237 1.19 0.0683
Fibria Cellulose ON BB to BB+ 12.30 12.74 12.40 0.87 0.00011* 1.29 0.0810
Hypermarcas ON BB to BB+ 8.08 6.99 7.10 0.83 0.000437* 0.83 0.00007*
ITAU Unibanco Holding PN BBB- to BB+ 18.58 14.23 13.37 1.14 0.001* 1.23 0.00001*
JBS On A- to A 6.67 6.73 6.93 0.33 0.5573 0.29 0.3646
Petrobras PET4 BB to BB+ 6.02 7.22 6.27 1.43 0.000087* 0.80 0.02675*
Petroleo Brasileiro PN BB to BB+ 17.50 20.87 15.50 1.40 0.00628* 1.27 0.0804
Sabesa ON BB to BB+ 10.35 11.64 10.46 1.13 0.00403* 0.68 0.0901
23
T Tests were conducted on average returns of all 20 firms for 10 days before the event and the returns on the day of the event. The higher the positive or negative magnitude of the ‘T’ test, the more chances are that the impact of an event such as a credit rating announcement has a significant impact on stock return prices and the alternative hypotheses being true. The closer the value of ‘t’ to ‘0’, the greater the chance that there is no significant impact on stock return prices due to credit rating announcements and that the null hypotheses is true. The value of ‘t stat’ was -0.07381 (close to zero) and that of ‘p’ was 0.47 (>0.05) indicating no significant difference between the returns.
The T test for returns corresponding to 10 days after the event and the returns on the day of the event gave a ‘t stat’ value of 1.59 (higher value of t) and a borderline significance ‘p’ value of 0.063 (significance value is < 0.05). This indicates that an upgrade results in moderately significant differences in stock returns after an upgrade announcement.
An analysis of the values of CAR Beta for returns 10 days prior to announcement date indicates 13 of the 20 firms have a high positive Beta greater than 1. This indicates a great deal of volatility in the Brazil stock market prior to the announcement with higher risks being associated with higher abnormal returns. There is only one negative Beta which implies that in general, the market follows index trends. For 14 firms, the value of ‘p’ is <0.05 indicating that the difference between abnormal returns and market returns is significant.
The values of CAR Beta for returns 10 days after the announcement that just 6 firms have a high positive Beta greater than 1. This suggests more caution in investor behavior in the immediate aftermath of an upgrade, with less speculation and less abnormal returns accruing. Just 8 firms have a value of ‘p’ <0.05 further indicating that for most upgraded firms there Is no significant difference between market returns and daily returns.
4.3. Analysis of Downgraded Firms
Tables 4.4 present the summaries of the of the analysis conducted on the data obtained from Datastream and Wharton database on 15 downgraded firms.
T Tests were conducted on average returns of all 15 firms for 10 days before the event and the returns on the day of the downgrade event. The value of ‘t stat’ was -0.14313 and that of ‘p’ was 0.44 indicating no significant difference between the returns before the event.
24
Table 4.4. Returns Data for Downgraded Firms
Note: Note: A regression analysis was done on the stock return values for each firm for the period (+10, -10) around the event date and values of Beta Coefficient and significance value of ‘p’ summarized in table 4.3. The Beta coefficient indicates the amount by which the stock return changes in response to an event announcement. The higher the value of beta (either positive or negative) the more the market reaction to an event such as credit rating announcement. If the value of ‘p’ is less than 0.05, this change is statistically significant. If the value of p is more than 0.05, changes in Beta are not statistically significant. In table 4.4, statistically significant values of ‘p’ are indicted through a * sign.
The T test for returns corresponding to 10 days after the event and the returns on the day of the downgrade event gave a ‘t stat’ value of 1.57 and a high significance ‘p’ value of 0.009. This indicates that a downgrade announcement results in significant differences in stock returns post the event.
An analysis of the values of CAR Beta for returns 10 days prior to announcement date indicates that none of the firms have a Beta greater than 1. This indicates a less volatility in the Brazil stock market prior to the announcement of a downgrade suggesting heightened investor caution just before the announcement of a downgrade. There is only no negative Betas which implies that in general, the market follows index trends. For 9 firms, the value of ‘p’ is >0.05 indicating that for most firms there are no significant abnormal returns accruing just before a downgrade announcement.
25
The values of CAR Beta for returns 10 days after the announcement indicates that as many as 10 firms have a high positive Beta greater than 1. Furthermore, for all the values of Beta, ‘p’ is < 0.05. This suggests more volatility in the market after the announcement of a downgrade with more significantly negative abnormal returns accruing after a downgrade announcement.
4.4. Discussion
The key inferences that may be made based on the findings in section 4.1 are as follows:
The high volatility that characterizes the Brazil stock market before the announcement of an upgrade and the low volatility in the market before the announcement of a downgrade suggests that the market acts strongly on rumors with information on possible rating changes being leaked to the financial markets even before the announcement date. This inference is arrived at as follows. Mathis (2009) pointed out that according to the information content hypotheses, abnormal stock price movements in the event of a credit rating announcement indicates that such events serve as an information source to investors indicating a change in credit worthiness of firms. The findings of Caton and Goh (2003), Lopes (2012), Steiner and Heinke (2001) & Freitas and Minardi (2013) suggest that abnormal stock price movements in response to credit rating changes occur more often in a low-disclosure environment than in an environment characterized by information transparency and disclosure. In the 10-day period prior to a credit rating upgrade in Brazil, it is observed that investors buy and sell in the hope of making larger returns indicating that there is some information source that is prompting them to take higher risks confidently. In the 10 days period prior to a credit rating downgrade, there is less speculation in the market indicating a more cautious reaction to sources of information that indicate the imminent downgrade.
The high volatility that characterizes the stock market before the announcement of a rating upgrade must be contrasted with the lack of significant differences in daily and market returns before and after the announcement. Miyamoto (2016) indicated that higher volatility in stock prices in reaction to rumors of a possible upgrade indicate that investors buy stocks even before the announcement irrespective of risk in the hopes of making abnormally high returns after the upgrade announcement is made.
However, an analysis of the stock returns in the immediate aftermath of an upgrade announcement indicates no significantly high abnormal returns accruing. This suggests that the Brazilian stock market is characterized by high degree of speculation. However, the
26
announcement of an upgrade acts more as an information source to investors that has the immediate impact of reducing speculation and consequently lesser instances of abnormal return accruals.
This finding corroborates that of Steiner and Heinke (2001) who pointed out that upgrades do not result in abnormal returns and do not compel investors to buy. It however contravenes the views of Rao (2013) according to which upgrades merely incorporate existing information in stock prices. It is evident that in Brazil credit rating upgrades serve to inject more accurate and correct information than is prevalent in a highly speculative market and this has the effect of investors investing with more caution in the stock market.
In the event of a downgrade, findings indicate that there is no significant difference in returns immediately before the announcement and the event date. Such investor behavior is opposite to the high volatility that is manifest before an upgrade further corroborating the inference that the investors in Brazil react strongly to rumors and that market leak of information takes place quite often in the Brazilian market.
Due to fears of a downgrade, there is no buying or selling taking place and this is reflected by fewer abnormal returns before the announcements. This finding matches the views of Elayan et al., (2003) and Linciano (2004) both of whom indicated how credit ratings primarily act as an information sources facilitating investors to make appropriate investing decisions. In the case of Brazil, even the rumor of a downgrade has the effect of cautioning an otherwise very volatile stock market.
Unlike the immediate aftermath of an upgrade where there are no significant abnormal returns occurring, the period immediately following a downgrade announcement results in significantly higher market volatility and statistically significant negative abnormal returns. The high positive values of Beta suggest investor behavior that is highly correlated with the index with negative abnormal returns accruing in the immediate aftermath of a downgrade rating announcement.
This finding matches with that of Frietas and Minardi (2013) and of Li et al., (2015) who pointed out that downgrade credit rating announcements significantly and negatively impact stock market prices. It may be inferred from this that credit rating announcements convey new information to capital markets with negative investor sentiment prevailing after the announcement of a downgrade.
27
The above findings can be used to validate the two hypotheses proposed by this research. The null hypotheses states that there are no abnormal returns in stock prices of Brazilian companies after the announcement of a credit rating announcement. However, the findings suggest that downgrades result in significant abnormal returns accrual. Even the rumor of an upgrade has the impact of creating market volatility with investors making significantly high abnormal returns. Hence the null hypotheses may be rejected. However, the alternative hypotheses that there are significantly large abnormal returns in stock prices of Brazilian companies after the announcement of a credit rating announcement must be accepted with caution. There are no abnormal return accruals in the case of an upgrade announcement but only with the announcement of a downgrade.
28 Chapter 5 – Conclusion
My main aim in doing this research is to examine how stock market prices of firms that are listed in Bovespa respond to credit ratings assigned to them. This is important given the conflicting findings in the literature with respect to stock market reactions to credit ratings. It is important to understand if Bovespa reflects all market information which pre-empts market speculation and there is little chance of speculation that leads to abnormal return accruals. This is because Brazil is an emerging economy where the public has just begun to invest in the stock market in larger numbers and hence investor behavior has to be guided if mistakes and consequently large losses are to be avoided. A key finding that I made is that the Brazilian stock market is extremely sensitive to information. This suggests a lack of disclosure in the market that results in investors acting on any source of information that they can find. Even the rumor of an upgrade leads to speculation and consequently abnormal returns accrual even before the announcement of an upgrade and much more cautious investor behavior in the days leading to a downgrade. However, in the period following an upgrade announcement, there are fewer instances of abnormal returns occurring which suggests that the upgrade announcement serves as an information source that results in much more cautious investor behavior. The period following a downgrade is characterized by statistically significant negative accruals which suggests that the Brazil stock market, like most stock markets, reacts negatively to downgrades. This validates the information content hypotheses rather than any of the other hypotheses discussed in the literature.
This has serious implications for the Brazilian market which has already been suffered two downgrades over the period 2016 – 2017. Such negative credit ratings indicate that Brazil is not an investor friendly market, prompting withdrawal of existing investment out of the country and hampering fresh investments. I also identified that these downgrades are reflective not of firm performance but rather to factors occurring in the political and economic spheres of the country. This means that unless there is more political stability in the country in the near future coupled with economic and structural reform, investor sentiment and investor returns will continue to remain weak in the country.
29 References
Abdeldayem, M.M., Nekhili, R. (2016) Credit rating changes and stock market reaction in the kingdom of Bahrain. International Journal of Economics and Finance, Vol.8, pp.23-32.
Archana, H.N., Jayanna, S. (2015) Impact of rating changes on stock prices in India – A Study of banking sector. International Multidisciplinary Journal, Vol. 4, pp.254-271.
Barron, M.J., Clare, A.D., Thomas, S.H. (2013) The effect of bond rating changes and new ratings on UK stock returns. Journal of Business Finance & Accounting, Vol. 24(3), pp. 497-509.
Becker, B., Milbourn, T. (2011) How did increased competition affect credit ratings? Journal of
Financial Economics, Vol.101(3), pp.493-514.
Calderoni, F., Colla, P. and Gatti, S. (2009) Rating Changes: The European Evidence, Carefin
Research Paper Series.
Caton, G.L, Goh, J. (2003) “Are All Rivals Affected Equally by Bond Rating Downgrades?”,
Review of Quantitative Finance and Accounting, Vol. 20. Pp. 49–62.
Credit Suisse (2016) Brazil Sovereign Credit Rating. Brazil, Credit Suisse. De Jong, F. (2007) Event Studies Methodology. Tilburg, Tilburg University.
Dichev, I.D., Piotroski, J. (2011), The long-run stock returns following bond ratings Changes, Journal
of Finance, Vol. 56, p. 173-203.
Elayan, F. A., Hsu, W. H., Meyer, T. O. (2003) The informational content of credit rating announcements for share prices in a small market. Journal of Economics and Finance, Vol.27(3), pp.337-356.
Fama F. Eugene, (1970) Efficient Capital Markets: a review of theory and empirical work. New York, ACM Press.
Freitas, A.P., Minardi, A.M. (2013) The Impact of Credit Rating Changes in Latin American Stock Markets. Brazilian Administration Review. pp. 439-461,
Gonzalez, F., Haas, F., Johannes, R., Persson, M., Toledo, L., Violi, R., Wieland, M. (2004) Market dynamics associated with credit ratings: a literature review. Financial Stability Review, Vol.(4), pp.53-76.
Haan, J de and Amtenbrink, F. (2011). Credit Rating Agencies. DNB Working Paper, No. 278, January.
Holthausen, R. W. Leftwich, R. (1986) The effect of bond rating changes on common stock prices.
Journal of Financial Economics, Vol.1 7(1), pp. 57-89.
Katsuda, H., Ma, W. (2007) Comparison study on the communicative function of rating information of American and Japanese rating agencies. Ikoma Journal of Economics, Vol.4(3), pp.371-389. Linciano, N. (2004), The reaction of stock prices to rating changes, Working Paper, Rome. Lopes, A. B. (2002) The accounting information and the capital market (translated), Sao Paulo: Pioneira Thomson Learning.
30
Ma, W., Katsuda, H., Araki, T. (2004) Analysis of the effect of rating change on stock prices. The
Business Review of Kansai University, Vol.49(1), pp.127-147.
Mathis, J., McAndrews, J., & Rochet, J.-C. (2009) Rating the raters: are reputation concerns powerful enough to discipline rating agencies? Journal of Monetary Economics, 56(5), 657-674.
Morana, C., Beltratti, A., (2008) Comovements in international stock markets. Journal of
International Financial Markets, Institutions & Money, Vol. 18. Pp. 31-45.
Miyamoto, M. (2016) Event Study of Credit Rating Announcement in the Tokyo Stock Market
Journal of Economics, Business and Management, Vol. 4(2), pp.1-6.
Neto, A. (2008) Mercado Financeiro. Sao Paulo, Atlas.
Norden, L. and Weber, M. (2014) Informational Efficiency of Credit Default Swap and Stock arkets:
The Impact of Credit Rating Announcements, Working paper, University of Mannheim.
Pacheco, L. (2012) Moody’s credit ratings and the stock market performance of Portuguese rated firms. Journal of Advanced Studies in Finance, ASERS Publishing, Vol.No.1, pp. 68-83.
Poon, W. P. H. and Chan, K. C. (2008), An Empirical Examination of the Informational Content of Credit Ratings in China, Journal of Business Research, Vol. 61 (7), pp. 790-797.
Rao, S. (2013) The impact of Standards & Poor’s Credit Rating of the Indian Economy. International
Journal of Social Science & Interdisciplinary Research. Vol.2(1), pp. 1-7.
Sabel, L.G. (2012) The effect of credit ratings on stock returns in Scandinavia. Norwegian School of Economics.
Singh, M. (2016) Investing in Brazil. Available at <https://www.investopedia.com/contributors/334/> Accessed on 10th November 2017.
Steiner, M., Heinke,V. (2001) Event study concerning international bond price effects of rating actions. International Journal of Finance & Economics, Vol.6(2), pp.139-157.
Timmermans, M.A.J. (2012) How credit rating changes affect stock prices in the European market. Master Thesis Finance, Tilburg University.
Tripathi, T. (2017) Impact of Credit rating changes on stock prices in India. Vol.3(3), pp. 1-8. Welch, I. (2004) Capital structure and stock returns. Journal of Political Economy, 112(1), 106-132.
31
Appendix 1 – List of Brazilian Firms Downgraded by Fitch (2007 – 2015)
Firm Name
Date of Downgrade
Type of
Downgrade Industry Type
BRASIL TELEC ON CRT - Value - Shareholder
Loyalty/Fitch Credit Rating Feb 06, 2012
C to D
Telecommunications BROOKFIELD ON - Value - Shareholder
Loyalty/Fitch Credit Rating Feb 28, 2012
B1 to B2 Infrastructure (Real Estate) CENTRAIS ELETR BRAS- - Value - Shareholder
Loyalty/Fitch Credit Rating Sept 06, 2012
BBB+ to
BBB Infrastructure (Energy) COMPANHIA SIDERURGICA - Value - Shareholder
Loyalty/Fitch Credit Rating July 20, 2012
B to CCC+
Manufacturing (Steel) COSAN INDUSTRIA E - Value - Shareholder
Loyalty/Fitch Credit Rating Oct 14, 2011
B+ to B
Manufacturing (Steel) ELETROBRAS ELE5 - Value - Shareholder
Loyalty/Fitch Credit Rating Aug 21, 2012
B to B-
Infrastructure (Energy) GOL LINHAS AEREAS - Value - Shareholder
Loyalty/Fitch Credit Rating July 11, 2011
CC to CC-
Airlines MARFRIG FRIGORIFICOS ON - Value - Shareholder
Loyalty/Fitch Credit Rating Feb 28, 2013
B to B-
Food OGX PETROLEO E GAS - Value - Shareholder
Loyalty/Fitch Credit Rating Mar 27, 2012
B+ to B
Oil and Gas OI PN - Value - Shareholder Loyalty/Fitch Credit
Rating Feb 27, 2012
CCC to
CCC- Telecommunications REDECARD ON - Value - Shareholder Loyalty/Fitch
Credit Rating April 8, 2011
AAA to AA+
Finance SID NACIONAL PNA - Value - Shareholder
Loyalty/Fitch Credit Rating Sept 11, 2012
B+ to B
Manufacturing (Steel) SIDER.NACIONAL PNB 1000 - Value - Shareholder
Loyalty/Fitch Credit Rating Oct 10, 2012
AAA to AA+
Manufacturing (Steel) TELEFONICA DR3 - Value - Shareholder
Loyalty/Fitch Credit Rating Jul 30, 2010
B+ to B
Telecommunications USIMINAS-USINAS SIDERURG - Value -
Shareholder Loyalty/Fitch Credit Rating Nov 29, 2011
B+ to B
32
Appendix 2 - List of Brazilian Firms Upgraded by Fitch (2007 – 2015)
Firm Name Date
Type of
Upgrade Industry Type
AMBEV ON - Value - Shareholder Loyalty/Fitch Credit Rating May 6, 2011
BBB to A Manufacturing (Brewing) BANCO BRADESCO PN - Value - Shareholder Loyalty/Fitch
Credit Rating Oct 26, 2011
BB- to
BB+ Financial Service BANCO DO BRASIL ON - Value - Shareholder Loyalty/Fitch
Credit Rating July 4, 2011
BB- to
BB+ Financial Service BANCO ITAU 4 - Value - Shareholder Loyalty/Fitch Credit
Rating
Sept 14, 2012
BBB to
BBB+ Financial Service BANCO SANTANDER BRASIL - Value - Shareholder
Loyalty/Fitch Credit Rating April 9, 2012
B to B+
Financial Service BR MALLS PARTICIPACOES - Value - Shareholder
Loyalty/Fitch Credit Rating Oct 18, 2012
BB to BB+ Infrastructure (Real Estate) BR PROPERTIES ON - Value - Shareholder Loyalty/Fitch
Credit Rating
March 21, 2013
BB- to BB Infrastructure (Real Estate) BRADESCO ON - Value - Shareholder Loyalty/Fitch Credit
Rating
May 22, 2012
A- to A
Financial Service BRASKEM PN SERIES 'A' - Value - Shareholder
Loyalty/Fitch Credit Rating Dec 31, 2011
A+ to AA-
Oil and Gas COMPANHIA VALE DO RIO - Value - Shareholder
Loyalty/Fitch Credit Rating Nov 20, 2010
BBB to
BB+ Manufacturing CPAD.SANMT.BASICO DE - Value - Shareholder
Loyalty/Fitch Credit Rating Nov 17, 2012
A- to A
Food ENGIE BRASIL ENERGIA ON - Value - Shareholder
Loyalty/Fitch Credit Rating July 20, 2012
BBB- to
BB+ Infrastructure (Energy) FIBRIA CELULOSE ON - Value - Shareholder Loyalty/Fitch
Credit Rating
March 1, 2011
BB to BB+
Manufacturing HYPERMARCAS ON - Value - Shareholder Loyalty/Fitch
Credit Rating Dec 27, 2012
BB to BB+ Manufacturing (Pharmaceuticals) ITAU UNIBANCO HOLDING PN - Value - Shareholder
Loyalty/Fitch Credit Rating Nov 15, 2010
BBB- to
BB+ Financial Service JBS ON - Value - Shareholder Loyalty/Fitch Credit Rating
Sept 10, 2010
A- to A
Manufacturing PETROBRAS PET4 - Value - Shareholder Loyalty/Fitch Credit
Rating Nov 12, 2008
BB to BB+
Oil and Gas PETROLEO BRASILEIRO PN - Value - Shareholder
Loyalty/Fitch Credit Rating Nov 17, 2008
BB to BB+
Oil and Gas SABESP ON - Value - Shareholder Loyalty/Fitch Credit Rating
May 25, 2012
BB to BB+
Infrastructure SUL AMERICA UNITS - Value - Shareholder Loyalty/Fitch
Credit Rating Jan 1st, 2012
BB to BB+
Financial Service VCP IN - Value - Shareholder Loyalty/Fitch Credit Rating
March 1, 2012 BB+ to BBB- Infrastructure (Airports)
33
Appendix 3 – T Tests for Upgrades & Downgrades
Upgrade – 10 days before Event Variable 1 Variable 2
Mean 11.53454534 11.55738039
Variance 62.8110784 63.37991592
Observations 20 20
Pearson Correlation 0.9848388
Hypothesized Mean Difference 0
df 19
t Stat -0.073806197
P(T<=t) one-tail 0.470968046
Upgrade – 10 days After Event Variable 1 Variable 2
Mean 11.53454534 11.07740989
Variance 62.8110784 59.47776524
Observations 20 20
Pearson Correlation 0.986968246
Hypothesized Mean Difference 0
df 19
t Stat 1.597124735
P(T<=t) one-tail 0.063368072
Downgrade – 10 days before Event Variable 1 Variable 2
Mean 88.23541447 88.26452313
Variance 71367.0019 71392.63742
Observations 15 15
Pearson Correlation 0.99999567
Hypothesized Mean Difference 0
df 14
t Stat -0.143126212
P(T<=t) one-tail 0.444115048
Downgrade – 10 days after Event Variable 1 Variable 2
Mean 88.23541447 86.26073849
Variance 71367.0019 68837.91162
Observations 15 15
Pearson Correlation 0.999993987
Hypothesized Mean Difference 0
df 14
t Stat 1.572451953
